Intelligent Testing with Interpretable and High-Capacity Agents by Embedding Logical Structure in Neural Architectures

Stanford Investigators

• Marco Pavone

Partner University Investigators

• Necmiye Ozay [Michigan]
• Sanjit Seshia [Berkeley]

TRI Investigators

• Nikos Arechiga
• Andrew Best
• Jin Ge
• Jonathan DeCastro

Project Summary

Building on recent success in developing a differentiable temporal logic,  this project seeks to bridge the gap between current deep learning approaches and classical, logic-based AI. This work represents a unique take into the burgeoning trend in neurosymbolic AI.