# Foresight Residual RL for Long-Horizon Robot Manipulation with Vision-Language-Action Models

arXiv: https://arxiv.org/abs/2607.16506 · HTML: https://mlsanity.com/paper/2607.16506/

## Key facts

- Cited by 2 papers in the map (data through 14 Oct 2026).
- Placed in Reinforcement learning and post-training, one of 13 directions in Vision-language-action models.
- Builds on 23 of its 31 references that are in the map.
- First posted on arXiv 17 Jul 2026.

## Placed in

- Vision-language-action models / [Reinforcement learning and post-training](https://mlsanity.com/vla/reinforcement-learning-and-post-training/)

## Builds on (most cited first, 23 of 23)

- [Proximal Policy Optimization Algorithms](https://mlsanity.com/paper/1707.06347/) (2017)
- [LoRA: Low-Rank Adaptation of Large Language Models](https://mlsanity.com/paper/2106.09685/) (2021)
- [π0: A Vision-Language-Action Flow Model for General Robot Control](https://mlsanity.com/paper/2410.24164/) (2024)
- [Octo: An Open-Source Generalist Robot Policy](https://mlsanity.com/paper/2405.12213/) (2024)
- [Isaac Gym: High Performance GPU-Based Physics Simulation For Robot Learning](https://mlsanity.com/paper/2108.10470/) (2021)
- [Fine-Tuning Vision-Language-Action Models: Optimizing Speed and Success](https://mlsanity.com/paper/2502.19645/) (2025)
- [CALVIN: A Benchmark for Language-Conditioned Policy Learning for Long-Horizon Robot Manipulation Tasks](https://mlsanity.com/paper/2112.03227/) (2021)
- [The Option-Critic Architecture](https://mlsanity.com/paper/1609.05140/) (2016)
- [RoboCasa: Large-Scale Simulation of Everyday Tasks for Generalist Robots](https://mlsanity.com/paper/2406.02523/) (2024)
- [MimicGen: A Data Generation System for Scalable Robot Learning using Human Demonstrations](https://mlsanity.com/paper/2310.17596/) (2023)
- [Residual Reinforcement Learning for Robot Control](https://mlsanity.com/paper/1812.03201/) (2018)
- [Reward Machines: Exploiting Reward Function Structure in Reinforcement Learning](https://mlsanity.com/paper/2010.03950/) (2020)
- [Residual Policy Learning](https://mlsanity.com/paper/1812.06298/) (2018)
- [FurnitureBench: Reproducible Real-World Benchmark for Long-Horizon Complex Manipulation](https://mlsanity.com/paper/2305.12821/) (2023)
- [From Imitation to Refinement -- Residual RL for Precise Assembly](https://mlsanity.com/paper/2407.16677/) (2024)
- [Self-Improving Vision-Language-Action Models with Data Generation via Residual RL](https://mlsanity.com/paper/2511.00091/) (2025)
- [Residual Off-Policy RL for Finetuning Behavior Cloning Policies](https://mlsanity.com/paper/2509.19301/) (2025)
- [Plan-Seq-Learn: Language Model Guided RL for Solving Long Horizon Robotics Tasks](https://mlsanity.com/paper/2405.01534/) (2024)
- [Sequential Dexterity: Chaining Dexterous Policies for Long-Horizon Manipulation](https://mlsanity.com/paper/2309.00987/) (2023)
- [Adversarial Skill Chaining for Long-Horizon Robot Manipulation via Terminal State Regularization](https://mlsanity.com/paper/2111.07999/) (2021)
- [Reward-Respecting Subtasks for Model-Based Reinforcement Learning](https://mlsanity.com/paper/2202.03466/) (2022)
- [Failure Forecasting Boosts Robustness of Sim2Real Rhythmic Insertion Policies](https://mlsanity.com/paper/2507.06519/) (2025)
- [Composing Option Sequences by Adaptation: Initial Results](https://mlsanity.com/paper/2409.08195/) (2024)

## Built on by (most cited first, 2 of 2)

- [Don't Drop the BATON: Long-Horizon Robot Manipulation via Agentic Subtask Exploration and Transition-aware Memory](https://mlsanity.com/paper/2608.16889/) (2026)
- [RouteRLT: Learning When and Which RL Specialist Should Control a Vision-Language-Action Policy](https://mlsanity.com/paper/2609.26467/) (2026)

## Abstract (arXiv)

Vision-Language-Action (VLA) policies offer strong general-purpose manipulation priors, but often fail on tight-tolerance, contact-rich assembly due to long-horizon credit assignment and subtask coupling: a state that is geometrically successful for the current skill can be brittle for downstream skills. We show this failure mode in residual reinforcement learning (RL) over a frozen VLA base policy: constant sparse success rewards improve each subtask in isolation yet yield little or no gain when skills are chained, because terminal state quality is uncontrolled. We propose Foresight Residual RL, which optimizes handoff quality by augmenting each subtask's sparse success reward with an offline-estimated foresight value -- the probability of future subtask success conditioned on the terminal state of the current subtask. Concretely, we (i) train a visual foresight predictor from images of terminal states of the base policy, labeled using downstream rollout statistics, and (ii) train residual policies via backward foresight induction, using the predictor output as a reward multiplier. On a three-phase wrench-based nut-tightening assembly task in Isaac Gym (grasp, move-insert, rotate), our method achieves 85.6% full-task success, outperforming standard subtask residual RL (54.5%) and VLA baselines, while leaving per-subtask success unchanged. These results highlight that improving long-horizon performance requires shaping which successful states are produced at each sub-task, not only whether success occurs.