Showing 1–29 of 29 results for author: Feldman, J

Reasoning or Simply Next Token Prediction? A Benchmark for Stress-Testing Large Language Models

Authors: Wentian Wang, Paul Kantor, Jacob Feldman, Lazaros Gallos, Hao Wang

Abstract: We propose MMLU-SR, a novel dataset designed to measure the true comprehension abilities of Large Language Models (LLMs) by challenging their performance in question-answering tasks with modified terms. We reasoned that an agent that ``truly'' understands a concept can still evaluate it when key terms are replaced by suitably defined alternate terms, and sought to differentiate such comprehension… ▽ More We propose MMLU-SR, a novel dataset designed to measure the true comprehension abilities of Large Language Models (LLMs) by challenging their performance in question-answering tasks with modified terms. We reasoned that an agent that ``truly'' understands a concept can still evaluate it when key terms are replaced by suitably defined alternate terms, and sought to differentiate such comprehension from mere text replacement. In our study, we modified standardized test questions by replacing a key term with a dummy word along with its definition. The key term could be in the context of questions, answers, or both questions and answers. Notwithstanding the high scores achieved by recent popular LLMs on the MMLU leaderboard, we found a substantial reduction in model performance after such replacement, suggesting poor comprehension. This new benchmark provides a rigorous benchmark for testing true model comprehension, and poses a challenge to the broader scientific community. △ Less

Submitted 15 June, 2024; originally announced June 2024.

Emotion Granularity from Text: An Aggregate-Level Indicator of Mental Health

Authors: Krishnapriya Vishnubhotla, Daniela Teodorescu, Mallory J. Feldman, Kristen A. Lindquist, Saif M. Mohammad

Abstract: We are united in how emotions are central to shaping our experiences; and yet, individuals differ greatly in how we each identify, categorize, and express emotions. In psychology, variation in the ability of individuals to differentiate between emotion concepts is called emotion granularity (determined through self-reports of one's emotions). High emotion granularity has been linked with better me… ▽ More We are united in how emotions are central to shaping our experiences; and yet, individuals differ greatly in how we each identify, categorize, and express emotions. In psychology, variation in the ability of individuals to differentiate between emotion concepts is called emotion granularity (determined through self-reports of one's emotions). High emotion granularity has been linked with better mental and physical health; whereas low emotion granularity has been linked with maladaptive emotion regulation strategies and poor health outcomes. In this work, we propose computational measures of emotion granularity derived from temporally-ordered speaker utterances in social media (in lieu of self-reports that suffer from various biases). We then investigate the effectiveness of such text-derived measures of emotion granularity in functioning as markers of various mental health conditions (MHCs). We establish baseline measures of emotion granularity derived from textual utterances, and show that, at an aggregate level, emotion granularities are significantly lower for people self-reporting as having an MHC than for the control population. This paves the way towards a better understanding of the MHCs, and specifically the role emotions play in our well-being. △ Less

Submitted 4 March, 2024; originally announced March 2024.

Comments: 9 pages plus appendices

Stretch with Stretch: Physical Therapy Exercise Games Led by a Mobile Manipulator

Authors: Matthew Lamsey, You Liang Tan, Meredith D. Wells, Madeline Beatty, Zexuan Liu, Arjun Majumdar, Kendra Washington, Jerry Feldman, Naveen Kuppuswamy, Elizabeth Nguyen, Arielle Wallenstein, Madeleine E. Hackney, Charles C. Kemp

Abstract: Physical therapy (PT) is a key component of many rehabilitation regimens, such as treatments for Parkinson's disease (PD). However, there are shortages of physical therapists and adherence to self-guided PT is low. Robots have the potential to support physical therapists and increase adherence to self-guided PT, but prior robotic systems have been large and immobile, which can be a barrier to use… ▽ More Physical therapy (PT) is a key component of many rehabilitation regimens, such as treatments for Parkinson's disease (PD). However, there are shortages of physical therapists and adherence to self-guided PT is low. Robots have the potential to support physical therapists and increase adherence to self-guided PT, but prior robotic systems have been large and immobile, which can be a barrier to use in homes and clinics. We present Stretch with Stretch (SWS), a novel robotic system for leading stretching exercise games for older adults with PD. SWS consists of a compact and lightweight mobile manipulator (Hello Robot Stretch RE1) that visually and verbally guides users through PT exercises. The robot's soft end effector serves as a target that users repetitively reach towards and press with a hand, foot, or knee. For each exercise, target locations are customized for the individual via a visually estimated kinematic model, a haptically estimated range of motion, and the person's exercise performance. The system includes sound effects and verbal feedback from the robot to keep users engaged throughout a session and augment physical exercise with cognitive exercise. We conducted a user study for which people with PD (n=10) performed 6 exercises with the system. Participants perceived the SWS to be useful and easy to use. They also reported mild to moderate perceived exertion (RPE). △ Less

Submitted 21 December, 2023; v1 submitted 20 December, 2023; originally announced December 2023.

