# Distilling and Retrieving Generalizable Knowledge for Robot Manipulation via Language Corrections

Source: [Distilling and Retrieving Generalizable Knowledge for Robot Manipulation via Language Corrections](<https://www.sota2.com/research/paper/68fa72931e209131cea573da>)

Collection index: [llms.txt](<https://www.sota2.com/research/llms.txt>)

Data updated: 2025-10-26T02:28:52.799Z

Authors: Lihan Zha, Yuchen Cui, Li-Heng Lin, Minae Kwon, Montserrat Gonzalez Arenas, Andy Zeng, Fei Xia, Dorsa Sadigh

Published: 2023-11-17T18:00:20.000Z

## Abstract

Today's robot policies exhibit subpar performance when faced with the challenge of generalizing to novel environments. Human corrective feedback is a crucial form of guidance to enable such generalization. However, adapting to and learning from online human corrections is a non-trivial endeavor: not only do robots need to remember human feedback over time to retrieve the right information in new settings and reduce the intervention rate, but also they would need to be able to respond to feedback that can be arbitrary corrections about high-level human preferences to low-level adjustments to skill parameters. In this work, we present Distillation and Retrieval of Online Corrections (DROC), a large language model (LLM)-based system that can respond to arbitrary forms of language feedback, distill generalizable knowledge from corrections, and retrieve relevant past experiences based on textual and visual similarity for improving performance in novel settings. DROC is able to respond to a sequence of online language corrections that address failures in both high-level task plans and low-level skill primitives. We demonstrate that DROC effectively distills the relevant information from the sequence of online corrections in a knowledge base and retrieves that knowledge in settings with new task or object instances. DROC outperforms other techniques that directly generate robot code via LLMs by using only half of the total number of corrections needed in the first round and requires little to no corrections after two iterations. We show further results, videos, prompts and code on https://sites.google.com/stanford.edu/droc .

## Recorded benchmark results

A research result applies to its stated dataset, split, metric and evaluation configuration. Compare like-for-like settings, preserve the metric's unit and direction, and distinguish paper-reported results from independent reproduction. The best result recorded here is not a claim of exhaustive, current global coverage.

- [Close the bottle on Real-world tasks](<https://www.sota2.com/research/sota/close-the-bottle-on-real-world-tasks.md>): Close the bottle; Real-world tasks; Success Rate: 80.

- [Empty the cabinet on Real-world tasks](<https://www.sota2.com/research/sota/empty-the-cabinet-on-real-world-tasks.md>): Empty the cabinet; Real-world tasks; Success Rate: 60.

- [Placement of Apple on Table on Real-world tasks](<https://www.sota2.com/research/sota/placement-of-apple-on-table-on-real-world-tasks.md>): Place the apple on the table; Real-world tasks; Success Percentage: 80.

- [Pour the can on Real-world tasks](<https://www.sota2.com/research/sota/pour-the-can-on-real-world-tasks.md>): Pour the can; Real-world tasks; Success Rate: 60.

- [Putting Food in Oven on Real-world tasks](<https://www.sota2.com/research/sota/putting-food-in-oven-on-real-world-tasks.md>): Put the food in the oven; Real-world tasks; Success Rate: 20.

## Original sources and implementation

[Original paper](<https://arxiv.org/abs/2311.10678>)

[PDF](<https://arxiv.org/pdf/2311.10678.pdf>)