AI Paper Atlas

Research map / Label-efficient and robust learning / Recognition and long-tail

AI for Interpretable Chemistry: Predicting Radical Mechanistic Pathways via Contrastive Learning

Mohammadamin Tavakoli, Pierre Baldi, Ann Marie Carlton, Yinting Chiu, Alexander Shmakov, David Van Vranken

NeurIPS 2023 · Label-efficient and robust learning › Recognition and long-tail · approach: Prototype and concept dictionary models · 3 citations (OpenAlex)

Read the paperCheck your idea against 137,217 papers

AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
use GNN instead of atom descriptor for reactive sitesreactive site identification · RMechDB combined testtop-5 accuracy: 89.3 -> 92.1Improved
use DRFP instead of pre-defined feature vectors for plausibility rankingplausibility ranker · RMechDB combined testtop-1 accuracy: 73.1 -> 78.6Improved
use DRFP instead of pretrained rxnfpplausibility ranker · RMechDB combined testtop-3 accuracy: 86.2 -> 90.2Improved
use AP reactionFP instead of Morgan2 / TT fingerprintsreactionFP ranker · RMechDB combined testtop-1 accuracy: 74.3 -> 74.6No clear effect
replace atom descriptor with RxnHypergraph in contrastive model (ablation)contrastive model · RMechDB Coretop-1 accuracy: 62.9 -> 64.3Improved
replace atom descriptor with RxnHypergraph in contrastive model (ablation)contrastive model · RMechDB Atmospherictop-1 accuracy: 61.0 -> 62.1Improved
contrastive model instead of two-step methodreaction predictor · RMechDB Coretop-1 accuracy: 62.4 -> 64.3Improved
contrastive atom descriptor instead of RxnHypergraph for inference speedcontrastive model · RMechDBinference time: 1.45 -> 0.08Improved
contrastive atom descriptor instead of two-step for inference speedreaction predictor · RMechDBinference time: 1.38 -> 0.08Improved
fine-tune molecular transformer on RMechDBtext-based model · RMechDB Coretop-1 accuracy: 58.2 -> 57.7Hurt
fine-tune molecular transformer on RMechDBtext-based model · RMechDB Atmospherictop-1 accuracy: 58.0 -> 57.1Hurt
graph-based contrastive model instead of text-based pretrained transformerreaction predictor · RMechDB Coretop-5 accuracy: 84.2 -> 95.1Improved

Closest papers by meaning

Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy