Showing 151–200 of 983 results for author: Basu, S

On the complexity of analyticity in semi-definite optimization

Authors: Saugata Basu, Ali Mohammad-Nezhad

Abstract: It is well-known that the central path of semi-definite optimization, unlike linear optimization, has no analytic extension to $μ= 0$ in the absence of the strict complementarity condition. In this paper, we show the existence of a positive integer $ρ$ by which the reparametrization $μ\mapsto μ^ρ$ recovers the analyticity of the central path at $μ= 0$. We investigate the complexity of computing… ▽ More It is well-known that the central path of semi-definite optimization, unlike linear optimization, has no analytic extension to $μ= 0$ in the absence of the strict complementarity condition. In this paper, we show the existence of a positive integer $ρ$ by which the reparametrization $μ\mapsto μ^ρ$ recovers the analyticity of the central path at $μ= 0$. We investigate the complexity of computing $ρ$ using algorithmic real algebraic geometry and the theory of complex algebraic curves. We prove that the optimal $ρ$ is bounded by $2^{O(m^2+n^2m+n^4)}$, where $n$ is the matrix size and $m$ is the number of affine constraints. Our approach leads to a symbolic algorithm, based on the Newton-Puiseux algorithm, which computes a feasible $ρ$ using $2^{O(m+n^2)}$ arithmetic operations. △ Less

Submitted 7 February, 2023; v1 submitted 15 January, 2023; originally announced January 2023.

MSC Class: 14P10; 90C22; 90C51

Multi-epoch radio source structure analysis of 11 calibrators at 2.3 and 8.4 GHz in the south

Authors: Sanway Chatterjee, Sayan Basu, Dan MacMillan

Abstract: We present the source structure analysis of 11 calibrator sources in the Southern Hemisphere at 2.3 (S-band) and 8.4 GHz (X-band). We used multi-epoch very long baseline interferometry source maps available in the radio fundamental catalog to analyse jet-structure variability and also used fluxes from the Goddard Space Flight Center database to see whether these two complement each other or not. A… ▽ More We present the source structure analysis of 11 calibrator sources in the Southern Hemisphere at 2.3 (S-band) and 8.4 GHz (X-band). We used multi-epoch very long baseline interferometry source maps available in the radio fundamental catalog to analyse jet-structure variability and also used fluxes from the Goddard Space Flight Center database to see whether these two complement each other or not. Also, total fluxes from the maps were plotted with the fluxes from the database. The S/X-band light curve analysis provides a more clear picture of the structural variability at the S/X-band also indicates the possibility of the "core-shift" phenomenon. We found jet-like structures in the majority of the sources in the sample. △ Less

Submitted 12 January, 2023; v1 submitted 12 January, 2023; originally announced January 2023.

Comments: 9 pages, 5 figures, 2 tables

MSC Class: 85-11 ACM Class: E.0

On the Challenges of using Reinforcement Learning in Precision Drug Dosing: Delay and Prolongedness of Action Effects

Authors: Sumana Basu, Marc-André Legault, Adriana Romero-Soriano, Doina Precup

Abstract: Drug dosing is an important application of AI, which can be formulated as a Reinforcement Learning (RL) problem. In this paper, we identify two major challenges of using RL for drug dosing: delayed and prolonged effects of administering medications, which break the Markov assumption of the RL framework. We focus on prolongedness and define PAE-POMDP (Prolonged Action Effect-Partially Observable Ma… ▽ More Drug dosing is an important application of AI, which can be formulated as a Reinforcement Learning (RL) problem. In this paper, we identify two major challenges of using RL for drug dosing: delayed and prolonged effects of administering medications, which break the Markov assumption of the RL framework. We focus on prolongedness and define PAE-POMDP (Prolonged Action Effect-Partially Observable Markov Decision Process), a subclass of POMDPs in which the Markov assumption does not hold specifically due to prolonged effects of actions. Motivated by the pharmacology literature, we propose a simple and effective approach to converting drug dosing PAE-POMDPs into MDPs, enabling the use of the existing RL algorithms to solve such problems. We validate the proposed approach on a toy task, and a challenging glucose control task, for which we devise a clinically-inspired reward function. Our results demonstrate that: (1) the proposed method to restore the Markov assumption leads to significant improvements over a vanilla baseline; (2) the approach is competitive with recurrent policies which may inherently capture the prolonged effect of actions; (3) it is remarkably more time and memory efficient than the recurrent baseline and hence more suitable for real-time dosing control systems; and (4) it exhibits favorable qualitative behavior in our policy analysis. △ Less

Submitted 1 January, 2023; originally announced January 2023.

Comments: Accepted to AAAI 2023

MetaCLUE: Towards Comprehensive Visual Metaphors Research

Authors: Arjun R. Akula, Brendan Driscoll, Pradyumna Narayana, Soravit Changpinyo, Zhiwei Jia, Suyash Damle, Garima Pruthi, Sugato Basu, Leonidas Guibas, William T. Freeman, Yuanzhen Li, Varun Jampani

Abstract: Creativity is an indispensable part of human cognition and also an inherent part of how we make sense of the world. Metaphorical abstraction is fundamental in communicating creative ideas through nuanced relationships between abstract concepts such as feelings. While computer vision benchmarks and approaches predominantly focus on understanding and generating literal interpretations of images, met… ▽ More Creativity is an indispensable part of human cognition and also an inherent part of how we make sense of the world. Metaphorical abstraction is fundamental in communicating creative ideas through nuanced relationships between abstract concepts such as feelings. While computer vision benchmarks and approaches predominantly focus on understanding and generating literal interpretations of images, metaphorical comprehension of images remains relatively unexplored. Towards this goal, we introduce MetaCLUE, a set of vision tasks on visual metaphor. We also collect high-quality and rich metaphor annotations (abstract objects, concepts, relationships along with their corresponding object boxes) as there do not exist any datasets that facilitate the evaluation of these tasks. We perform a comprehensive analysis of state-of-the-art models in vision and language based on our annotations, highlighting strengths and weaknesses of current approaches in visual metaphor Classification, Localization, Understanding (retrieval, question answering, captioning) and gEneration (text-to-image synthesis) tasks. We hope this work provides a concrete step towards developing AI systems with human-like creative capabilities. △ Less

Submitted 2 June, 2023; v1 submitted 19 December, 2022; originally announced December 2022.

Comments: Accepted in CVPR 2023. Project page: https://metaclue.github.io/ , Video summary: https://youtu.be/V3TmeNETL-o

Bayesian Arc Length Survival Analysis Model (BALSAM): Theory and Application to an HIV/AIDS Clinical Trial

Authors: Yan Gao, Rodney A. Sparapani, Sanjib Basu

Abstract: Stochastic volatility often implies increasing risks that are difficult to capture given the dynamic nature of real-world applications. We propose using arc length, a mathematical concept, to quantify cumulative variations (the total variability over time) to more fully characterize stochastic volatility. The hazard rate, as defined by the Cox proportional hazards model in survival analysis, is as… ▽ More Stochastic volatility often implies increasing risks that are difficult to capture given the dynamic nature of real-world applications. We propose using arc length, a mathematical concept, to quantify cumulative variations (the total variability over time) to more fully characterize stochastic volatility. The hazard rate, as defined by the Cox proportional hazards model in survival analysis, is assumed to be impacted by the instantaneous value of a longitudinal variable. However, when cumulative variations pose a significant impact on the hazard, this assumption is questionable. Our proposed Bayesian Arc Length Survival Analysis Model (BALSAM) infuses arc length into a united statistical framework by synthesizing three parallel components (joint models, distributed lag models, and arc length). We illustrate the use of BALSAM in simulation studies and also apply it to an HIV/AIDS clinical trial to assess the impact of cumulative variations of CD4 count (a critical longitudinal biomarker) on mortality while accounting for measurement errors and relevant variables. △ Less

Submitted 12 December, 2022; originally announced December 2022.

Training-Free Structured Diffusion Guidance for Compositional Text-to-Image Synthesis

Authors: Weixi Feng, Xuehai He, Tsu-Jui Fu, Varun Jampani, Arjun Akula, Pradyumna Narayana, Sugato Basu, Xin Eric Wang, William Yang Wang

Abstract: Large-scale diffusion models have achieved state-of-the-art results on text-to-image synthesis (T2I) tasks. Despite their ability to generate high-quality yet creative images, we observe that attribution-binding and compositional capabilities are still considered major challenging issues, especially when involving multiple objects. In this work, we improve the compositional skills of T2I models, s… ▽ More Large-scale diffusion models have achieved state-of-the-art results on text-to-image synthesis (T2I) tasks. Despite their ability to generate high-quality yet creative images, we observe that attribution-binding and compositional capabilities are still considered major challenging issues, especially when involving multiple objects. In this work, we improve the compositional skills of T2I models, specifically more accurate attribute binding and better image compositions. To do this, we incorporate linguistic structures with the diffusion guidance process based on the controllable properties of manipulating cross-attention layers in diffusion-based T2I models. We observe that keys and values in cross-attention layers have strong semantic meanings associated with object layouts and content. Therefore, we can better preserve the compositional semantics in the generated image by manipulating the cross-attention representations based on linguistic insights. Built upon Stable Diffusion, a SOTA T2I model, our structured cross-attention design is efficient that requires no additional training samples. We achieve better compositional skills in qualitative and quantitative results, leading to a 5-8% advantage in head-to-head user comparison studies. Lastly, we conduct an in-depth analysis to reveal potential causes of incorrect image compositions and justify the properties of cross-attention layers in the generation process. △ Less

Submitted 28 February, 2023; v1 submitted 9 December, 2022; originally announced December 2022.

