Implementing Symbols and Rules with Neural Networks

When: 
Thursday, September 9, 2021 - 7:00pm
Room: 
This seminar will be webcast on Zoom
Lecturer(s): 
Ellie Pavlick
Lecturer Photo

Register in advance for this webinar at
https://acm-org.zoom.us/webinar/register/6016263861654/WN_kBPd0z0MR7CDqJ...

After registering, you will receive a confirmation email containing information about joining the webinar.

Many aspects of human language and reasoning are well explained in terms of symbols and rules. However, state-of-the-art computational models are based on large neural networks which lack explicit symbolic representations of the type frequently used in cognitive theories. One response has been the development of neuro-symbolic models which introduce explicit representations of symbols into neural network architectures or loss functions. In terms of Marr's levels of analysis, such approaches achieve symbolic reasoning at the computational level ("what the system does and why") by introducing symbols and rules at the implementation and algorithmic levels. In this talk, I will consider an alternative: can neural networks (without any explicit symbolic components) nonetheless implement symbolic reasoning at the computational level? I will describe several diagnostic tests of "symbolic" and "rule-governed" behavior and use these tests to analyze neural models of visual and language processing. Our results show that on many counts, neural models appear to encode symbol-like concepts (e.g., conceptual representations that are abstract, systematic, and modular), but not perfectly so. Analysis of the failure cases reveals that future work is needed on methodological tools for analyzing neural networks, as well as refinement of models of hybrid neuro-symbolic reasoning in humans, in order to determine whether neural networks' deviations from the symbolic paradigm are a feature or a bug.

Ellie Pavlick is the Manning Assistant Professor of Computer Science at Brown University and a Research Scientist at Google AI.

Ellie received her PhD in Computer and Information Science from University of Pennsylvania in 2017. In 2012, she received a Bachelor of Arts in Economics from Johns Hopkins University and a Bachelor of Music in Saxophone Performance from the Peabody Conservatory. Ellie's current research is in Natural Language Processing, specifically on computational models of semantics and pragmatics which emulate human inferences. She is interested in building better computational models of natural language semantics and pragmatics: how does language work, and how can we get computers to understand it the way humans do?

This joint meeting of the Boston Chapter of the IEEE Computer Society and GBC/ACM will be online only due to the COVID-19 lockdown.

Up-to-date information about this and other talks is available online at https://ewh.ieee.org/r1/boston/computer/. You can sign up to receive updated status information about this talk and informational emails about future talks at https://mailman.mit.edu/mailman/listinfo/gbc-acm, our self-administered mailing list.