Comments: This work has been submitted to the IEEE for possible publication. Copyright may be transferred without notice, after which this version may no longer be accessible

Socially Cognizant Robotics for a Technology Enhanced Society

Authors: Kristin J. Dana, Clinton Andrews, Kostas Bekris, Jacob Feldman, Matthew Stone, Pernille Hemmer, Aaron Mazzeo, Hal Salzman, Jingang Yi

Abstract: Emerging applications of robotics, and concerns about their impact, require the research community to put human-centric objectives front-and-center. To meet this challenge, we advocate an interdisciplinary approach, socially cognizant robotics, which synthesizes technical and social science methods. We argue that this approach follows from the need to empower stakeholder participation (from synchr… ▽ More Emerging applications of robotics, and concerns about their impact, require the research community to put human-centric objectives front-and-center. To meet this challenge, we advocate an interdisciplinary approach, socially cognizant robotics, which synthesizes technical and social science methods. We argue that this approach follows from the need to empower stakeholder participation (from synchronous human feedback to asynchronous societal assessment) in shaping AI-driven robot behavior at all levels, and leads to a range of novel research perspectives and problems both for improving robots' interactions with individuals and impacts on society. Drawing on these arguments, we develop best practices for socially cognizant robot design that balance traditional technology-based metrics (e.g. efficiency, precision and accuracy) with critically important, albeit challenging to measure, human and society-based metrics. △ Less

Submitted 27 October, 2023; originally announced October 2023.

A Novel Site-Agnostic Multimodal Deep Learning Model to Identify Pro-Eating Disorder Content on Social Media

Authors: Jonathan Feldman

Abstract: Over the last decade, there has been a vast increase in eating disorder diagnoses and eating disorder-attributed deaths, reaching their zenith during the Covid-19 pandemic. This immense growth derived in part from the stressors of the pandemic but also from increased exposure to social media, which is rife with content that promotes eating disorders. This study aimed to create a multimodal deep le… ▽ More Over the last decade, there has been a vast increase in eating disorder diagnoses and eating disorder-attributed deaths, reaching their zenith during the Covid-19 pandemic. This immense growth derived in part from the stressors of the pandemic but also from increased exposure to social media, which is rife with content that promotes eating disorders. This study aimed to create a multimodal deep learning model that can determine if a given social media post promotes eating disorders based on a combination of visual and textual data. A labeled dataset of Tweets was collected from Twitter, recently rebranded as X, upon which twelve deep learning models were trained and evaluated. Based on model performance, the most effective deep learning model was the multimodal fusion of the RoBERTa natural language processing model and the MaxViT image classification model, attaining accuracy and F1 scores of 95.9% and 0.959, respectively. The RoBERTa and MaxViT fusion model, deployed to classify an unlabeled dataset of posts from the social media sites Tumblr and Reddit, generated results akin to those of previous research studies that did not employ artificial intelligence-based techniques, indicating that deep learning models can develop insights congruent to those of researchers. Additionally, the model was used to conduct a time-series analysis of yet unseen Tweets from eight Twitter hashtags, uncovering that, since 2014, the relative abundance of content that promotes eating disorders has decreased drastically within those communities. Despite this reduction, by 2018, content that promotes eating disorders had either stopped declining or increased in ampleness anew on those hashtags. △ Less

Submitted 5 November, 2023; v1 submitted 6 July, 2023; originally announced July 2023.

RevealED: Uncovering Pro-Eating Disorder Content on Twitter Using Deep Learning

Authors: Jonathan Feldman

Abstract: The Covid-19 pandemic induced a vast increase in adolescents diagnosed with eating disorders and hospitalized due to eating disorders. This immense growth stemmed partially from the stress of the pandemic but also from increased exposure to content that promotes eating disorders via social media, which, within the last decade, has become plagued by pro-eating disorder content. This study aimed to… ▽ More The Covid-19 pandemic induced a vast increase in adolescents diagnosed with eating disorders and hospitalized due to eating disorders. This immense growth stemmed partially from the stress of the pandemic but also from increased exposure to content that promotes eating disorders via social media, which, within the last decade, has become plagued by pro-eating disorder content. This study aimed to create a deep learning model capable of determining whether a given social media post promotes eating disorders based solely on image data. Tweets from hashtags that have been documented to promote eating disorders along with Tweets from unrelated hashtags were collected. After prepossessing, these images were labeled as either pro-eating disorder or not based on which Twitter hashtag they were scraped from. Several deep-learning models were trained on the scraped dataset and were evaluated based on their accuracy, F1 score, precision, and recall. Ultimately, the Vision Transformer model was determined to be the most accurate, attaining an F1 score of 0.877 and an accuracy of 86.7% on the test set. The model, which was applied to unlabeled Twitter image data scraped from "#selfie", uncovered seasonal fluctuations in the relative abundance of pro-eating disorder content, which reached its peak in the summertime. These fluctuations correspond not only to the seasons, but also to stressors, such as the Covid-19 pandemic. Moreover, the Twitter image data indicated that the relative amount of pro-eating disorder content has been steadily rising over the last five years and is likely to continue increasing in the future. △ Less

Submitted 1 January, 2023; v1 submitted 28 December, 2022; originally announced December 2022.

On the Evolution of Subjective Experience

Authors: Jerome A. Feldman

Abstract: Subjective Experience (SE) is part of the ancient mind-body problem, which continues to be one of deepest mysteries of science. Despite major advances in many fields, there is still no plausible causal link between SE and its realization in the body. The core issue is the incompatibility of objective (3rd person) public science with subjective (1st person) private experience. Any scientific approa… ▽ More Subjective Experience (SE) is part of the ancient mind-body problem, which continues to be one of deepest mysteries of science. Despite major advances in many fields, there is still no plausible causal link between SE and its realization in the body. The core issue is the incompatibility of objective (3rd person) public science with subjective (1st person) private experience. Any scientific approach to SE assumes that it arose from extended evolutionary processes and that examining evolutionary history should help us understand it. While the core mystery remains, converging evidence from theoretical, experimental, and computational studies yields strong constraints on SE and some suggestions for further research. All animals confront many of the same fitness challenges. They all need some kind of internal model to relate their life goals and actionable sensed information to action. We understand the evolution of the bodily aspects of human perception and emotion, but not the SE. The first evolutionary evidence for SE appears in vertebrates and much of its neural substrate and simulation mechanism is preserved in mammals and humans. People exhibit the same phenomena, but there are remaining mysteries of everyday experience that are demonstrably incompatible with current neuroscience. In spite of this limitation, there is considerable progress on understanding the role of SE in the success of prostheses. △ Less

Submitted 25 March, 2022; v1 submitted 18 August, 2020; originally announced August 2020.

