Investigating the Effectiveness of a Socratic Chain-of-Thoughts Reasoning Method for Task Planning in Robotics, A Case Study

Veronica Augustina Bot, Zheyuan Xu · 2025

Read Investigating the Effectiveness of a Socratic Chain-of-Thoughts Reasoning Method for Task Planning in Robotics, A Case Study by Veronica Augustina Bot, Zheyuan Xu (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2503-08174.txt.

Description

Veronica Bot and Zheyuan Xu test whether questioning a robot's plan Socratically makes the plan better, in a 2025 case study. Language models can draft task plans for robots from words alone, no retraining needed. Whether step-by-step reasoning, and the ancient trick of self-questioning, improve real spatial tasks was open.

A simulated service robot searched for objects in a physics engine, driven by a leading language model. Three strategies competed across twenty trials: direct planning, chain-of-thought listing, and their Socratic variant, which interrogates each step before acting. Success rate and completion time scored the contest. The paper reports preliminary evidence on which prompting style survives contact with a physical task. The question matters because zero-shot planning, if reliable, would spare every robot deployment a costly programming phase.

Catalog record

CREATORVeronica Augustina Bot, Zheyuan Xu
YEAR2025
COLLECTIONPAPERS
FORMATTXT
FILE SIZE22 KB
VERSION1.0
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2503-08174

Browse the full library · About the Tower of Babel