AI Paper Atlas

Research map / Reinforcement learning / Visual and representation RL

Towards Better Interpretability in Deep Q-Networks

Raghuram Mandyam Annasamy, Katia Sycara

AAAI 2019 · Reinforcement learning › Visual and representation RL · approach: Scalable RL systems and environments · 62 citations (OpenAlex)

Read the paperCheck your idea against 137,217 papers

AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
Use directed uncertainty-based exploration instead of epsilon-greedy explorationi-DQN vs DDQN · Atari games, 10M framestraining score: Alien: 1,533.45 -> 2,380.72; Freeway: 22.5 -> 28.79; Frostbite: 754.48 -> 3,968.45; Gravitar: 279.89 -> 517.33; MsPacmanImproved
Add reconstruction loss with weight 0.5i-DQN · MsPacmantraining score: 3,245.3Hurt
Add reconstruction loss with weight 5.0i-DQN · MsPacmantraining score: 3,013.1Hurt
Increase Bellman and distributional loss weights to 10.0i-DQN · MsPacmantraining score: 5,267.15Hurt
Increase distributional-loss weight to 10.0 while keeping Bellman weight at 1.0i-DQN · MsPacmantraining score: 4,916.0Hurt
Increase Bellman-loss weight to 10.0 while keeping distributional weight at 1.0i-DQN · MsPacmantraining score: 4,053.6Hurt
Vary exploration factor across 0.1, 0.01, and 0.001i-DQN · Atari gamestraining scoreNo clear effect
Add distributional Bellman loss to diversity plus Bellman losses (ablation)i-DQN · PongNoFrameskip-v4training rewardMixed
Add reconstruction loss to diversity, Bellman, and distributional losses (ablation)i-DQN · PongNoFrameskip-v4training rewardMixed
Remove diversity loss (ablation)i-DQN · PongNoFrameskip-v4attention diversityHurt
Use VAE (β=0) instead of AE for key reconstruction (ablation)i-DQN · MsPacman (Color)Agreement: 30.76 -> 29.78Hurt
Use VAE (β=0.01) instead of AE for key reconstruction (ablation)i-DQN · MsPacman (Color)Agreement: 30.76 -> 16.8Hurt

Closest papers by meaning

Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy