Arrow Research search

Author name cluster

Soumya Ray

Possible papers associated with this exact author name in Arrow. This page groups case-insensitive exact name matches and is not a full identity disambiguation profile.

21 papers
2 author rows

Possible papers

21

AAAI Conference 2018 Conference Paper

Gesture Annotation With a Visual Search Engine for Multimodal Communication Research

  • Sergiy Turchyn
  • Inés Olza Moreno
  • Cristóbal Pagán Cánovas
  • Francis Steen
  • Mark Turner
  • Javier Valenzuela
  • Soumya Ray

Human communication is multimodal and includes elements such as gesture and facial expression along with spoken language. Modern technology makes it feasible to capture all such aspects of communication in natural settings. As a result, similar to fields such as genetics, astronomy and neuroscience, scholars in areas such as linguistics and communication studies are on the verge of a data-driven revolution in their fields. These new approaches require analytical support from machine learning and artificial intelligence to develop tools to help process the vast data repositories. The Distributed Little Red Hen Lab project is an international team of interdisciplinary researchers building a large-scale infrastructure for data-driven multimodal communications research. In this paper, we describe a machine learning system developed to automatically annotate a large database of television program videos as part of this project. The annotations mark regions where people or speakers are on screen along with body part motions including head, hand and shoulder motion. We also annotate a specific class of gestures known as timeline gestures. An existing gesture annotation tool, ELAN, can be used with these annotations to quickly locate gestures of interest. Finally, we provide an update mechanism for the system based on human feedback. We empirically evaluate the accuracy of the system as well as present data from pilot human studies to show its effectiveness at aiding gesture scholars in their work.

PDF Details

IJCAI Conference 2016 Conference Paper

A Unifying Framework for Learning Bag Labels from Generalized Multiple-Instance Data

  • Gary Doran
  • Andrew Latham
  • Soumya Ray

We study the problem of bag-level classification from generalized multiple-instance (GMI) data. GMI learning is an extension of the popular multiple-instance setting. In GMI data, bags are labeled positive if they contain instances of certain types, and avoid instances of other types. For example, an image of a "sunny beach"' should contain sand and sea, but not clouds. We formulate a novel generative process for the GMI setting in which bags are distributions over instances. In this model, we show that a broad class of distribution-distance kernels is sufficient to represent arbitrary GMI concepts. Further, we show that a variety of previously proposed kernel approaches to the standard MI and GMI settings can be unified under the distribution kernel framework. We perform an extensive empirical study which indicates that the family of distribution distance kernels is accurate for a wide variety of real-world MI and GMI tasks as well as efficient when compared to a large set of baselines. Our theoretical and empirical results indicate that distribution-distance kernels can serve as a unifying framework for learning bag labels from GMI (and therefore MI) problems.

PDF Details

JMLR Journal 2016 Journal Article

Multiple-Instance Learning from Distributions

  • Gary Doran
  • Soumya Ray

We propose a new theoretical framework for analyzing the multiple-instance learning (MIL) setting. In MIL, training examples are provided to a learning algorithm in the form of labeled sets, or "bags," of instances. Applications of MIL include 3-D quantitative structure--activity relationship prediction for drug discovery and content-based image retrieval for web search. The goal of an algorithm is to learn a function that correctly labels new bags or a function that correctly labels new instances. We propose that bags should be treated as latent distributions from which samples are observed. We show that it is possible to learn accurate instance- and bag-labeling functions in this setting as well as functions that correctly rank bags or instances under weak assumptions. Additionally, our theoretical results suggest that it is possible to learn to rank efficiently using traditional, well-studied "supervised" learning approaches. We perform an extensive empirical evaluation that supports the theoretical predictions entailed by the new framework. The proposed theoretical framework leads to a better understanding of the relationship between the MI and standard supervised learning settings, and it provides new methods for learning from MI data that are more accurate, more efficient, and have better understood theoretical properties than existing MI-specific algorithms. [abs] [ pdf ][ bib ] &copy JMLR 2016. ( edit, beta )

PDF Details

ICML Conference 2016 Conference Paper

On the Consistency of Feature Selection With Lasso for Non-linear Targets

  • Yue Zhang
  • Weihong Guo
  • Soumya Ray

An important question in feature selection is whether a selection strategy recovers the “true” set of features, given enough data. We study this question in the context of the popular Least Absolute Shrinkage and Selection Operator (Lasso) feature selection strategy. In particular, we consider the scenario when the model is misspecified so that the learned model is linear while the underlying real target is nonlinear. Surprisingly, we prove that under certain conditions, Lasso is still able to recover the correct features in this case. We also carry out numerical studies to empirically verify the theoretical results and explore the necessity of the conditions under which the proof holds.

