Pith. sign in

REVIEW 20 cited by

Implementation Matters in Deep Policy Gradients: A Case Study on PPO and TRPO

Not yet reviewed by Pith; the record is open.

This paper has not been read by Pith yet. Machine review is queued; the pith claim, tier, and objections will appear here once it completes.

SPECIMEN: schema-true, not a live event

T0 review · schema-true

One-sentence machine reading of the paper's core claim.

pith:XXXXXXXX · record.json · timestamp

arxiv 2005.12729 v1 pith:MBXQASGD submitted 2020-05-25 cs.LG cs.ROstat.ML

Implementation Matters in Deep Policy Gradients: A Case Study on PPO and TRPO

classification cs.LG cs.ROstat.ML
keywords policydeeptrpoalgorithmalgorithmscaseimportanceoptimization
verification ladder T0 review T1 audit T2 compute T3 formal T4 reserved
0 comments
read the original abstract

We study the roots of algorithmic progress in deep policy gradient algorithms through a case study on two popular algorithms: Proximal Policy Optimization (PPO) and Trust Region Policy Optimization (TRPO). Specifically, we investigate the consequences of "code-level optimizations:" algorithm augmentations found only in implementations or described as auxiliary details to the core algorithm. Seemingly of secondary importance, such optimizations turn out to have a major impact on agent behavior. Our results show that they (a) are responsible for most of PPO's gain in cumulative reward over TRPO, and (b) fundamentally change how RL methods function. These insights show the difficulty and importance of attributing performance gains in deep reinforcement learning. Code for reproducing our results is available at https://github.com/MadryLab/implementation-matters .

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Forward citations

Cited by 20 Pith papers

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score.

  1. Does Synthetic Data Help? Empirical Evidence from Deep Learning Time Series Forecasters

    cs.LG 2026-05 accept novelty 7.0

    Synthetic data augmentation helps channel-mixing time series models but degrades channel-independent ones, with reliable gains only from seasonal-trend generators and gradual schedules in low-resource settings.

  2. Bounded Ratio Reinforcement Learning

    cs.LG 2026-04 conditional novelty 7.0

    BRRL derives an analytic optimal policy for regularized constrained RL that guarantees monotonic improvement and yields the BPO algorithm that matches or exceeds PPO.

  3. Multi-Turn On-Policy Distillation with Prefix Replay

    cs.LG 2026-07 conditional novelty 6.0

    ReOPD offline-distills multi-turn agentic LLMs via teacher-prefix replay plus step-decay sampling, matching online OPD accuracy at ≥4× speed with zero tool calls.

  4. Understanding electricity consumption behaviour through Inverse Reinforcement Learning

    cs.LG 2026-07 conditional novelty 6.0

    AIRL-recovered reward functions show that the 2022 energy crisis and heatwave reshaped Italian households' cooling responses heterogeneously, sometimes durably, with time-of-use as an independent dimension.

  5. LOLLA: Deep Reinforcement Learning for Closed-Loop Link Adaptation Towards a GPU-Accelerated AI-RAN

    eess.SP 2026-06 unverdicted novelty 6.0

    LOLLA is a PPO-based DRL framework that learns continuous SINR offsets for 5G link adaptation, delivering 15-92% throughput gains over OLLA while meeting tunable BLER targets and running with sub-500us latency on GPU-...

  6. Response Time Enhances Alignment with Heterogeneous Preferences

    cs.LG 2026-05 unverdicted novelty 6.0

    Response times modeled as drift-diffusion processes enable consistent estimation of population-average preferences from heterogeneous anonymous binary choices.

  7. ANO: A Principled Approach to Robust Policy Optimization

    cs.AI 2026-05 unverdicted novelty 6.0

    ANO derives a robust policy optimizer from geometric principles that replaces clipping with a smooth redescending gradient, showing better performance and stability than PPO, SPO, and GRPO in MuJoCo, Atari, and RLHF e...

  8. Application of Deep Reinforcement Learning to Event-Triggered Control for Networked Artificial Pancreas Systems

    eess.SY 2026-04 unverdicted novelty 6.0

    A DRL-based event-triggered controller for networked artificial pancreas systems uses blood glucose change rules to formulate control as a semi-Markov decision process, improving communication efficiency.

