AndroidWorld is a dynamic, reproducible Android benchmark that generates unlimited natural-language tasks for autonomous agents and shows current agents succeed on only 30.6 percent of them.
How Many Random Seeds? Statistical Power Analysis in Deep Reinforcement Learning Experiments
7 Pith papers cite this work. Polarity classification is still indexing.
7
Pith papers citing it
abstract
Consistently checking the statistical significance of experimental results is one of the mandatory methodological steps to address the so-called "reproducibility crisis" in deep reinforcement learning. In this tutorial paper, we explain how the number of random seeds relates to the probabilities of statistical errors. For both the t-test and the bootstrap confidence interval test, we recall theoretical guidelines to determine the number of random seeds one should use to provide a statistically significant comparison of the performance of two algorithms. Finally, we discuss the influence of deviations from the assumptions usually made by statistical tests. We show that they can lead to inaccurate evaluations of statistical errors and provide guidelines to counter these negative effects. We make our code available to perform the tests.
representative citing papers
EvoTrainer co-evolves LLM policies and training harnesses via empirical feedback to match or exceed human-engineered RL on math reasoning, code generation, and long-horizon software engineering.
SOCIA-EVO generates statistically consistent simulators by separating structural refinement from parameter calibration via bi-level optimization and falsifying strategies through execution feedback in a Bayesian-weighted playbook.
A black-box evaluation framework with multi-run metrics reveals that LLM-based DSM generation is sensitive to phrasing, exhibits hallucination, and fails to consistently acknowledge knowledge boundaries.
Stochastic MeanFlow Policies enable one-step generative control in off-policy mirror descent by mapping noise through a MeanFlow transform, yielding tractable entropy and improved MuJoCo performance over Gaussian and generative baselines.
Proposes min-max IPR and percentile highlighting to evaluate run-to-run performance variation in deep RL, with case studies on normalizations in PPO/SAC, algorithm comparisons, and DQN/Rainbow on Atari.
citing papers explorer
-
AndroidWorld: A Dynamic Benchmarking Environment for Autonomous Agents
AndroidWorld is a dynamic, reproducible Android benchmark that generates unlimited natural-language tasks for autonomous agents and shows current agents succeed on only 30.6 percent of them.
-
EvoTrainer: Co-Evolving LLM Policies and Training Harnesses for Autonomous Agentic Reinforcement Learning
EvoTrainer co-evolves LLM policies and training harnesses via empirical feedback to match or exceed human-engineered RL on math reasoning, code generation, and long-horizon software engineering.
-
SOCIA-EVO: Automated Simulator Construction via Dual-Anchored Bi-Level Optimization
SOCIA-EVO generates statistically consistent simulators by separating structural refinement from parameter calibration via bi-level optimization and falsifying strategies through execution feedback in a Bayesian-weighted playbook.
-
Auto-DSM Under the Lens: A Black-Box Evaluation Framework for LLM-Based DSM Generation
A black-box evaluation framework with multi-run metrics reveals that LLM-based DSM generation is sensitive to phrasing, exhibits hallucination, and fails to consistently acknowledge knowledge boundaries.
-
Stochastic MeanFlow Policies: One-Step Generative Control with Entropic Mirror Descent
Stochastic MeanFlow Policies enable one-step generative control in off-policy mirror descent by mapping noise through a MeanFlow transform, yielding tractable entropy and improved MuJoCo performance over Gaussian and generative baselines.
-
Performance Variation in Deep Reinforcement Learning
Proposes min-max IPR and percentile highlighting to evaluate run-to-run performance variation in deep RL, with case studies on normalizations in PPO/SAC, algorithm comparisons, and DQN/Rainbow on Atari.
- On Effectiveness and Efficiency of Agentic Tool-calling and RL Training