Senior Computer Scientist, Artificial Intelligence Center
Andrew Silberfarb, Ph.D., is Senior Computer Scientist in the Advanced Analytics group of the Artificial Intelligence Center at SRI International. His research focuses primarily on development and application of novel artificial intelligence techniques. He applies these techniques to real world problems across a variety of application areas including image analysis, radar processing, atmospheric science, biological systems, industrial applications, and multi-source data fusion.
Andrew leads implementation of SRI’s Deep Adaptive Semantic Logic (DASL) machine learning library that integrates logical reasoning with Deep Learning. He currently is the technical leader for multiple novel applications of machine learning including early identification of viral adaptation to humans, characterization of drones based on post-processing of radar detections, and automated generation of synthetic measurements of the upper atmosphere.
Prior to joining SRI in 2016, Andrew was technical staff at MIT Lincoln Laboratory working on multi-sensor data analysis and fusion using both supervised and unsupervised machine learning techniques where he developed techniques for detecting relevant signals in radar, video, radio frequency, social media, and other data. He also developed novel algorithms in support of battle management, all-source intelligence analysis and sociological analysis.
Andrew holds a BS in physics from the California Institute of Technology (CalTech) and a Ph.D. in physics (quantum measurement and control) from the University of New Mexico. Before joining MITLL he was post-doctoral scholar at CalTech.
We build a Transformer-based molecule encoder and property predictor network with novel input featurization that performs significantly better than existing methods.
We introduce Deep Adaptive Semantic Logic (DASL), a novel framework for automating the generation of deep neural networks that incorporates user-provided formal knowledge to improve learning from data.
We survey viral uses of SLiMs to mimic host proteins, and information resources available for motif discovery.