method · first seen 4 Oct 2026
proximal policy optimization
Also: PPO-Clip
A policy gradient reinforcement learning algorithm that uses clipped probability ratios for stable on-policy updates.
Proximal policy optimization (PPO) is a policy gradient reinforcement learning algorithm, introduced by OpenAI in 2017, that limits each update by clipping the probability ratio between new and old policies. The sources do not show anything new about PPO itself. They use it as a default baseline or training method across many applications: underwater ROV control in Isaac Sim, a JAX Factorio-style simulator, chip detailed placement, Franka arm assembly and reaching, Mario generalization, Titan probe mission planning, and safe RL on humanoids. Variants such as a tracking-value PPO for language model reasoning and a maskable physics-constrained PPO appear. Usage is broad and routine.
PPO remains the standard on-policy workhorse for robotics, games and LLM post-training, so its frequent mentions reflect how widely it is used rather than a new trend. Variants built on it, such as expert-guided or safety-constrained PPO, are more likely to carry new signal than PPO itself.
Analyst note (established), 17 Nov 2026.
Mentions per week
PapersCodeDiscussion
Rank 140 on 14 Nov 2026 · 51 mentions in 28 days · 51 independent groups · papers → code · first seen 6 weeks ago · acceleration -0.06
Earliest sources
- A Unified Dynamics Framework for Reinforcement Learning and Classical Control of a Six-DOF Pipeline-Tracking ROV in NVIDIA Isaac SimPapers · 4 Oct 2026 · Cheng Siong Chin, M. Venkateshkumar, Jianhua Zhang
- Factoriax: A GPU-Accelerated Factorio-Style Simulator for Reinforcement LearningPapers · 4 Oct 2026 · Mickey Beurskens, Tristan Tomilin, Thiago D. Simão
- Adaptive Expert Guidance for Efficient On-Policy Reinforcement LearningPapers · 5 Oct 2026 · Daniele Affinita, Ming Xu, Rudolf Reiter, Davide Scaramuzza, Pascal Fua
- GPlaceRL: An Open-Source Graph Reinforcement Learning Framework for Detailed PlacementPapers · 5 Oct 2026 · Pavlos Stoikos, Foteini Oikonomou, Christos Poulos, Maria Pantazi-Kypriou, Athan…
- Learned Adaptive Multiresolution Diffusion ImagingPapers · 6 Oct 2026 · Christian Tantardini, Stig Rune Jensen, Roberto Di Remigio Eikås, Joakim Henrik…
- Revisiting Numerical Forecasting Models for Language-Based Trajectory PredictionPapers · 6 Oct 2026 · JunGyu Lee, Inhwan Bae, Hae-Gon Jeon
Latest sources
- Dynamic Shaping: Acquiring Interfaces and Constructing Execution over Frozen Pretrained ComponentsPapers · 16 Nov 2026 · Nathaniel Oh
- GNN-Based Deep Reinforcement Learning for Coastal Emergency Response under Extreme Water LevelsPapers · 15 Nov 2026 · Marcello Sano, Samuele Casagrande, Davide Ferrario, Sebastiano Vascon, Silvia To…
- Towards Probabilistically Safe Reinforcement Learning: Theory and Deployment on Humanoid and Aerial RobotsPapers · 14 Nov 2026 · Pol Mestres, Varun Gabbita, Jason H. K. Siu, Sergio A. Esteban, Hampus Serneke,…
- Enrichment-Oriented Virtual Screening of Perovskite Solar Cells through Reinforcement LearningPapers · 13 Nov 2026 · Yezeyuan Zheng, Ziyang Hu, Maojun Sun. Ruijian Han, Jian Huang, Guanhua Chen, Ji…
- Robot Navigation and Obstacle Avoidance from Randomly Positioned External CamerasPapers · 12 Nov 2026 · Raul Steinmetz, Jair Augusto Bottega, Ayanori Yorozu, Akihisa Ohya
- Stack more levels: how to get general and human-like Mario playingPapers · 12 Nov 2026 · Shinichi Arimura, Yuchen Li, Julian Togelius
- Embedded Sim-to-Real Visual Servoing: Learned Reaching and Force-Guided PlacementPapers · 11 Nov 2026 · Dezhi Sun, Duy Khang Banh, Kenan Niu
- A physics-constrained deep reinforcement learning method for autonomous mission planning of Titan deep space probesPapers · 11 Nov 2026 · Hanmin Li, Tao Wang, Shuangliang Liu, Zhe Zhang, Maodeng Li
- Conformal Risk Certificates for Portfolio Optimization Under Uncertainty: A Gated Reinforcement-Learning ApproachPapers · 10 Nov 2026 · Mohammad Zarei, Amirhossein MohammadpourAzari, Moslem Habibi
- RedirectMimic: Local Motion Content Redaction for Humanoid Whole-Body ControlPapers · 10 Nov 2026 · Zepeng Wang, Xiaochuan Shi
- DeRep: Learning Depth Representations for End-to-End Navigation in Cluttered EnvironmentsPapers · 10 Nov 2026 · Hanyu Jin, Kenji Shimada
- HalfVulpes/sts2aiCode · 9 Nov 2026 · HalfVulpes · 44 stars