Comments: ICLR 2023 Camera Ready version

Depth from Defocus Technique Applied to Unsteady Shock-Drop Secondary atomization

Authors: Shubham Sharma, Saini Jatin Rao, Navin Kumar Chandra, Aloke Kumar, Saptarshi Basu, Cameron Tropea

Abstract: The two-sensor depth from defocus (DFD) technique for the measurement of drop sizes in a spray is further developed to achieve higher spatial and temporal resolution, to improve estimates of size and number concentration, and to provide additional guidelines for the calibration and design of the optical system for a specific application. The technique and these improvements are demonstrated using… ▽ More The two-sensor depth from defocus (DFD) technique for the measurement of drop sizes in a spray is further developed to achieve higher spatial and temporal resolution, to improve estimates of size and number concentration, and to provide additional guidelines for the calibration and design of the optical system for a specific application. The technique and these improvements are demonstrated using the case of secondary atomization when a shock wave interacts with a single drop. This is an application in which both high spatially and temporally resolved number density and size distributions of secondary droplets generated in the wake of the original drop are necessary. △ Less

Submitted 6 December, 2022; originally announced December 2022.

Experimental investigation on the performance of thermosyphon charging of a single-medium stratified storage system for concentrated solar power applications

Authors: Dipti Ranjan Parida, Saptarshi Basu, Dhanush A P

Abstract: Concentrated solar power (CSP) plants utilize two-tank, sensible-heat thermal energy storage (TES) for uninterrupted electricity generation. However, the cost for the design and operation of TES is expensive. Therefore, researchers are focusing on implementing single-tank storage. Additional cutbacks can be made by utilizing pump-less thermosyphon charging for the TES. But prior thermosyphon resea… ▽ More Concentrated solar power (CSP) plants utilize two-tank, sensible-heat thermal energy storage (TES) for uninterrupted electricity generation. However, the cost for the design and operation of TES is expensive. Therefore, researchers are focusing on implementing single-tank storage. Additional cutbacks can be made by utilizing pump-less thermosyphon charging for the TES. But prior thermosyphon researches for TES are related to domestic water-heating systems of small-capacity (<100 liters) and low-temperature (<100 °C). Thus, investigations into thermosyphon charging for high-temperature storage are desired. This study focuses on thermosyphon-charging and storing of a single-medium stratified TES. The experiments were conducted on a 370 liters cylindrical storage (aspect ratio 4:1) with a heat-pipe system (3-liter volume) acting as a collector. Dowtherm-A oil was used as the heat transfer fluid (HTF), and the thermal expansion of HTF was accommodated in an expansion tank via two different designs (top and bottom connections from storage tank to expansion tank). Moreover, continuous and pulsatile charging are investigated for low (150 °C) and high (250 and 300 °C) temperatures. The results indicate that the maximum HTF temperature coming out of the heating pipes is ~25 °C more for the bottom-expansion design. Furthermore, it results in higher charging efficiency than the top-expansion setup for high-temperature studies. Finally, it is revealed that under design conditions, there are limits on the degree of thermal stratification achieved in the charging cycle and the maximum layover time allowable for interrupted charging. These results provide insights into the operational strategy of thermosyphon-charging stratified storage for CSP applications. △ Less

Submitted 30 November, 2022; originally announced November 2022.

Topic Segmentation in the Wild: Towards Segmentation of Semi-structured & Unstructured Chats

Authors: Reshmi Ghosh, Harjeet Singh Kajal, Sharanya Kamath, Dhuri Shrivastava, Samyadeep Basu, Soundararajan Srinivasan

Abstract: Breaking down a document or a conversation into multiple contiguous segments based on its semantic structure is an important and challenging problem in NLP, which can assist many downstream tasks. However, current works on topic segmentation often focus on segmentation of structured texts. In this paper, we comprehensively analyze the generalization capabilities of state-of-the-art topic segmentat… ▽ More Breaking down a document or a conversation into multiple contiguous segments based on its semantic structure is an important and challenging problem in NLP, which can assist many downstream tasks. However, current works on topic segmentation often focus on segmentation of structured texts. In this paper, we comprehensively analyze the generalization capabilities of state-of-the-art topic segmentation models on unstructured texts. We find that: (a) Current strategies of pre-training on a large corpus of structured text such as Wiki-727K do not help in transferability to unstructured texts. (b) Training from scratch with only a relatively small-sized dataset of the target unstructured domain improves the segmentation results by a significant margin. △ Less

Submitted 27 November, 2022; originally announced November 2022.

Comments: NeurIPS 2022 : ENLSP

Topological features of Haldane model on a dice lattice: role of flat band on transport properties

Authors: Sayan Mondal, Saurabh Basu

Abstract: We study the topological properties of a Haldane model on a band deformed dice lattice, which has three atoms per unit cell (call them as A, B and C) and the spectrum comprises of three bands, including a flat band. The bands are systematically deformed with an aim to study the evolution of topology and the transport properties. The deformations are induced through hopping anisotropies and are ach… ▽ More We study the topological properties of a Haldane model on a band deformed dice lattice, which has three atoms per unit cell (call them as A, B and C) and the spectrum comprises of three bands, including a flat band. The bands are systematically deformed with an aim to study the evolution of topology and the transport properties. The deformations are induced through hopping anisotropies and are achieved in two distinct ways. In one of them, the hopping amplitudes between the sites of B and C sublattices and those between A and B sublattices are varied along a particular direction, and in the other, the hopping between the sites of A and B sublattices are varied (keeping B-C hopping unaltered) along the same direction. The first case retains some of the spectral features of the familiar dice lattice and yields Chern insulating lobes in the phase diagram with $C=\pm2$ till a certain critical deformation. The topological features are supported by the presence of a pair of chiral edge modes at each edge of a ribbon and the plateaus observed in the anomalous Hall conductivity support the above scenario. Whereas, a selective tuning of only the A-B hopping amplitudes distorts the flat band and has important ramifications on the topological properties of the system. The insulating lobes in the phase diagram have distinct features compared to the case above, and there are dips observed in the Hall conductivity near the zero bias. The dip widens as the hopping anisotropy is made larger, and thus the scenario registers significant deviation from the familiar plateau structure observed in the anomalous Hall conductivity. However, a phase transition from a topological to a trivial insulating region demonstrated by the Chern number changing discontinuously from $\pm2$ to zero beyond a certain critical hopping anisotropy remains a common feature in the two cases. △ Less

Submitted 24 November, 2022; originally announced November 2022.

Effects of Non-Vanishing Net Charge in Balance Functions

Authors: Claude Pruneau, Victor Gonzalez, Brian Hanley, Ana Marin, Sumit Basu

Abstract: We investigate the impact of non-vanishing net-charge in collision systems on measurements of balance functions and their integrals. We show that the nominal balance function definition yields integrals that deviate from unity because of the non-vanishing net-charge. However, the integral of unified balance functions is shown to appropriately converge to unity when measured in a sufficiently wide… ▽ More We investigate the impact of non-vanishing net-charge in collision systems on measurements of balance functions and their integrals. We show that the nominal balance function definition yields integrals that deviate from unity because of the non-vanishing net-charge. However, the integral of unified balance functions is shown to appropriately converge to unity when measured in a sufficiently wide experimental acceptance. We furthermore explore the rate of convergence of unified balance functions integrals and study distortions imparted on the shape of balance functions when measurements are carried out in limited transverse momentum ($p_{\rm T}$) and rapidity ($y$) acceptances, such as those featured by experiments at the Large Hadron Collider or the Relativistic Heavy Ion Collider. We show that the shape and integral of unified balance functions may be strongly biased by reductions in the rapidity and transverse momentum acceptances of existing experiments. △ Less

Submitted 11 December, 2022; v1 submitted 19 November, 2022; originally announced November 2022.

Comments: 12 pages, 13 figures, regular article to be submitted to PRC

Improved effective Łojasiewicz inequality and applications

Authors: Saugata Basu, Ali Mohammad-Nezhad

Abstract: Let $\mathrm{R}$ be a real closed field. Given a closed and bounded semi-algebraic set $A \subset \mathrm{R}^n$ and semi-algebraic continuous functions $f,g:A \rightarrow \mathrm{R}$, such that $f^{-1}(0) \subset g^{-1}(0)$, there exist $N$ and $c \in \mathrm{R}$, such that the inequality (Łojasiewicz inequality) $|g(x)|^N \le c \cdot |f(x)|$ holds for all $x \in A$. In this paper we consider the… ▽ More Let $\mathrm{R}$ be a real closed field. Given a closed and bounded semi-algebraic set $A \subset \mathrm{R}^n$ and semi-algebraic continuous functions $f,g:A \rightarrow \mathrm{R}$, such that $f^{-1}(0) \subset g^{-1}(0)$, there exist $N$ and $c \in \mathrm{R}$, such that the inequality (Łojasiewicz inequality) $|g(x)|^N \le c \cdot |f(x)|$ holds for all $x \in A$. In this paper we consider the case when $A$ is defined by a quantifier-free formula with atoms of the form $P = 0, P >0, P \in \mathcal{P}$ for some finite subset of polynomials $\mathcal{P} \subset \mathrm{R}[X_1,\ldots,X_n]_{\leq d}$, and the graphs of $f,g$ are also defined by quantifier-free formulas with atoms of the form $Q = 0, Q >0, Q \in \mathcal{Q}$, for some finite set $\mathcal{Q} \subset \mathrm{R}[X_1,\ldots,X_n,Y]_{\leq d}$. We prove that the Łojasiewicz exponent $N$ in this case is bounded by $(8 d)^{2(n+7)}$. Our bound depends on $d$ and $n$, but is independent of the combinatorial parameters, namely the cardinalities of $\mathcal{P}$ and $\mathcal{Q}$. As a consequence we improve the current best error bounds for polynomial systems under some conditions. Finally, as an abstraction of the notion of independence of the Łojasiewicz exponent from the combinatorial parameters occurring in the descriptions of the given pair of functions, we prove a version of Łojasiewicz inequality in polynomially bounded o-minimal structures. We prove the existence of a common Łojasiewicz exponent for certain combinatorially defined infinite (but not necessarily definable) families of pairs of functions. △ Less

Submitted 1 December, 2022; v1 submitted 18 November, 2022; originally announced November 2022.

