# AssayRouter: Historical Utility Priors for Frozen Molecular Predictor Routing

arXiv: https://arxiv.org/abs/2609.37285 · HTML: https://mlsanity.com/paper/2609.37285/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Builds on 6 of its 37 references that are in the map.
- First posted on arXiv 29 Sep 2026.

## Builds on (most cited first, 6 of 6)

- [ChemBERTa-2: Towards Chemical Foundation Models](https://mlsanity.com/paper/2209.01712/) (2022)
- [Deep Learning Foundation Models from Classical Molecular Descriptors](https://mlsanity.com/paper/2506.15792/) (2025)
- [In-Context Learning for Few-Shot Molecular Property Prediction](https://mlsanity.com/paper/2310.08863/) (2023)
- [UniMatch: Universal Matching from Atom to Task for Few-Shot Drug Discovery](https://mlsanity.com/paper/2502.12453/) (2025)
- [Multitask finetuning and acceleration of chemical pretrained models for small molecule drug property prediction](https://mlsanity.com/paper/2510.12719/) (2025)
- [Probabilistic Contrastive Pretraining for Multi-task ADME Property Prediction](https://mlsanity.com/paper/2606.11508/) (2026)

## Abstract (arXiv)

Laboratories often face a new molecular assay with 16-64 labels and a bank of predictors whose training data and parameters are unavailable. The practical question is which frozen outputs to include in a small local model. AssayRouter treats completed assays as pseudo-targets and labels each candidate by its post-fit utility: the reduction in held-out discovery loss when the candidate is added to the local target predictor. A shared regressor learns to predict this utility from candidate behavior on the support set, without source identity; on a new assay, one frozen ranking selects four sources and separate labels fit a convex combiner. We train only on completed ChEMBL-MT assays and evaluate 24 external regression assays across six frozen interface families. AssayRouter-C lowers strict four-call negative log-likelihood (NLL) by 0.0409 relative to Support-CV@4. Frozen candidate-label permutations confirm that candidate-utility correspondence carries the transferred information, and leave-one-interface-out training shows that the mapping generalizes to unseen predictor families. Completed assays therefore provide transferable supervision for scarce-label routing through frozen prediction interfaces.