English (United Kingdom)

https://curated-unify.zendy.io/wp-json/zendy-region/v1/featured_content/oa?rat=en

https://curated-unify.zendy.io/wp-json/zendy-region/v1/highlighted_journal/

Global: Zendy Tools annual

Presents the keyphrase highlighting as premium feature

Keyphrase Highlighting

Presents the summarisation as premium feature

Summarisation

Insights

Presents the pdf analysis as premium feature

PDF Analysis

Presents the zaia usage as premium feature

ZAIA

Global: Zendy Tools monthly

Global: Zendy Plus monthly

Presents the access of premium content as premium feature

Premium Content

Global: Zendy Plus annual

Global: Zendy Free Plan

How to accurately predict the properties of molecules is an essential problemin AI-driven drug discovery, which generally requires a large amount ofannotation for training deep learning models. Annotating molecules, however, isquite costly because it requires lab experiments conducted by experts. Toreduce annotation cost, deep Active Learning (AL) methods are developed toselect only the most representative and informative data for annotating.However, existing best deep AL methods are mostly developed for a single typeof learning task (e.g., single-label classification), and hence may not performwell in molecular property prediction that involves various task types. In thispaper, we propose a Task-type-generic active learning framework (termed Tyger)that is able to handle different types of learning tasks in a unified manner.The key is to learn a chemically-meaningful embedding space and perform activeselection fully based on the embeddings, instead of relying ontask-type-specific heuristics (e.g., class-wise prediction probability) as donein existing works. Specifically, for learning the embedding space, weinstantiate a querying module that learns to translate molecule graphs intocorresponding SMILES strings. Furthermore, to ensure that samples selected fromthe space are both representative and informative, we propose to shape theembedding space by two learning objectives, one based on domain knowledge andthe other leveraging feedback from the task learner (i.e., model that performsthe learning task at hand). We conduct extensive experiments on benchmarkdatasets of different task types. Experimental results show that Tygerconsistently achieves high AL performance on molecular property prediction,outperforming baselines by a large margin. We also perform ablative experimentsto verify the effectiveness of each component in Tyger.

Tyger: Task-Type-Generic Active Learning for Molecular Property Prediction