zacharyfralish

DeepDelta

DeepDelta is a pairwise deep learning approach based on Chemprop that processes two molecules simultaneously and learns to predict property differences between two molecules.

ActiveDelta

ActiveDelta applies our pairwise deep learning approach to predict property differences to molecular active learning tasks.

DeltaClassifier

DeltaClassifier is a machine learning method to directly train upon and learn molecular changes to classify molecular improvements in potency using inexact regression data.

active-subsampling

Using active learning for data curation

CoAggregators

active-subsampling

Using active learning for data curation

MolALKit

Benchmark for molecular active learning.

YokedLearning

:woman_teacher: :arrow_right: :woman_student: Active learning teaching other models

zacharyfralish/CoAggregators

zacharyfralish/active-subsampling

Using active learning for data curation