end-to-end learning
A modeling approach where the entire workflow from input to output is handled in one continuous pipeline, allowing for joint optimization of components.
- Beyond Higher Rank: Token-wise Input-Output Projections for Efficient Low-Rank Adaptation
- Beyond Pairwise Connections: Extracting High-Order Functional Brain Network Structures under Global Constraints
- DiffE2E: Rethinking End-to-End Driving with a Hybrid Diffusion-Regression-Classification Policy
- Improving Formal Reasoning of Transformer with State Stack
- Knowledge Insulating Vision-Language-Action Models: Train Fast, Run Fast, Generalize Better
- Neural Attention Search
- RIGNO: A Graph-based Framework For Robust And Accurate Operator Learning For PDEs On Arbitrary Domains
- Squared families are useful conjugate priors