Comments: 49 pages 5 figures. This 7/22/2021 version preserves all the content of the previous version and adds additional discussion (in italics). It also includes several new references to connect with current literature. A companion arXiv article has also been updated

Cross-layer Design and SDR Implementation of DSA, Backpressure Routing and Network Coding

Authors: Sohraab Soltani, Yalin E. Sagduyu, Sean Scanlon, Yi Shi, Jason H. Li, Jared Feldman, John D. Matyjas

Abstract: A cross-layer cognitive radio system is designed to support unicast and multicast traffic with integration of dynamic spectrum access (DSA), backpressure algorithm, and network coding for multi-hop networking. The full protocol stack that operates with distributed coordination and local information exchange is implemented with software-defined radios (SDRs) and assessed in a realistic test and eva… ▽ More A cross-layer cognitive radio system is designed to support unicast and multicast traffic with integration of dynamic spectrum access (DSA), backpressure algorithm, and network coding for multi-hop networking. The full protocol stack that operates with distributed coordination and local information exchange is implemented with software-defined radios (SDRs) and assessed in a realistic test and evaluation (T\&E) system based on a network emulation testbed. Without a common control channel, each SDR performs neighborhood discovery, spectrum sensing and channel estimation, and executes a distributed extension of backpressure algorithm that optimizes the spectrum utility (that represents link rates and traffic congestion) with joint DSA and routing. The backpressure algorithm is extended to support multicast traffic with network coding deployed over virtual queues (for multicast destinations). In addition to full rank decoding at destinations, rank deficient decoding is also considered to reduce the delay. Cognitive network functionalities are programmed with GNU Radio and Python modules are developed for different layers. USRP radios are used as RF front ends. A wireless network T\&E system is presented to execute emulation tests, where radios communicate with each other through a wireless network emulator that controls physical channels according to path loss, fading, and topology effects. Emulation tests are presented for different topologies to evaluate the throughput, backlog and energy consumption. Results verify the SDR implementation and the joint effect of DSA, backpressure routing and network coding under realistic channel and radio hardware effects. △ Less

Submitted 11 December, 2019; originally announced December 2019.

Commonsense Knowledge Mining from Pretrained Models

Authors: Joshua Feldman, Joe Davison, Alexander M. Rush

Abstract: Inferring commonsense knowledge is a key challenge in natural language processing, but due to the sparsity of training data, previous work has shown that supervised methods for commonsense knowledge mining underperform when evaluated on novel data. In this work, we develop a method for generating commonsense knowledge using a large, pre-trained bidirectional language model. By transforming relatio… ▽ More Inferring commonsense knowledge is a key challenge in natural language processing, but due to the sparsity of training data, previous work has shown that supervised methods for commonsense knowledge mining underperform when evaluated on novel data. In this work, we develop a method for generating commonsense knowledge using a large, pre-trained bidirectional language model. By transforming relational triples into masked sentences, we can use this model to rank a triple's validity by the estimated pointwise mutual information between the two entities. Since we do not update the weights of the bidirectional model, our approach is not biased by the coverage of any one commonsense knowledge base. Though this method performs worse on a test set than models explicitly trained on a corresponding training set, it outperforms these methods when mining commonsense knowledge from new sources, suggesting that unsupervised techniques may generalize better than current supervised approaches. △ Less

Submitted 1 September, 2019; originally announced September 2019.

High Degree Vertices and Spread of Infections in Spatially Modelled Social Networks

Authors: Joshua Feldman, Jeannette Janssen

Abstract: We examine how the behaviour of high degree vertices in a network affects whether an infection spreads through communities or jumps between them. We study two stochastic susceptible-infected-recovered (SIR) processes and represent our network with a spatial preferential attachment (SPA) network. In one of the two epidemic scenarios we adjust the contagiousness of high degree vertices so that they… ▽ More We examine how the behaviour of high degree vertices in a network affects whether an infection spreads through communities or jumps between them. We study two stochastic susceptible-infected-recovered (SIR) processes and represent our network with a spatial preferential attachment (SPA) network. In one of the two epidemic scenarios we adjust the contagiousness of high degree vertices so that they are less contagious. We show that, for this scenario, the infection travels through communities rather than jumps between them. We conjecture that this is not the case in the other scenario, when contagion is independent of the degree of the originating vertex. Our theoretical results and conjecture are supported by simulations. △ Less

Submitted 28 February, 2019; originally announced March 2019.