Details

AAAI Conference 2014 Conference Paper

Learning Instance Concepts from Multiple-Instance Data with Bags as Distributions

  • Gary Doran
  • Soumya Ray

We analyze and evaluate a generative process for multipleinstance learning (MIL) in which bags are distributions over instances. We show that our generative process contains as special cases generative models explored in prior work, while excluding scenarios known to be hard for MIL. Further, under the mild assumption that every negative instance is observed with nonzero probability in some negative bag, we show that it is possible to learn concepts that accurately label instances from MI data in this setting. Finally, we show that standard supervised approaches can learn concepts with low area-under-ROC error from MI data in this setting. We validate this surprising result with experiments using several synthetic and real-world MI datasets that have been annotated with instance labels.

PDF Details

AAAI Conference 2013 Conference Paper

SMILe: Shuffled Multiple-Instance Learning

  • Gary Doran
  • Soumya Ray

Resampling techniques such as bagging are often used in supervised learning to produce more accurate classifiers. In this work, we show that multiple-instance learning admits a different form of resampling, which we call “shuffling. ” In shuffling, we resample instances in such a way that the resulting bags are likely to be correctly labeled. We show that resampling results in both a reduction of bag label noise and a propagation of additional informative constraints to a multiple-instance classifier. We empirically evaluate shuffling in the context of multiple-instance classification and multiple-instance active learning and show that the approach leads to significant improvements in accuracy.

PDF Details

NeurIPS Conference 2012 Conference Paper

Bayesian Hierarchical Reinforcement Learning

  • Feng Cao
  • Soumya Ray

We describe an approach to incorporating Bayesian priors in the maxq framework for hierarchical reinforcement learning (HRL). We define priors on the primitive environment model and on task pseudo-rewards. Since models for composite tasks can be complex, we use a mixed model-based/model-free learning approach to find an optimal hierarchical policy. We show empirically that (i) our approach results in improved convergence over non-Bayesian baselines, given sensible priors, (ii) task hierarchies and Bayesian priors can be complementary sources of information, and using both sources is better than either alone, (iii) taking advantage of the structural decomposition induced by the task hierarchy significantly reduces the computational cost of Bayesian reinforcement learning and (iv) in this framework, task pseudo-rewards can be learned instead of being manually specified, leading to automatic learning of hierarchically optimal rather than recursively optimal policies.

PDF Details

JMLR Journal 2009 Journal Article

Exploiting Product Distributions to Identify Relevant Variables of Correlation Immune Functions

  • Lisa Hellerstein
  • Bernard Rosell
  • Eric Bach
  • Soumya Ray
  • David Page

A Boolean function f is correlation immune if each input variable is independent of the output, under the uniform distribution on inputs. For example, the parity function is correlation immune. We consider the problem of identifying relevant variables of a correlation immune function, in the presence of irrelevant variables. We address this problem in two different contexts. First, we analyze Skewing, a heuristic method that was developed to improve the ability of greedy decision tree algorithms to identify relevant variables of correlation immune Boolean functions, given examples drawn from the uniform distribution (Page and Ray, 2003). We present theoretical results revealing both the capabilities and limitations of skewing. Second, we explore the problem of identifying relevant variables in the Product Distribution Choice (PDC) learning model, a model in which the learner can choose product distributions and obtain examples from them. We prove a lemma establishing a property of Boolean functions that may be of independent interest. Using this lemma, we give two new algorithms for finding relevant variables of correlation immune functions in the PDC model. [abs] [ pdf ][ bib ] &copy JMLR 2009. ( edit, beta )

PDF Details

UAI Conference 2007 Conference Paper

Learning Bayesian Network Structure from Correlation-Immune Data

  • Eric Lantz
  • Soumya Ray
  • David Page

Searching the complete space of possible Bayesian networks is intractable for problems of interesting size, so Bayesian network structure learning algorithms, such as the commonly used Sparse Candidate algorithm, employ heuristics. However, these heuristics also restrict the types of relationships that can be learned exclusively from data. They are unable to learn relationships that exhibit "correlation-immunity", such as parity. To learn Bayesian networks in the presence of correlation-immune relationships, we extend the Sparse Candidate algorithm with a technique called "skewing". This technique uses the observation that relationships that are correlation-immune under a specific input distribution may not be correlation-immune under another, sufficiently different distribution. We show that by extending Sparse Candidate with this technique we are able to discover relationships between random variables that are approximately correlation-immune, with a significantly lower computational cost than the alternative of considering multiple parents of a node at a time.

