# CellScientist: From Execution Feedback to Auditable Model-Revision Trajectories for Cellular Perturbation Prediction

arXiv: https://arxiv.org/abs/2605.07335 · HTML: https://mlsanity.com/paper/2605.07335/

## Key facts

- Cited by 1 papers in the map (data through 14 Oct 2026).
- Builds on 17 of its 54 references that are in the map.
- First posted on arXiv 8 May 2026.

## Builds on (most cited first, 17 of 17)

- [The AI Scientist: Towards Fully Automated Open-Ended Scientific Discovery](https://mlsanity.com/paper/2408.06292/) (2024)
- [MLE-bench: Evaluating Machine Learning Agents on Machine Learning Engineering](https://mlsanity.com/paper/2410.07095/) (2024)
- [MLAgentBench: Evaluating Language Agents on Machine Learning Experimentation](https://mlsanity.com/paper/2310.03302/) (2023)
- [AIDE: AI-Driven Exploration in the Space of Code](https://mlsanity.com/paper/2502.13138/) (2025)
- [FLAML: A Fast and Lightweight AutoML Library](https://mlsanity.com/paper/1911.04706/) (2019)
- [Better by Default: Strong Pre-Tuned MLPs and Boosted Trees on Tabular Data](https://mlsanity.com/paper/2407.04491/) (2024)
- [TabR: Tabular Deep Learning Meets Nearest Neighbors in 2023](https://mlsanity.com/paper/2307.14338/) (2023)
- [MLE-STAR: Machine Learning Engineering Agent via Search and Targeted Refinement](https://mlsanity.com/paper/2506.15692/) (2025)
- [AMLB: an AutoML Benchmark](https://mlsanity.com/paper/2207.12560/) (2022)
- [PerturBench: Benchmarking Machine Learning Models for Cellular Perturbation Analysis](https://mlsanity.com/paper/2408.10609/) (2024)
- [CellForge: Agentic Design of Virtual Cell Models](https://mlsanity.com/paper/2508.02276/) (2025)
- [BioMARS: A Multi-Agent Robotic System for Autonomous Biological Experiments](https://mlsanity.com/paper/2507.01485/) (2025)
- [VCWorld: A Biological World Model for Virtual Cell Simulation](https://mlsanity.com/paper/2512.00306/) (2025)
- [TusoAI: Agentic Optimization for Scientific Methods](https://mlsanity.com/paper/2509.23986/) (2025)
- [HarmonyCell: Automating Single-Cell Perturbation Modeling under Semantic and Distribution Shifts](https://mlsanity.com/paper/2603.01396/) (2026)
- [Large Language Models Meet Virtual Cell: A Survey](https://mlsanity.com/paper/2510.07706/) (2025)
- [Learning Cell-Aware Hierarchical Multi-Modal Representations for Robust Molecular Modeling](https://mlsanity.com/paper/2511.21120/) (2025)

## Built on by (most cited first, 1 of 1)

- [Discover, Falsify, Revise: Auditing Input-Use Claims from Source Code to Predictive Contribution in Agent-Discovered Cell Models](https://mlsanity.com/paper/2609.27234/) (2026)

## Abstract (arXiv)

Cellular perturbation-response modeling requires coordinated choices of representations, fusion mechanisms, objectives, and training procedures. Large language models (LLMs) can propose executable candidates, but unconstrained revision can produce invalid implementations, change task semantics, or discard useful components. We present CellScientist, a protocol-constrained workflow that converts execution and validation feedback into auditable model-revision trajectories. It records design states and outcomes, routes discrepancies to specific components, and applies local revisions under a fixed task contract. A matched-budget study fixes the candidate language, predictor, fitting, and evaluator: structured revision finds better held-out predictors at small budgets under two LLM backbones. Operational audits link history to fewer repeated proposals, contract checks to contained violations, and discrepancy routing to targeted repairs. Refits of two frozen designs on an independently acquired cohort evaluate external predictive utility. Open-workflow trajectories retain improvements, regressions, and failures, while transcriptomic and single-cell searches extend application to additional response spaces. CellScientist produces both a selected predictor and an inspectable record of its development. Project page: https://limengran98.github.io/CellScientist/.