Journal ref: In: Bonato A., Chung Graham F., Prałat P. (eds) Algorithms and Models for the Web Graph. WAW 2017. LNCS 10519. Springer, pp. 60-74

Towards a Science of Mind

Authors: Jerome Feldman

Abstract: The ancient mind/body problem continues to be one of deepest mysteries of science and of the human spirit. Despite major advances in many fields, there is still no plausible link between subjective experience (qualia) and its realization in the body. This paper outlines some of the elements of a rigorous science of mind (SoM) - key ideas include scientific realism of mind, agnostic mysterianism, c… ▽ More The ancient mind/body problem continues to be one of deepest mysteries of science and of the human spirit. Despite major advances in many fields, there is still no plausible link between subjective experience (qualia) and its realization in the body. This paper outlines some of the elements of a rigorous science of mind (SoM) - key ideas include scientific realism of mind, agnostic mysterianism, careful attention to language, and a focus on concrete (touchstone) questions and results. A core suggestion is to focus effort on the (still mysterious) mapping from neural activity to subjective experience. △ Less

Submitted 29 July, 2019; v1 submitted 6 November, 2018; originally announced November 2018.

Comments: 18 pages and 1 Figure. The ancient mind/body remains a scientific and existential mystery. This article develops a methodology for an incremental Science of Mind and describes some ongoing prospects and successes. Updates include additional phenomena, including emotions,and several more references. A major addition is a postulated general (mysterious) brain-mind mapping

Processing Natural Language About Ongoing Actions

Authors: Steve Doubleday, Sean Trott, Jerome Feldman

Abstract: Actions may not proceed as planned; they may be interrupted, resumed or overridden. This is a challenge to handle in a natural language understanding system. We describe extensions to an existing implementation for the control of autonomous systems by natural language, to enable such systems to handle incoming language requests regarding actions. Language Communication with Autonomous Systems (LCA… ▽ More Actions may not proceed as planned; they may be interrupted, resumed or overridden. This is a challenge to handle in a natural language understanding system. We describe extensions to an existing implementation for the control of autonomous systems by natural language, to enable such systems to handle incoming language requests regarding actions. Language Communication with Autonomous Systems (LCAS) has been extended with support for X-nets, parameterized executable schemas representing actions. X-nets enable the system to control actions at a desired level of granularity, while providing a mechanism for language requests to be processed asynchronously. Standard semantics supported include requests to stop, continue, or override the existing action. The specific domain demonstrated is the control of motion of a simulated robot, but the approach is general, and could be applied to other domains. △ Less

Submitted 30 July, 2016; v1 submitted 22 July, 2016; originally announced July 2016.

Comments: 6 pages, 8 figures. Updated with PIPE citations

Mysteries of Visual Experience

Authors: Jerome Feldman

Abstract: Science is a crowning glory of the human spirit and its applications remain our best hope for social progress. But there are limitations to current science and perhaps to any science. The general mind-body problem is known to be intractable and currently mysterious. This is one of many deep problems that are universally agreed to be beyond the current purview of Science, including quantum phenomen… ▽ More Science is a crowning glory of the human spirit and its applications remain our best hope for social progress. But there are limitations to current science and perhaps to any science. The general mind-body problem is known to be intractable and currently mysterious. This is one of many deep problems that are universally agreed to be beyond the current purview of Science, including quantum phenomena, etc. But all of these famous unsolved problems are either remote from everyday experience (entanglement, dark matter) or are hard to even define sharply (phenomenology, consciousness, etc.). An updated summary of this work has been published as: Feldman, J. (2022). Computation, perception, and mind. Behavioral and Brain Sciences, 45, E48. doi:10.1017/S0140525X21001886 A more readable, open access, version is: https://escholarship.org/uc/item/6cs78450 △ Less

Submitted 25 March, 2022; v1 submitted 28 April, 2016; originally announced April 2016.

Comments: This 3/27/2022 revision retains all the original text but adds new comments (in italics) There are several new references, connecting with current research

Journal ref: Behavioral and Brain Sciences, 45, E48 (2022)

Exploiting Deep Semantics and Compositionality of Natural Language for Human-Robot-Interaction

Authors: Manfred Eppe, Sean Trott, Jerome Feldman

Abstract: We develop a natural language interface for human robot interaction that implements reasoning about deep semantics in natural language. To realize the required deep analysis, we employ methods from cognitive linguistics, namely the modular and compositional framework of Embodied Construction Grammar (ECG) [Feldman, 2009]. Using ECG, robots are able to solve fine-grained reference resolution proble… ▽ More We develop a natural language interface for human robot interaction that implements reasoning about deep semantics in natural language. To realize the required deep analysis, we employ methods from cognitive linguistics, namely the modular and compositional framework of Embodied Construction Grammar (ECG) [Feldman, 2009]. Using ECG, robots are able to solve fine-grained reference resolution problems and other issues related to deep semantics and compositionality of natural language. This also includes verbal interaction with humans to clarify commands and queries that are too ambiguous to be executed safely. We implement our NLU framework as a ROS package and present proof-of-concept scenarios with different robots, as well as a survey on the state of the art. △ Less

Submitted 22 April, 2016; originally announced April 2016.

The Role of Typicality in Object Classification: Improving The Generalization Capacity of Convolutional Neural Networks

Authors: Babak Saleh, Ahmed Elgammal, Jacob Feldman

Abstract: Deep artificial neural networks have made remarkable progress in different tasks in the field of computer vision. However, the empirical analysis of these models and investigation of their failure cases has received attention recently. In this work, we show that deep learning models cannot generalize to atypical images that are substantially different from training images. This is in contrast to t… ▽ More Deep artificial neural networks have made remarkable progress in different tasks in the field of computer vision. However, the empirical analysis of these models and investigation of their failure cases has received attention recently. In this work, we show that deep learning models cannot generalize to atypical images that are substantially different from training images. This is in contrast to the superior generalization ability of the visual system in the human brain. We focus on Convolutional Neural Networks (CNN) as the state-of-the-art models in object recognition and classification; investigate this problem in more detail, and hypothesize that training CNN models suffer from unstructured loss minimization. We propose computational models to improve the generalization capacity of CNNs by considering how typical a training image looks like. By conducting an extensive set of experiments we show that involving a typicality measure can improve the classification results on a new set of images by a large margin. More importantly, this significant improvement is achieved without fine-tuning the CNN model on the target image set. △ Less

Submitted 9 February, 2016; originally announced February 2016.

Comments: In Submission

Toward a Taxonomy and Computational Models of Abnormalities in Images

Authors: Babak Saleh, Ahmed Elgammal, Jacob Feldman, Ali Farhadi

Abstract: The human visual system can spot an abnormal image, and reason about what makes it strange. This task has not received enough attention in computer vision. In this paper we study various types of atypicalities in images in a more comprehensive way than has been done before. We propose a new dataset of abnormal images showing a wide range of atypicalities. We design human subject experiments to dis… ▽ More The human visual system can spot an abnormal image, and reason about what makes it strange. This task has not received enough attention in computer vision. In this paper we study various types of atypicalities in images in a more comprehensive way than has been done before. We propose a new dataset of abnormal images showing a wide range of atypicalities. We design human subject experiments to discover a coarse taxonomy of the reasons for abnormality. Our experiments reveal three major categories of abnormality: object-centric, scene-centric, and contextual. Based on this taxonomy, we propose a comprehensive computational model that can predict all different types of abnormality in images and outperform prior arts in abnormality recognition. △ Less

Submitted 4 December, 2015; originally announced December 2015.

Comments: To appear in the Thirtieth AAAI Conference on Artificial Intelligence (AAAI 2016)

Multiplicative Bidding in Online Advertising

Authors: MohammadHossein Bateni, Jon Feldman, Vahab Mirrokni, Sam Chiu-wai Wong

Abstract: In this paper, we initiate the study of the multiplicative bidding language adopted by major Internet search companies. In multiplicative bidding, the effective bid on a particular search auction is the product of a base bid and bid adjustments that are dependent on features of the search (for example, the geographic location of the user, or the platform on which the search is conducted). We consi… ▽ More In this paper, we initiate the study of the multiplicative bidding language adopted by major Internet search companies. In multiplicative bidding, the effective bid on a particular search auction is the product of a base bid and bid adjustments that are dependent on features of the search (for example, the geographic location of the user, or the platform on which the search is conducted). We consider the task faced by the advertiser when setting these bid adjustments, and establish a foundational optimization problem that captures the core difficulty of bidding under this language. We give matching algorithmic and approximation hardness results for this problem; these results are against an information-theoretic bound, and thus have implications on the power of the multiplicative bidding language itself. Inspired by empirical studies of search engine price data, we then codify the relevant restrictions of the problem, and give further algorithmic and hardness results. Our main technical contribution is an $O(\log n)$-approximation for the case of multiplicative prices and monotone values. We also provide empirical validations of our problem restrictions, and test our algorithms on real data against natural benchmarks. Our experiments show that they perform favorably compared with the baseline. △ Less

Submitted 27 April, 2014; originally announced April 2014.

Comments: 25 pages; accepted to EC'14

Yield Optimization of Display Advertising with Ad Exchange

Authors: Santiago Balseiro, Jon Feldman, Vahab Mirrokni, S. Muthukrishnan

Abstract: In light of the growing market of Ad Exchanges for the real-time sale of advertising slots, publishers face new challenges in choosing between the allocation of contract-based reservation ads and spot market ads. In this setting, the publisher should take into account the tradeoff between short-term revenue from an Ad Exchange and quality of allocating reservation ads. In this paper, we formalize… ▽ More In light of the growing market of Ad Exchanges for the real-time sale of advertising slots, publishers face new challenges in choosing between the allocation of contract-based reservation ads and spot market ads. In this setting, the publisher should take into account the tradeoff between short-term revenue from an Ad Exchange and quality of allocating reservation ads. In this paper, we formalize this combined optimization problem as a stochastic control problem and derive an efficient policy for online ad allocation in settings with general joint distribution over placement quality and exchange bids. We prove asymptotic optimality of this policy in terms of any trade-off between quality of delivered reservation ads and revenue from the exchange, and provide a rigorous bound for its convergence rate to the optimal policy. We also give experimental results on data derived from real publisher inventory, showing that our policy can achieve any pareto-optimal point on the quality vs. revenue curve. Finally, we study a parametric training-based algorithm in which instead of learning the dual variables from a sample data (as is done in non-parametric training-based algorithms), we learn the parameters of the distribution and construct those dual variables from the learned parameter values. We compare parametric and non-parametric ways to estimate from data both analytically and experimentally in the special case without the ad exchange, and show that though both methods converge to the optimal policy as the sample size grows, our parametric method converges faster, and thus performs better on smaller samples. △ Less

Submitted 21 September, 2012; v1 submitted 12 February, 2011; originally announced February 2011.

