We develop data deep learning methods containing geometric, topological, dynamic systems-based constructs for discovery from scientific and biomedical data.
Today, I’m proud to share our latest work published in @NatureBiotech describing MELD, a #MachineLearning algorithm for #SingleCell perturbation analysis.
Read this #tweetorial to learn about the work led by @dbburkhardt and Jay Stanley 🥳🎉🧪
https://t.co/yquu66G8No
(1/16)
(1/15) Today, I'm proud to share Neurospectrum, a #DeepLearning framework for neural dynamics, out in Neurocomputing! Neurospectrum is a graph-based network that represents brain activity as graph wavelets and then uses an autoencoder to dimensionality-reduce this representation while preserving geometry and trajectories. The trajectories are then characterized by a series of geometric and topological descriptors for regression and classification tasks, including classification of OCD.
https://t.co/zrX5jPQa26
(13/n) After the task, trajectories showed more frequent, localized sharp turns. This curvature increase was significantly greater in OCD than HC across both sexes, especially in males. Learned attention highlighted the Postcentral Gyrus and Putamen in females and the Superior Frontal Sulcus in males, consistent with regions previously implicated in OCD.
Interested to learn about Yale’s #PhD program in Computational Biology and Biomedical Informatics (CBB)? Join our virtual information session for interested applicants on Tuesday, November 10, 2026, 8–9PM ET.
👉Register at: https://t.co/h0jZii6Tj6
@YaleCBB@YALEBBS