Comments: 29 pages. Theorem 1 expanded to include a bound on the constant $c$ in the integer case. Comments welcome

MSC Class: Primary 14P10; Secondary 03C64; 90C23

On a model-based analysis of vortex formations and decay in flows through bio-inspired T-shaped cavities

Authors: Sneham Das, Saikat Basu

Abstract: Fluidic transport in inverted T-shaped cavities with the flow entering through the top and exiting from the two bottom outlets experiences an interesting phenomenon that causes particles having density lower than that of the fluid medium to get trapped at the junction, in a horizontal formation. However, this only occurs across a small range of Reynolds numbers and that too in the laminar regime.… ▽ More Fluidic transport in inverted T-shaped cavities with the flow entering through the top and exiting from the two bottom outlets experiences an interesting phenomenon that causes particles having density lower than that of the fluid medium to get trapped at the junction, in a horizontal formation. However, this only occurs across a small range of Reynolds numbers and that too in the laminar regime. The unexpected phenomenon is conjectured to be modulated by formation of vortex tubes in the flow. Interestingly enough, such T-shaped (or, more generically, Y-shaped) cavities are also seen quite widely in anatomic pathways, e.g., in animal upper airways. It can be hypothesized that this trapping phenomenon can emerge in such geometries as well. Our current model simulates the occurrence with water as the ambient fluid medium passing through an idealized T-shaped space and measures the length of the vortices in each of the arms of the junction. In our study, we have conducted the investigation using a Reynolds number of 400 that lies in the laminar regime and have estimated the vortex tube lengths using the concept of nodal maximum velocities that the streamwise flow would attain on vortex dissipation. Vorticity and helicity contour variations as one goes further away from the junction have also been reported. △ Less

Submitted 13 November, 2022; originally announced November 2022.

Comments: 11 pages, 5 figures, 1 table

Spectral Triadic Decompositions of Real-World Networks

Authors: Sabyasachi Basu, Suman Kalyan Bera, C. Seshadhri

Abstract: A fundamental problem in mathematics and network analysis is to find conditions under which a graph can be partitioned into smaller pieces. The most important tool for this partitioning is the Fiedler vector or discrete Cheeger inequality. These results relate the graph spectrum (eigenvalues of the normalized adjacency matrix) to the ability to break a graph into two pieces, with few edge deletion… ▽ More A fundamental problem in mathematics and network analysis is to find conditions under which a graph can be partitioned into smaller pieces. The most important tool for this partitioning is the Fiedler vector or discrete Cheeger inequality. These results relate the graph spectrum (eigenvalues of the normalized adjacency matrix) to the ability to break a graph into two pieces, with few edge deletions. An entire subfield of mathematics, called spectral graph theory, has emerged from these results. Yet these results do not say anything about the rich community structure exhibited by real-world networks, which typically have a significant fraction of edges contained in numerous densely clustered blocks. Inspired by the properties of real-world networks, we discover a new spectral condition that relates eigenvalue powers to a network decomposition into densely clustered blocks. We call this the \emph{spectral triadic decomposition}. Our relationship exactly predicts the existence of community structure, as commonly seen in real networked data. Our proof provides an efficient algorithm to produce the spectral triadic decomposition. We observe on numerous social, coauthorship, and citation network datasets that these decompositions have significant correlation with semantically meaningful communities. △ Less

Submitted 8 May, 2024; v1 submitted 11 November, 2022; originally announced November 2022.

A fast-rotating pear-shaped nucleus

Authors: A. Karmakar, P. Datta, Soumik Bhattacharya, Shabir Dar, S. Bhattacharya, G. Mukherjee, H. Pai, S. Basu, S. Nandi, S. S. Nayak, Sneha Das, R. Raut, S. S. Ghugre, R. Banik, Sajad Ali, W. Shaikh, G. Gangopadhyay, S. Chattopadhyay

Abstract: The spectroscopic studies have identified few Actinide and Lanthanide nuclei of the periodic table, which can assume the pear shape. The low frequency collective rotation of these nuclei has been established by determining the band structure of the excited levels and their gamma decay rates. In this article, we report the rotation of a pear-shaped nucleus in 100 Ru, which rotates nearly three time… ▽ More The spectroscopic studies have identified few Actinide and Lanthanide nuclei of the periodic table, which can assume the pear shape. The low frequency collective rotation of these nuclei has been established by determining the band structure of the excited levels and their gamma decay rates. In this article, we report the rotation of a pear-shaped nucleus in 100 Ru, which rotates nearly three times faster than the previously known cases. The three novel consequences of this fast rotation are: the realization of a pear-shape in an excited state, its moment of inertia becoming a constant of motion and its shape evolution. These inferences have been arrived at by comparing the characteristics of the rotational band of 100 Ru with three of the best-known examples of a rotating pear-shaped nucleus and through the theoretical interpretation of the data. The observation of a pear-shaped nucleus in the lighter mass region opens up the possibility of a systematic study of the effects of fast rotation on this shape. △ Less

Submitted 11 November, 2022; originally announced November 2022.

RadFormer: Transformers with Global-Local Attention for Interpretable and Accurate Gallbladder Cancer Detection

Authors: Soumen Basu, Mayank Gupta, Pratyaksha Rana, Pankaj Gupta, Chetan Arora

Abstract: We propose a novel deep neural network architecture to learn interpretable representation for medical image analysis. Our architecture generates a global attention for region of interest, and then learns bag of words style deep feature embeddings with local attention. The global, and local feature maps are combined using a contemporary transformer architecture for highly accurate Gallbladder Cance… ▽ More We propose a novel deep neural network architecture to learn interpretable representation for medical image analysis. Our architecture generates a global attention for region of interest, and then learns bag of words style deep feature embeddings with local attention. The global, and local feature maps are combined using a contemporary transformer architecture for highly accurate Gallbladder Cancer (GBC) detection from Ultrasound (USG) images. Our experiments indicate that the detection accuracy of our model beats even human radiologists, and advocates its use as the second reader for GBC diagnosis. Bag of words embeddings allow our model to be probed for generating interpretable explanations for GBC detection consistent with the ones reported in medical literature. We show that the proposed model not only helps understand decisions of neural network models but also aids in discovery of new visual features relevant to the diagnosis of GBC. Source-code and model will be available at https://github.com/sbasu276/RadFormer △ Less

Submitted 9 November, 2022; originally announced November 2022.

Comments: To Appear in Elsevier Medical Image Analysis

CPL: Counterfactual Prompt Learning for Vision and Language Models

Authors: Xuehai He, Diji Yang, Weixi Feng, Tsu-Jui Fu, Arjun Akula, Varun Jampani, Pradyumna Narayana, Sugato Basu, William Yang Wang, Xin Eric Wang

Abstract: Prompt tuning is a new few-shot transfer learning technique that only tunes the learnable prompt for pre-trained vision and language models such as CLIP. However, existing prompt tuning methods tend to learn spurious or entangled representations, which leads to poor generalization to unseen concepts. Towards non-spurious and efficient prompt learning from limited examples, this paper presents a no… ▽ More Prompt tuning is a new few-shot transfer learning technique that only tunes the learnable prompt for pre-trained vision and language models such as CLIP. However, existing prompt tuning methods tend to learn spurious or entangled representations, which leads to poor generalization to unseen concepts. Towards non-spurious and efficient prompt learning from limited examples, this paper presents a novel \underline{\textbf{C}}ounterfactual \underline{\textbf{P}}rompt \underline{\textbf{L}}earning (CPL) method for vision and language models, which simultaneously employs counterfactual generation and contrastive learning in a joint optimization framework. Particularly, CPL constructs counterfactual by identifying minimal non-spurious feature change between semantically-similar positive and negative samples that causes concept change, and learns more generalizable prompt representation from both factual and counterfactual examples via contrastive learning. Extensive experiments demonstrate that CPL can obtain superior few-shot performance on different vision and language tasks than previous prompt tuning methods on CLIP. On image classification, we achieve 3.55\% average relative improvement on unseen classes across seven datasets; on image-text retrieval and visual question answering, we gain up to 4.09\% and 25.08\% relative improvements across three few-shot scenarios on unseen test sets respectively. △ Less

Submitted 4 November, 2022; v1 submitted 19 October, 2022; originally announced October 2022.

Equi-Tuning: Group Equivariant Fine-Tuning of Pretrained Models

Authors: Sourya Basu, Prasanna Sattigeri, Karthikeyan Natesan Ramamurthy, Vijil Chenthamarakshan, Kush R. Varshney, Lav R. Varshney, Payel Das

Abstract: We introduce equi-tuning, a novel fine-tuning method that transforms (potentially non-equivariant) pretrained models into group equivariant models while incurring minimum $L_2$ loss between the feature representations of the pretrained and the equivariant models. Large pretrained models can be equi-tuned for different groups to satisfy the needs of various downstream tasks. Equi-tuned models benef… ▽ More We introduce equi-tuning, a novel fine-tuning method that transforms (potentially non-equivariant) pretrained models into group equivariant models while incurring minimum $L_2$ loss between the feature representations of the pretrained and the equivariant models. Large pretrained models can be equi-tuned for different groups to satisfy the needs of various downstream tasks. Equi-tuned models benefit from both group equivariance as an inductive bias and semantic priors from pretrained models. We provide applications of equi-tuning on three different tasks: image classification, compositional generalization in language, and fairness in natural language generation (NLG). We also provide a novel group-theoretic definition for fairness in NLG. The effectiveness of this definition is shown by testing it against a standard empirical method of fairness in NLG. We provide experimental results for equi-tuning using a variety of pretrained models: Alexnet, Resnet, VGG, and Densenet for image classification; RNNs, GRUs, and LSTMs for compositional generalization; and GPT2 for fairness in NLG. We test these models on benchmark datasets across all considered tasks to show the generality and effectiveness of the proposed method. △ Less

Submitted 4 February, 2023; v1 submitted 13 October, 2022; originally announced October 2022.

