3

An earlier speech JEPA study using fixed soft clustering targets to stabilize representation learning.

Accounting for ambiguous ground truth when comparing AI systems with experts.

We develop a framework for evaluating AI behavior with realistic situational judgment tests and structured personas, connecting AI evaluation with psychometrics.

A review connecting AI with cognitive science, with a focus on memory, reasoning, representation, and the architectural questions behind more general intelligence.

Reducing parameter costs with linear transformations that preserve the structure of multidimensional inputs.

A new framework, which resolves some of the known drawbacks of the Information Bottleneck. We provide a theoretical analysis of the framework, finding the structure of its solutions and present a novel variational formulation for DNNs.