-
Minimising changes to audit when updating decision trees
Authors:
Anj Simmons,
Scott Barnett,
Anupam Chaudhuri,
Sankhya Singh,
Shangeetha Sivasothy
Abstract:
Interpretable models are important, but what happens when the model is updated on new training data? We propose an algorithm for updating a decision tree while minimising the number of changes to the tree that a human would need to audit. We achieve this via a greedy approach that incorporates the number of changes to the tree as part of the objective function. We compare our algorithm to existing…
▽ More
Interpretable models are important, but what happens when the model is updated on new training data? We propose an algorithm for updating a decision tree while minimising the number of changes to the tree that a human would need to audit. We achieve this via a greedy approach that incorporates the number of changes to the tree as part of the objective function. We compare our algorithm to existing methods and show that it sits in a sweet spot between final accuracy and number of changes to audit.
△ Less
Submitted 29 August, 2024;
originally announced August 2024.
-
SPICED: Syntactical Bug and Trojan Pattern Identification in A/MS Circuits using LLM-Enhanced Detection
Authors:
Jayeeta Chaudhuri,
Dhruv Thapar,
Arjun Chaudhuri,
Farshad Firouzi,
Krishnendu Chakrabarty
Abstract:
Analog and mixed-signal (A/MS) integrated circuits (ICs) are crucial in modern electronics, playing key roles in signal processing, amplification, sensing, and power management. Many IC companies outsource manufacturing to third-party foundries, creating security risks such as stealthy analog Trojans. Traditional detection methods, including embedding circuit watermarks or conducting hardware-base…
▽ More
Analog and mixed-signal (A/MS) integrated circuits (ICs) are crucial in modern electronics, playing key roles in signal processing, amplification, sensing, and power management. Many IC companies outsource manufacturing to third-party foundries, creating security risks such as stealthy analog Trojans. Traditional detection methods, including embedding circuit watermarks or conducting hardware-based monitoring, often impose significant area and power overheads, and may not effectively identify all types of Trojans. To address these shortcomings, we propose SPICED, a Large Language Model (LLM)-based framework that operates within the software domain, eliminating the need for hardware modifications for Trojan detection and localization. This is the first work using LLM-aided techniques for detecting and localizing syntactical bugs and analog Trojans in circuit netlists, requiring no explicit training and incurring zero area overhead. Our framework employs chain-of-thought reasoning and few-shot examples to teach anomaly detection rules to LLMs. With the proposed method, we achieve an average Trojan coverage of 93.32% and an average true positive rate of 93.4% in identifying Trojan-impacted nodes for the evaluated analog benchmark circuits. These experimental results validate the effectiveness of LLMs in detecting and locating both syntactical bugs and Trojans within analog netlists.
△ Less
Submitted 25 August, 2024;
originally announced August 2024.
-
Bayesian learning with Gaussian processes for low-dimensional representations of time-dependent nonlinear systems
Authors:
Shane A. McQuarrie,
Anirban Chaudhuri,
Karen E. Willcox,
Mengwu Guo
Abstract:
This work presents a data-driven method for learning low-dimensional time-dependent physics-based surrogate models whose predictions are endowed with uncertainty estimates. We use the operator inference approach to model reduction that poses the problem of learning low-dimensional model terms as a regression of state space data and corresponding time derivatives by minimizing the residual of reduc…
▽ More
This work presents a data-driven method for learning low-dimensional time-dependent physics-based surrogate models whose predictions are endowed with uncertainty estimates. We use the operator inference approach to model reduction that poses the problem of learning low-dimensional model terms as a regression of state space data and corresponding time derivatives by minimizing the residual of reduced system equations. Standard operator inference models perform well with accurate training data that are dense in time, but producing stable and accurate models when the state data are noisy and/or sparse in time remains a challenge. Another challenge is the lack of uncertainty estimation for the predictions from the operator inference models. Our approach addresses these challenges by incorporating Gaussian process surrogates into the operator inference framework to (1) probabilistically describe uncertainties in the state predictions and (2) procure analytical time derivative estimates with quantified uncertainties. The formulation leads to a generalized least-squares regression and, ultimately, reduced-order models that are described probabilistically with a closed-form expression for the posterior distribution of the operators. The resulting probabilistic surrogate model propagates uncertainties from the observed state data to reduced-order predictions. We demonstrate the method is effective for constructing low-dimensional models of two nonlinear partial differential equations representing a compressible flow and a nonlinear diffusion-reaction process, as well as for estimating the parameters of a low-dimensional system of nonlinear ordinary differential equations representing compartmental models in epidemiology.
△ Less
Submitted 6 August, 2024;
originally announced August 2024.
-
Packing and ejection dynamics of polymers: Role of confinement, polymer stiffness and activity
Authors:
Gokul Upadhyay,
Rajeev Kapri,
Anil Kumar Dasanna,
Abhishek Chaudhuri
Abstract:
The translocation of biopolymers, such as DNA and proteins, across cellular or nuclear membranes is essential for numerous biological processes. The translocation dynamics are influenced by the properties of the polymers, such as polymer stiffness, and the geometry of the capsid. In our study, we aim to investigate the impact of polymer stiffness, activity, and different capsid geometries on the p…
▽ More
The translocation of biopolymers, such as DNA and proteins, across cellular or nuclear membranes is essential for numerous biological processes. The translocation dynamics are influenced by the properties of the polymers, such as polymer stiffness, and the geometry of the capsid. In our study, we aim to investigate the impact of polymer stiffness, activity, and different capsid geometries on the packing and ejection dynamics of both passive and active polymers. We employ Langevin dynamics simulations for a systematic investigation. We observe that flexible polymers exhibit packing times that are faster than those of their semi-flexible counterparts. Interestingly, for large polymers compared to the capsid size, sphere facilitates faster packing and unpacking compared to ellipsoid, mimicking the cell nucleus and suggesting a geometrical advantage for biopolymer translocation. In summary, we observe that increasing activity accelerates both the packing and ejection processes for both flexible and semi-flexible polymers. However, the effect is significantly more pronounced for semi-flexible polymers, highlighting the crucial role of polymer flexibility in these dynamics. These findings deepen our understanding of the intricate interplay between polymer flexibility, capsid geometry, and activity, providing valuable insight into the dynamics of polymer packing and ejection processes.
△ Less
Submitted 26 July, 2024;
originally announced July 2024.
-
On the Design of Ethereum Data Availability Sampling: A Comprehensive Simulation Study
Authors:
Arunima Chaudhuri,
Sudipta Basak,
Csaba Kiraly,
Dmitriy Ryajov,
Leonardo Bautista-Gomez
Abstract:
This paper presents an in-depth exploration of Data Availability Sampling (DAS) and sharding mechanisms within decentralized systems through simulation-based analysis. DAS, a pivotal concept in blockchain technology and decentralized networks, is thoroughly examined to unravel its intricacies and assess its impact on system performance. Through the development of a simulator tailored explicitly fo…
▽ More
This paper presents an in-depth exploration of Data Availability Sampling (DAS) and sharding mechanisms within decentralized systems through simulation-based analysis. DAS, a pivotal concept in blockchain technology and decentralized networks, is thoroughly examined to unravel its intricacies and assess its impact on system performance. Through the development of a simulator tailored explicitly for DAS, we embark on a comprehensive investigation into the parameters that influence system behavior and efficiency. A series of experiments are conducted within the simulated environment to validate theoretical formulations and dissect the interplay of DAS parameters. This includes an exploration of approaches such as custody by row, variations in validators per node, and malicious nodes. The outcomes of these experiments furnish insights into the efficacy of DAS protocols and pave the way for the formulation of optimization strategies geared towards enhancing decentralized network performance. Moreover, the findings serve as guidelines for future research endeavors, offering a nuanced understanding of the complexities inherent in decentralized systems. This study not only contributes to the theoretical understanding of DAS but also offers practical implications for the design, implementation, and optimization of decentralized systems.
△ Less
Submitted 25 July, 2024;
originally announced July 2024.
-
Inertia and Activity: Spiral transitions in semi-flexible, self-avoiding polymers
Authors:
Chitrak Karan,
Abhishek Chaudhuri,
Debasish Chaudhuri
Abstract:
We consider a two-dimensional, tangentially active, semi-flexible, self-avoiding polymer to find a dynamical re-entrant transition between motile open chains and spinning achiral spirals with increasing activity. Utilizing probability distributions of the turning number, we ascertain the comparative stability of the spiral structure and present a detailed phase diagram within the activity inertia…
▽ More
We consider a two-dimensional, tangentially active, semi-flexible, self-avoiding polymer to find a dynamical re-entrant transition between motile open chains and spinning achiral spirals with increasing activity. Utilizing probability distributions of the turning number, we ascertain the comparative stability of the spiral structure and present a detailed phase diagram within the activity inertia plane. The onset of spiral formation at low activity levels is governed by a torque balance and is independent of inertia. At higher activities, however, inertial effects lead to spiral destabilization, an effect absent in the overdamped limit. We further delineate alterations in size and shape by analyzing the end-to-end distance distribution and the radius of gyration tensor. The Kullback-Leibler divergence from equilibrium distributions exhibits a non-monotonic relationship with activity, reaching a peak at the most compact spirals characterized by the most persistent spinning. As inertia increases, this divergence from equilibrium diminishes.
△ Less
Submitted 24 April, 2024;
originally announced April 2024.
