Associate Technical Director, Artificial Intelligence Center
John Byrnes, Ph.D., is an Associate Technical Director in the Advanced Analytics group of the Artificial Intelligence Center at SRI International. His research focuses primarily on statistical pattern detection as a technique for the discovery of hidden structure in large amounts of data. He applies these techniques to the self-organization of collections of natural language text in support of machine understanding of language. He has also applied pattern recognition techniques to formalized knowledge and is interested in automated machine reasoning.
Byrnes is currently a principal investigator on multiple projects involving the analysis of scientific and technical publications and patents in order to track and forecast the emergence of technical capabilities and to carry out retrospective analyses of technical achievement.
Prior to joining SRI in 2009, Byrnes carried out research at Fair Isaac Corporation including the application of natural language clustering techniques to automate the generation of proof plans for automated reasoning and novel language compression techniques to develop content-based routing at fiber-optic data rates. He researched novel signal processing representations and algorithms at Kromos Technology from 2000 to 2002.
Byrnes holds a BS in mathematics and a Ph.D. in pure and applied logic from Carnegie Mellon University.
Recent publications
more +-
Transformer Based Molecule Encoding for Property Prediction
We build a Transformer-based molecule encoder and property predictor network with novel input featurization that performs significantly better than existing methods.
-
Deep Adaptive Semantic Logic (DASL): Compiling Declarative Knowledge into Deep Neural Networks
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.
-
Application of Text Analytics to Extract and Analyze Material–Application Pairs from a Large Scientific Corpus
In this work, we have successfully extracted material–application pairs and ranked them on their importance. This method provides a novel way to map scientific advances in a particular material to the application for which it is used.