# Robotic Long-Horizon Manipulation with Progressive In-Context Code Generation and Episodic Feedback

arXiv: https://arxiv.org/abs/2503.21969 · HTML: https://mlsanity.com/paper/2503.21969/

## Key facts

- Cited by 3 papers in the map (data through 14 Oct 2026).
- Builds on 18 of its 22 references that are in the map.
- First posted on arXiv 27 Mar 2025.

## Builds on (most cited first, 18 of 18)

- [GPT-4 Technical Report](https://mlsanity.com/paper/2303.08774/) (2023)
- [Do As I Can, Not As I Say: Grounding Language in Robotic Affordances](https://mlsanity.com/paper/2204.01691/) (2022)
- [Code as Policies: Language Model Programs for Embodied Control](https://mlsanity.com/paper/2209.07753/) (2022)
- [Inner Monologue: Embodied Reasoning through Planning with Language Models](https://mlsanity.com/paper/2207.05608/) (2022)
- [Grounded SAM: Assembling Open-World Models for Diverse Visual Tasks](https://mlsanity.com/paper/2401.14159/) (2024)
- [VoxPoser: Composable 3D Value Maps for Robotic Manipulation with Language Models](https://mlsanity.com/paper/2307.05973/) (2023)
- [CLIPort: What and Where Pathways for Robotic Manipulation](https://mlsanity.com/paper/2109.12098/) (2021)
- [CALVIN: A Benchmark for Language-Conditioned Policy Learning for Long-Horizon Robot Manipulation Tasks](https://mlsanity.com/paper/2112.03227/) (2021)
- [Relay Policy Learning: Solving Long-Horizon Tasks via Imitation and Reinforcement Learning](https://mlsanity.com/paper/1910.11956/) (2019)
- [ReKep: Spatio-Temporal Reasoning of Relational Keypoint Constraints for Robotic Manipulation](https://mlsanity.com/paper/2409.01652/) (2024)
- [RT-H: Action Hierarchies Using Language](https://mlsanity.com/paper/2403.01823/) (2024)
- [GenSim: Generating Robotic Simulation Tasks via Large Language Models](https://mlsanity.com/paper/2310.01361/) (2023)
- [DoReMi: Grounding Language Model by Detecting and Recovering from Plan-Execution Misalignment](https://mlsanity.com/paper/2307.00329/) (2023)
- [RoboCodeX: Multimodal Code Generation for Robotic Behavior Synthesis](https://mlsanity.com/paper/2402.16117/) (2024)
- [RoboScript: Code Generation for Free-Form Manipulation Tasks across Real and Simulation](https://mlsanity.com/paper/2402.14623/) (2024)
- [Bridging Language and Action: A Survey of Language-Conditioned Robot Manipulation](https://mlsanity.com/paper/2312.10807/) (2023)
- [LoHoRavens: A Long-Horizon Language-Conditioned Benchmark for Robotic Tabletop Manipulation](https://mlsanity.com/paper/2310.12020/) (2023)
- [LEMMo-Plan: LLM-Enhanced Learning from Multi-Modal Demonstration for Planning Sequential Contact-Rich Manipulation Tasks](https://mlsanity.com/paper/2409.11863/) (2024)

## Built on by (most cited first, 3 of 3)

- [LoHoVLA: A Unified Vision-Language-Action Model for Long-Horizon Embodied Tasks](https://mlsanity.com/paper/2506.00411/) (2025)
- [RoboInspector: Unveiling the Unreliability of Policy Code for LLM-enabled Robotic Manipulation](https://mlsanity.com/paper/2508.21378/) (2025)
- [CodeActionBench: Evaluating Agentic Code-as-Policy for Embodied Manipulation](https://mlsanity.com/paper/2609.33807/) (2026)

## Abstract (arXiv)

Robotic long-horizon manipulation requires robots to compose perception, reasoning, and action over extended task sequences, yet existing language-conditioned frameworks often rely on dense step-wise feedback, learned action policies, or unstructured prompting, which limits robustness and generalization. We propose DAHLIA, a data-agnostic code-generation framework that treats long-horizon manipulation as episodic task planning and evaluation. DAHLIA uses progressively organized in-context examples with chain-of-thought reasoning to guide an LLM in synthesizing executable code plans from language instructions. To improve robustness under partial observability, a VLM-based reporter evaluates task outcomes only after each execution episode and provides structured feedback for replanning, avoiding the overhead of per-step inference. Experiments on LoHoRavens, CALVIN, Franka Kitchen, and real-world manipulation tasks show that DAHLIA achieves strong performance on over 30 long-horizon tasks and improves generalization to unseen scenarios, including cluttered scenes and occluded objects.