-
The N2HDM, Entropy Production and Stochastic Gravitational Waves
Authors:
Arnab Chaudhuri,
Kazunori Kohri
Abstract:
This study undertakes a reconsideration of the potential for a first-order electroweak phase transition, focusing on the next-to-minimal two Higgs doublet model (N2HDM). Our exploration spans diverse parameter spaces associated with the phase transition, with a particular emphasis on examining the generation of stochastic Gravitational Waves (GW) resulting from this transition. The obtained result…
▽ More
This study undertakes a reconsideration of the potential for a first-order electroweak phase transition, focusing on the next-to-minimal two Higgs doublet model (N2HDM). Our exploration spans diverse parameter spaces associated with the phase transition, with a particular emphasis on examining the generation of stochastic Gravitational Waves (GW) resulting from this transition. The obtained results are meticulously compared against data from prominent gravitational wave observatories, and the possibility of their detection in the future GW observations have been established. In passing by we analyse the strength of the phase transition through the production of entropy during the electroweak phase transition.
△ Less
Submitted 16 April, 2024;
originally announced April 2024.
-
Impact of chirality on active Brownian particle: Exact moments in two and three dimensions
Authors:
Anweshika Pattanayak,
Amir Shee,
Debasish Chaudhuri,
Abhishek Chaudhuri
Abstract:
In this work, we investigate the effects of chirality, accounting for translational diffusion, on active Brownian particles in two and three dimensions. Despite the inherent complexity in solving the Fokker-Planck equation, we demonstrate a Laplace transform method for precisely calculating the temporal evolution of various dynamic moments. Our analysis yields explicit expressions for multiple mom…
▽ More
In this work, we investigate the effects of chirality, accounting for translational diffusion, on active Brownian particles in two and three dimensions. Despite the inherent complexity in solving the Fokker-Planck equation, we demonstrate a Laplace transform method for precisely calculating the temporal evolution of various dynamic moments. Our analysis yields explicit expressions for multiple moments, such as the second and fourth moments of displacement, revealing the impact of persistence and chirality. These moments exhibit oscillatory behaviour, and excess kurtosis indicates deviations from the Gaussian distribution during intermediate time intervals.
△ Less
Submitted 15 April, 2024;
originally announced April 2024.
-
Round Robin Active Sequential Change Detection for Dependent Multi-Channel Data
Authors:
Anamitra Chaudhuri,
Georgios Fellouris,
Ali Tajer
Abstract:
This paper considers the problem of sequentially detecting a change in the joint distribution of multiple data sources under a sampling constraint. Specifically, the channels or sources generate observations that are independent over time, but not necessarily independent at any given time instant. The sources follow an initial joint distribution, and at an unknown time instant, the joint distribut…
▽ More
This paper considers the problem of sequentially detecting a change in the joint distribution of multiple data sources under a sampling constraint. Specifically, the channels or sources generate observations that are independent over time, but not necessarily independent at any given time instant. The sources follow an initial joint distribution, and at an unknown time instant, the joint distribution of an unknown subset of sources changes. Importantly, there is a hard constraint that only a fixed number of sources are allowed to be sampled at each time instant. The goal is to sequentially observe the sources according to the constraint, and stop sampling as quickly as possible after the change while controlling the false alarm rate below a user-specified level. The sources can be selected dynamically based on the already collected data, and thus, a policy for this problem consists of a joint sampling and change-detection rule. A non-randomized policy is studied, and an upper bound is established on its worst-case conditional expected detection delay with respect to both the change point and the observations from the affected sources before the change.
△ Less
Submitted 24 March, 2024;
originally announced March 2024.
-
Multifidelity linear regression for scientific machine learning from scarce data
Authors:
Elizabeth Qian,
Dayoung Kang,
Vignesh Sella,
Anirban Chaudhuri
Abstract:
Machine learning (ML) methods, which fit to data the parameters of a given parameterized model class, have garnered significant interest as potential methods for learning surrogate models for complex engineering systems for which traditional simulation is expensive. However, in many scientific and engineering settings, generating high-fidelity data on which to train ML models is expensive, and the…
▽ More
Machine learning (ML) methods, which fit to data the parameters of a given parameterized model class, have garnered significant interest as potential methods for learning surrogate models for complex engineering systems for which traditional simulation is expensive. However, in many scientific and engineering settings, generating high-fidelity data on which to train ML models is expensive, and the available budget for generating training data is limited, so that high-fidelity training data are scarce. ML models trained on scarce data have high variance, resulting in poor expected generalization performance. We propose a new multifidelity training approach for scientific machine learning via linear regression that exploits the scientific context where data of varying fidelities and costs are available: for example, high-fidelity data may be generated by an expensive fully resolved physics simulation whereas lower-fidelity data may arise from a cheaper model based on simplifying assumptions. We use the multifidelity data within an approximate control variate framework to define new multifidelity Monte Carlo estimators for linear regression models. We provide bias and variance analysis of our new estimators that guarantee the approach's accuracy and improved robustness to scarce high-fidelity data. Numerical results demonstrate that our multifidelity training approach achieves similar accuracy to the standard high-fidelity only approach with orders-of-magnitude reduced high-fidelity data requirements.
△ Less
Submitted 1 July, 2024; v1 submitted 13 March, 2024;
originally announced March 2024.
-
Quantifying Manifolds: Do the manifolds learned by Generative Adversarial Networks converge to the real data manifold
Authors:
Anupam Chaudhuri,
Anj Simmons,
Mohamed Abdelrazek
Abstract:
This paper presents our experiments to quantify the manifolds learned by ML models (in our experiment, we use a GAN model) as they train. We compare the manifolds learned at each epoch to the real manifolds representing the real data. To quantify a manifold, we study the intrinsic dimensions and topological features of the manifold learned by the ML model, how these metrics change as we continue t…
▽ More
This paper presents our experiments to quantify the manifolds learned by ML models (in our experiment, we use a GAN model) as they train. We compare the manifolds learned at each epoch to the real manifolds representing the real data. To quantify a manifold, we study the intrinsic dimensions and topological features of the manifold learned by the ML model, how these metrics change as we continue to train the model, and whether these metrics convergence over the course of training to the metrics of the real data manifold.
△ Less
Submitted 7 March, 2024;
originally announced March 2024.
-
Learning Conditional Invariances through Non-Commutativity
Authors:
Abhra Chaudhuri,
Serban Georgescu,
Anjan Dutta
Abstract:
Invariance learning algorithms that conditionally filter out domain-specific random variables as distractors, do so based only on the data semantics, and not the target domain under evaluation. We show that a provably optimal and sample-efficient way of learning conditional invariances is by relaxing the invariance criterion to be non-commutatively directed towards the target domain. Under domain…
▽ More
Invariance learning algorithms that conditionally filter out domain-specific random variables as distractors, do so based only on the data semantics, and not the target domain under evaluation. We show that a provably optimal and sample-efficient way of learning conditional invariances is by relaxing the invariance criterion to be non-commutatively directed towards the target domain. Under domain asymmetry, i.e., when the target domain contains semantically relevant information absent in the source, the risk of the encoder $\varphi^*$ that is optimal on average across domains is strictly lower-bounded by the risk of the target-specific optimal encoder $Φ^*_τ$. We prove that non-commutativity steers the optimization towards $Φ^*_τ$ instead of $\varphi^*$, bringing the $\mathcal{H}$-divergence between domains down to zero, leading to a stricter bound on the target risk. Both our theory and experiments demonstrate that non-commutative invariance (NCI) can leverage source domain samples to meet the sample complexity needs of learning $Φ^*_τ$, surpassing SOTA invariance learning algorithms for domain adaptation, at times by over $2\%$, approaching the performance of an oracle. Implementation is available at https://github.com/abhrac/nci.
△ Less
Submitted 18 February, 2024;
originally announced February 2024.
-
GPU Acceleration of a Conjugate Exponential Model for Cancer Tissue Heterogeneity
Authors:
Anik Chaudhuri,
Anwoy Mohanty,
Manoranjan Satpathy
Abstract:
Heterogeneity in the cell population of cancer tissues poses many challenges in cancer diagnosis and treatment. Studying the heterogeneity in cell populations from gene expression measurement data in the context of cancer research is a problem of paramount importance. In addition, reducing the computation time of the algorithms that deal with high volumes of data has its obvious merits. Paralleliz…
▽ More
Heterogeneity in the cell population of cancer tissues poses many challenges in cancer diagnosis and treatment. Studying the heterogeneity in cell populations from gene expression measurement data in the context of cancer research is a problem of paramount importance. In addition, reducing the computation time of the algorithms that deal with high volumes of data has its obvious merits. Parallelizable models using Markov chain Monte Carlo methods are typically slow. This paper shows a novel, computationally efficient, and parallelizable model to analyze heterogeneity in cancer tissues using GPUs. Because our model is parallelizable, the input data size does not affect the computation time much, provided the hardware resources are not exhausted. Our model uses qPCR (quantitative polymerase chain reaction) gene expression measurements to study heterogeneity in cancer tissue. We compute the cell proportion breakup by accelerating variational methods on a GPU. We test this model on synthetic and real-world gene expression data collected from fibroblasts and compare the performance of our algorithm with those of MCMC and Expectation Maximization. Our new model is computationally less complex and faster than existing Bayesian models for cancer tissue heterogeneity.
△ Less
Submitted 18 January, 2024;
originally announced January 2024.
