• Publications
  • Influence
Breast cancer classification and prognosis based on gene expression profiles from a population-based study
Comprehensive gene expression patterns generated from cDNA microarrays were correlated with detailed clinico-pathological characteristics and clinical outcome in an unselected group of 99Expand
  • 1,951
  • 89
  • PDF
Benign overfitting in linear regression
TLDR
The phenomenon of benign overfitting is one of the key mysteries uncovered by deep learning methodology: deep neural networks seem to predict well, even with a perfect fit to noisy training data. Expand
  • 129
  • 28
  • PDF
The Relaxed Online Maximum Margin Algorithm
TLDR
We describe a new incremental algorithm for training linear threshold functions: the Relaxed Online Maximum Margin Algorithm, or ROMMA. Expand
  • 233
  • 21
  • PDF
The Power of Localization for Efficiently Learning Linear Separators with Noise
TLDR
We introduce a new approach for designing computationally efficient learning algorithms that are tolerant to noise, and we demonstrate its effectiveness by designing algorithms with improved noise tolerance guarantees for learning linear separators in the presence of malicious noise or adversarial label noise. Expand
  • 111
  • 19
  • PDF
Comparative full-length genome sequence analysis of 14 SARS coronavirus isolates and common mutations associated with putative origins of infection
Summary Background The cause of severe acute respiratory syndrome (SARS) has been identified as a new coronavirus. Whole genomeExpand
  • 405
  • 18
  • PDF
The Singular Values of Convolutional Layers
TLDR
We characterize the singular values of the linear transformation associated with a standard 2D multi-channel convolutional layer, enabling their efficient computation. Expand
  • 92
  • 18
  • PDF
Improved bounds on the sample complexity of learning
TLDR
We present a new general upper bound on the number of examples required to estimate all of the expectations of a set of random variables uniformly well using Haussler's decision theoretic model. Expand
  • 162
  • 16
  • PDF
Random classification noise defeats all convex potential boosters
TLDR
A broad class of boosting algorithms can be interpreted as performing coordinate-wise gradient descent to minimize some potential function of the margins of a data set. Expand
  • 160
  • 15
  • PDF
Comment on " 'Stemness': Transcriptional Profiling of Embryonic and Adult Stem Cells" and "A Stem Cell Molecular Signature" (I)
Ramalho-Santos et al . ([ 1 ][1]) and Ivanova et al. ([ 2 ][2]), comparing the same three “stem cells”— embryonic stem cells (ESCs); neural stem cells (NSCs), referred to as neural progenitor/stemExpand
  • 361
  • 14
  • PDF
On the difficulty of approximately maximizing agreements
TLDR
We address the computational complexity of learning in the agnostic framework. Expand
  • 137
  • 12
  • PDF