Statistical Learning
Subsidiary Prototype Alignment for Universal Domain Adaptation
A major problem in UniDA is negative transfer, i.e .
Towards a Unified Information-Theoretic Framework for Generalization
In this work, we investigate the expressiveness of the "conditional mutual information" (CMI) framework of Steinke and Zakynthinou [1] and the prospect of using it
Contrastive and Non-Contrastive Self-Supervised Learning Recover Global and Local Spectral Embedding Methods
Self-Supervised Learning (SSL) surmises that inputs and pairwise positive relationships are enough to learn meaningful representations.
TREC: Transient Redundancy Elimination-based Convolution Jiawei Guan
TREC removes over 96% computations and achieves 3.51
Manifold GPL VMs for discovering non-Euclidean latent structure in neural data
Often, these latent variables are internal constructs not directly accessible to the experimenter.