-
CLIPDrawX: Primitive-based Explanations for Text Guided Sketch Synthesis
Authors:
Nityanand Mathur,
Shyam Marjit,
Abhra Chaudhuri,
Anjan Dutta
Abstract:
With the goal of understanding the visual concepts that CLIP associates with text prompts, we show that the latent space of CLIP can be visualized solely in terms of linear transformations on simple geometric primitives like circles and straight lines. Although existing approaches achieve this by sketch-synthesis-through-optimization, they do so on the space of Bézier curves, which exhibit a waste…
▽ More
With the goal of understanding the visual concepts that CLIP associates with text prompts, we show that the latent space of CLIP can be visualized solely in terms of linear transformations on simple geometric primitives like circles and straight lines. Although existing approaches achieve this by sketch-synthesis-through-optimization, they do so on the space of Bézier curves, which exhibit a wastefully large set of structures that they can evolve into, as most of them are non-essential for generating meaningful sketches. We present CLIPDrawX, an algorithm that provides significantly better visualizations for CLIP text embeddings, using only simple primitive shapes like straight lines and circles. This constrains the set of possible outputs to linear transformations on these primitives, thereby exhibiting an inherently simpler mathematical form. The synthesis process of CLIPDrawX can be tracked end-to-end, with each visual concept being explained exclusively in terms of primitives. Implementation will be released upon acceptance. Project Page: $\href{https://clipdrawx.github.io/}{\text{https://clipdrawx.github.io/}}$.
△ Less
Submitted 4 December, 2023;
originally announced December 2023.
-
Detecting out-of-distribution text using topological features of transformer-based language models
Authors:
Andres Pollano,
Anupam Chaudhuri,
Anj Simmons
Abstract:
To safeguard machine learning systems that operate on textual data against out-of-distribution (OOD) inputs that could cause unpredictable behaviour, we explore the use of topological features of self-attention maps from transformer-based language models to detect when input text is out of distribution. Self-attention forms the core of transformer-based language models, dynamically assigning vecto…
▽ More
To safeguard machine learning systems that operate on textual data against out-of-distribution (OOD) inputs that could cause unpredictable behaviour, we explore the use of topological features of self-attention maps from transformer-based language models to detect when input text is out of distribution. Self-attention forms the core of transformer-based language models, dynamically assigning vectors to words based on context, thus in theory our methodology is applicable to any transformer-based language model with multihead self-attention. We evaluate our approach on BERT and compare it to a traditional OOD approach using CLS embeddings. Our results show that our approach outperforms CLS embeddings in distinguishing in-distribution samples from far-out-of-domain samples, but struggles with near or same-domain datasets.
△ Less
Submitted 18 July, 2024; v1 submitted 21 November, 2023;
originally announced November 2023.
-
Smart Traffic Management of Vehicles using Faster R-CNN based Deep Learning Method
Authors:
Arindam Chaudhuri
Abstract:
With constant growth of civilization and modernization of cities all across the world since past few centuries smart traffic management of vehicles is one of the most sorted after problem by research community. It is a challenging problem in computer vision and artificial intelligence domain. Smart traffic management basically involves segmentation of vehicles, estimation of traffic density and tr…
▽ More
With constant growth of civilization and modernization of cities all across the world since past few centuries smart traffic management of vehicles is one of the most sorted after problem by research community. It is a challenging problem in computer vision and artificial intelligence domain. Smart traffic management basically involves segmentation of vehicles, estimation of traffic density and tracking of vehicles. The vehicle segmentation from traffic videos helps realization of niche applications such as monitoring of speed and estimation of traffic. When occlusions, background with clutters and traffic with density variations are present, this problem becomes more intractable in nature. Keeping this motivation in this research work, we investigate Faster R-CNN based deep learning method towards segmentation of vehicles. This problem is addressed in four steps viz minimization with adaptive background model, Faster R-CNN based subnet operation, Faster R-CNN initial refinement and result optimization with extended topological active nets. The computational framework uses ideas of adaptive background modeling. It also addresses shadow and illumination related issues. Higher segmentation accuracy is achieved through topological active net deformable models. The topological and extended topological active nets help to achieve stated deformations. Mesh deformation is achieved with minimization of energy. The segmentation accuracy is improved with modified version of extended topological active net. The experimental results demonstrate superiority of this computational framework
△ Less
Submitted 3 November, 2023;
originally announced November 2023.
-
ChipNeMo: Domain-Adapted LLMs for Chip Design
Authors:
Mingjie Liu,
Teodor-Dumitru Ene,
Robert Kirby,
Chris Cheng,
Nathaniel Pinckney,
Rongjian Liang,
Jonah Alben,
Himyanshu Anand,
Sanmitra Banerjee,
Ismet Bayraktaroglu,
Bonita Bhaskaran,
Bryan Catanzaro,
Arjun Chaudhuri,
Sharon Clay,
Bill Dally,
Laura Dang,
Parikshit Deshpande,
Siddhanth Dhodhi,
Sameer Halepete,
Eric Hill,
Jiashang Hu,
Sumit Jain,
Ankit Jindal,
Brucek Khailany,
George Kokai
, et al. (17 additional authors not shown)
Abstract:
ChipNeMo aims to explore the applications of large language models (LLMs) for industrial chip design. Instead of directly deploying off-the-shelf commercial or open-source LLMs, we instead adopt the following domain adaptation techniques: domain-adaptive tokenization, domain-adaptive continued pretraining, model alignment with domain-specific instructions, and domain-adapted retrieval models. We e…
▽ More
ChipNeMo aims to explore the applications of large language models (LLMs) for industrial chip design. Instead of directly deploying off-the-shelf commercial or open-source LLMs, we instead adopt the following domain adaptation techniques: domain-adaptive tokenization, domain-adaptive continued pretraining, model alignment with domain-specific instructions, and domain-adapted retrieval models. We evaluate these methods on three selected LLM applications for chip design: an engineering assistant chatbot, EDA script generation, and bug summarization and analysis. Our evaluations demonstrate that domain-adaptive pretraining of language models, can lead to superior performance in domain related downstream tasks compared to their base LLaMA2 counterparts, without degradations in generic capabilities. In particular, our largest model, ChipNeMo-70B, outperforms the highly capable GPT-4 on two of our use cases, namely engineering assistant chatbot and EDA scripts generation, while exhibiting competitive performance on bug summarization and analysis. These results underscore the potential of domain-specific customization for enhancing the effectiveness of large language models in specialized applications.
△ Less
Submitted 4 April, 2024; v1 submitted 31 October, 2023;
originally announced November 2023.
-
Transitivity Recovering Decompositions: Interpretable and Robust Fine-Grained Relationships
Authors:
Abhra Chaudhuri,
Massimiliano Mancini,
Zeynep Akata,
Anjan Dutta
Abstract:
Recent advances in fine-grained representation learning leverage local-to-global (emergent) relationships for achieving state-of-the-art results. The relational representations relied upon by such methods, however, are abstract. We aim to deconstruct this abstraction by expressing them as interpretable graphs over image views. We begin by theoretically showing that abstract relational representati…
▽ More
Recent advances in fine-grained representation learning leverage local-to-global (emergent) relationships for achieving state-of-the-art results. The relational representations relied upon by such methods, however, are abstract. We aim to deconstruct this abstraction by expressing them as interpretable graphs over image views. We begin by theoretically showing that abstract relational representations are nothing but a way of recovering transitive relationships among local views. Based on this, we design Transitivity Recovering Decompositions (TRD), a graph-space search algorithm that identifies interpretable equivalents of abstract emergent relationships at both instance and class levels, and with no post-hoc computations. We additionally show that TRD is provably robust to noisy views, with empirical evidence also supporting this finding. The latter allows TRD to perform at par or even better than the state-of-the-art, while being fully interpretable. Implementation is available at https://github.com/abhrac/trd.
△ Less
Submitted 24 October, 2023;
originally announced October 2023.
-
Sarcasm in Sight and Sound: Benchmarking and Expansion to Improve Multimodal Sarcasm Detection
Authors:
Swapnil Bhosale,
Abhra Chaudhuri,
Alex Lee Robert Williams,
Divyank Tiwari,
Anjan Dutta,
Xiatian Zhu,
Pushpak Bhattacharyya,
Diptesh Kanojia
Abstract:
The introduction of the MUStARD dataset, and its emotion recognition extension MUStARD++, have identified sarcasm to be a multi-modal phenomenon -- expressed not only in natural language text, but also through manners of speech (like tonality and intonation) and visual cues (facial expression). With this work, we aim to perform a rigorous benchmarking of the MUStARD++ dataset by considering state-…
▽ More
The introduction of the MUStARD dataset, and its emotion recognition extension MUStARD++, have identified sarcasm to be a multi-modal phenomenon -- expressed not only in natural language text, but also through manners of speech (like tonality and intonation) and visual cues (facial expression). With this work, we aim to perform a rigorous benchmarking of the MUStARD++ dataset by considering state-of-the-art language, speech, and visual encoders, for fully utilizing the totality of the multi-modal richness that it has to offer, achieving a 2\% improvement in macro-F1 over the existing benchmark. Additionally, to cure the imbalance in the `sarcasm type' category in MUStARD++, we propose an extension, which we call \emph{MUStARD++ Balanced}, benchmarking the same with instances from the extension split across both train and test sets, achieving a further 2.4\% macro-F1 boost. The new clips were taken from a novel source -- the TV show, House MD, which adds to the diversity of the dataset, and were manually annotated by multiple annotators with substantial inter-annotator agreement in terms of Cohen's kappa and Krippendorf's alpha. Our code, extended data, and SOTA benchmark models are made public.