  9. Personalized Observation Normalization for Federated Reinforcement Learning in Simulation Environments with Heterogeneity

    cs.LG 2026-04 conditional novelty 6.0

    Personalized local running-mean/variance observation normalization prevents weight-norm overshadowing in FedAvg and improves FedRL-PPO on heterogeneous MuJoCo morphology variants.

  10. SERA: Soft-Verified Efficient Repository Agents

    cs.CL 2026-01 conditional novelty 6.0

    SERA trains competitive coding agents via supervised finetuning on synthetic trajectories verified only by patch-overlap recall, cutting data-generation cost by roughly two orders of magnitude.

  11. Agent Q: Advanced Reasoning and Learning for Autonomous AI Agents

    cs.AI 2024-08 unverdicted novelty 6.0

    Agent Q integrates MCTS-guided search, self-critique, and off-policy DPO to train LLM agents that outperform behavior cloning and reinforced fine-tuning baselines in WebShop and achieve up to 95.4% success in real-wor...

  12. Multi-Turn On-Policy Distillation with Prefix Replay

    cs.LG 2026-07 conditional novelty 5.0

    By replaying teacher prefixes with a step-decay schedule, multi-turn on-policy distillation can run without live environment interaction, matching or slightly beating online OPD accuracy.

  13. Distribution-Agnostic Robust Trajectory Optimization via Chance-Constrained Reinforcement Learning

    math.OC 2026-06 unverdicted novelty 5.0

    A distribution-agnostic robust trajectory optimization framework uses chance-constrained reinforcement learning with rollout-based quantiles to enforce probabilistic feasibility on nominal trajectories via affine corrections.

  14. Ratio-Variance Regularized Policy Optimization

    cs.LG 2026-05 unverdicted novelty 5.0

    R²VPO uses ratio-variance regularization as a distributional soft brake on policy updates, claiming better performance than PPO on math reasoning and robotic control without hard clipping.

  15. TOPPO: Rethinking PPO for Multi-Task Reinforcement Learning with Critic Balancing

    cs.AI 2026-05 unverdicted novelty 5.0

    TOPPO reformulates PPO with critic balancing to address gradient ill-conditioning in multi-task RL and reports stronger mean and tail performance than SAC baselines on Meta-World+ using fewer parameters and steps.

  16. Application of Deep Reinforcement Learning to Event-Triggered Control for Networked Artificial Pancreas Systems

    eess.SY 2026-04 unverdicted novelty 5.0

    A DRL-based event-triggered controller for artificial pancreas systems uses blood glucose change rules to reduce communication frequency while maintaining control performance via an SMDP formulation.

  17. Preventing Learning Stagnation in PPO by Scaling to 1 Million Parallel Environments

    cs.LG 2026-03 conditional novelty 5.0

    PPO plateaus can be avoided by increasing the number of parallel environments, which reduces both the outer-loop step size and update noise; scaling to 1M environments sustained improvement to 1T transitions.

  18. RAFT: Reward rAnked FineTuning for Generative Foundation Model Alignment

    cs.LG 2023-04 unverdicted novelty 5.0

    RAFT aligns generative models by ranking samples with a reward model and fine-tuning only on the top-ranked outputs, reporting gains on reward scores and automated metrics for LLMs and diffusion models.

  19. ODYSSE: Episode-wise Policy Optimization for Personalized Agentic Reasoning

    cs.AI 2026-07 conditional novelty 4.0

    An episode-level variant of GRPO (ESPO) improves personalized GUI-agent reasoning on the 102-episode SmartSpot benchmark, outperforming step-wise and outcome-only training baselines.

  20. Dynamic Multi-Pair Trading Strategy in Cryptocurrency Markets with Deep Reinforcement Learning

    cs.LG 2026-06 unverdicted novelty 4.0

    A hybrid DRL system for multi-pair crypto trading with deterministic risk shielding outperforms a heuristic baseline at 10% significance on Binance futures data.