Research map / Reinforcement learning / Policy gradient and actor-critic
Fully Parameterized Quantile Function for Distributional Reinforcement Learning
Read the paperCheck your idea against 108,967 papers
AI-extracted summary. Written by AI from the paper's title only (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 changed | Where | Measured | Effect |
|---|---|---|---|
| Parameterize quantile fractions along with quantile values instead of fixing or randomly sampling fractions | Distributional RL on Atari · Atari-55 | mean human-normalized score: 1112% -> 1426% | Helped |
| Parameterize quantile fractions along with quantile values instead of fixing or randomly sampling fractions | Distributional RL on Atari · Atari-55 | median human-normalized score: 218% -> 272% | Helped |
| Use adjustable rather than randomly sampled quantile fractions | FQF and IQN · 6 Atari games, N=8 | mean human-normalized score: 60.2 -> 83.2 | Helped |
| Use adjustable rather than randomly sampled quantile fractions | FQF and IQN · 6 Atari games, N=32 | mean human-normalized score: 91.5 -> 124.6 | Helped |
| Use adjustable rather than randomly sampled quantile fractions | FQF and IQN · 6 Atari games, N=64 | mean human-normalized score: 64.4 -> 69.5 | Helped |
| Add the fraction proposal network | IQN-style distributional RL · Atari training | training speed: roughly 20% | Hurt |
| Increase the number of quantile fractions | FQF · 6 Atari games | mean human-normalized score: 83.2 -> 124.6 -> 69.5 | Mixed |
Closest papers by meaning
- Distributional Reinforcement Learning With Quantile Regression (AAAI 2018 · similarity 0.94)
- Implicit Quantile Networks for Distributional Reinforcement Learning (ICML 2018 · similarity 0.93)
- Non-Crossing Quantile Regression for Distributional Reinforcement Learning (NeurIPS 2020 · similarity 0.92)
- Distributional Reinforcement Learning for Efficient Exploration (ICML 2019 · similarity 0.91)
- Quantile Geometry Regularization for Distributional Reinforcement Learning (NeurIPS 2026 · similarity 0.91)
- Tilted Quantile Gradient Updates for Quantile-Constrained Reinforcement Learning (AAAI 2025 · similarity 0.87)
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.