Online Stochastic Packing Applied to Display Ad Allocation

Authors: Jon Feldman, Monika Henzinger, Nitish Korula, Vahab S. Mirrokni, Cliff Stein

Abstract: Inspired by online ad allocation, we study online stochastic packing linear programs from theoretical and practical standpoints. We first present a near-optimal online algorithm for a general class of packing linear programs which model various online resource allocation problems including online variants of routing, ad allocations, generalized assignment, and combinatorial auctions. As our main… ▽ More Inspired by online ad allocation, we study online stochastic packing linear programs from theoretical and practical standpoints. We first present a near-optimal online algorithm for a general class of packing linear programs which model various online resource allocation problems including online variants of routing, ad allocations, generalized assignment, and combinatorial auctions. As our main theoretical result, we prove that a simple primal-dual training-based algorithm achieves a (1 - o(1))-approximation guarantee in the random order stochastic model. This is a significant improvement over logarithmic or constant-factor approximations for the adversarial variants of the same problems (e.g. factor 1 - 1/e for online ad allocation, and \log m for online routing). We then focus on the online display ad allocation problem and study the efficiency and fairness of various training-based and online allocation algorithms on data sets collected from real-life display ad allocation system. Our experimental evaluation confirms the effectiveness of training-based primal-dual algorithms on real data sets, and also indicate an intrinsic trade-off between fairness and efficiency. △ Less

Submitted 16 February, 2010; v1 submitted 27 January, 2010; originally announced January 2010.

Comments: 19 pages, 3 figures, 3 tables. The new version generalizes results to a broader class of packing problems, and gives additional applications

ACM Class: F.2.2

Online Stochastic Matching: Beating 1-1/e

Authors: Jon Feldman, Aranyak Mehta, Vahab Mirrokni, S. Muthukrishnan

Abstract: We study the online stochastic bipartite matching problem, in a form motivated by display ad allocation on the Internet. In the online, but adversarial case, the celebrated result of Karp, Vazirani and Vazirani gives an approximation ratio of $1-1/e$. In the online, stochastic case when nodes are drawn repeatedly from a known distribution, the greedy algorithm matches this approximation ratio, b… ▽ More We study the online stochastic bipartite matching problem, in a form motivated by display ad allocation on the Internet. In the online, but adversarial case, the celebrated result of Karp, Vazirani and Vazirani gives an approximation ratio of $1-1/e$. In the online, stochastic case when nodes are drawn repeatedly from a known distribution, the greedy algorithm matches this approximation ratio, but still, no algorithm is known that beats the $1 - 1/e$ bound. Our main result is a 0.67-approximation online algorithm for stochastic bipartite matching, breaking this $1 - {1/e}$ barrier. Furthermore, we show that no online algorithm can produce a $1-ε$ approximation for an arbitrarily small $ε$ for this problem. We employ a novel application of the idea of the power of two choices from load balancing: we compute two disjoint solutions to the expected instance, and use both of them in the online algorithm in a prescribed preference order. To identify these two disjoint solutions, we solve a max flow problem in a boosted flow graph, and then carefully decompose this maximum flow to two edge-disjoint (near-)matchings. These two offline solutions are used to characterize an upper bound for the optimum in any scenario. This is done by identifying a cut whose value we can bound under the arrival distribution. △ Less

Submitted 25 May, 2009; originally announced May 2009.

Algorithmic Methods for Sponsored Search Advertising

Authors: Jon Feldman, S. Muthukrishnan

Abstract: Modern commercial Internet search engines display advertisements along side the search results in response to user queries. Such sponsored search relies on market mechanisms to elicit prices for these advertisements, making use of an auction among advertisers who bid in order to have their ads shown for specific keywords. We present an overview of the current systems for such auctions and also d… ▽ More Modern commercial Internet search engines display advertisements along side the search results in response to user queries. Such sponsored search relies on market mechanisms to elicit prices for these advertisements, making use of an auction among advertisers who bid in order to have their ads shown for specific keywords. We present an overview of the current systems for such auctions and also describe the underlying game-theoretic aspects. The game involves three parties--advertisers, the search engine, and search users--and we present example research directions that emphasize the role of each. The algorithms for bidding and pricing in these games use techniques from three mathematical areas: mechanism design, optimization, and statistical estimation. Finally, we present some challenges in sponsored search advertising. △ Less

Submitted 12 May, 2008; originally announced May 2008.

Online Ad Slotting With Cancellations

Authors: Florin Constantin, Jon Feldman, S. Muthukrishnan, Martin Pal

Abstract: Many advertisers buy advertisements (ads) on the Internet or on traditional media and seek simple, online mechanisms to reserve ad slots in advance. Media publishers represent a vast and varying inventory, and they too seek automatic, online mechanisms for pricing and allocating such reservations. In this paper, we present and study a simple model for auctioning such ad slots in advance. Bidders… ▽ More Many advertisers buy advertisements (ads) on the Internet or on traditional media and seek simple, online mechanisms to reserve ad slots in advance. Media publishers represent a vast and varying inventory, and they too seek automatic, online mechanisms for pricing and allocating such reservations. In this paper, we present and study a simple model for auctioning such ad slots in advance. Bidders arrive sequentially and report which slots they are interested in. The seller must decide immediately whether or not to grant a reservation. Our model allows a seller to accept reservations, but possibly cancel the allocations later and pay the bidder a cancellation compensation (bump payment). Our main result is an online mechanism to derive prices and bump payments that is efficient to implement. This mechanism has many desirable properties. It is individually rational; winners have an incentive to be honest and bidding one's true value dominates any lower bid. Our mechanism's efficiency is within a constant fraction of the a posteriori optimally efficient solution. Its revenue is within a constant fraction of the a posteriori revenue of the Vickrey-Clarke-Groves mechanism. Our results make no assumptions about the order of arrival of bids or the value distribution of bidders and still hold if the items for sale are elements of a matroid, a more general setting than slot allocation. △ Less

Submitted 8 May, 2008; originally announced May 2008.