△ Less
Submitted 29 September, 2023;
originally announced October 2023.
-
Parallel transport on a Lie 2-group bundle over a Lie groupoid along Haefliger paths
Authors:
Saikat Chatterjee,
Adittya Chaudhuri
Abstract:
We prove a Lie 2-group torsor version of the well-known one-one correspondence between fibered categories and pseudofunctors. Consequently, we obtain a weak version of the principal Lie group bundle over a Lie groupoid. The correspondence also enables us to extend a particular class of principal 2-bundles to be defined over differentiable stacks. We show that the differential geometric connection…
▽ More
We prove a Lie 2-group torsor version of the well-known one-one correspondence between fibered categories and pseudofunctors. Consequently, we obtain a weak version of the principal Lie group bundle over a Lie groupoid. The correspondence also enables us to extend a particular class of principal 2-bundles to be defined over differentiable stacks. We show that the differential geometric connection structures introduced in the authors' previous work, combine nicely with the underlying fibration structure of a principal 2-bundle over a Lie groupoid. This interrelation allows us to derive a notion of parallel transport in the framework of principal 2-bundles over Lie groupoids along a particular class of Haefliger paths. The corresponding parallel transport functor is shown to be smooth. We apply our results to examine the parallel transport on an associated VB-groupoid.
△ Less
Submitted 11 September, 2023;
originally announced September 2023.
-
Hydrodynamics of passive and active rods in a temperature gradient
Authors:
Anweshika Pattanayak,
Abhishek Chaudhuri
Abstract:
Temperature plays a very important role in various biological processes like the evolution of life, as it is anticipated that early life existed in a very hot environment that eventually cooled down with time. In vitro experiments, conducted with various active matter systems in anisotropic temperature environments, show several interesting outcomes. Motivated by these experiments and some simulat…
▽ More
Temperature plays a very important role in various biological processes like the evolution of life, as it is anticipated that early life existed in a very hot environment that eventually cooled down with time. In vitro experiments, conducted with various active matter systems in anisotropic temperature environments, show several interesting outcomes. Motivated by these experiments and some simulation results, we study the collective behavior of self-propelled hard rods with excluded volume interactions, moving on a substrate in two dimensions and subjected to a small but finite external temperature gradient. Using a coarse-grained hydrodynamic approach, we explore the behavior of the system in the steady state and the stability both in the passive and active limits.
△ Less
Submitted 6 September, 2023;
originally announced September 2023.
-
Predictive Digital Twin for Optimizing Patient-Specific Radiotherapy Regimens under Uncertainty in High-Grade Gliomas
Authors:
Anirban Chaudhuri,
Graham Pash,
David A. Hormuth II,
Guillermo Lorenzo,
Michael Kapteyn,
Chengyue Wu,
Ernesto A. B. F. Lima,
Thomas E. Yankeelov,
Karen Willcox
Abstract:
We develop a methodology to create data-driven predictive digital twins for optimal risk-aware clinical decision-making. We illustrate the methodology as an enabler for an anticipatory personalized treatment that accounts for uncertainties in the underlying tumor biology in high-grade gliomas, where heterogeneity in the response to standard-of-care (SOC) radiotherapy contributes to sub-optimal pat…
▽ More
We develop a methodology to create data-driven predictive digital twins for optimal risk-aware clinical decision-making. We illustrate the methodology as an enabler for an anticipatory personalized treatment that accounts for uncertainties in the underlying tumor biology in high-grade gliomas, where heterogeneity in the response to standard-of-care (SOC) radiotherapy contributes to sub-optimal patient outcomes. The digital twin is initialized through prior distributions derived from population-level clinical data in the literature for a mechanistic model's parameters. Then the digital twin is personalized using Bayesian model calibration for assimilating patient-specific magnetic resonance imaging data and used to propose optimal radiotherapy treatment regimens by solving a multi-objective risk-based optimization under uncertainty problem. The solution leads to a suite of patient-specific optimal radiotherapy treatment regimens exhibiting varying levels of trade-off between the two competing clinical objectives: (i) maximizing tumor control (characterized by minimizing the risk of tumor volume growth) and (ii) minimizing the toxicity from radiotherapy. The proposed digital twin framework is illustrated by generating an in silico cohort of 100 patients with high-grade glioma growth and response properties typically observed in the literature. For the same total radiation dose as the SOC, the personalized treatment regimens lead to median increase in tumor time to progression of around six days. Alternatively, for the same level of tumor control as the SOC, the digital twin provides optimal treatment options that lead to a median reduction in radiation dose by 16.7% (10 Gy) compared to SOC total dose of 60 Gy. The range of optimal solutions also provide options with increased doses for patients with aggressive cancer, where SOC does not lead to sufficient tumor control.
△ Less
Submitted 23 August, 2023;
originally announced August 2023.
-
Stochastic resonance in a model of a periodically driven DNA : Multiple transitions, scaling and sequence dependence
Authors:
Ramu Kumar Yadav,
M. Suman Kalyan,
Rajeev Kapri,
Abhishek Chaudhuri
Abstract:
We numerically study stochastic resonance in the unzipping of a model double-stranded DNA by a periodic force. We observe multiple peaks in stochastic resonance in the output signal as the driving force frequency is varied for different force amplitudes, temperature, chain length, and chain heterogeneity. Multiple peaks point to the existence of multiple stable and metastable states, which corresp…
▽ More
We numerically study stochastic resonance in the unzipping of a model double-stranded DNA by a periodic force. We observe multiple peaks in stochastic resonance in the output signal as the driving force frequency is varied for different force amplitudes, temperature, chain length, and chain heterogeneity. Multiple peaks point to the existence of multiple stable and metastable states, which correspond to dynamical states of partially zipped and unzipped conformations and transitions between them. We quantify such transitions by looking at the time evolution of the fraction of bound base pairs. We obtain phase diagrams in the force amplitude-temperature plane both in the resonance frequency of the primary peak and the output signal at the peak value. We further obtain an excellent scaling behavior of the output signal for changing lengths of the DNA. Resonance behavior is also affected by chain heterogeneity as it depends strongly on which base pair the periodic forcing is applied.
△ Less
Submitted 19 August, 2023;
originally announced August 2023.
-
Dynamics of a Passive Droplet in Active Turbulence
Authors:
Chamkor Singh,
Abhishek Chaudhuri
Abstract:
We numerically study the effect of an active turbulent environment on a passive deformable droplet. The system is simulated using coupled hydrodynamic and nematodynamic equations for nematic liquid crystals with an active stress which is non-zero outside the droplet, and is zero inside. The droplet undergoes deformation fluctuations and its movement shows periods of ``runs" and ``stays". The mean…
▽ More
We numerically study the effect of an active turbulent environment on a passive deformable droplet. The system is simulated using coupled hydrodynamic and nematodynamic equations for nematic liquid crystals with an active stress which is non-zero outside the droplet, and is zero inside. The droplet undergoes deformation fluctuations and its movement shows periods of ``runs" and ``stays". The mean square displacement of the geometric center of the droplet shows an extended ballistic regime and a transition to normal diffusive regime which depends on the size of the droplet. We relate this transition with a temporal scale associated with velocity autocorrelation function of the droplet trajectories, and with a spatial scale associated with one-time two-point velocity correlation function of the surrounding active medium. As the radius is decreased below the integral length scale, the velocity autocorrelation time of trajectories increases and the transition to normal diffusion is delayed.
△ Less
Submitted 24 July, 2023;
originally announced July 2023.
-
A Zero-Threshold PT-Symmetric Polariton-Raman Laser
Authors:
Avijit Dhara,
Devarshi Chakrabarty,
Pritam Das,
Kritika Ghosh,
Ayan Roy Chaudhuri,
Sajal Dhara
Abstract:
Parity-time (PT)-symmetry in the classical regime has been realized in optics by introducing loss and gain in electromagnetic wave propagation which has yielded numerous applications like nonreciprocal propagation and finite threshold single-mode lasers. However, PT-symmetry in the quantum regime so far remains elusive. Here, we demonstrate a PT-symmetric zero-threshold polariton-Raman laser by ut…
▽ More
Parity-time (PT)-symmetry in the classical regime has been realized in optics by introducing loss and gain in electromagnetic wave propagation which has yielded numerous applications like nonreciprocal propagation and finite threshold single-mode lasers. However, PT-symmetry in the quantum regime so far remains elusive. Here, we demonstrate a PT-symmetric zero-threshold polariton-Raman laser by utilizing stimulated resonant Raman scattering of polarized exciton-polaritons. By pumping resonantly at the exceptional point of polariton bands with non-Hermitian topology, a quantum PT-symmetric phase is realized when the Raman-active phonon mode frequencies match with the polariton mode frequency difference. The PT-symmetric phase corresponding to zero-threshold lasing can be switched to PT broken phase showing a finite threshold via cavity detuning by the variation of bath temperature or pump polarization. Our realization of PT-symmetry in the quantum regime and consequently the zero-threshold laser can open up applications in quantum information and stimulate new research activities in cavity quantum electrodynamics.
