closed-form expression
- Asymptotics of SGD in Sequence-Single Index Models and Single-Layer Attention Networks
- Decomposing Interventional Causality into Synergistic, Redundant, and Unique Components
- Diffusion Beats Autoregressive in Data-Constrained Settings
- Local Curvature Descent: Squeezing More Curvature out of Standard and Polyak Gradient Descent
- Matchings Under Biased and Correlated Evaluations
- Maximizing the Value of Predictions in Control: Accuracy Is Not Enough