Journal ref: AAAI 2023

Non-rough norms and dentability in spaces of operators

Authors: Susmita Seal, Sudeshna Basu, Julio Becerra Guerrero, Juan Miguel Villegas Yeguas

Abstract: In this work, we study non-rough norms in L(X,Y), the space of bounded linear operators between Banach spaces X and Y. We prove that L(X,Y) has non-rough norm if and only if X* and Y have non-rough norm. We show that the injective tensor product of X and Y has non-rough norm if and only if both X and Y have non-rough norm. We also give an example to show that non-rough norms are not stable under p… ▽ More In this work, we study non-rough norms in L(X,Y), the space of bounded linear operators between Banach spaces X and Y. We prove that L(X,Y) has non-rough norm if and only if X* and Y have non-rough norm. We show that the injective tensor product of X and Y has non-rough norm if and only if both X and Y have non-rough norm. We also give an example to show that non-rough norms are not stable under projective tensor product. We also study a related concept namely the small diameter properties in the context of L(X,Y)*. These results leads to a discussion on stability of the small diameter properties for projective and injective tensor product spaces. △ Less

Submitted 18 October, 2022; v1 submitted 11 October, 2022; originally announced October 2022.

Comments: 13 pages

MSC Class: 46B20; 46B28

Generalization Properties of Retrieval-based Models

Authors: Soumya Basu, Ankit Singh Rawat, Manzil Zaheer

Abstract: Many modern high-performing machine learning models such as GPT-3 primarily rely on scaling up models, e.g., transformer networks. Simultaneously, a parallel line of work aims to improve the model performance by augmenting an input instance with other (labeled) instances during inference. Examples of such augmentations include task-specific prompts and similar examples retrieved from the training… ▽ More Many modern high-performing machine learning models such as GPT-3 primarily rely on scaling up models, e.g., transformer networks. Simultaneously, a parallel line of work aims to improve the model performance by augmenting an input instance with other (labeled) instances during inference. Examples of such augmentations include task-specific prompts and similar examples retrieved from the training data by a nonparametric component. Remarkably, retrieval-based methods have enjoyed success on a wide range of problems, ranging from standard natural language processing and vision tasks to protein folding, as demonstrated by many recent efforts, including WebGPT and AlphaFold. Despite growing literature showcasing the promise of these models, the theoretical underpinning for such models remains underexplored. In this paper, we present a formal treatment of retrieval-based models to characterize their generalization ability. In particular, we focus on two classes of retrieval-based classification approaches: First, we analyze a local learning framework that employs an explicit local empirical risk minimization based on retrieved examples for each input instance. Interestingly, we show that breaking down the underlying learning task into local sub-tasks enables the model to employ a low complexity parametric component to ensure good overall accuracy. The second class of retrieval-based approaches we explore learns a global model using kernel methods to directly map an input instance and retrieved examples to a prediction, without explicitly solving a local learning task. △ Less

Submitted 5 October, 2022; originally announced October 2022.

Towards Safe Mechanical Ventilation Treatment Using Deep Offline Reinforcement Learning

Authors: Flemming Kondrup, Thomas Jiralerspong, Elaine Lau, Nathan de Lara, Jacob Shkrob, My Duc Tran, Doina Precup, Sumana Basu

Abstract: Mechanical ventilation is a key form of life support for patients with pulmonary impairment. Healthcare workers are required to continuously adjust ventilator settings for each patient, a challenging and time consuming task. Hence, it would be beneficial to develop an automated decision support tool to optimize ventilation treatment. We present DeepVent, a Conservative Q-Learning (CQL) based offli… ▽ More Mechanical ventilation is a key form of life support for patients with pulmonary impairment. Healthcare workers are required to continuously adjust ventilator settings for each patient, a challenging and time consuming task. Hence, it would be beneficial to develop an automated decision support tool to optimize ventilation treatment. We present DeepVent, a Conservative Q-Learning (CQL) based offline Deep Reinforcement Learning (DRL) agent that learns to predict the optimal ventilator parameters for a patient to promote 90 day survival. We design a clinically relevant intermediate reward that encourages continuous improvement of the patient vitals as well as addresses the challenge of sparse reward in RL. We find that DeepVent recommends ventilation parameters within safe ranges, as outlined in recent clinical trials. The CQL algorithm offers additional safety by mitigating the overestimation of the value estimates of out-of-distribution states/actions. We evaluate our agent using Fitted Q Evaluation (FQE) and demonstrate that it outperforms physicians from the MIMIC-III dataset. △ Less

Submitted 5 October, 2022; originally announced October 2022.

Comments: to be published in IAAI (Innovative Applications of Artificial Intelligence) 2023

Mode Mixing and Rotational Splittings: I. Near-Degeneracy Effects Revisited

Authors: J. M. Joel Ong, Lisa Bugnet, Sarbani Basu

Abstract: Rotation is typically assumed to induce strictly symmetric rotational splitting into the rotational multiplets of pure p- and g-modes. However, for evolved stars exhibiting mixed modes, avoided crossings between different multiplet components are known to yield asymmetric rotational splitting, particularly for near-degenerate mixed-mode pairs, where notional pure p-modes are fortuitiously in reson… ▽ More Rotation is typically assumed to induce strictly symmetric rotational splitting into the rotational multiplets of pure p- and g-modes. However, for evolved stars exhibiting mixed modes, avoided crossings between different multiplet components are known to yield asymmetric rotational splitting, particularly for near-degenerate mixed-mode pairs, where notional pure p-modes are fortuitiously in resonance with pure g-modes. These near-degeneracy effects have been described in subgiants, but their consequences for the characterisation of internal rotation in red giants has not previously been investigated in detail, in part owing to theoretical intractability. We employ new developments in the analytic theory of mixed-mode coupling to study these near-resonance phenomena. In the vicinity of the most p-dominated mixed modes, the near-degenerate intrinsic asymmetry from pure rotational splitting increases dramatically over the course of stellar evolution, and depends strongly on the mode mixing fraction $ζ$. We also find that a linear treatment of rotation remains viable for describing the underlying p- and g-modes, even when it does not for the resulting mixed modes undergoing these avoided crossings. We explore observational consequences for potential measurements of asymmetric mixed-mode splitting, which has been proposed as a magnetic-field diagnostic. Finally, we propose improved measurement techniques for rotational characterisation, exploiting the linearity of rotational effects on the underlying p/g modes, while still accounting for these mixed-mode coupling effects. △ Less

Submitted 4 October, 2022; originally announced October 2022.

Comments: 21 pages, 13 figures. Accepted to ApJ

Improved asteroseismic inversions for red-giant surface rotation rates

Authors: F. Ahlborn, E. P. Bellinger, S. Hekker, S. Basu, D. Mokrytska

Abstract: Asteroseismic observations of internal stellar rotation have indicated a substantial lack of angular momentum transport in theoretical models of subgiant and red-giant stars. Accurate core and surface rotation rate measurements are therefore needed to constrain internal transport processes included in the models. We eliminate substantial systematic errors of asteroseismic surface rotation rates fo… ▽ More Asteroseismic observations of internal stellar rotation have indicated a substantial lack of angular momentum transport in theoretical models of subgiant and red-giant stars. Accurate core and surface rotation rate measurements are therefore needed to constrain internal transport processes included in the models. We eliminate substantial systematic errors of asteroseismic surface rotation rates found in previous studies. We propose a new objective function for the Optimally Localized Averages method of rotational inversions for red-giant stars, which results in more accurate envelope rotation rate estimates obtained from the same data. We use synthetic observations from stellar models across a range of evolutionary stages and masses to demonstrate the improvement. We find that our new inversion technique allows us to obtain estimates of the surface rotation rate that are independent of the core rotation. For a star at the base of the red-giant branch, we reduce the systematic error from about 20% to a value close to 0, assuming constant envelope rotation. We also show the equivalence between this method and the method of linearised rotational splittings. Our new rotational inversion method substantially reduces the systematic errors of red-giant surface rotation rates. In combination with independent measures of the surface rotation rate, this will allow better constraints to be set on the internal rotation profile. This will be a very important probe to further constrain the internal angular momentum transport along the lower part of the red-giant branch. △ Less

Submitted 28 September, 2022; originally announced September 2022.

Comments: 17 pages, 22 figures, accepted for publication in Astronomy and Astrophysics

Journal ref: A&A 668, A98 (2022)

Properties of the non-Hermitian SSH model: role of PT-symmetry

Authors: Dipendu Halder, Sudin Ganguly, Saurabh Basu

Abstract: The present work addresses the distinction between the topological properties of PT symmetric and non-PT symmetric scenarios for the non-Hermitian Su-Schrieffer-Heeger (SSH) model. The non-PT symmetric case is represented by non-reciprocity in both the inter- and the intra-cell hopping amplitudes, while the one with PT symmetry is modeled by a complex on-site staggered potential. In particular, we… ▽ More The present work addresses the distinction between the topological properties of PT symmetric and non-PT symmetric scenarios for the non-Hermitian Su-Schrieffer-Heeger (SSH) model. The non-PT symmetric case is represented by non-reciprocity in both the inter- and the intra-cell hopping amplitudes, while the one with PT symmetry is modeled by a complex on-site staggered potential. In particular, we study the loci of the exceptional points, the winding numbers, band structures, and explore the breakdown of bulk-boundary correspondence (BBC). We further study the interplay of the dimerization strengths on the observables for these cases. The non-PT symmetric case denotes a more familiar situation, where the winding number abruptly changes by a half-integer through tuning of the non-reciprocity parameters, and demonstrates a complete breakdown of BBC, thereby showing the non-Hermitian skin effect. The topological nature of the PT symmetric case appears to follow closely to its Hermitian analogue, except that it shows unbroken (broken) regions with complex (purely real) energy spectra, while another variant of the winding number exhibits a continuous behavior as a function of the strength of the potential, while the conventional BBC is preserved. △ Less

Submitted 21 December, 2022; v1 submitted 28 September, 2022; originally announced September 2022.