△ Less
Submitted 27 July, 2023; v1 submitted 27 May, 2023;
originally announced May 2023.
-
Anisotropic exciton polariton pairs as a platform for PT-symmetric non-Hermitian physics
Authors:
Devarshi Chakrabarty,
Avijit Dhara,
Pritam Das,
Kritika Ghosh,
Ayan Roy Chaudhuri,
Sajal Dhara
Abstract:
Non-Hermitian systems with parity-time (PT) symmetry have been realized using optical constructs in the classical domain, leading to a plethora of non-intuitive phenomena. However, PT-symmetry in purely quantum non-Hermitian systems like microcavity exciton-polaritons has not been realized so far. Here we show how a pair of nearly orthogonal sets of anisotropic exciton-polaritons can offer a versa…
▽ More
Non-Hermitian systems with parity-time (PT) symmetry have been realized using optical constructs in the classical domain, leading to a plethora of non-intuitive phenomena. However, PT-symmetry in purely quantum non-Hermitian systems like microcavity exciton-polaritons has not been realized so far. Here we show how a pair of nearly orthogonal sets of anisotropic exciton-polaritons can offer a versatile platform for realizing multiple spectral degeneracies called Exceptional Points (EPs) and propose a roadmap to achieve a PT-symmetric system. Polarization-tunable coupling strength creates one class of EPs, while Voigt EPs are observed for specific orientations where splitting of polariton modes due to birefringence is compensated by Transverse Electric (TE) -Transverse Magnetic (TM) mode splitting. Thus, paired sets of polarized anisotropic microcavity exciton-polaritons can offer a promising platform not only for fundamental research in non-Hermitian quantum physics and topological polaritons, but also, we propose that it will be critical for realizing zero threshold lasers.
△ Less
Submitted 31 May, 2023; v1 submitted 27 May, 2023;
originally announced May 2023.
-
Emergent room-temperature ferroelectricity in spark-plasma sintered DyCrO$_3$ and LaCrO$_3$
Authors:
Suryakanta Mishra,
Keerthana,
Krishna Rudrapal,
Biswajit Jana,
Kazi Parvez Islam,
Archna Sagdeo,
Ayan Roy Chaudhuri,
Venimadhav Adyam,
Debraj Choudhury
Abstract:
Identification of novel multiferroic materials with high-ordering temperatures remains at the forefront of condensed matter physics research. In this regard, the antiferromagnetic RCrO$_3$ compounds (like GdCrO$_3$) constitute a promising class of multiferroic compounds, which, however, mostly become ferroelectric concomitant with the antiferromagnetic ordering much below room-temperature, arising…
▽ More
Identification of novel multiferroic materials with high-ordering temperatures remains at the forefront of condensed matter physics research. In this regard, the antiferromagnetic RCrO$_3$ compounds (like GdCrO$_3$) constitute a promising class of multiferroic compounds, which, however, mostly become ferroelectric concomitant with the antiferromagnetic ordering much below room-temperature, arising from a subtle competition between the ferroelectric off-centering mode and a non-polar antiferrodistortive rotation mode that inhibits ferroelectricity. Recently, room-temperature ferroelectricity of structural origin, arising from off-centering displacements of R and Cr ions, has been identified in spark-plasma sintered GdCrO$_3$ [Suryakanta Mishra et al., Phys. Rev. B 104, L180101 (2021)]. Interestingly, some of the experimentally observed non-ferroelectric RCrO$_3$ compounds have been theoretically predicted to host similar ferroelectric instabilities. Here, we have identified two such non-ferroelectric RCrO3 compounds, one DyCrO$_3$ (which is reported as a quantum paraelectric) and another LaCrO$_3$ (which is paraelectric), and using a modified synthesis protocol involving spark-plasma-sintering (SPS), we have been successful in engineering an intrinsic room-temperature ferroelectricity in the paramagnetic state, driven by noncentrosymmetric structural phase in both SPS sintered DyCrO$_3$ and LaCrO$_3$, in contrast to room-temperature paraelectricity in solid-state synthesized DyCrO$_3$ and LaCrO$_3$. While the ferroelectricity in SPS-prepared DyCrO$_3$ and LaCrO$_3$ is stable at room-temperature, it undergoes an irreversible transition from a ferroelectric (Pna2$_1$) phase to a paraelectric (Pbnm) phase at 440 K. Significantly, SPS-sintered LaCrO$_3$, which undergoes antiferromagnetic ordering at 290 K, emerges as a promising near room-temperature multiferroic material.
△ Less
Submitted 24 June, 2023; v1 submitted 19 May, 2023;
originally announced May 2023.
-
Data-Free Sketch-Based Image Retrieval
Authors:
Abhra Chaudhuri,
Ayan Kumar Bhunia,
Yi-Zhe Song,
Anjan Dutta
Abstract:
Rising concerns about privacy and anonymity preservation of deep learning models have facilitated research in data-free learning (DFL). For the first time, we identify that for data-scarce tasks like Sketch-Based Image Retrieval (SBIR), where the difficulty in acquiring paired photos and hand-drawn sketches limits data-dependent cross-modal learning algorithms, DFL can prove to be a much more prac…
▽ More
Rising concerns about privacy and anonymity preservation of deep learning models have facilitated research in data-free learning (DFL). For the first time, we identify that for data-scarce tasks like Sketch-Based Image Retrieval (SBIR), where the difficulty in acquiring paired photos and hand-drawn sketches limits data-dependent cross-modal learning algorithms, DFL can prove to be a much more practical paradigm. We thus propose Data-Free (DF)-SBIR, where, unlike existing DFL problems, pre-trained, single-modality classification models have to be leveraged to learn a cross-modal metric-space for retrieval without access to any training data. The widespread availability of pre-trained classification models, along with the difficulty in acquiring paired photo-sketch datasets for SBIR justify the practicality of this setting. We present a methodology for DF-SBIR, which can leverage knowledge from models independently trained to perform classification on photos and sketches. We evaluate our model on the Sketchy, TU-Berlin, and QuickDraw benchmarks, designing a variety of baselines based on state-of-the-art DFL literature, and observe that our method surpasses all of them by significant margins. Our method also achieves mAPs competitive with data-dependent approaches, all the while requiring no training data. Implementation is available at \url{https://github.com/abhrac/data-free-sbir}.
△ Less
Submitted 14 March, 2023;
originally announced March 2023.
-
Dark matter production from two evaporating PBH distributions
Authors:
Arnab Chaudhuri,
Baradhwaj Coleppa,
Kousik Loho
Abstract:
Particulate Dark Matter (DM), completely isolated from the Standard Model particle sector, can be produced in the early universe from Primordial Black Hole (PBH) evaporation. However, Big Bang Nucleosynthesis (BBN) observations put an upper bound on the initial mass of PBH requiring the PBH to evaporate completely before the advent of BBN. DM particles in the mass range $\sim(1-10^9)$ GeV can not…
▽ More
Particulate Dark Matter (DM), completely isolated from the Standard Model particle sector, can be produced in the early universe from Primordial Black Hole (PBH) evaporation. However, Big Bang Nucleosynthesis (BBN) observations put an upper bound on the initial mass of PBH requiring the PBH to evaporate completely before the advent of BBN. DM particles in the mass range $\sim(1-10^9)$ GeV can not explain the observed relic abundance for an early matter dominated universe due to this BBN constraint. However, this assumes the presence of only one monochromatic PBH mass distribution in the early universe. In this work, we explore the simple possibility of achieving the observed relic with DM masses from the above mentioned range for an early matter dominated era with two monochromatic evaporating PBH mass distributions and demonstrate that the fermionic DM masses consistent with BBN change slightly.
△ Less
Submitted 18 August, 2023; v1 submitted 20 January, 2023;
originally announced January 2023.
-
Nostradamus: Weathering Worth
Authors:
Alapan Chaudhuri,
Zeeshan Ahmed,
Ashwin Rao,
Shivansh Subramanian,
Shreyas Pradhan,
Abhishek Mittal
Abstract:
Nostradamus, inspired by the French astrologer and reputed seer, is a detailed study exploring relations between environmental factors and changes in the stock market. In this paper, we analyze associative correlation and causation between environmental elements (including natural disasters, climate and weather conditions) and stock prices, using historical stock market data, historical climate da…
▽ More
Nostradamus, inspired by the French astrologer and reputed seer, is a detailed study exploring relations between environmental factors and changes in the stock market. In this paper, we analyze associative correlation and causation between environmental elements (including natural disasters, climate and weather conditions) and stock prices, using historical stock market data, historical climate data, and various climate indicators such as carbon dioxide emissions. We have conducted our study based on the US financial market, global climate trends, and daily weather records to demonstrate a significant relationship between climate and stock price fluctuation. Our analysis covers both short-term and long-term rises and dips in company stock performances. Lastly, we take four natural disasters as a case study to observe the effect they have on people's emotional state and their influence on the stock market.
△ Less
Submitted 17 January, 2023; v1 submitted 8 December, 2022;
originally announced December 2022.
