scholarly journals Inspiration of Bayesian decision theory for action anticipation in complex decision making in sports: Taking tennis and soccer as examples

2021 ◽  
Vol 29 (7) ◽  
pp. 1300
Author(s):  
Ze-Jun WANG ◽  
Xin-Yu CHU
Author(s):  
Peter Dayan ◽  
Jonathan P. Roiser ◽  
Essi Viding

That we shape our environment, and our environment shapes us, are truisms with deep and complicated consequences. The resulting feedback interaction leads to a substantial form of what is known as path dependency. This is that small initial variations, stemming from individual differences or even just the vicissitudes of chance, can potentially result in large and persistent divergence in outcomes. This has implications for the nature and interpretation of adaptive and maladaptive choice. This chapter offers a simple formulation in terms of active observers—a formalization of decision-making problems in which actors have the choice of whether and how to gather information to improve what happens. The chapter notes that, according to Bayesian decision theory, it is often optimal for active observers to remain incorrectly calibrated with their surroundings; it explores consequences of this in non-interactive environments, and environments containing other people who might compete or cooperate. The chapter draws loose parallels with the literature on active and evocative gene–environment correlations.


2002 ◽  
Vol 18 (2) ◽  
pp. 303-328 ◽  
Author(s):  
Igor Douven

Bayesian decision theory operates under the fiction that in any decision-making situation the agent is simply given the options from which he is to choose. It thereby sets aside some characteristics of the decision-making situation that are pre-analytically of vital concern to the verdict on the agent's eventual decision. In this paper it is shown that and how these characteristics can be accommodated within a still recognizably Bayesian account of rational agency.


2020 ◽  
Vol 43 ◽  
Author(s):  
Peter Dayan

Abstract Bayesian decision theory provides a simple formal elucidation of some of the ways that representation and representational abstraction are involved with, and exploit, both prediction and its rather distant cousin, predictive coding. Both model-free and model-based methods are involved.


2021 ◽  
Vol 35 (2) ◽  
Author(s):  
Nicolas Bougie ◽  
Ryutaro Ichise

AbstractDeep reinforcement learning methods have achieved significant successes in complex decision-making problems. In fact, they traditionally rely on well-designed extrinsic rewards, which limits their applicability to many real-world tasks where rewards are naturally sparse. While cloning behaviors provided by an expert is a promising approach to the exploration problem, learning from a fixed set of demonstrations may be impracticable due to lack of state coverage or distribution mismatch—when the learner’s goal deviates from the demonstrated behaviors. Besides, we are interested in learning how to reach a wide range of goals from the same set of demonstrations. In this work we propose a novel goal-conditioned method that leverages very small sets of goal-driven demonstrations to massively accelerate the learning process. Crucially, we introduce the concept of active goal-driven demonstrations to query the demonstrator only in hard-to-learn and uncertain regions of the state space. We further present a strategy for prioritizing sampling of goals where the disagreement between the expert and the policy is maximized. We evaluate our method on a variety of benchmark environments from the Mujoco domain. Experimental results show that our method outperforms prior imitation learning approaches in most of the tasks in terms of exploration efficiency and average scores.


Sign in / Sign up

Export Citation Format

Share Document