Comments: 15 pages, 15 figures

Journal ref: Journal of Physics: Condensed Matter, Volume 35, Number 10, Published 29 December 2022, 16 pages

Higher order topology in a Creutz ladder

Authors: Srijata Lahiri, Saurabh Basu

Abstract: A Creutz ladder, is a quasi one dimensional system hosting robust topological phases with localized edge modes protected by different symmetries such as inversion, chiral and particle-hole symmetries. Non-trivial topology is observed in a large region of the parameter space defined by the horizontal, diagonal and vertical hopping ampitudes and a transverse magnetic flux that threads through the la… ▽ More A Creutz ladder, is a quasi one dimensional system hosting robust topological phases with localized edge modes protected by different symmetries such as inversion, chiral and particle-hole symmetries. Non-trivial topology is observed in a large region of the parameter space defined by the horizontal, diagonal and vertical hopping ampitudes and a transverse magnetic flux that threads through the ladder. In this work, we investigate higher order topology in a two dimensional extrapolated version of the Creutz ladder. To explore the topological phases, we consider two different configurations, namely a torus (periodic boundary) and a ribbon (open boundary) to look for hints of gap closing phase transitions. We also associate suitable topological invariants to characterize the non-trivial sectors. Further, we find that the resultant phase diagram hosts two different topological phases, one where the higher order topological excitations are realized in the form of robust corner modes, along with (usual) first order excitations demonstrated via the presence of edge modes in a finite lattice, for the other. △ Less

Submitted 22 September, 2022; originally announced September 2022.

Comments: 9 pages, 16 figures

Unified Balance Functions

Authors: Claude Pruneau, Victor Gonzales, Brian Hanley, Ana Marin, Sumit Basu

Abstract: The use of charge balance functions in heavy-ion collision studies was initially proposed as a probe of delayed hadronization and two-stage quark production in these collisions. It later emerged that general balance functions can also serve as a probe of the diffusivity of light quarks as well as the evolution of the systems formed in heavy-ion collisions. In this work, we reexamine the formulatio… ▽ More The use of charge balance functions in heavy-ion collision studies was initially proposed as a probe of delayed hadronization and two-stage quark production in these collisions. It later emerged that general balance functions can also serve as a probe of the diffusivity of light quarks as well as the evolution of the systems formed in heavy-ion collisions. In this work, we reexamine the formulation of general balance functions and consider how to best define and measure these correlation functions in terms of differences of conditional densities of unlike-sign and like-sign particle pairs. We define general balance functions in terms of associated particle functions and show these obey a simple sum rule. We additionally proceed to distinguish between balance functions expressed as differences of conditional densities valid irrespective of experimental acceptance boundaries and bound balance functions that explicitly account for the limited acceptance of experiments. General balance functions are additionally extended to accommodate strange, baryon, as well as charm and bottom quantum numbers based on the densities of these quantum numbers. △ Less

Submitted 21 September, 2022; originally announced September 2022.

Comments: 19 pages. 3 figures

Critical and Topological Phases of Dimerized Kitaev Chain in Presence of Quasiperiodic Potential

Authors: Shilpi Roy, Sk Noor Nabi, Saurabh Basu

Abstract: We investigate localization and topological properties of a dimerized Kitaev chain with p-wave superconducting correlations and a quasiperiodically modulated chemical potential. With regard to the localization studies, we demonstrate the existence of distinct phases, such as, the extended phase, the critical (intermediate) phase, and the localized phase that arise due to the competition between th… ▽ More We investigate localization and topological properties of a dimerized Kitaev chain with p-wave superconducting correlations and a quasiperiodically modulated chemical potential. With regard to the localization studies, we demonstrate the existence of distinct phases, such as, the extended phase, the critical (intermediate) phase, and the localized phase that arise due to the competition between the dimerization and the onsite quasiperiodic potential. Most interestingly, the critical phase comprises of two different mobility edges that are found to exist between the extended to the localized phase, and between the critical (multifractal) and localized phases. We perform our analysis employing the inverse and the normalized participation ratios, fractal dimension, and the level spacing. Subsequently, a finite-size analysis is done to provide support of our findings. Furthermore, we study the topological properties of the zero-energy edge modes via computing the real-space winding number and number of the Majorana zero modes present in the system. We specifically illustrate that our model exhibits a phase transition from a topologically trivial to a non-trivial phase (topological Anderson phase) beyond a critical dimerization strength under the influence of the quasiperiodic potential strength. Finally, in presence of a large potential, we demonstrate that the system undergoes yet another transition from the topologically non-trivial to an Anderson localized phase. Thus, we believe that our results will aid exploration of fundamentally different physics pertaining to the critical and the topological Anderson phases. △ Less

Submitted 13 September, 2022; originally announced September 2022.

Comments: Comments are welcome

Magnetic Effects Promote Supermassive Star Formation in Metal-enriched Atomic-cooling Halos

Authors: Shingo Hirano, Masahiro N. Machida, Shantanu Basu

Abstract: Intermediate-mass black holes (with $\geq\!10^5\,M_\odot$) are promising candidates for the origin of supermassive black holes (with $\sim\!10^9\,M_\odot$) in the early universe (redshift $z\sim6$). Chon & Omukai (2020) firstly pointed out the direct collapse black hole (DCBH) formation in metal-enriched atomic-cooling halos (ACHs), which relaxes the DCBH formation criterion. On the other hand, Hi… ▽ More Intermediate-mass black holes (with $\geq\!10^5\,M_\odot$) are promising candidates for the origin of supermassive black holes (with $\sim\!10^9\,M_\odot$) in the early universe (redshift $z\sim6$). Chon & Omukai (2020) firstly pointed out the direct collapse black hole (DCBH) formation in metal-enriched atomic-cooling halos (ACHs), which relaxes the DCBH formation criterion. On the other hand, Hirano et al. (2021) showed that the magnetic effects promote the DCBH formation in metal-free ACHs. We perform a set of magnetohydrodynamical simulations to investigate star formation in the magnetized ACHs with metallicities $Z/Z_\odot = 0$, $10^{-5}$, and $10^{-4}$. Our simulations show that the mass accretion rate onto the protostars becomes lower in metal-enriched ACHs than that of metal-free ACHs. However, many protostars form from gravitationally and thermally unstable metal-enriched gas clouds. Under such circumstances, the magnetic field rapidly increases as the magnetic field lines wind up due to the spin of protostars. The region with the amplified magnetic field expands outwards due to the orbital motion of protostars and the rotation of the accreting gas. The amplified magnetic field extracts the angular momentum from the accreting gas, promotes the coalescence of the low-mass protostars, and increases the mass growth rate of the primary protostar. We conclude that the magnetic field amplification is always realized in the metal-enriched ACHs regardless of the initial magnetic field strength, which affects the DCBH formation criterion. In addition, we find a qualitatively different trend from the previous unmagnetized simulations in that the mass growth rate is maximal for the extremely metal-poor ACHs with $Z/Z_\odot = 10^{-5}$. △ Less

Submitted 13 June, 2023; v1 submitted 8 September, 2022; originally announced September 2022.

Comments: 14 pages, 10 figures, 1 table, Accepted for publication in ApJ

Primordial Dusty Rings and Episodic Outbursts in Protoplanetary Discs

Authors: Kundan Kadam, Eduard Vorobyov, Shantanu Basu

Abstract: We investigate the formation and evolution of "primordial" dusty rings occurring in the inner regions of protoplanetary discs, with the help of long-term, coupled dust-gas, magnetohydrodynamic simulations. The simulations are global and start from the collapse phase of the parent cloud core, while the dead zone is calculated via an adaptive $α$ formulation by taking into account the local ionizati… ▽ More We investigate the formation and evolution of "primordial" dusty rings occurring in the inner regions of protoplanetary discs, with the help of long-term, coupled dust-gas, magnetohydrodynamic simulations. The simulations are global and start from the collapse phase of the parent cloud core, while the dead zone is calculated via an adaptive $α$ formulation by taking into account the local ionization balance. The evolution of the dusty component includes its growth and back reaction on to the gas. Previously, using simulations with only a gas component, we showed that dynamical rings form at the inner edge of the dead zone. We find that when dust evolution as well as magnetic field evolution in the flux-freezing limit are included, the dusty rings formed are more numerous and span a larger radial extent in the inner disc, while the dead zone is more robust and persists for a much longer time. We show that these dynamical rings concentrate enough dust mass to become streaming unstable, which should result in rapid planetesimal formation even in the embedded phases of the system. The episodic outbursts caused by the magnetorotational instability have significant impact on the evolution of the rings. The outbursts drain the inner disc of grown dust, however, the period between bursts is sufficiently long for the planetesimal growth via streaming instability.The dust mass contained within the rings is large enough to ultimately produce planetary systems with the core accretion scenario. The low mass systems rarely undergo outbursts and thus, the conditions around such stars can be especially conducive for planet formation. △ Less

Submitted 25 August, 2022; originally announced August 2022.

Comments: 21 pages, 14 figures

Dynamic State Estimation-Based Protection for Induction Motor Loads

Authors: Arthur K. Barnes, Sarbajit Basu, Adam Mate

Abstract: Ensuring protective device coordination is critical to maintain the resilience and improve the reliability of large microgrids. Inverter-interfaced generation, however, poses significant challenges when designing protection systems. Traditional time-overcurrent protective devices are unsuitable on account of the lack of fault current. Present industry practice is to force all inverters to shut dow… ▽ More Ensuring protective device coordination is critical to maintain the resilience and improve the reliability of large microgrids. Inverter-interfaced generation, however, poses significant challenges when designing protection systems. Traditional time-overcurrent protective devices are unsuitable on account of the lack of fault current. Present industry practice is to force all inverters to shut down during faults, which prevents large microgrids from operating in a resilient and reliable manner. Dynamic state estimation (DSE) has been proposed for line protection, and more recently for the protection of load buses or downstream radial portions of microgrids. However, only passive loads with series resistive-inductive loads have been tested with DSE, even though the behavior of dynamic loads -- such as induction motors or power electronics -- may differ significantly during faults. This paper considers the case of applying DSE to protecting a load bus serving a three-phase induction motor. △ Less

Submitted 30 January, 2023; v1 submitted 18 August, 2022; originally announced August 2022.

