Vikash K. Mansinghka
[my initials] at mit dot edu
My work is aimed at making computer systems that infer the causes underlying data and use that knowledge to make good decisions. It has yielded systems that analyze tables, graphs, text and images.

I do this by scaling probabilistic reasoning --- making probabilistic models that are sufficiently rich and realistic to support reliable inferences, and making inference, estimation and prediction efficient through massive parallelism. To make rich, realistic models manageable and fast inference possible, I have worked on universal probabilistic programming languages, efficient universal inference engines, special purpose randomized algorithms for inference, estimation and prediction, the computational complexity of inference, and (where necessary) natively probabilistic hardware.

The scientific foundations of this work lie in artificial intelligence and probabilistic computation. The long-term goal is to build computing systems that make good guesses under uncertainty as naturally as traditional computing systems do logic and arithmetic, to use these systems to help explain the mind and brain, and to close the efficiency gap between natural and artificial computation.

I do this work through Navia Systems, a company I co-founded, and MIT (chiefly the Computer Science and Artificial Intelligence Laboratory and the Brain and Cognitive Sciences Department) where I received my PhD in 2009 and was advised by Professor Joshua Tenenbaum.
Some of my publications, in reverse chronological order::
Feel free to read my old homepage.