-
Evolution of Resistive Switching Characteristics in WO3-x-based MIM Devices by Tailoring Oxygen Deficiency
Authors:
Krishna Rudrapal,
Biswajit Jana,
Venimadhav Adyam,
Ayan Roy Chaudhuri
Abstract:
We report on resistive switching (RS) characteristics of W/WO3-x/Pt-based thin film memristors modulated by precisely controlled oxygen non-stoichiometry. RS properties of the devices with varied oxygen vacancy (VO) concentration have been studied by measuring their DC current voltage properties. Switchability of the resistance states in the memristors have been found to depend strongly on the VOs…
▽ More
We report on resistive switching (RS) characteristics of W/WO3-x/Pt-based thin film memristors modulated by precisely controlled oxygen non-stoichiometry. RS properties of the devices with varied oxygen vacancy (VO) concentration have been studied by measuring their DC current voltage properties. Switchability of the resistance states in the memristors have been found to depend strongly on the VOs concentration in the WO3-x layer. Depending on x, the memristors exhibited forming-free bipolar, forming-required bipolar and non-formable characteristics. Devices with high VOs concentration (~1*1021 cm-3) exhibited lower initial resistance and memory window of only 15, which has been increased to ~6500 with reducing VOs concentration to ~5.8*1020 cm-3. Forming-free, stable RS with memory window of ~2000 have been realized for a memristor possessing VOs concentration of ~6.2*1020 cm-3. Investigation of the conduction mechanism suggests that tailoring VOs concentration modifies the formation and dimension of the conducting filaments as well as the Schottky barrier height at WO3-x/Pt interface which deterministically modulates RS characteristics of the WO3-x based memristors.
△ Less
Submitted 27 November, 2022;
originally announced November 2022.
-
A Dynamic Weighted Federated Learning for Android Malware Classification
Authors:
Ayushi Chaudhuri,
Arijit Nandi,
Buddhadeb Pradhan
Abstract:
Android malware attacks are increasing daily at a tremendous volume, making Android users more vulnerable to cyber-attacks. Researchers have developed many machine learning (ML)/ deep learning (DL) techniques to detect and mitigate android malware attacks. However, due to technological advancement, there is a rise in android mobile devices. Furthermore, the devices are geographically dispersed, re…
▽ More
Android malware attacks are increasing daily at a tremendous volume, making Android users more vulnerable to cyber-attacks. Researchers have developed many machine learning (ML)/ deep learning (DL) techniques to detect and mitigate android malware attacks. However, due to technological advancement, there is a rise in android mobile devices. Furthermore, the devices are geographically dispersed, resulting in distributed data. In such scenario, traditional ML/DL techniques are infeasible since all of these approaches require the data to be kept in a central system; this may provide a problem for user privacy because of the massive proliferation of Android mobile devices; putting the data in a central system creates an overhead. Also, the traditional ML/DL-based android malware classification techniques are not scalable. Researchers have proposed federated learning (FL) based android malware classification system to solve the privacy preservation and scalability with high classification performance. In traditional FL, Federated Averaging (FedAvg) is utilized to construct the global model at each round by merging all of the local models obtained from all of the customers that participated in the FL. However, the conventional FedAvg has a disadvantage: if one poor-performing local model is included in global model development for each round, it may result in an under-performing global model. Because FedAvg favors all local models equally when averaging. To address this issue, our main objective in this work is to design a dynamic weighted federated averaging (DW-FedAvg) strategy in which the weights for each local model are automatically updated based on their performance at the client. The DW-FedAvg is evaluated using four popular benchmark datasets, Melgenome, Drebin, Kronodroid and Tuandromd used in android malware classification research.
△ Less
Submitted 23 November, 2022;
originally announced November 2022.
-
Cross-Modal Fusion Distillation for Fine-Grained Sketch-Based Image Retrieval
Authors:
Abhra Chaudhuri,
Massimiliano Mancini,
Yanbei Chen,
Zeynep Akata,
Anjan Dutta
Abstract:
Representation learning for sketch-based image retrieval has mostly been tackled by learning embeddings that discard modality-specific information. As instances from different modalities can often provide complementary information describing the underlying concept, we propose a cross-attention framework for Vision Transformers (XModalViT) that fuses modality-specific information instead of discard…
▽ More
Representation learning for sketch-based image retrieval has mostly been tackled by learning embeddings that discard modality-specific information. As instances from different modalities can often provide complementary information describing the underlying concept, we propose a cross-attention framework for Vision Transformers (XModalViT) that fuses modality-specific information instead of discarding them. Our framework first maps paired datapoints from the individual photo and sketch modalities to fused representations that unify information from both modalities. We then decouple the input space of the aforementioned modality fusion network into independent encoders of the individual modalities via contrastive and relational cross-modal knowledge distillation. Such encoders can then be applied to downstream tasks like cross-modal retrieval. We demonstrate the expressive capacity of the learned representations by performing a wide range of experiments and achieving state-of-the-art results on three fine-grained sketch-based image retrieval benchmarks: Shoe-V2, Chair-V2 and Sketchy. Implementation is available at https://github.com/abhrac/xmodal-vit.
△ Less
Submitted 19 October, 2022;
originally announced October 2022.
-
Relational Proxies: Emergent Relationships as Fine-Grained Discriminators
Authors:
Abhra Chaudhuri,
Massimiliano Mancini,
Zeynep Akata,
Anjan Dutta
Abstract:
Fine-grained categories that largely share the same set of parts cannot be discriminated based on part information alone, as they mostly differ in the way the local parts relate to the overall global structure of the object. We propose Relational Proxies, a novel approach that leverages the relational information between the global and local views of an object for encoding its semantic label. Star…
▽ More
Fine-grained categories that largely share the same set of parts cannot be discriminated based on part information alone, as they mostly differ in the way the local parts relate to the overall global structure of the object. We propose Relational Proxies, a novel approach that leverages the relational information between the global and local views of an object for encoding its semantic label. Starting with a rigorous formalization of the notion of distinguishability between fine-grained categories, we prove the necessary and sufficient conditions that a model must satisfy in order to learn the underlying decision boundaries in the fine-grained setting. We design Relational Proxies based on our theoretical findings and evaluate it on seven challenging fine-grained benchmark datasets and achieve state-of-the-art results on all of them, surpassing the performance of all existing works with a margin exceeding 4% in some cases. We also experimentally validate our theory on fine-grained distinguishability and obtain consistent results across multiple benchmarks. Implementation is available at https://github.com/abhrac/relational-proxies.
△ Less
Submitted 5 October, 2022;
originally announced October 2022.
-
ProtoBandit: Efficient Prototype Selection via Multi-Armed Bandits
Authors:
Arghya Roy Chaudhuri,
Pratik Jawanpuria,
Bamdev Mishra
Abstract:
In this work, we propose a multi-armed bandit-based framework for identifying a compact set of informative data instances (i.e., the prototypes) from a source dataset $S$ that best represents a given target set $T$. Prototypical examples of a given dataset offer interpretable insights into the underlying data distribution and assist in example-based reasoning, thereby influencing every sphere of h…
▽ More
In this work, we propose a multi-armed bandit-based framework for identifying a compact set of informative data instances (i.e., the prototypes) from a source dataset $S$ that best represents a given target set $T$. Prototypical examples of a given dataset offer interpretable insights into the underlying data distribution and assist in example-based reasoning, thereby influencing every sphere of human decision-making. Current state-of-the-art prototype selection approaches require $O(|S||T|)$ similarity comparisons between source and target data points, which becomes prohibitively expensive for large-scale settings. We propose to mitigate this limitation by employing stochastic greedy search in the space of prototypical examples and multi-armed bandits for reducing the number of similarity comparisons. Our randomized algorithm, ProtoBandit, identifies a set of $k$ prototypes incurring $O(k^3|S|)$ similarity comparisons, which is independent of the size of the target set. An interesting outcome of our analysis is for the $k$-medoids clustering problem $T = S$ setting) in which we show that our algorithm ProtoBandit approximates the BUILD step solution of the partitioning around medoids (PAM) method in $O(k^3|S|)$ complexity. Empirically, we observe that ProtoBandit reduces the number of similarity computation calls by several orders of magnitudes ($100-1000$ times) while obtaining solutions similar in quality to those from state-of-the-art approaches.
△ Less
Submitted 23 August, 2023; v1 submitted 4 October, 2022;
originally announced October 2022.
-
Dark Matter from Evaporating Primordial Black Holes in the Early Universe
Authors:
Pratik Chattopadhyay,
Arnab Chaudhuri,
Maxim Yu. Khlopov
Abstract:
Primordial Black Holes (PBH) could dominate in the early universe and, evaporating before Big bang Nucleosynthesis, can provide new freeze in mechanism of dark matter (DM) production. The proposed scenario is considered for two possible mechanisms of PBH formation and the corresponding continuous PBH mass spectra so that the effect of non-single PBH mass spectrum is taken into account in the resul…
▽ More
Primordial Black Holes (PBH) could dominate in the early universe and, evaporating before Big bang Nucleosynthesis, can provide new freeze in mechanism of dark matter (DM) production. The proposed scenario is considered for two possible mechanisms of PBH formation and the corresponding continuous PBH mass spectra so that the effect of non-single PBH mass spectrum is taken into account in the results of PBH evaporation, by which PBH dominance in the early universe ends. We specify the conditions under which the proposed scenario can explain production of dark matter in very early Universe.
△ Less
Submitted 10 June, 2024; v1 submitted 22 September, 2022;
originally announced September 2022.
