David Chiang
Lunch at 12:30pm, talk at 1pm, in 148 Fitzpatrick
Title: Transformers and Formal Logic
Abstract: Characterizing neural networks in terms of better-understood formal systems has the potential to yield new insights into the power and limitations of these networks. Doing so for transformers remains an active area of research. Previous work has related RNNs and CNNs to finite automata, and transformers to a variety of models: counter machines, Boolean circuits, programming languages, and formal logic. I will talk about some of our recent work relating transformers to a variant of first-order logic and some open questions that remain.