Accelerated Learning on Large-Scale Screens using Generative Library Models

Eli N Weinstein, Andrei Slabodkin, Mattia Gollub, Elizabeth Baker Wood
Proceedings of The 29th International Conference on Artificial Intelligence and Statistics, PMLR 300:4771-4779, 2026.

Abstract

Biological machine learning is often bottlenecked by a lack of scaled data. One promising route to relieving data bottlenecks is through high-throughput screens, which can experimentally test the activity of $10^6-10^{12}$ protein sequences in parallel. In this article, we introduce algorithms to optimize high throughput screens for data creation and model training. We focus on the large-scale regime, where dataset sizes are limited by the cost of measurement and sequencing. We show that when active sequences are rare, we maximize information gain if we only collect positive examples of active sequences, i.e. $x$ with $y>0$. We can correct for the missing negative examples using a generative model of the library, producing a consistent and efficient estimate of the true $p(y\mid x)$. We demonstrate this approach in simulation and on a large-scale screen of antibodies. Overall, co-design of experiments and inference lets us accelerate learning dramatically.

Cite this Paper

BibTeX
@InProceedings{pmlr-v300-weinstein26a, title = { Accelerated Learning on Large-Scale Screens using Generative Library Models }, author = {Weinstein, Eli N and Slabodkin, Andrei and Gollub, Mattia and Wood, Elizabeth Baker}, booktitle = {Proceedings of The 29th International Conference on Artificial Intelligence and Statistics}, pages = {4771--4779}, year = {2026}, editor = {Khan, Emtiyaz and Li, Yingzhen and Solin, Arno and Ramdas, Aaditya}, volume = {300}, series = {Proceedings of Machine Learning Research}, month = {02--05 May}, publisher = {PMLR}, pdf = {https://raw.githubusercontent.com/mlresearch/v300/main/assets/weinstein26a/weinstein26a.pdf}, url = {https://proceedings.mlr.press/v300/weinstein26a.html}, abstract = { Biological machine learning is often bottlenecked by a lack of scaled data. One promising route to relieving data bottlenecks is through high-throughput screens, which can experimentally test the activity of $10^6-10^{12}$ protein sequences in parallel. In this article, we introduce algorithms to optimize high throughput screens for data creation and model training. We focus on the large-scale regime, where dataset sizes are limited by the cost of measurement and sequencing. We show that when active sequences are rare, we maximize information gain if we only collect positive examples of active sequences, i.e. $x$ with $y>0$. We can correct for the missing negative examples using a generative model of the library, producing a consistent and efficient estimate of the true $p(y\mid x)$. We demonstrate this approach in simulation and on a large-scale screen of antibodies. Overall, co-design of experiments and inference lets us accelerate learning dramatically. } }
Endnote
%0 Conference Paper %T Accelerated Learning on Large-Scale Screens using Generative Library Models %A Eli N Weinstein %A Andrei Slabodkin %A Mattia Gollub %A Elizabeth Baker Wood %B Proceedings of The 29th International Conference on Artificial Intelligence and Statistics %C Proceedings of Machine Learning Research %D 2026 %E Emtiyaz Khan %E Yingzhen Li %E Arno Solin %E Aaditya Ramdas %F pmlr-v300-weinstein26a %I PMLR %P 4771--4779 %U https://proceedings.mlr.press/v300/weinstein26a.html %V 300 %X Biological machine learning is often bottlenecked by a lack of scaled data. One promising route to relieving data bottlenecks is through high-throughput screens, which can experimentally test the activity of $10^6-10^{12}$ protein sequences in parallel. In this article, we introduce algorithms to optimize high throughput screens for data creation and model training. We focus on the large-scale regime, where dataset sizes are limited by the cost of measurement and sequencing. We show that when active sequences are rare, we maximize information gain if we only collect positive examples of active sequences, i.e. $x$ with $y>0$. We can correct for the missing negative examples using a generative model of the library, producing a consistent and efficient estimate of the true $p(y\mid x)$. We demonstrate this approach in simulation and on a large-scale screen of antibodies. Overall, co-design of experiments and inference lets us accelerate learning dramatically.
APA
Weinstein, E.N., Slabodkin, A., Gollub, M. & Wood, E.B.. (2026). Accelerated Learning on Large-Scale Screens using Generative Library Models . Proceedings of The 29th International Conference on Artificial Intelligence and Statistics, in Proceedings of Machine Learning Research 300:4771-4779 Available from https://proceedings.mlr.press/v300/weinstein26a.html.

Related Material