Derives tighter Koopman-based generalization bounds for multi-task DNNs by assuming small weight-matrix condition numbers and using a tailored Sobolev space as hypothesis class.
Norm-Based Capacity Control in Neural Networks
2 Pith papers cite this work. Polarity classification is still indexing.
2
Pith papers citing it
abstract
We investigate the capacity, convexity and characterization of a general family of norm-constrained feed-forward networks.
fields
cs.LG 2verdicts
UNVERDICTED 2representative citing papers
MILD reformulates two-stage learning to defer as cost-sensitive learning over the input-expert domain and derives new margin-based losses with guarantees, yielding better performance than baselines on image classification and LLM routing tasks.
citing papers explorer
-
On the Koopman-Based Generalization Bounds for Multi-Task Deep Learning
Derives tighter Koopman-based generalization bounds for multi-task DNNs by assuming small weight-matrix condition numbers and using a tailored Sobolev space as hypothesis class.
-
Optimized Deferral for Imbalanced Settings
MILD reformulates two-stage learning to defer as cost-sensitive learning over the input-expert domain and derives new margin-based losses with guarantees, yielding better performance than baselines on image classification and LLM routing tasks.