Comments: 5 pages. 5 figures. 2 tables

Report number: LA-UR-22-27020

Journal ref: Proceedings of the 2022 North American Power Symposium (NAPS)

A note on sets avoiding rational distances in category bases

Authors: Sanjib Basu, Abhit Chandra Pramanik

Abstract: Michalski gave a short and elegant proof of a theorem of A. Kumar which states that for each set A in R, there exists a subset B of A which is full in A and such that no distance between points in B is a rational number. He also proved a similar theorem for sets in the Euclidean plane. In this paper, we generalize these results in some special types of category bases. Michalski gave a short and elegant proof of a theorem of A. Kumar which states that for each set A in R, there exists a subset B of A which is full in A and such that no distance between points in B is a rational number. He also proved a similar theorem for sets in the Euclidean plane. In this paper, we generalize these results in some special types of category bases. △ Less

Submitted 15 August, 2022; originally announced August 2022.

Double Auctions with Two-sided Bandit Feedback

Authors: Soumya Basu, Abishek Sankararaman

Abstract: Double Auction enables decentralized transfer of goods between multiple buyers and sellers, thus underpinning functioning of many online marketplaces. Buyers and sellers compete in these markets through bidding, but do not often know their own valuation a-priori. As the allocation and pricing happens through bids, the profitability of participants, hence sustainability of such markets, depends cru… ▽ More Double Auction enables decentralized transfer of goods between multiple buyers and sellers, thus underpinning functioning of many online marketplaces. Buyers and sellers compete in these markets through bidding, but do not often know their own valuation a-priori. As the allocation and pricing happens through bids, the profitability of participants, hence sustainability of such markets, depends crucially on learning respective valuations through repeated interactions. We initiate the study of Double Auction markets under bandit feedback on both buyers' and sellers' side. We show with confidence bound based bidding, and `Average Pricing' there is an efficient price discovery among the participants. In particular, the regret on combined valuation of the buyers and the sellers -- a.k.a. the social regret -- is $O(\log(T)/Δ)$ in $T$ rounds, where $Δ$ is the minimum price gap. Moreover, the buyers and sellers exchanging goods attain $O(\sqrt{T})$ regret, individually. The buyers and sellers who do not benefit from exchange in turn only experience $O(\log{T}/ Δ)$ regret individually in $T$ rounds. We augment our upper bound by showing that $ω(\sqrt{T})$ individual regret, and $ω(\log{T})$ social regret is unattainable in certain Double Auction markets. Our paper is the first to provide decentralized learning algorithms in a two-sided market where \emph{both sides have uncertain preference} that need to be learned. △ Less

Submitted 10 July, 2023; v1 submitted 12 August, 2022; originally announced August 2022.

$p$-local decompositions of projective Stiefel manifolds

Authors: Samik Basu, Debanil Dasgupta, Shilpa Gondhali, Swagata Sarkar

Abstract: The main objective of this paper is to analyze the $p$-local homotopy type of the complex projective Stiefel manifolds, and other analogous quotients of Stiefel manifolds. We take the cue from a result of Yamaguchi about the $p$-regularity of the complex Stiefel manifolds which lays down some hypotheses under which the Stiefel manifold is $p$-locally a product of odd dimensional spheres. We show t… ▽ More The main objective of this paper is to analyze the $p$-local homotopy type of the complex projective Stiefel manifolds, and other analogous quotients of Stiefel manifolds. We take the cue from a result of Yamaguchi about the $p$-regularity of the complex Stiefel manifolds which lays down some hypotheses under which the Stiefel manifold is $p$-locally a product of odd dimensional spheres. We show that in many cases, the projective Stiefel manifolds are $p$-locally a product of a complex projective space and some odd dimensional spheres. As an application, we prove that in these cases, the $p$-regularity result of Yamaguchi is also $S^1$-equivariant. △ Less

Submitted 12 August, 2022; originally announced August 2022.

Learning Financial Networks with High-frequency Trade Data

Authors: Kara Karpman, Sumanta Basu, David Easley

Abstract: Financial networks are typically estimated by applying standard time series analyses to price-based economic variables collected at low-frequency (e.g., daily or monthly stock returns or realized volatility). These networks are used for risk monitoring and for studying information flows in financial markets. High-frequency intraday trade data sets may provide additional insights into network linka… ▽ More Financial networks are typically estimated by applying standard time series analyses to price-based economic variables collected at low-frequency (e.g., daily or monthly stock returns or realized volatility). These networks are used for risk monitoring and for studying information flows in financial markets. High-frequency intraday trade data sets may provide additional insights into network linkages by leveraging high-resolution information. However, such data sets pose significant modeling challenges due to their asynchronous nature, nonlinear dynamics, and nonstationarity. To tackle these challenges, we estimate financial networks using random forests. The edges in our network are determined by using microstructure measures of one firm to forecast the sign of the change in a market measure (either realized volatility or returns kurtosis) of another firm. We first investigate the evolution of network connectivity in the period leading up to the U.S. financial crisis of 2007-09. We find that the networks have the highest density in 2007, with high degree connectivity associated with Lehman Brothers in 2006. A second analysis into the nature of linkages among firms suggests that larger firms tend to offer better predictive power than smaller firms, a finding qualitatively consistent with prior works in the market microstructure literature. △ Less

Submitted 6 August, 2022; originally announced August 2022.

Solar-like oscillations and ellipsoidal variations in TESS observations of the binary 12 Boötis

Authors: Warrick H. Ball, Andrea Miglio, William J. Chaplin, Keivan G. Stassun, Rafael García, Lucia González-Cuesta, Savita Mathur, Thierry Appourchaux, Othman Benomar, Derek L. Buzasi, Chen Jiang, Cenk Kayhan, Sibel Örtel, Zeynep Çelik Orhan, Mutlu Yıldız, J. M. Joel Ong, Sarbani Basu

Abstract: Binary stars in which oscillations can be studied in either or both components can provide powerful constraints on our understanding of stellar physics. The bright binary 12 Boötis (12 Boo) is a particularly promising system because the primary is roughly 60 per cent brighter than the secondary despite being only a few per cent more massive. Both stars have substantial surface convection zones and… ▽ More Binary stars in which oscillations can be studied in either or both components can provide powerful constraints on our understanding of stellar physics. The bright binary 12 Boötis (12 Boo) is a particularly promising system because the primary is roughly 60 per cent brighter than the secondary despite being only a few per cent more massive. Both stars have substantial surface convection zones and are therefore, presumably, solar-like oscillators. We report here the first detection of solar-like oscillations and ellipsoidal variations in the TESS light curve of 12 Boo. Though the solar-like oscillations are not clear enough to unambiguously measure individual mode frequencies, we combine global asteroseismic parameters and a precise fit to the spectral energy distribution (SED) to provide new constraints on the properties of the system that are several times more precise than values in the literature. The SED fit alone provides new effective temperatures, luminosities and radii of $6115\pm45\,\mathrm{K}$, $7.531\pm0.110\,\mathrm{L}_\odot$ and $2.450\pm0.045\,\mathrm{R}_\odot$ for 12 Boo A and $6200\pm60\,\mathrm{K}$, $4.692\pm0.095\,\mathrm{L}_\odot$ and $1.901\pm0.045\,\mathrm{R}_\odot$ for 12 Boo B. When combined with our asteroseismic constraints on 12 Boo A, we obtain an age of $2.67^{+0.12}_{-0.16}\,\mathrm{Gyr}$, which is consistent with that of 12 Boo B. △ Less

Submitted 3 August, 2022; originally announced August 2022.

Comments: Accepted for publication in MNRAS. 6 pages, 4 figures

Sequents, barcodes, and homology

Authors: Saugata Basu, Negin Karisani, Laxmi Parida

Abstract: We consider the problem of generating hypothesis from data based on ideas from logic. We introduce a notion of barcodes, which we call sequent barcodes, that mirrors the barcodes in persistent homology theory in topological data analysis. We prove a theoretical result on the stability of these barcodes in analogy with similar results in persistent homology theory. Additionally we show that our new… ▽ More We consider the problem of generating hypothesis from data based on ideas from logic. We introduce a notion of barcodes, which we call sequent barcodes, that mirrors the barcodes in persistent homology theory in topological data analysis. We prove a theoretical result on the stability of these barcodes in analogy with similar results in persistent homology theory. Additionally we show that our new notion of barcodes can be interpreted in terms of a persistent homology of a particular filtration of topological spaces induced by the data. Finally, we discuss a concrete application of the sequent barcodes in a discovery problem arising from the area of cancer genomics. △ Less

Submitted 2 August, 2022; originally announced August 2022.

Comments: 26 pages, 10 figures

The index of equidimensional flag manifolds

Authors: Samik Basu, Bikramjit Kundu

Abstract: In this paper, we consider the flag manifold of $p$ orthogonal subspaces of equal dimension which carries an action of the cyclic group of order $p$. We provide a complete calculation of the associated Fadell-Husseini index. This may be thought of as an odd primary version of the computations of Baralić et al [Forum Math., 30 (2018), pp. 1539--1572] for the Grassmann manifold… ▽ More In this paper, we consider the flag manifold of $p$ orthogonal subspaces of equal dimension which carries an action of the cyclic group of order $p$. We provide a complete calculation of the associated Fadell-Husseini index. This may be thought of as an odd primary version of the computations of Baralić et al [Forum Math., 30 (2018), pp. 1539--1572] for the Grassmann manifold $G_n(\mathbb{R}^{2n})$. These results have geometric consequences for $p$-fold orthogonal shadows of a convex body. △ Less

Submitted 28 July, 2022; v1 submitted 27 July, 2022; originally announced July 2022.

