Search papers, labs, and topics across Lattice.
5
1
7
Expert imbalance can cripple learning-to-defer systems, but a novel cost-sensitive margin-based loss function can restore performance.
Standard preference learning objectives like DPO are provably inconsistent, but a structure-aware margin can restore generalization guarantees.
Get the best of both worlds: Linear-Core Surrogates offer the fast optimization of smooth losses and the statistical efficiency of margin-based losses, without sacrificing differentiability.
Bounded context windows in next-token prediction models can be fundamentally incompatible with low adversarial regret, even with long context lengths.
Modular generative models can theoretically and empirically outperform monolithic models, offering a robust alternative to resource-intensive retraining on aggregate data.
