Corpus ID: 12642009

Active Sensing as Bayes-Optimal Sequential Decision Making

  title={Active Sensing as Bayes-Optimal Sequential Decision Making},
  author={Sheeraz Ahmad and Angela J. Yu},
Sensory inference under conditions of uncertainty is a major problem in both machine learning and computational neuroscience. An important but poorly understood aspect of sensory processing is the role of active sensing. Here, we present a Bayes-optimal inference and control framework for active sensing, C-DAC (Context-Dependent Active Controller). Unlike previously proposed algorithms that optimize abstract statistical objectives such as information maximization (Infomax) [Butko & Movellan… Expand
Context-sensitive active sensing in humans
Novel experimental results are provided that differentiate theoretical models for human active sensing, as well as a novel active sensing algorithm that retains the context-sensitivity of the optimal controller while achieving significant computational savings. Expand
Inverse Active Sensing: Modeling and Understanding Timely Decision-Making
This paper develops an expressive, unified framework for the general setting of evidence-based decision-making under endogenous, context-dependent time pressure---which requires negotiating (subjective) tradeoffs between accuracy, speediness, and cost of information. Expand
Adaptive Sampling of Information in Perceptual Decision-Making
The optimal decision procedure for two-alternative choice tasks in which the different options are sampled one at a time, sources vary in the quality of the information they provide, and the available time is fixed is derived. Expand
Efficient Resource Allocation for Sparse Multiple Object Tracking
This work proposes a resource constrained Partially Observable Markov Decision Process (POMDP) formulation that allows for real-time on-line planning of multiple target tracking and shows that directing the attentional focci to relevant image sub-regions allows for large detection speed-ups and improvements on tracking precision. Expand
Robust Decision Making with the Same-Decision Probability
  • S. Chen
  • Computer Science, Mathematics
  • 2015
This thesis shows how the SDP can be used to be an effective stopping criterion, and compares it to traditional criteria to demonstrate how it provides a fresh perspective in decision making under uncertainty, and develops the first exact algorithm to compute the SDR so that it may be used as a stopping criterion. Expand
Toward predictive machine learning for active vision
Though optimizing future posterior entropy over the actions set is shown enough to attain locally optimal action selection, offline calculation using class-specific saliency maps is shown better for it saves processing costs through saccades pathways pre-processing, with a negligible effect on the recognition/compression rates. Expand
Yu Lab Research Overview
    Our lab is interested in understanding computational principles and neural processes underlying the brain's marvelous ability to infer hidden causes, process a cacophony of inputs in a focusedExpand
    On the advantages of foveal mechanisms for active stereo systems in visual search tasks
    This work investigates an Upper Confidence Bound algorithm for the task of simultaneously finding the closest object to the observer and learning the surrounding environment 3D map (mapping) and proposes a probabilistic observational model for a stereo system that relies on the Unscented Transform. Expand
    Active Perception for Cyber Intrusion Detection and Defense
    • R. Goldman, M. Burstein, +6 authors R. Bobrow
    • Computer Science
    • 2015 IEEE International Conference on Self-Adaptive and Self-Organizing Systems Workshops
    • 2015
    An automated process of active perception for cyber defense that allows the deployment of sensors to build structured interpretations of situations with calibrated responses proportional to the likelihood of true detection and degree of threat is described. Expand
    Knowledge Representation for Dynamic Formation of Sensing Goals in Cyber Defense
    This paper describes how a cognitive agent for cyber defense forms sensing goals through a process of active perception. Loosely following a beliefs-desires-intentions (BDI) model (Bratman, 1987), weExpand


    Near-Optimal Bayesian Active Learning with Noisy Observations
    EC2 is developed, a novel, greedy active learning algorithm and it is proved that it is competitive with the optimal policy, thus obtaining the first competitiveness guarantees for Bayesian active learning with noisy observations. Expand
    Active M-ary sequential hypothesis testing
    This paper considers a generalized sequential hypothesis testing problem in which a decision maker not only can sequentially trade off the sensing cost with the declaration precision, but also canExpand
    A Decision-Making Framework for Control Strategies in Probabilistic Search
    • T. Chung, J. Burdick
    • Computer Science
    • Proceedings 2007 IEEE International Conference on Robotics and Automation
    • 2007
    A Bayesian framework in which the objective is to improve the decision, rather than the sensing, using different control policies is presented, which enables the study and comparison of the role of control for decision-making applications. Expand
    Reinforcement learning for sensing strategies
    • Cody C. T. Kwok, D. Fox
    • Engineering, Computer Science
    • 2004 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS) (IEEE Cat. No.04CH37566)
    • 2004
    Experimental results suggest the least squares reinforcement learning approach is able to learn highly effective sensing strategies in the RoboCup domain, where the robot attempts to reach a ball and accurately kick it into the goal. Expand
    Anytime Point-Based Approximations for Large POMDPs
    The point selection procedure is combined with point-based value backups to form an effective anytime POMDP algorithm called Point-Based Value Iteration (PBVI), and a theoretical analysis justifying the choice of belief selection technique is presented. Expand
    Nonmyopic Multiaspect Sensing With Partially Observable Markov Decision Processes
    The problem of sensing a concealed or distant target by interrogation from multiple sensors situated on a single platform is considered, and a partially observable Markov decision process (POMDP) is modeled, in which the belief state at any point is mapped to a corresponding action. Expand
    Infomax Control of Eye Movements
    It is shown how an optimal eye movement controller can be learned from subjective experiences of information gathering, and in simulation properties of the optimal controller this controller outperforms other eye movement strategies proposed in the literature. Expand
    A Closer Look at MOMDPs
    A complementary analysis and discussion about MOMDPs, a formalism that exploits the fact that the state space may be factored in one visible part and one hidden part, and how classical algorithms can be adapted to this factorization. Expand
    Computationally Feasible Bounds for Partially Observed Markov Decision Processes
    • W. Lovejoy
    • Mathematics, Computer Science
    • Oper. Res.
    • 1991
    This paper explains how to approximate the state space by a finite grid of points, and use that grid to construct upper and lower value function bounds, generate approximate nonstationary and stationary policies, and bound the value loss relative to optimal for using these policies in the decision problem. Expand
    Bayesian surprise attracts human attention
    A formal Bayesian definition of surprise is proposed to capture subjective aspects of sensory information and it is shown that Bayesian surprise is a strong attractor of human attention, with 72% of all gaze shifts directed towards locations more surprising than the average. Expand