Comments: 23 pages

MSC Class: Primary: 55M20; 55M35; Secondary: 52A35; 55N91

Unsupervised Contrastive Learning of Image Representations from Ultrasound Videos with Hard Negative Mining

Authors: Soumen Basu, Somanshu Singla, Mayank Gupta, Pratyaksha Rana, Pankaj Gupta, Chetan Arora

Abstract: Rich temporal information and variations in viewpoints make video data an attractive choice for learning image representations using unsupervised contrastive learning (UCL) techniques. State-of-the-art (SOTA) contrastive learning techniques consider frames within a video as positives in the embedding space, whereas the frames from other videos are considered negatives. We observe that unlike multi… ▽ More Rich temporal information and variations in viewpoints make video data an attractive choice for learning image representations using unsupervised contrastive learning (UCL) techniques. State-of-the-art (SOTA) contrastive learning techniques consider frames within a video as positives in the embedding space, whereas the frames from other videos are considered negatives. We observe that unlike multiple views of an object in natural scene videos, an Ultrasound (US) video captures different 2D slices of an organ. Hence, there is almost no similarity between the temporally distant frames of even the same US video. In this paper we propose to instead utilize such frames as hard negatives. We advocate mining both intra-video and cross-video negatives in a hardness-sensitive negative mining curriculum in a UCL framework to learn rich image representations. We deploy our framework to learn the representations of Gallbladder (GB) malignancy from US videos. We also construct the first large-scale US video dataset containing 64 videos and 15,800 frames for learning GB representations. We show that the standard ResNet50 backbone trained with our framework improves the accuracy of models pretrained with SOTA UCL techniques as well as supervised pretrained models on ImageNet for the GB malignancy detection task by 2-6%. We further validate the generalizability of our method on a publicly available lung US image dataset of COVID-19 pathologies and show an improvement of 1.5% compared to SOTA. Source code, dataset, and models are available at https://gbc-iitd.github.io/usucl. △ Less

Submitted 26 July, 2022; originally announced July 2022.

Comments: ACCEPTED for publication at MICCAI 2022

Exploring Financial Networks Using Quantile Regression and Granger Causality

Authors: Kara Karpman, Samriddha Lahiry, Diganta Mukherjee, Sumanta Basu

Abstract: In the post-crisis era, financial regulators and policymakers are increasingly interested in data-driven tools to measure systemic risk and to identify systemically important firms. Granger Causality (GC) based techniques to build networks among financial firms using time series of their stock returns have received significant attention in recent years. Existing GC network methods model conditiona… ▽ More In the post-crisis era, financial regulators and policymakers are increasingly interested in data-driven tools to measure systemic risk and to identify systemically important firms. Granger Causality (GC) based techniques to build networks among financial firms using time series of their stock returns have received significant attention in recent years. Existing GC network methods model conditional means, and do not distinguish between connectivity in lower and upper tails of the return distribution - an aspect crucial for systemic risk analysis. We propose statistical methods that measure connectivity in the financial sector using system-wide tail-based analysis and is able to distinguish between connectivity in lower and upper tails of the return distribution. This is achieved using bivariate and multivariate GC analysis based on regular and Lasso penalized quantile regressions, an approach we call quantile Granger causality (QGC). By considering centrality measures of these financial networks, we can assess the build-up of systemic risk and identify risk propagation channels. We provide an asymptotic theory of QGC estimators under a quantile vector autoregressive model, and show its benefit over regular GC analysis on simulated data. We apply our method to the monthly stock returns of large U.S. firms and demonstrate that lower tail based networks can detect systemically risky periods in historical data with higher accuracy than mean-based networks. In a similar analysis of large Indian banks, we find that upper and lower tail networks convey different information and have the potential to distinguish between periods of high connectivity that are governed by positive vs negative news in the market. △ Less

Submitted 26 July, 2022; v1 submitted 21 July, 2022; originally announced July 2022.

Computing the homology functor on semi-algebraic maps and diagrams

Authors: Saugata Basu, Negin Karisani

Abstract: Developing an algorithm for computing the Betti numbers of semi-algebraic sets with singly exponential complexity has been a holy grail in algorithmic semi-algebraic geometry and only partial results are known. In this paper we consider the more general problem of computing the image under the homology functor of a semi-algebraic map $f:X \rightarrow Y$ between closed and bounded semi-algebraic se… ▽ More Developing an algorithm for computing the Betti numbers of semi-algebraic sets with singly exponential complexity has been a holy grail in algorithmic semi-algebraic geometry and only partial results are known. In this paper we consider the more general problem of computing the image under the homology functor of a semi-algebraic map $f:X \rightarrow Y$ between closed and bounded semi-algebraic sets. For every fixed $\ell \geq 0$ we give an algorithm with singly exponential complexity that computes bases of the homology groups $\mathrm{H}_i(X), \mathrm{H}_i(Y)$ (with rational coefficients) and a matrix with respect to these bases of the induced linear maps $\mathrm{H}_i(f):\mathrm{H}_i(X) \rightarrow \mathrm{H}_i(Y), 0 \leq i \leq \ell$. We generalize this algorithm to more general (zigzag) diagrams of maps between closed and bounded semi-algebraic sets and give a singly exponential algorithm for computing the homology functors on such diagrams. This allows us to give an algorithm with singly exponential complexity for computing barcodes of semi-algebraic zigzag persistent homology in small dimensions. △ Less

Submitted 21 July, 2022; originally announced July 2022.

Comments: 28 pages, 2 figures. Comments most welcome. arXiv admin note: text overlap with arXiv:2009.13365

MSC Class: Primary 14F25; 55N31; Secondary 68W30

Respiration driven CO2 pulses dominate Australia's flux variability

Authors: Eva-Marie Metz, Sanam N. Vardag, Sourish Basu, Martin Jung, Bernhard Ahrens, Tarek El-Madany, Stephen Sitch, Vivek K. Arora, Peter R. Briggs, Pierre Friedlingstein, Daniel S. Goll, Atul K. Jain, Etsushi Kato, Danica Lombardozzi, Julia E. M. S. Nabel, Benjamin Poulter, Roland Séférian, Hanqin Tian, Andrew Wiltshire, Wenping Yuan, Xu Yue, Sönke Zaehle, Nicholas M. Deutscher, David W. T. Griffith, André Butz

Abstract: The Australian continent contributes substantially to the year-to-year variability of the global terrestrial carbon dioxide (CO2) sink. However, the scarcity of in-situ observations in remote areas prevents deciphering the processes that force the CO2 flux variability. Here, examining atmospheric CO2 measurements from satellites in the period 2009-2018, we find recurrent end-of-dry-season CO2 puls… ▽ More The Australian continent contributes substantially to the year-to-year variability of the global terrestrial carbon dioxide (CO2) sink. However, the scarcity of in-situ observations in remote areas prevents deciphering the processes that force the CO2 flux variability. Here, examining atmospheric CO2 measurements from satellites in the period 2009-2018, we find recurrent end-of-dry-season CO2 pulses over the Australian continent. These pulses largely control the year-to-year variability of Australia's CO2 balance, due to 2-3 times higher seasonal variations compared to previous top-down inversions and bottom-up estimates. The CO2 pulses occur shortly after the onset of rainfall and are driven by enhanced soil respiration preceding photosynthetic uptake in Australia's semi-arid regions. The suggested continental-scale relevance of soil rewetting processes has large implications for our understanding and modelling of global climate-carbon cycle feedbacks. △ Less

Submitted 30 November, 2022; v1 submitted 14 July, 2022; originally announced July 2022.

Comments: 28 pages (including supplementary materials), 3 main figures, 7 supplementary figures; v2 changes: Last name of first author changed

Segmentation of Blood Vessels, Optic Disc Localization, Detection of Exudates and Diabetic Retinopathy Diagnosis from Digital Fundus Images

Authors: Soham Basu, Sayantan Mukherjee, Ankit Bhattacharya, Anindya Sen

Abstract: Diabetic Retinopathy (DR) is a complication of long-standing, unchecked diabetes and one of the leading causes of blindness in the world. This paper focuses on improved and robust methods to extract some of the features of DR, viz. Blood Vessels and Exudates. Blood vessels are segmented using multiple morphological and thresholding operations. For the segmentation of exudates, k-means clustering a… ▽ More Diabetic Retinopathy (DR) is a complication of long-standing, unchecked diabetes and one of the leading causes of blindness in the world. This paper focuses on improved and robust methods to extract some of the features of DR, viz. Blood Vessels and Exudates. Blood vessels are segmented using multiple morphological and thresholding operations. For the segmentation of exudates, k-means clustering and contour detection on the original images are used. Extensive noise reduction is performed to remove false positives from the vessel segmentation algorithm's results. The localization of Optic Disc using k-means clustering and template matching is also performed. Lastly, this paper presents a Deep Convolutional Neural Network (DCNN) model with 14 Convolutional Layers and 2 Fully Connected Layers, for the automatic, binary diagnosis of DR. The vessel segmentation, optic disc localization and DCNN achieve accuracies of 95.93%, 98.77% and 75.73% respectively. The source code and pre-trained model are available https://github.com/Sohambasu07/DR_2021 △ Less

Submitted 9 July, 2022; originally announced July 2022.

