Molecule sequence models learn vast molecular spaces, but how to navigate them efficiently? We explored multi-objective RL, SFT, merging, but these fall short in balancing control and diversity. We introduce **Mol-MoE**: a mixture of experts for controllable molecule generation🧵 - ThreadSky

diadochus.bsky.social • 7 days ago

Molecule sequence models learn vast molecular spaces, but how to navigate them efficiently? We explored multi-objective RL, SFT, merging, but these fall short in balancing control and diversity. We introduce **Mol-MoE**: a mixture of experts for controllable molecule generation🧵

Comments

diadochus.bsky.social•7 days ago

Drug discovery inherently involves multi-objective optimization, requiring candidate molecules to not only bind effectively to target proteins, triggering a specific function, but also to meet safety and compatibility criteria to become drugs. Is supervised learning sufficient?

diadochus.bsky.social•7 days ago

In our ablation studies, instruction-tuned models struggle with higher property values due to lack of explicit optimization. Even RL fine-tuning on multiple objectives can hit performance plateaus or declines, and balancing objectives requires re-training, limiting steerability.

diadochus.bsky.social•7 days ago

Think, think, think... what if we trained experts on single properties separately and leveraged model merging techniques to obtain a multi-property model? We re-implement rewarded soups and obtain a robust baseline capable of generating high-quality, out-of-distribution samples.

diadochus.bsky.social•7 days ago

But the relationship between interpolation coefficients and properties isn’t strictly linear, needing a calibration function. Mol-MoE addresses this by training only the routers to predict optimal merging weights from prompts, enabling more precise control and less interference.

diadochus.bsky.social•7 days ago

The model we obtain does achieve a smaller mean absolute error in generating compounds according to the provided properties, surpassing the alternative methods. Arguably, the learned routing functions can tackle task interference.

diadochus.bsky.social•7 days ago

Mol-MoE improves with more property experts with a larger gain than classic merging and overall, it achieves the highest scores. Simple reward scalarization here does not work. We aim at further calibrating Mol-MoE and testing the performance on larger sets of objectives.

Comments

Posting Rules

Reply