# Stage-Transition Dense Reward Modeling for Reinforcement Learning

arXiv: https://arxiv.org/abs/2606.31377 · HTML: https://mlsanity.com/paper/2606.31377/

## Key facts

- Cited by 1 papers in the map (data through 14 Oct 2026).
- Placed in Reward modeling and RLHF, one of 10 directions in LLM alignment / RLHF / preference optimisation.
- Builds on 21 of its 31 references that are in the map.
- First posted on arXiv 30 Jun 2026.

## Placed in

- LLM alignment / RLHF / preference optimisation / [Reward modeling and RLHF](https://mlsanity.com/alignment/reward-modeling-and-rlhf/)
- Deep reinforcement learning methods / [RL for robot locomotion and manipulation](https://mlsanity.com/deeprl/rl-for-robot-locomotion-and-manipulation/)

## Builds on (most cited first, 21 of 21)

- [Deep Residual Learning for Image Recognition](https://mlsanity.com/paper/1512.03385/) (2015)
- [Proximal Policy Optimization Algorithms](https://mlsanity.com/paper/1707.06347/) (2017)
- [Qwen3 Technical Report](https://mlsanity.com/paper/2505.09388/) (2025)
- [Diffusion Policy: Visuomotor Policy Learning via Action Diffusion](https://mlsanity.com/paper/2303.04137/) (2023)
- [FiLM: Visual Reasoning with a General Conditioning Layer](https://mlsanity.com/paper/1709.07871/) (2017)
- [Concrete Problems in AI Safety](https://mlsanity.com/paper/1606.06565/) (2016)
- [D4RL: Datasets for Deep Data-Driven Reinforcement Learning](https://mlsanity.com/paper/2004.07219/) (2020)
- [End-to-End Training of Deep Visuomotor Policies](https://mlsanity.com/paper/1504.00702/) (2015)
- [Learning Complex Dexterous Manipulation with Deep Reinforcement Learning and Demonstrations](https://mlsanity.com/paper/1709.10087/) (2017)
- [Time-Contrastive Networks: Self-Supervised Learning from Video](https://mlsanity.com/paper/1704.06888/) (2017)
- [VIP: Towards Universal Visual Reward and Representation via Value-Implicit Pre-Training](https://mlsanity.com/paper/2210.00030/) (2022)
- [Overcoming Exploration in Reinforcement Learning with Demonstrations](https://mlsanity.com/paper/1709.10089/) (2017)
- [ManiSkill2: A Unified Benchmark for Generalizable Manipulation Skills](https://mlsanity.com/paper/2302.04659/) (2023)
- [Extrapolating Beyond Suboptimal Demonstrations via Inverse Reinforcement Learning from Observations](https://mlsanity.com/paper/1904.06387/) (2019)
- [Precise and Dexterous Robotic Manipulation via Human-in-the-Loop Reinforcement Learning](https://mlsanity.com/paper/2410.21845/) (2024)
- [LIV: Language-Image Representations and Rewards for Robotic Control](https://mlsanity.com/paper/2306.00958/) (2023)
- [Learning by Playing - Solving Sparse Reward Tasks from Scratch](https://mlsanity.com/paper/1802.10567/) (2018)
- [ReWiND: Language-Guided Rewards Teach Robot Policies without New Demonstrations](https://mlsanity.com/paper/2505.10911/) (2025)
- [Meta-World+: An Improved, Standardized, RL Benchmark](https://mlsanity.com/paper/2505.11289/) (2025)
- [Fast Lifelong Adaptive Inverse Reinforcement Learning from Demonstrations](https://mlsanity.com/paper/2209.11908/) (2022)
- [Multi-Stage Manipulation with Demonstration-Augmented Reward, Policy, and World Model Learning](https://mlsanity.com/paper/2503.01837/) (2025)

## Built on by (most cited first, 1 of 1)

- [ARS: Agentic Reward System for Robot Learning](https://mlsanity.com/paper/2609.34484/) (2026)

## Abstract (arXiv)

Reinforcement learning for long-horizon robotic manipulation is often limited by sparse and delayed rewards, while manually designing dense shaping signals is costly and brittle to changes in environments and object configurations. This work proposes Stage-Transition Dense Reward (STDR), a visual reward-learning framework that converts unstructured expert videos into logically grounded dense rewards for training RL agents from scratch. STDR leverages semantic understanding to infer a task's stage structure from demonstrations, and delivers two complementary learning signals during online training: (i) stage-transition feedback that provides goal-directed reward, and (ii) within-stage progress feedback that supplies fine-grained guidance toward completing each stage. Furthermore, an out-of-distribution (OOD) detection mechanism and a grasping regulation module are integrated to enhance robustness and prevent reward hacking. Experiments on 14 manipulation tasks across MetaWorld, ManiSkill, and Franka Kitchen show that STDR consistently improves sample efficiency and success rates over multiple baselines, and matches or surpasses handcrafted dense rewards on several challenging tasks. Real-robot evaluations further indicate that STDR assigns stable, progress-aligned rewards on successful executions while producing appropriately low rewards for failures, suggesting robustness to visual noise and better-calibrated reward assignment across settings.