Cost-sensitive Bayesian control policy in human active sensing
نویسندگان
چکیده
An important but poorly understood aspect of sensory processing is the role of active sensing, the use of self-motion such as eye or head movements to focus sensing resources on the most rewarding or informative aspects of the sensory environment. Here, we present behavioral data from a visual search experiment, as well as a Bayesian model of within-trial dynamics of sensory processing and eye movements. Within this Bayes-optimal inference and control framework, which we call C-DAC (Context-Dependent Active Controller), various types of behavioral costs, such as temporal delay, response error, and sensor repositioning cost, are explicitly minimized. This contrasts with previously proposed algorithms that optimize abstract statistical objectives such as anticipated information gain (Infomax) (Butko and Movellan, 2010) and expected posterior maximum (greedy MAP) (Najemnik and Geisler, 2005). We find that C-DAC captures human visual search dynamics better than previous models, in particular a certain form of "confirmation bias" apparent in the way human subjects utilize prior knowledge about the spatial distribution of the search target to improve search speed and accuracy. We also examine several computationally efficient approximations to C-DAC that may present biologically more plausible accounts of the neural computations underlying active sensing, as well as practical tools for solving active sensing problems in engineering applications. To summarize, this paper makes the following key contributions: human visual search behavioral data, a context-sensitive Bayesian active sensing model, a comparative study between different models of human active sensing, and a family of efficient approximations to the optimal model.
منابع مشابه
Development and Validation of Active Performance Indicators of Electrical Safety Using Bow-Tie and Bayesian Network Techniques Case Study: Oil and Gas Industries Construction Projects
Background: With the developing use of electricity in all aspects of human life, electricity accidents have also increased. One of the main components of the for the prevention policy, is the safety performance assessment of the organization's or industry's by using appropriate performance indicators with related operations. Method: This study was a descriptive-analytical of 6 steps inc...
متن کاملContext-sensitive active sensing in humans
Humans and animals readily utilize active sensing, or the use of self-motion, to focus sensory and cognitive resources on the behaviorally most relevant stimuli and events in the environment. Understanding the computational basis of natural active sensing is important both for advancing brain sciences and for developing more powerful artificial systems. Recently, we proposed a goal-directed, co...
متن کاملVariational Bayesian Optimization for Runtime Risk-Sensitive Control
We present a new Bayesian policy search algorithm suitable for problems with policy-dependent cost variance, a property present in many robot control tasks. We extend recent work on variational heteroscedastic Gaussian processes to the optimization case to achieve efficient minimization of very noisy cost signals. In contrast to most policy search algorithms, our method explicitly models the co...
متن کاملDesign and development of ShrewdShoe, a smart pressure sensitive wearable platform
This study introduces a wearable in-shoe system for real-time monitoring and measurement of the plantar pressure distribution of the foot using eleven sensing elements. The sensing elements utilized in ShrewdShoe have been designed in an innovative way, they are based on a barometric pressure sensor covered with a silicon coating. The presented sensing element has great linearity up to 300...
متن کاملAn Iterative Decision Rule to minimize cost of Acceptance Sampling Plan in Machine Replacement Problem
In this paper, we presented an optimal iterative decision rule for minimizing total cost in designing a sampling plan for machine replacement problem using the approach of dynamic programming and Bayesian inferences. Cost of replacing the machine and cost of defectives produced by machine has been considered in model. Concept of control threshold policy has been applied for decision making. If ...
متن کامل