Details

ICML Conference 2007 Conference Paper

Multi-task reinforcement learning: a hierarchical Bayesian approach

  • Aaron Wilson
  • Alan Fern
  • Soumya Ray
  • Prasad Tadepalli

We consider the problem of multi-task reinforcement learning, where the agent needs to solve a sequence of Markov Decision Processes (MDPs) chosen randomly from a fixed but unknown distribution. We model the distribution over MDPs using a hierarchical Bayesian infinite mixture model. For each novel MDP, we use the previously learned distribution as an informed prior for modelbased Bayesian reinforcement learning. The hierarchical Bayesian framework provides a strong prior that allows us to rapidly infer the characteristics of new environments based on previous environments, while the use of a nonparametric model allows us to quickly adapt to environments we have not encountered before. In addition, the use of infinite mixtures allows for the model to automatically learn the number of underlying MDP components. We evaluate our approach and show that it leads to significant speedups in convergence to an optimal policy after observing only a small number of tasks.

Details

NeurIPS Conference 2007 Conference Paper

Multiple-Instance Active Learning

  • Burr Settles
  • Mark Craven
  • Soumya Ray

In a multiple instance (MI) learning problem, instances are naturally organized into bags and it is the bags, instead of individual instances, that are labeled for training. MI learners assume that every instance in a bag labeled negative is actually negative, whereas at least one instance in a bag labeled positive is actually positive. We present a framework for active learning in the multiple-instance setting. In particular, we consider the case in which an MI learner is allowed to selectively query unlabeled instances in positive bags. This approach is well motivated in domains in which it is inexpensive to acquire bag labels and possible, but expensive, to acquire instance labels. We describe a method for learning from labels at mixed levels of granularity, and introduce two active query selection strategies motivated by the MI setting. Our experiments show that learning from instance labels can significantly improve performance of a basic MI learning algorithm in two multiple-instance domains: content-based image recognition and text classification.

PDF Details

ICAPS Conference 2007 Conference Paper

Online Planning for Resource Production in Real-Time Strategy Games

  • Hei Chan
  • Alan Fern
  • Soumya Ray
  • Nick Wilson
  • Chris Ventura

Planning in domains with temporal and numerical properties is an important research problem. One application of this is the resource production problem in real-time strategy (RTS) games, where players attempt to achieve the goal of producing a certain amount of resources as fast as possible. In this paper, we develop an online planner for resource production in the RTS game of Wargus, where the preconditions and effects of the actions obey many properties that are common across RTS games. Our planner is based on a computationally efficient action-selection mechanism, which at each decision epoch creates a possibly sub-optimal concurrent plan from the current state to the goal and then begins executing the initial set of actions. The plan is formed via a combination of means-ends analysis, scheduling, and a bounded search over sub-goals that are not required for goal achievement but may improve makespan. Experiments in the RTS game of Wargus show that the online planner is highly competitive with a human expert and often performs significantly better than state-of-the-art planning algorithms for this domain.

Details

IJCAI Conference 2003 Conference Paper

Hierarchical Hidden Markov Models for Information Extraction

  • Marios Skounakis
  • Mark Craven
  • Soumya Ray

Information extraction can be defined as the task of automatically extracting instances of specified classes or relations from text. We consider the case of using machine learning methods to induce models for extracting relation instances from biomedical articles. We propose and evaluate an approach that is based on using hierarchical hidden Markov models to represent the grammatical structure of the sentences being processed. Our approach first uses a shallow parser to construct a multi-level representation of each sentence being processed. Then we train hierarchical H M M s to capture the regularities of the parses for both positive and negative sentences. We evaluate our method by inducing models to extract binary relations in three biomedical domains. Our experiments indicate that our approach results in more accurate models than several baseline H M M approaches.

PDF

IJCAI Conference 2003 Conference Paper

Skewing: An Efficient Alternative to Lookahead for Decision Tree Induction

  • David Page
  • Soumya Ray

This paper presents a novel, promising approach that allows greedy decision tree induction algorithms to handle problematic functions such as parity functions. Lookahead is the standard approach to addressing difficult functions for greedy decision tree learners. Nevertheless, this approach is limited to very small problematic functions or subfunctions (2 or 3 variables), because the time complexity grows more than exponentially with the depth of lookahead. In contrast, the approach presented in this paper carries only a constant run-time penalty. Experiments indicate that the approach is effective with only modest amounts of data for problematic functions or subfunctions of up to six or seven variables, where the examples themselves may contain numerous other (irrelevant) variables as well.

PDF