Comments: 10 pages, 1 figure

Sponsored Search Auctions with Markovian Users

Authors: Gagan Aggarwal, Jon Feldman, S. Muthukrishnan, Martin Pal

Abstract: Sponsored search involves running an auction among advertisers who bid in order to have their ad shown next to search results for specific keywords. Currently, the most popular auction for sponsored search is the "Generalized Second Price" (GSP) auction in which advertisers are assigned to slots in the decreasing order of their "score," which is defined as the product of their bid and click-thro… ▽ More Sponsored search involves running an auction among advertisers who bid in order to have their ad shown next to search results for specific keywords. Currently, the most popular auction for sponsored search is the "Generalized Second Price" (GSP) auction in which advertisers are assigned to slots in the decreasing order of their "score," which is defined as the product of their bid and click-through rate. In the past few years, there has been significant research on the game-theoretic issues that arise in an advertiser's interaction with the mechanism as well as possible redesigns of the mechanism, but this ranking order has remained standard. From a search engine's perspective, the fundamental question is: what is the best assignment of advertisers to slots? Here "best" could mean "maximizing user satisfaction," "most efficient," "revenue-maximizing," "simplest to interact with," or a combination of these. To answer this question we need to understand the behavior of a search engine user when she sees the displayed ads, since that defines the commodity the advertisers are bidding on, and its value. Most prior work has assumed that the probability of a user clicking on an ad is independent of the other ads shown on the page. We propose a simple Markovian user model that does not make this assumption. We then present an algorithm to determine the most efficient assignment under this model, which turns out to be different than that of GSP. A truthful auction then follows from an application of the Vickrey-Clarke-Groves (VCG) mechanism. Further, we show that our assignment has many of the desirable properties of GSP that makes bidding intuitive. At the technical core of our result are a number of insights about the structure of the optimal assignment. △ Less

Submitted 6 May, 2008; originally announced May 2008.

A Truthful Mechanism for Offline Ad Slot Scheduling

Authors: Jon Feldman, S. Muthukrishnan, Evdokia Nikolova, Martin Pal

Abstract: We consider the "Offline Ad Slot Scheduling" problem, where advertisers must be scheduled to "sponsored search" slots during a given period of time. Advertisers specify a budget constraint, as well as a maximum cost per click, and may not be assigned to more than one slot for a particular search. We give a truthful mechanism under the utility model where bidders try to maximize their clicks, s… ▽ More We consider the "Offline Ad Slot Scheduling" problem, where advertisers must be scheduled to "sponsored search" slots during a given period of time. Advertisers specify a budget constraint, as well as a maximum cost per click, and may not be assigned to more than one slot for a particular search. We give a truthful mechanism under the utility model where bidders try to maximize their clicks, subject to their personal constraints. In addition, we show that the revenue-maximizing mechanism is not truthful, but has a Nash equilibrium whose outcome is identical to our mechanism. As far as we can tell, this is the first treatment of sponsored search that directly incorporates both multiple slots and budget constraints into an analysis of incentives. Our mechanism employs a descending-price auction that maintains a solution to a certain machine scheduling problem whose job lengths depend on the price, and hence is variable over the auction. The price stops when the set of bidders that can afford that price pack exactly into a block of ad slots, at which point the mechanism allocates that block and continues on the remaining slots. To prove our result on the equilibrium of the revenue-maximizing mechanism, we first show that a greedy algorithm suffices to solve the revenue-maximizing linear program; we then use this insight to prove that bidders allocated in the same block of our mechanism have no incentive to deviate from bidding the fixed price of that block. △ Less

Submitted 18 January, 2008; originally announced January 2008.

Budget Optimization in Search-Based Advertising Auctions

Authors: Jon Feldman, S. Muthukrishnan, Martin Pal, Cliff Stein

Abstract: Internet search companies sell advertisement slots based on users' search queries via an auction. While there has been a lot of attention on the auction process and its game-theoretic aspects, our focus is on the advertisers. In particular, the advertisers have to solve a complex optimization problem of how to place bids on the keywords of their interest so that they can maximize their return (t… ▽ More Internet search companies sell advertisement slots based on users' search queries via an auction. While there has been a lot of attention on the auction process and its game-theoretic aspects, our focus is on the advertisers. In particular, the advertisers have to solve a complex optimization problem of how to place bids on the keywords of their interest so that they can maximize their return (the number of user clicks on their ads) for a given budget. We model the entire process and study this budget optimization problem. While most variants are NP hard, we show, perhaps surprisingly, that simply randomizing between two uniform strategies that bid equally on all the keywords works well. More precisely, this strategy gets at least 1-1/e fraction of the maximum clicks possible. Such uniform strategies are likely to be practical. We also present inapproximability results, and optimal algorithms for variants of the budget optimization problem. △ Less

Submitted 8 December, 2006; originally announced December 2006.

On the Complexity of Processing Massive, Unordered, Distributed Data

Authors: Jon Feldman, S. Muthukrishnan, Anastasios Sidiropoulos, Cliff Stein, Zoya Svitkina

Abstract: An existing approach for dealing with massive data sets is to stream over the input in few passes and perform computations with sublinear resources. This method does not work for truly massive data where even making a single pass over the data with a processor is prohibitive. Successful log processing systems in practice such as Google's MapReduce and Apache's Hadoop use multiple machines. They… ▽ More An existing approach for dealing with massive data sets is to stream over the input in few passes and perform computations with sublinear resources. This method does not work for truly massive data where even making a single pass over the data with a processor is prohibitive. Successful log processing systems in practice such as Google's MapReduce and Apache's Hadoop use multiple machines. They efficiently perform a certain class of highly distributable computations defined by local computations that can be applied in any order to the input. Motivated by the success of these systems, we introduce a simple algorithmic model for massive, unordered, distributed (mud) computation. We initiate the study of understanding its computational complexity. Our main result is a positive one: any unordered function that can be computed by a streaming algorithm can also be computed with a mud algorithm, with comparable space and communication complexity. We extend this result to some useful classes of approximate and randomized streaming algorithms. We also give negative results, using communication complexity arguments to prove that extensions to private randomness, promise problems and indeterminate functions are impossible. We believe that the line of research we introduce in this paper has the potential for tremendous impact. The distributed systems that motivate our work successfully process data at an unprecedented scale, distributed over hundreds or even thousands of machines, and perform hundreds of such analyses each day. The mud model (and its generalizations) inspire a set of complexity-theoretic questions that lie at their heart. △ Less

Submitted 22 May, 2007; v1 submitted 21 November, 2006; originally announced November 2006.