-
Joint Sequential Detection and Isolation for Dependent Data Streams
Authors:
Anamitra Chaudhuri,
Georgios Fellouris
Abstract:
The problem of joint sequential detection and isolation is considered in the context of multiple, not necessarily independent, data streams. A multiple testing framework is proposed, where each hypothesis corresponds to a different subset of data streams, the sample size is a stopping time of the observations, and the probabilities of four kinds of error are controlled below distinct, user-specifi…
▽ More
The problem of joint sequential detection and isolation is considered in the context of multiple, not necessarily independent, data streams. A multiple testing framework is proposed, where each hypothesis corresponds to a different subset of data streams, the sample size is a stopping time of the observations, and the probabilities of four kinds of error are controlled below distinct, user-specified levels. Two of these errors reflect the detection component of the formulation, whereas the other two the isolation component. The optimal expected sample size is characterized to a first-order asymptotic approximation as the error probabilities go to 0. Different asymptotic regimes, expressing different prioritizations of the detection and isolation tasks, are considered. A novel, versatile family of testing procedures is proposed, in which two distinct, in general, statistics are computed for each hypothesis, one addressing the detection task and the other the isolation task. Tests in this family, of various computational complexities, are shown to be asymptotically optimal under different setups. The general theory is applied to the detection and isolation of anomalous, not necessarily independent, data streams, as well as to the detection and isolation of an unknown dependence structure.
△ Less
Submitted 30 June, 2022;
originally announced July 2022.
-
Study of entropy production due to electroweak phase transition in $Z_2$ symmetric extension of the Standard Model
Authors:
Arnab Chaudhuri,
Jaydeb Das
Abstract:
In this work we consider the simple $Z_2$ symmetric extension to the Standard Model (SM) and proceed to study the nature of electroweak phase transition (EWPT) in the early universe. We show that the nature of the phase transition changes from a smooth crossover in the SM to a strong first order with this addition of the real scalar. Furthermore, we show the entropy release in this scenario is hig…
▽ More
In this work we consider the simple $Z_2$ symmetric extension to the Standard Model (SM) and proceed to study the nature of electroweak phase transition (EWPT) in the early universe. We show that the nature of the phase transition changes from a smooth crossover in the SM to a strong first order with this addition of the real scalar. Furthermore, we show the entropy release in this scenario is higher than that of the SM. This can lead to a strong dilution of frozen out dark matter particles and baryon asymmetry, if something existed before the onset of the phase transition.
△ Less
Submitted 5 August, 2022; v1 submitted 17 June, 2022;
originally announced June 2022.
-
Driven translocation of a semiflexible polymer through a conical channel in the presence of attractive surface interactions
Authors:
Andri Sharma,
Rajeev Kapri,
Abhishek Chaudhuri
Abstract:
We study the translocation of a semiflexible polymer through a conical channel with attractive surface interactions and a driving force which varies spatially inside the channel. Using the results of the translocation dynamics of a flexible polymer through an extended channel as control, we first show that the asymmetric shape of the channel gives rise to non-monotonic features in the total transl…
▽ More
We study the translocation of a semiflexible polymer through a conical channel with attractive surface interactions and a driving force which varies spatially inside the channel. Using the results of the translocation dynamics of a flexible polymer through an extended channel as control, we first show that the asymmetric shape of the channel gives rise to non-monotonic features in the total translocation time as a function of the apex angle of the channel. The waiting time distributions of individual monomer beads inside the channel show unique features strongly dependent on the driving force and the surface interactions. Polymer stiffness results in longer translocation times for all angles of the channel. Further, non-monotonic features in the translocation time as a function of the channel angle changes substantially as the polymer becomes stiffer, which is reflected in the changing features of the waiting time distributions. We construct a free energy description of the system incorporating entropic and energetic contributions in the low force regime to explain the simulation results.
△ Less
Submitted 14 November, 2022; v1 submitted 26 May, 2022;
originally announced May 2022.
-
Nitrogen doped In$_2$O$_3$-ZnO nanocomposite thin film based sensitive and selective ethanol sensor
Authors:
P. K. Shihabudeen,
Ayan Roy Chaudhuri
Abstract:
Nanocomposite metal oxide thin films exhibit assuring qualities in the field of gas sensors because of the collective opportunities provided by the heterointerface formation. In this work, we present the synthesis of nitrogen doped mesoporous In$_2$O$_3$-ZnO nano-composite thin films by simple wet chemical method using urea as the nitrogen precursor. SEM investigation suggests formation of mesopor…
▽ More
Nanocomposite metal oxide thin films exhibit assuring qualities in the field of gas sensors because of the collective opportunities provided by the heterointerface formation. In this work, we present the synthesis of nitrogen doped mesoporous In$_2$O$_3$-ZnO nano-composite thin films by simple wet chemical method using urea as the nitrogen precursor. SEM investigation suggests formation of mesoporous nano-composite thin films, where the uniformity of surface pore distribution depends on the relative proportion of In$_2$O$_3$ and ZnO in the composites. HRTEM investigation suggest formation of sharp interfaces between N-In$_2$O$_3$ and N-ZnO grains in the nano-composite thin films. The nano-composite thin films have been tested for their ethanol sensing performance over an extensive range of temperature, ethanol vapor concentration and relative humidity. Nitrogen doped nano-composite thin film with equal proportion of In$_2$O$_3$ and ZnO exhibits excellent ethanol sensing performance at a reasonable operating temperature (~94 % at 200 °C for 50 ppm of ethanol), fast response time (~two seconds), stability over time, enhanced resilience against humidity and selectivity to ethanol over various other volatile organic compounds.
△ Less
Submitted 24 January, 2022;
originally announced February 2022.
-
Extension and dynamical phases in random walkers depositing and following chemical trails
Authors:
Subhashree Subhrasmita Khuntia,
Abhishek Chaudhuri,
Debasish Chaudhuri
Abstract:
Active walker models have proved to be extremely effective in understanding the evolution of a large class of systems in biology like ant trail formation and pedestrian trails. We propose a simple model of a random walker which modifies its local environment that in turn influences the motion of the walker at a {\em later} time. We perform direct numerical simulations of the walker in a discrete l…
▽ More
Active walker models have proved to be extremely effective in understanding the evolution of a large class of systems in biology like ant trail formation and pedestrian trails. We propose a simple model of a random walker which modifies its local environment that in turn influences the motion of the walker at a {\em later} time. We perform direct numerical simulations of the walker in a discrete lattice with the walker actively depositing a chemical which attracts the walker trajectory and also evaporates in time. We propose a method to look at the structural transitions of the trajectory using radius of gyration for finite time walks. The extension over a definite time-window shows a non-monotonic change with the deposition rate characteristic of a coil-globule transition. At certain regions of the parameter space of the chemical deposition and evaporation rates, the extensions of the walker shows a re-entrant behavior. The dynamics, characterised by the mean-squared displacement, shows deviation from diffusive scaling at intermediate time scales, returning to diffusive behavior asymptotically. A mean field theory captures the variation of the asymptotic diffusivity.
△ Less
Submitted 17 January, 2023; v1 submitted 1 February, 2022;
originally announced February 2022.
-
A Cross Validation Framework for Signal Denoising with Applications to Trend Filtering, Dyadic CART and Beyond
Authors:
Anamitra Chaudhuri,
Sabyasachi Chatterjee
Abstract:
This paper formulates a general cross validation framework for signal denoising. The general framework is then applied to nonparametric regression methods such as Trend Filtering and Dyadic CART. The resulting cross validated versions are then shown to attain nearly the same rates of convergence as are known for the optimally tuned analogues. There did not exist any previous theoretical analyses o…
▽ More
This paper formulates a general cross validation framework for signal denoising. The general framework is then applied to nonparametric regression methods such as Trend Filtering and Dyadic CART. The resulting cross validated versions are then shown to attain nearly the same rates of convergence as are known for the optimally tuned analogues. There did not exist any previous theoretical analyses of cross validated versions of Trend Filtering or Dyadic CART. To illustrate the generality of the framework we also propose and study cross validated versions of two fundamental estimators; lasso for high dimensional linear regression and singular value thresholding for matrix estimation. Our general framework is inspired by the ideas in Chatterjee and Jafarov (2015) and is potentially applicable to a wide range of estimation methods which use tuning parameters.
△ Less
Submitted 3 May, 2023; v1 submitted 7 January, 2022;
originally announced January 2022.
-
Nitrogen incorporated Zinc oxide thin film for efficient ethanol detection
Authors:
P. K. Shihabudeen,
Mina Yaghoobi Notash,
Jaber Jahanbin Sardroodi,
Ayan Roy Chaudhuri
Abstract:
Zinc oxide which is a n-type semiconducting metal oxide (SMO) has been a promising material for detecting ethanol vapor. However, pure ZnO based ethanol sensors often suffer from high working temperature, cross sensitivity towards methanol and poor stability against humidity. Doping ZnO with various metal ions has been widely explored as a proficient approach to improve its ethanol sensing propert…
▽ More
Zinc oxide which is a n-type semiconducting metal oxide (SMO) has been a promising material for detecting ethanol vapor. However, pure ZnO based ethanol sensors often suffer from high working temperature, cross sensitivity towards methanol and poor stability against humidity. Doping ZnO with various metal ions has been widely explored as a proficient approach to improve its ethanol sensing properties, while anionic dopants have been rarely considered. Here in we demonstrate the effect of nitrogen doping on the ethanol sensing characteristics of ZnO thin films. Nitrogen doped ZnO (N-ZnO) thin films have been synthesized following sol-gel technique with urea as nitrogen precursor. Ethanol sensing characteristics of the N-ZnO thin film has been com-pared with pure ZnO sensor over a wide range of temperature and relative humidity conditions. The N-ZnO sensor exhibits significantly large ethanol sensing response at a lower operating temperature (~99 % at 225 °C vs ~81 % at 250 °C for ZnO), faster response time (12 s vs 33 s for ZnO), long term stability, improved resilience against humidity and selectivity towards ethanol over methanol and acetone. The experimental observations have been supplemented by estimating the adsorption energies of ethanol on ZnO and N-ZnO surface using density functional theory (DFT) calculations. We discuss that the microscopic origin of improved ethanol sensing of N-ZnO is related to the facile adsorption of ethanol molecules on the oxide surface which is promoted by modification of electronic properties of ZnO by the nitrogen dopant atoms.
