Reinforcement Learning with Simple Sequence Priors

Authors: Tankred Saanum, Noémi Éltető, Peter Dayan, Marcel Binz, Eric Schulz

NeurIPS 2023 | Conference PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Experimental We show that the resulting RL algorithm leads to faster learning, and attains higher returns than state-of-the-art model-free approaches in a series of continuous control tasks from the Deep Mind Control Suite. These priors also produce a powerful informationregularized agent that is robust to noisy observations and can perform open-loop control.
Researcher Affiliation Academia Tankred Saanum1 Noémi Éltet o1 Peter Dayan1,2 Marcel Binz1 Eric Schulz1 1Max Planck Institute for Biological Cybernetics, 2University of Tübingen
Pseudocode Yes Algorithm 1 LZ4 pseudo-code
Open Source Code Yes Code: https://github.com/tankred-saanum/simple_priors
Open Datasets Yes We evaluated the agents described in Section 3 on eight continuous control tasks from the Deep Mind Control Suite [34].
Dataset Splits No The paper describes training steps (e.g., '1 million environment steps') and evaluation episodes, but does not provide specific train/validation/test dataset splits in terms of percentages or counts for a fixed dataset, which is common in supervised learning but less so for reinforcement learning environments.
Hardware Specification No The paper does not provide specific hardware details (e.g., GPU/CPU models, memory specifications) used for running the experiments.
Software Dependencies No The paper mentions software like 'Py Torch' and 'Adam optimizer', and refers to specific algorithms like 'LZ4'. However, it does not provide specific version numbers for these software components.
Experiment Setup Yes Table 1: Hyperparameters used for SAC, MIRACLE, LZ-SAC, and SPAC. Table 2: Transformer hyperparameters.