RENLABENDOSCOPIC AUTONOMY Chi Kit Ng
← All research projects

Physical-evidence-driven policy research 2026 · Preprint

EndoRSI

Physical-Evidence-Driven Recursive Improvement of Endoscopic Robot Policies

Every experiment informs what to try next.

Chi Kit Ng*, Yidong Zhang*, Tianchun Wu, Siu Hing Lui, Hanlin Zhao, Huxin Gao, Hongliang Ren†.

* Equal contribution · † Corresponding author

EndoRSI task, research, execution and evidence-driven policy improvement loop.
The EndoRSI research loop and its embedded EndoForge coding agent. Figure 1 of the preprint. Source ↗

01 / The idea

Every experiment informs what to try next.

A coding agent analyses experiments and revises candidate policies. Human-authorised trials run in a separate safety-constrained runtime, and physical outcomes feed back into the next research cycle.

2 → 17insertion successes out of 21
4policy generations
6independent ex-vivo specimens

Insertion improves from 2/21 in generation 1 to 17/21 in generation 4. The final frozen policy is evaluated separately for transfer. Evaluation details ↗

02 / The approach

How it works

From observations to a physical outcome.

01

Propose & implement

The agent inspects evidence and develops a candidate policy.

02

Test under human control

An isolated runtime executes the authorised experiment.

03

Learn from outcomes

Failures and feedback shape the next policy revision.

Research decisions, execution boundaries and the experimental progression. Figure 2 of the preprint.
Research decisions, execution boundaries and the experimental progression. Figure 2 of the preprint. Source ↗

03 / Physical evidence

Inside the experiments

Full results in the paper ↗
Policy improvement informed by physical outcomes across generations. Figure 5 of the preprint.
Policy improvement informed by physical outcomes across generations. Figure 5 of the preprint. Source ↗

04 / Reference

Build on this work

@misc{ng2026endorsi,
  title = {EndoRSI: Physical-Evidence-Driven Recursive Improvement of Endoscopic Robot Policies},
  author = {Chi Kit Ng and Yidong Zhang and Tianchun Wu and Siu Hing Lui and Hanlin Zhao and Huxin Gao and Hongliang Ren},
  year = {2026},
  eprint = {2610.13376},
  archivePrefix = {arXiv},
  primaryClass = {cs.RO},
  url = {https://arxiv.org/abs/2610.13376}
}

Chi Kit Ng · CUHK

Get in touch.

For research, student projects and opportunities.

Open email app ↗