Research map / Learning theory and algorithms / Sample complexity
Sample Efficient Reinforcement Learning via Low-Rank Matrix Estimation
Devavrat Shah, Dogyoon Song, Zhi Xu, Yuzhe Yang
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 changed | Where | Measured | Effect |
|---|---|---|---|
| Use low-rank matrix estimation instead of exploring and updating every state-action pair | Low-rank RL · Inverted Pendulum and other stochastic control tasks | samples to reach a similar error | Improved |
| Use the proposed matrix estimation method instead of Nuclear Norm | Low-rank RL · Inverted Pendulum | angular deviation: 3.1 { $\pm$ .3} -> 3.4 { $\pm$ .7} | Hurt |
| Use the proposed matrix estimation method instead of Nuclear Norm | Low-rank RL · Mountain Car | time-to-goal: 92.4 { $\pm$ 2.8} -> 91.8 { $\pm$ 7.2} | Improved |
| Use the proposed matrix estimation method instead of USVT | Low-rank RL · Double Integrator | time-to-goal: 200.0 { $\pm$ .4} -> 199.7 { $\pm$ .4} | Improved |
| Use the proposed matrix estimation method instead of Soft-Impute | Low-rank RL · Cart-Pole | angular deviation: 10.4 { $\pm$ .1} -> 10.2 { $\pm$ .2} | Improved |
| Use the proposed matrix estimation method instead of Nuclear Norm | Low-rank RL · Acrobot | angular deviation: 5.1 { $\pm$ .8} -> 6.2 { $\pm$ 1.0} | Hurt |
Closest papers by meaning
- Tightening the Dependence on Horizon in the Sample Complexity of Q-Learning (ICML 2021 · similarity 0.91)
- Sample-Efficient Reinforcement Learning Is Feasible for Linearly Realizable MDPs with Limited Revisiting (NeurIPS 2021 · similarity 0.90)
- Model-free Low-Rank Reinforcement Learning via Leveraged Entry-wise Matrix Estimation (NeurIPS 2024 · similarity 0.90)
- Sample Efficient Reinforcement Learning with Partial Dynamics Knowledge (AAAI 2024 · similarity 0.88)
- Linear $Q$-Learning Does Not Diverge in $L^2$: Convergence Rates to a Bounded Set (ICML 2025 · similarity 0.88)
- Shift Before You Learn: Enabling Low-Rank Representations in Reinforcement Learning (NeurIPS 2025 · similarity 0.88)
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.
