MassSpecGym in the Wild: Uncovering and Correcting Evaluation Pitfalls in AI-Driven Molecule Discovery
cs.LG
Submitted: 2026-06-17
Updated: 2026-10-29
Code: https://github.com/pluskal-lab/MassSpecGym
Terminology
Sources
- Natural Emergent Misalignment from Reward Hacking in Production RL
- Revealing data leakage in protein interaction benchmarks
- Comparative Analysis of Formula and Structure Prediction from Tandem Mass Spectra
- On objective mismatch in molecular retrieval from tandem mass spectrometry
- FlexMS: A Unified Public Benchmark for Molecule Tandem Mass Spectrum Prediction
- One Small Step with Fingerprints, One Giant Leap for De Novo Molecule Generation from Mass Spectra
- FRIGID: Scaling Diffusion-Based Molecular Generation from Mass Spectra at Training and Inference Time
- ChemBERTa-2: Towards Chemical Foundation Models
- Test-Time Tuned Language Models Enable End-to-end De Novo Molecular Structure Generation from MS/MS Spectra
Related papers
- A Physics-Regulated Neural Framework for Learning 3D Grain Growth Dynamics
- FlexMoE: Learning One-for-All Multi-Budget Expert Pruning for Flexible MoE Deployment
- OptiMind: Improving Small Language Models for Optimization Modeling through Formulation Guidance and Multi-Turn Learning
- Staying in the Sweet Spot: Boost Reasoning Evolution via Responsive Rollout Difficulty Manipulation
- Examining the Impact of Data and Training Methods on Emergent and Subliminal Misalignment
- Rethinking Black-Box On-Policy Distillation: A Rubric-based Analysis