Rules, Exceptions, and Exploration: The Secret to EXPLORER’s Success
1 Apr 2025
EXPLORER outperforms baselines by combining neural exploration and symbolic reasoning, excelling in text-based games with unseen entities.
Beyond Seen Worlds: EXPLORER’s Journey into Generalized Reasoning
1 Apr 2025
EXPLORER dynamically generalizes symbolic rules using WordNet hypernyms to improve performance on unseen entities in text-based games.
AI That Learns and Unlearns: The Exceptionally Smart EXPLORER
1 Apr 2025
EXPLORER uses Inductive Logic Programming and exception learning to iteratively refine symbolic policies for better adaptability in text-based games.
Learning Logic in Games: How EXPLORER Combines NLP and RL for
1 Apr 2025
EXPLORER combines neural exploration and symbolic reasoning using Answer Set Programming to enhance decision-making and generalization in text-based games.
Neuro-Symbolic Reasoning Meets RL: EXPLORER Outperforms in Text-World Games
1 Apr 2025
EXPLORER combines neural exploration and symbolic reasoning to achieve better generalization, interpretability, and performance in text-based games.
Continual Learning in Reinforcement Learning Agents with Pre-trained, Testing, and Untrained Groups
1 Jan 2025
Analyze how continual learning enhances RL agents' adaptability in market simulations, compared to testing and untrained groups, under real-world scenarios.
Reinforcement Learning Simulation Metrics: QQ plots, ACF graphs, and Volatility Analysis
1 Jan 2025
Additional simulation results, including QQ plots, ACF graphs, and volatility analysis, revealing insights into agent performance and market dynamics.
RL Agents Show Superior Realism and Adaptability Over Zero-Intelligence Agents
1 Jan 2025
Analyzing RL and ZI agents shows RL agents closely mimic real market patterns, including volatility clustering and return autocorrelations, over time.
Configuring Reinforcement Learning Simulation: Agent Settings, Hyper-Parameters, & Market Insights
1 Jan 2025
Detailed configurations of agents and simulations, including flash sales, informed LT setups, and market characteristics across different hyper-parameters.