His research interests include natural language and generation, machine reasoning, open-ended learning, and meta-learning. He was involved in, and on multiple occasions was the lead of, various projects such as the production of differentiable neural computers, data structures, and program interpreters; teaching artificial agents to play the 80s game NetHack; and examining whether neural networks could reliably solve logical or mathematical problems.
His life’s goal is to get computers to do the thinking as much as possible, so he can focus on the fun stuff.
Add comment