PAC Learning Mixtures of Axis-Aligned Gaussians with No Separation Assumption

Authors: Jon Feldman, Ryan O'Donnell, Rocco A. Servedio

Abstract: We propose and analyze a new vantage point for the learning of mixtures of Gaussians: namely, the PAC-style model of learning probability distributions introduced by Kearns et al. Here the task is to construct a hypothesis mixture of Gaussians that is statistically indistinguishable from the actual mixture generating the data; specifically, the KL-divergence should be at most epsilon. In this… ▽ More We propose and analyze a new vantage point for the learning of mixtures of Gaussians: namely, the PAC-style model of learning probability distributions introduced by Kearns et al. Here the task is to construct a hypothesis mixture of Gaussians that is statistically indistinguishable from the actual mixture generating the data; specifically, the KL-divergence should be at most epsilon. In this scenario, we give a poly(n/epsilon)-time algorithm that learns the class of mixtures of any constant number of axis-aligned Gaussians in n-dimensional Euclidean space. Our algorithm makes no assumptions about the separation between the means of the Gaussians, nor does it have any dependence on the minimum mixing weight. This is in contrast to learning results known in the ``clustering'' model, where such assumptions are unavoidable. Our algorithm relies on the method of moments, and a subalgorithm developed in previous work by the authors (FOCS 2005) for a discrete mixture-learning problem. △ Less

Submitted 16 September, 2006; originally announced September 2006.

Comments: 19 pages; no figures; preliminary abridged version appeared in Proceedings of 19th Annual Conference on Learning Theory (COLT) 2006

Journal ref: Proceedings of 19th Annual Conference on Learning Theory (COLT), Pittsburgh, PA, pp. 20--34, 2006

Bidding to the Top: VCG and Equilibria of Position-Based Auctions

Authors: Gagan Aggarwal, S. Muthukrishnan, Jon Feldman

Abstract: Many popular search engines run an auction to determine the placement of advertisements next to search results. Current auctions at Google and Yahoo! let advertisers specify a single amount as their bid in the auction. This bid is interpreted as the maximum amount the advertiser is willing to pay per click on its ad. When search queries arrive, the bids are used to rank the ads linearly on the s… ▽ More Many popular search engines run an auction to determine the placement of advertisements next to search results. Current auctions at Google and Yahoo! let advertisers specify a single amount as their bid in the auction. This bid is interpreted as the maximum amount the advertiser is willing to pay per click on its ad. When search queries arrive, the bids are used to rank the ads linearly on the search result page. The advertisers pay for each user who clicks on their ad, and the amount charged depends on the bids of all the advertisers participating in the auction. In order to be effective, advertisers seek to be as high on the list as their budget permits, subject to the market. We study the problem of ranking ads and associated pricing mechanisms when the advertisers not only specify a bid, but additionally express their preference for positions in the list of ads. In particular, we study "prefix position auctions" where advertiser $i$ can specify that she is interested only in the top $b_i$ positions. We present a simple allocation and pricing mechanism that generalizes the desirable properties of current auctions that do not have position constraints. In addition, we show that our auction has an "envy-free" or "symmetric" Nash equilibrium with the same outcome in allocation and pricing as the well-known truthful Vickrey-Clarke-Groves (VCG) auction. Furthermore, we show that this equilibrium is the best such equilibrium for the advertisers in terms of the profit made by each advertiser. We also discuss other position-based auctions. △ Less

Submitted 26 July, 2006; originally announced July 2006.

The Benefit of Thresholding in LP Decoding of LDPC Codes

Authors: Jon Feldman, Ralf Koetter, Pascal O. Vontobel

Abstract: Consider data transmission over a binary-input additive white Gaussian noise channel using a binary low-density parity-check code. We ask the following question: Given a decoder that takes log-likelihood ratios as input, does it help to modify the log-likelihood ratios before decoding? If we use an optimal decoder then it is clear that modifying the log-likelihoods cannot possibly help the decod… ▽ More Consider data transmission over a binary-input additive white Gaussian noise channel using a binary low-density parity-check code. We ask the following question: Given a decoder that takes log-likelihood ratios as input, does it help to modify the log-likelihood ratios before decoding? If we use an optimal decoder then it is clear that modifying the log-likelihoods cannot possibly help the decoder's performance, and so the answer is "no." However, for a suboptimal decoder like the linear programming decoder, the answer might be "yes": In this paper we prove that for certain interesting classes of low-density parity-check codes and large enough SNRs, it is advantageous to truncate the log-likelihood ratios before passing them to the linear programming decoder. △ Less

Submitted 2 August, 2005; originally announced August 2005.

Comments: To appear in Proc. 2005 IEEE International Symposium on Information Theory, Adelaide, Australia, September 4-9, 2005