Restricted gradient-descent algorithm for value-function approximation in reinforcement learning
نویسندگان
چکیده
منابع مشابه
Gradient Descent for General Reinforcement Learning
Andrew Moore [email protected] www.cs.cmu.edu/-awm Computer Science Department 5000 Forbes Avenue Carnegie Mellon University Pittsburgh, PA 15213-3891 A simple learning rule is derived, the VAPS algorithm, which can be instantiated to generate a wide range of new reinforcementlearning algorithms. These algorithms solve a number of open problems, define several new approaches to reinforcement learn...
متن کاملPolicy Gradient Methods for Reinforcement Learning with Function Approximation
Function approximation is essential to reinforcement learning, but the standard approach of approximating a value function and determining a policy from it has so far proven theoretically intractable. In this paper we explore an alternative approach in which the policy is explicitly represented by its own function approximator, independent of the value function, and is updated according to the ...
متن کاملManifold Representations for Value-Function Approximation in Reinforcement Learning
Reinforcement learning (RL) has shown itself to be a successful paradigm for solving optimal control problems. However, that success has been mostly limited to problems with a finite set of states and actions. The problem of extending reinforcement learning techniques to the continuous state case has received quite a bit of attention in the last few years. One approach to solving reinforcement ...
متن کاملExplicit Manifold Representations for Value-Function Approximation in Reinforcement Learning
We are interested in using reinforcement learning for large, real-world control problems. In particular, we are interested in problems with continuous, multi-dimensional state spaces, in which traditional reinforcement learning approached perform poorly. Value-function approximation addresses some of the problems of traditional algorithms (for example, continuous state spaces), and has been sho...
متن کاملذخیره در منابع من
با ذخیره ی این منبع در منابع من، دسترسی به آن را برای استفاده های بعدی آسان تر کنید
ژورنال
عنوان ژورنال: Artificial Intelligence
سال: 2008
ISSN: 0004-3702
DOI: 10.1016/j.artint.2007.08.001