A Direct Approximation of AIXI Using Logical State Abstractions
We propose a practical integration of logical state abstraction with AIXI, a Bayesian
optimality notion for reinforcement learning agents, to significantly expand the model class …
optimality notion for reinforcement learning agents, to significantly expand the model class …
Dynamic Knowledge Injection for AIXI Agents
Prior approximations of AIXI, a Bayesian optimality notion for general reinforcement
learning, can only approximate AIXI's Bayesian environment model using an a-priori defined …
learning, can only approximate AIXI's Bayesian environment model using an a-priori defined …
Monte carlo tree search with iteratively refining state abstractions
S Sokota, CY Ho, Z Ahmad… - Advances in Neural …, 2021 - proceedings.neurips.cc
Decision-time planning is the process of constructing a transient, local policy with the intent
of using it to make the immediate decision. Monte Carlo tree search (MCTS), which has …
of using it to make the immediate decision. Monte Carlo tree search (MCTS), which has …
Elastic monte carlo tree search with state abstraction for strategy game playing
L Xu, J Hurtado-Grueso, D Jeurissen… - … IEEE Conference on …, 2022 - ieeexplore.ieee.org
Strategy video games challenge AI agents with their combinatorial search space caused by
complex game elements. State abstraction is a popular technique that reduces the state …
complex game elements. State abstraction is a popular technique that reduces the state …
Selecting the partial state abstractions of MDPs: A metareasoning approach with deep reinforcement learning
Markov decision processes (MDPs) are a common general-purpose model used in robotics
for representing sequential decision-making problems. Given the complexity of robotics …
for representing sequential decision-making problems. Given the complexity of robotics …
Game state and action abstracting monte carlo tree search for general strategy game-playing
A Dockhorn, J Hurtado-Grueso… - … IEEE Conference on …, 2021 - ieeexplore.ieee.org
When implementing intelligent agents for strategy games, we observe that search-based
methods struggle with the complexity of such games. To tackle this problem, we propose a …
methods struggle with the complexity of such games. To tackle this problem, we propose a …
Elastic Monte Carlo Tree Search
L Xu, A Dockhorn… - IEEE Transactions on …, 2023 - ieeexplore.ieee.org
Strategy games are a challenge for the design of artificial intelligence agents due to their
complexity and the combinatorial search space they produce. State abstraction has been …
complexity and the combinatorial search space they produce. State abstraction has been …
[PDF][PDF] Goal Agnostic Learning and Planning without Reward Functions.
CK Robinson, J Lancaster - Adv. Artif. Intell. Mach. Learn., 2023 - oajaiml.com
In this paper we present an algorithm, the Goal Agnostic Planner (GAP), which combines
elements of Reinforcement Learning (RL) and Markov Decision Processes (MDPs) into an …
elements of Reinforcement Learning (RL) and Markov Decision Processes (MDPs) into an …
Learning to Solve Sequential Planning Problems Without Rewards
C Robinson - Proceedings of the Future Technologies Conference, 2022 - Springer
In this paper we present an algorithm, the Goal Agnostic Planner (GAP), which combines
elements of Reinforcement Learning (RL) and Markov Decision Processes (MDPs) into an …
elements of Reinforcement Learning (RL) and Markov Decision Processes (MDPs) into an …
On state aggregation in a class of cyber physical energy systems
Various cyber physical energy systems (CPES) are discrete event dynamic systems (DEDS),
in which the states are discrete and the dynamics of state transitions are triggered by events …
in which the states are discrete and the dynamics of state transitions are triggered by events …