Government
Adaptive Sampling for Efficient Softmax Approximation
The softmax function is ubiquitous in machine learning and optimization applications.
Navigating Extremes: Dynamic Sparsity in Large Output Spaces
In recent years, Dynamic Sparse Training (DST) has emerged as an alternative to post-training pruning for generating efficient models. In principle, DST allows for a more memory efficient training process, as it maintains sparsity throughout the entire training run. However, current DST implementations fail to capitalize on this in practice. Because sparse matrix multiplication is much less efficient than dense matrix multiplication on GPUs, most implementations simulate sparsity by masking weights.
Learning Low-Rank Feature for Thorax Disease Classification
We study thorax disease classification in this paper.
Uncovering Safety Risks of Large Language Models through Concept Activation Vector
Warning: This paper contains text examples that are offensive or harmful in nature.
Strategic Linear Contextual Bandits
Recommendation algorithms that select the most relevant item for sequentially arriving users or queries have become vital for navigating the internet and its many online platforms.
Label Noise: Ignorance Is Bliss
We establish a new theoretical framework for learning under multi-class, instance-dependent label noise. This framework casts learning with label noise as a form of domain adaptation, in particular, domain adaptation under posterior drift.