Comments: RAAI 2020, 11 pages, 12 figures, 2 tables

ACM Class: I.4.6; I.5.1; I.5.4

Journal ref: Proceedings of Research and Applications in Artificial Intelligence. Advances in Intelligent Systems and Computing, vol. 1355 (2021) pp.173-184. Springer, Singapore

Synthetic Polarization Maps of an Outflow Zone from Magnetohydrodynamic Simulations

Authors: Gianfranco Bino, Shantanu Basu, Masahiro N Machida, Aris Tritsis, Mahmoud Sharkawi, Kundan Kadam, Indrani Das

Abstract: The canonical theory of star formation in a magnetized environment predicts the formation of hourglass-shaped magnetic fields during the prestellar collapse phase. In protostellar cores, recent observations reveal complex and strongly distorted magnetic fields in the inner regions that are sculpted by rotation and outflows. We conduct resistive, nonideal magnetohydrodynamic (MHD) simulations of a… ▽ More The canonical theory of star formation in a magnetized environment predicts the formation of hourglass-shaped magnetic fields during the prestellar collapse phase. In protostellar cores, recent observations reveal complex and strongly distorted magnetic fields in the inner regions that are sculpted by rotation and outflows. We conduct resistive, nonideal magnetohydrodynamic (MHD) simulations of a protostellar core and employ the radiative transfer code POLARIS to produce synthetic polarization segment maps. Comparison of our mock-polarization maps based on the toroidal-dominated magnetic field in the outflow zone with the observed polarization vectors of SiO lines in Orion Source I shows a reasonable agreement when the magnetic axis is tilted at an angle $θ= 15^{\circ}$ with respect to the plane-of-sky and if the SiO lines have a net polarization parallel to the local magnetic field. Although the observed polarization is from SiO lines and our synthetic maps are due to polarized dust emission, a comparison is useful and allows us to resolve the ambiguity of whether the line polarization is parallel or perpendicular to the local magnetic field direction. △ Less

Submitted 22 August, 2022; v1 submitted 4 July, 2022; originally announced July 2022.

Journal ref: The Astrophysical Journal (2022)

Generalization of two theorems of Steinhaus in category bases

Authors: Sanjib Basu, Abhit Chandra Pramanik

Abstract: Here we unify two results of Steinhaus and their corresponding category analogues by extending them in the settings of category bases. We further show that in any perfect translation base, every abundant Baire set contains a full subset for which our second theorem fails. Here we unify two results of Steinhaus and their corresponding category analogues by extending them in the settings of category bases. We further show that in any perfect translation base, every abundant Baire set contains a full subset for which our second theorem fails. △ Less

Submitted 29 June, 2022; originally announced June 2022.

Modeling Multivariate Positive-Valued Time Series Using R-INLA

Authors: Chiranjit Dutta, Nalini Ravishanker, Sumanta Basu

Abstract: In this paper we describe fast Bayesian statistical analysis of vector positive-valued time series, with application to interesting financial data streams. We discuss a flexible level correlated model (LCM) framework for building hierarchical models for vector positive-valued time series. The LCM allows us to combine marginal gamma distributions for the positive-valued component responses, while a… ▽ More In this paper we describe fast Bayesian statistical analysis of vector positive-valued time series, with application to interesting financial data streams. We discuss a flexible level correlated model (LCM) framework for building hierarchical models for vector positive-valued time series. The LCM allows us to combine marginal gamma distributions for the positive-valued component responses, while accounting for association among the components at a latent level. We use integrated nested Laplace approximation (INLA) for fast approximate Bayesian modeling via the R-INLA package, building custom functions to handle this setup. We use the proposed method to model interdependencies between realized volatility measures from several stock indexes. △ Less

Submitted 2 July, 2022; v1 submitted 10 June, 2022; originally announced June 2022.

Comments: 19 pages, 1 figure

HMI Ring Diagram Analysis: Effects of tracking, noise and resolution

Authors: Sarbani Basu, Richard S. Bogart

Abstract: Ring diagram analysis is a standard local helioseismic technique. Data from the Helioseismic and Magnetic Imager (HMI) on the Solar Dynamics Observatory (SDO) are routinely used for Ring-diagram analysis, and fits to the power spectra as well as inversion results are standard data products. In this paper we examine the effects of different tracking rates, noise, and resolution on ring-diagram resu… ▽ More Ring diagram analysis is a standard local helioseismic technique. Data from the Helioseismic and Magnetic Imager (HMI) on the Solar Dynamics Observatory (SDO) are routinely used for Ring-diagram analysis, and fits to the power spectra as well as inversion results are standard data products. In this paper we examine the effects of different tracking rates, noise, and resolution on ring-diagram results. Most of the analysis is for $15^\circ$ tiles, but we also examine the effects of different tile sizes. The largest effect we find is that of resolution. Doppler noise has very little effect on the results, except perhaps at the deepest regions for which the tiles can give reliable results; variations in the tracking rate have a similar effect. △ Less

Submitted 1 June, 2022; originally announced June 2022.

Comments: Accepted for publication in Solar Physics

Shock induced aerobreakup of a polymeric droplet

Authors: Navin Kumar Chandra, Shubham Sharma, Saptarshi Basu, Aloke Kumar

Abstract: Droplet atomization through aerobreakup is omnipresent in various natural and industrial processes. Atomization of Newtonian droplets is a well-studied area; however, non-Newtonian droplets have received less attention despite their frequent encounters. By subjecting polymeric droplets of different concentrations to the induced airflow behind a moving shock wave, we explore the role of elasticity… ▽ More Droplet atomization through aerobreakup is omnipresent in various natural and industrial processes. Atomization of Newtonian droplets is a well-studied area; however, non-Newtonian droplets have received less attention despite their frequent encounters. By subjecting polymeric droplets of different concentrations to the induced airflow behind a moving shock wave, we explore the role of elasticity in modulating the aerobreakup of viscoelastic droplets. Three distinct modes of aerobreakup are identified for a wide range of Weber number ($\sim 10^2-10^4$) and Elasticity number ($\sim 10^{-4}-10^2$) variation; these modes are: vibrational, shear-induced entrainment and catastrophic breakup mode. Each mode is described as a three stage process. Stage-I is the droplet deformation, stage-II is the appearance and growth of hydrodynamic instabilities, and stage-III is the evolution of liquid mass morphology. It is observed that elasticity plays an insignificant role in the first two stages, but a dominant role in the final stage. The results are described with the support of adequate mathematical analysis. △ Less

Submitted 31 May, 2022; originally announced May 2022.

Comments: 32 pages, 10 figures

Equivariant Mesh Attention Networks

Authors: Sourya Basu, Jose Gallego-Posada, Francesco Viganò, James Rowbottom, Taco Cohen

Abstract: Equivariance to symmetries has proven to be a powerful inductive bias in deep learning research. Recent works on mesh processing have concentrated on various kinds of natural symmetries, including translations, rotations, scaling, node permutations, and gauge transformations. To date, no existing architecture is equivariant to all of these transformations. In this paper, we present an attention-ba… ▽ More Equivariance to symmetries has proven to be a powerful inductive bias in deep learning research. Recent works on mesh processing have concentrated on various kinds of natural symmetries, including translations, rotations, scaling, node permutations, and gauge transformations. To date, no existing architecture is equivariant to all of these transformations. In this paper, we present an attention-based architecture for mesh data that is provably equivariant to all transformations mentioned above. Our pipeline relies on the use of relative tangential features: a simple, effective, equivariance-friendly alternative to raw node positions as inputs. Experiments on the FAUST and TOSCA datasets confirm that our proposed architecture achieves improved performance on these benchmarks and is indeed equivariant, and therefore robust, to a wide variety of local/global transformations. △ Less

Submitted 27 August, 2022; v1 submitted 21 May, 2022; originally announced May 2022.

Comments: Published in Transactions on Machine Learning Research (08/2022). Official code made available at https://github.com/gallego-posada/eman - For the OpenReview entry, see https://openreview.net/forum?id=3IqqJh2Ycy

On the mechanics of droplet surface crater during impact on immiscible viscous liquid pool

Authors: Durbar Roy, Sophia M, Saptarshi Basu

Abstract: We study drop impacts on immiscible viscous liquid pool and investigate the formation of droplet surface craters using experimental and theoretical analysis. We attribute the formation of air craters to the rapid deceleration of the droplet due to viscous drag force. The droplet response to the external impulsive decelerating force induces oscillatory modes on the surface exposed to the air formin… ▽ More We study drop impacts on immiscible viscous liquid pool and investigate the formation of droplet surface craters using experimental and theoretical analysis. We attribute the formation of air craters to the rapid deceleration of the droplet due to viscous drag force. The droplet response to the external impulsive decelerating force induces oscillatory modes on the surface exposed to the air forming capillary waves that superimpose to form air craters of various shapes and sizes. We introduce a non-dimensional parameter ($Γ$), that is, the ratio of drag force to the capillary force acting on the droplet. We show that $Γ$ is directly proportional to the capillary number. We show that droplets forming air craters of significant depths have $Γ>1$. Further, we demonstrate that Legendre polynomials can locally approximate the central air crater jet profile. We also decipher that the air crater response time scale ($T$) varies as the square root of impact Weber number ($T{\sim}We^{1/2}$). Further, we generalize the local droplet response with a global response model for low-impact energies based on an eigenvalue problem. We represent the penetrating drop as a constrained Rayleigh drop problem with a dynamic contact line. The air-water interface dynamics is analyzed using an inviscid droplet deformation model for small deformation amplitudes. The local and global droplet response theory conforms with each other and depicts that the deformation profiles could be represented as a linear superposition of eigenmodes in Legendre polynomial basis. We unearth that the droplet response in an immiscible impact system differs from the miscible impact systems due to the presence of such a dynamic contact line. △ Less

Submitted 29 September, 2022; v1 submitted 11 May, 2022; originally announced May 2022.

A novel approach for deriving the stable boundary layer height and eddy viscosity profiles from the Ekman equations

Authors: Sukanta Basu, Albert A. M. Holtslag

Abstract: In this study, we utilize a novel approach to solve the Ekman equations for eddy viscosity profiles in the stable boundary layer. By doing so, a well-known expression for the stable boundary layer height by Zilitinkevich (Boundary-Layer Meteorology, 1972, Vol. 3, 141--145) is rediscovered. In this study, we utilize a novel approach to solve the Ekman equations for eddy viscosity profiles in the stable boundary layer. By doing so, a well-known expression for the stable boundary layer height by Zilitinkevich (Boundary-Layer Meteorology, 1972, Vol. 3, 141--145) is rediscovered. △ Less

Submitted 20 July, 2022; v1 submitted 2 May, 2022; originally announced May 2022.