Multimodal LLMs for health grounded in individual-specific data

no code implementations • 18 Jul 2023 • Anastasiya Belyaeva, Justin Cosentino, Farhad Hormozdiari, Krish Eswaran, Shravya Shetty, Greg Corrado, Andrew Carroll, Cory Y. McLean, Nicholas A. Furlotte

To effectively solve personalized health tasks, LLMs need the ability to ingest a diversity of data modalities that are relevant to an individual's health status.

Language Modelling Large Language Model +1

Large-scale machine learning-based phenotyping significantly improves genomic discovery for optic nerve head morphology

1 code implementation • 25 Nov 2020 • Babak Alipanahi, Farhad Hormozdiari, Babak Behsaz, Justin Cosentino, Zachary R. McCaw, Emanuel Schorsch, D. Sculley, Elizabeth H. Dorfman, Sonia Phene, Naama Hammel, Andrew Carroll, Anthony P. Khawaja, Cory Y. McLean

A GWAS of ML-based VCDR identified 299 independent genome-wide significant (GWS; $P\leq5\times10^{-8}$) hits in 156 loci.

BIG-bench Machine Learning

The Search for Sparse, Robust Neural Networks

1 code implementation • 5 Dec 2019 • Justin Cosentino, Federico Zaiter, Dan Pei, Jun Zhu

Recent work on deep neural network pruning has shown there exist sparse subnetworks that achieve equal or improved accuracy, training time, and loss using fewer network parameters when compared to their dense counterparts.

Network Pruning

Generative Well-intentioned Networks

no code implementations • NeurIPS 2019 • Justin Cosentino, Jun Zhu

We propose Generative Well-intentioned Networks (GWINs), a novel framework for increasing the accuracy of certainty-based, closed-world classifiers.