△ Less
Submitted 18 December, 2021;
originally announced December 2021.
-
Learning High-Dimensional Parametric Maps via Reduced Basis Adaptive Residual Networks
Authors:
Thomas O'Leary-Roseberry,
Xiaosong Du,
Anirban Chaudhuri,
Joaquim R. R. A. Martins,
Karen Willcox,
Omar Ghattas
Abstract:
We propose a scalable framework for the learning of high-dimensional parametric maps via adaptively constructed residual network (ResNet) maps between reduced bases of the inputs and outputs. When just few training data are available, it is beneficial to have a compact parametrization in order to ameliorate the ill-posedness of the neural network training problem. By linearly restricting high-dime…
▽ More
We propose a scalable framework for the learning of high-dimensional parametric maps via adaptively constructed residual network (ResNet) maps between reduced bases of the inputs and outputs. When just few training data are available, it is beneficial to have a compact parametrization in order to ameliorate the ill-posedness of the neural network training problem. By linearly restricting high-dimensional maps to informed reduced bases of the inputs, one can compress high-dimensional maps in a constructive way that can be used to detect appropriate basis ranks, equipped with rigorous error estimates. A scalable neural network learning framework is thus to learn the nonlinear compressed reduced basis mapping. Unlike the reduced basis construction, however, neural network constructions are not guaranteed to reduce errors by adding representation power, making it difficult to achieve good practical performance. Inspired by recent approximation theory that connects ResNets to sequential minimizing flows, we present an adaptive ResNet construction algorithm. This algorithm allows for depth-wise enrichment of the neural network approximation, in a manner that can achieve good practical performance by first training a shallow network and then adapting. We prove universal approximation of the associated neural network class for $L^2_ν$ functions on compact sets. Our overall framework allows for constructive means to detect appropriate breadth and depth, and related compact parametrizations of neural networks, significantly reducing the need for architectural hyperparameter tuning. Numerical experiments for parametric PDE problems and a 3D CFD wing design optimization parametric map demonstrate that the proposed methodology can achieve remarkably high accuracy for limited training data, and outperformed other neural network strategies we compared against.
△ Less
Submitted 15 November, 2022; v1 submitted 13 December, 2021;
originally announced December 2021.
-
Entopy release in Electroweak Phase Transition in 2HDM
Authors:
Arnab Chaudhuri,
Maxim Yu. Khlopov,
Shiladitya Porey
Abstract:
Electroweak phase transition in the simplest extension of the standard model namely two Higgs doublet model and entropy production within this framework is studied. We have considered several benchmark points which were called using BSMPT, a C++ package, within the limit of $vev/T_C>0.2$ are studied, and corresponding entropy productions are shown in this paper.
Electroweak phase transition in the simplest extension of the standard model namely two Higgs doublet model and entropy production within this framework is studied. We have considered several benchmark points which were called using BSMPT, a C++ package, within the limit of $vev/T_C>0.2$ are studied, and corresponding entropy productions are shown in this paper.
△ Less
Submitted 29 October, 2021;
originally announced November 2021.
-
Dark matter dilution scenarios in the early universe
Authors:
Arnab Chaudhuri,
Maxim Yu. Khlopov
Abstract:
When the vacuum like energy of the Higgs potential within the standard model undergoes electroweak phase transition, an influx of entropy into the primordial plasma can lead to a significant dilution of frozen out dark matter density that was already present before the onset of the phase transition. The same effect can take place, if the early Universe was dominated by primordial black holes of sm…
▽ More
When the vacuum like energy of the Higgs potential within the standard model undergoes electroweak phase transition, an influx of entropy into the primordial plasma can lead to a significant dilution of frozen out dark matter density that was already present before the onset of the phase transition. The same effect can take place, if the early Universe was dominated by primordial black holes of small mass, evaporating before the period of Big Bang Nucleosynthesis. In this paper we calculate the dilution factor for the above mentioned scenarios.
△ Less
Submitted 29 October, 2021;
originally announced October 2021.
-
Particle dark matter density and entropy production in the early universe
Authors:
Arnab Chaudhuri,
Maxim Yu. Khlopov,
Shiladitya Porey
Abstract:
Dark Matter (DM) density is reduced if entropy production takes place after DM particles abundance is frozen out in the early universe. We study a possibility of such reduction due to entropy production in the electroweak phase transition (EWPT). We compare scenarios of entropy production in the standard model (SM) and its simplest extension, the two-Higgs doublet model (2HDM). Assuming the EWPT i…
▽ More
Dark Matter (DM) density is reduced if entropy production takes place after DM particles abundance is frozen out in the early universe. We study a possibility of such reduction due to entropy production in the electroweak phase transition (EWPT). We compare scenarios of entropy production in the standard model (SM) and its simplest extension, the two-Higgs doublet model (2HDM). Assuming the EWPT is of second order in the SM scenario and the first order in the 2HDM, we calculate the entropy release in these scenarios and the corresponding dilution of preexisting DM density in the early universe. We find the effect of dilution in EWPT significant for confrontation with observations of any form of possible DM (including primordial black holes (PBHs)), which is frozen out, decoupled, frozen in, or formed before EWPT.
△ Less
Submitted 30 January, 2022; v1 submitted 27 October, 2021;
originally announced October 2021.
-
Charge asymmetry of new stable quarks in baryon asymmetrical Universe
Authors:
Arnab Chaudhuri,
Maxim Yu. Khlopov
Abstract:
Effects of electroweak phase transition (EWPT) in balance between baryon excess and the excess of stable quarks of new generation is studied. With the conservation of SU(2) symmetry and other quantum numbers, it makes possible sphaleron transitions between baryons, leptons and new of leptons and quarks. A definite relationship between the excess relative to baryon asymmetry is established. In pass…
▽ More
Effects of electroweak phase transition (EWPT) in balance between baryon excess and the excess of stable quarks of new generation is studied. With the conservation of SU(2) symmetry and other quantum numbers, it makes possible sphaleron transitions between baryons, leptons and new of leptons and quarks. A definite relationship between the excess relative to baryon asymmetry is established. In passing by we also show the small, yet negligible dilution in the pre-existing dark matter density due the sphaleron transition.
△ Less
Submitted 19 October, 2021;
originally announced October 2021.
-
B-Splines
Authors:
Arindam Chaudhuri
Abstract:
BSplines are one of the most promising curves in computer graphics. They are blessed with some superior geometric properties which make them an ideal candidate for several applications in computer aided design industry. In this article, some basic properties of B-Spline curves are presented. Two significant B-Spline properties viz convex hull property and repeated points effects are discussed. The…
▽ More
BSplines are one of the most promising curves in computer graphics. They are blessed with some superior geometric properties which make them an ideal candidate for several applications in computer aided design industry. In this article, some basic properties of B-Spline curves are presented. Two significant B-Spline properties viz convex hull property and repeated points effects are discussed. The BSplines computation in computational devices is also illustrated. An industry application based on image processing where B-Spline curve reconstructs the 3D surfaces for CT image datasets of inner organs further highlights the strength of these curves
△ Less
Submitted 14 August, 2021;
originally announced August 2021.
-
Atiyah sequence and Gauge transformations of a principal $2$-bundle over a Lie groupoid
Authors:
Saikat Chatterjee,
Adittya Chaudhuri,
Praphulla Koushik
Abstract:
In this paper, a notion of a principal $2$-bundle over a Lie groupoid has been introduced. For such principal $2$-bundles, we produced a short exact sequence of VB-groupoids, namely, the Atiyah sequence. Two notions of connection structures viz. strict connections and semi-strict connections on a principal $2$-bundle arising respectively, from a retraction of the Atiyah sequence and a retraction u…
▽ More
In this paper, a notion of a principal $2$-bundle over a Lie groupoid has been introduced. For such principal $2$-bundles, we produced a short exact sequence of VB-groupoids, namely, the Atiyah sequence. Two notions of connection structures viz. strict connections and semi-strict connections on a principal $2$-bundle arising respectively, from a retraction of the Atiyah sequence and a retraction up to a natural isomorphism have been introduced. We constructed a class of principal $\mathbb{G}=[G_1\rightrightarrows G_0]$-bundles and connections from a given principal $G_0$-bundle $E_0\rightarrow X_0$ over $[X_1\rightrightarrows X_0]$ with connection. An existence criterion for the connections on a principal $2$-bundle over a proper, étale Lie groupoid is proposed. The action of the $2$-group of gauge transformations on the category of strict and semi-strict connections has been studied. Finally we noted an extended symmetry of the category of semi-strict connections.
△ Less
Submitted 4 August, 2021; v1 submitted 29 July, 2021;
originally announced July 2021.