# A Self-feedback Knowledge Elicitation Approach for Chemical Reaction Predictions

Source: [A Self-feedback Knowledge Elicitation Approach for Chemical Reaction Predictions](<https://www.sota2.com/research/paper/68fa5f581e209131ce9e9189>)

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

Data updated: 2025-12-25T15:36:45.981Z

Authors: Pengfei Liu, Jun Tao, Zhixiang Ren

Published: 2024-04-15T09:26:33.000Z

## Abstract

The task of chemical reaction predictions (CRPs) plays a pivotal role in advancing drug discovery and material science. However, its effectiveness is constrained by the vast and uncertain chemical reaction space and challenges in capturing reaction selectivity, particularly due to existing methods' limitations in exploiting the data's inherent knowledge. To address these challenges, we introduce a data-curated self-feedback knowledge elicitation approach. This method starts from iterative optimization of molecular representations and facilitates the extraction of knowledge on chemical reaction types (RTs). Then, we employ adaptive prompt learning to infuse the prior knowledge into the large language model (LLM). As a result, we achieve significant enhancements: a 14.2% increase in retrosynthesis prediction accuracy, a 74.2% rise in reagent prediction accuracy, and an expansion in the model's capability for handling multi-task chemical reactions. This research offers a novel paradigm for knowledge elicitation in scientific research and showcases the untapped potential of LLMs in CRPs.

## 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.

- [Forward reaction prediction on Mol-Instructions (test)](<https://www.sota2.com/research/sota/forward-reaction-prediction-on-mol-instructions-test.md>): Forward reaction prediction; Mol-Instructions (test); EM Score: 0.945.

- [Retrosynthesis on Mol-Instructions (test)](<https://www.sota2.com/research/sota/retrosynthesis-on-mol-instructions-test.md>): Retrosynthesis; Mol-Instructions (test); EM Score: 0.757.

- [Multi-task Reaction Prediction on Aggregated Multi-task Chemical Reaction Dataset (test)](<https://www.sota2.com/research/sota/multi-task-reaction-prediction-on-aggregated-multi-task-chemical-reaction-dataset-test.md>): Multi-task Reaction Prediction; Aggregated Multi-task Chemical Reaction Dataset (test); BLEU Score: 0.879.

## Original sources and implementation

[Original paper](<https://arxiv.org/abs/2404.09606>)

[PDF](<https://arxiv.org/pdf/2404.09606.pdf>)

[Implementation](<https://github.com/AI-HPC-Research-Team/SLM4CRP>)