Functional Mechanism
Functional mechanism research investigates the inner workings of systems, aiming to understand how their components interact to produce observed behaviors. Current research focuses on identifying and characterizing these mechanisms across diverse fields, employing techniques like graph convolutional networks, attention mechanisms in transformers, and analysis of hidden states in large language models. This work is crucial for improving the reliability, robustness, and interpretability of complex systems, ranging from artificial intelligence models to biological processes and material properties, ultimately leading to more effective designs and predictions.
Papers
Mechanisms and Computational Design of Multi-Modal End-Effector with Force Sensing using Gated Networks
Yusuke Tanaka, Alvin Zhu, Richard Lin, Ankur Mehta, Dennis Hong
Mechanisms of Symbol Processing for In-Context Learning in Transformer Networks
Paul Smolensky, Roland Fernandez, Zhenghao Herbert Zhou, Mattia Opper, Jianfeng Gao