Search papers, labs, and topics across Lattice.
3
0
5
TL-ANDI transforms how Tabular Foundation Models handle transfer learning by optimizing context selection to prevent negative transfer.
Forget-critical experts in MoE models can be under-regularized during unlearning, but TRACE rebalances their activation to significantly boost utility.
Unlearning data doesn't have to mean retraining from scratch: this method gets you statistically optimal performance with just a fraction of the original data.
