Variance-based regularization with convex objectives
J Duchi, H Namkoong - Journal of Machine Learning Research, 2019 - jmlr.org
We develop an approach to risk minimization and stochastic optimization that provides a
convex surrogate for variance, allowing near-optimal and computationally efficient trading …
convex surrogate for variance, allowing near-optimal and computationally efficient trading …
Adaptive regularization of weight vectors
We present AROW, a new online learning algorithm that combines several properties of
successful: large margin training, confidence weighting, and the capacity to handle non …
successful: large margin training, confidence weighting, and the capacity to handle non …
Detecting concept change in dynamic data streams: A sequential approach based on reservoir sampling
In this research we present a novel approach to the concept change detection problem.
Change detection is a fundamental issue with data stream mining as classification models …
Change detection is a fundamental issue with data stream mining as classification models …
Adaptive regularization of weight vectors
We present AROW, an online learning algorithm for binary and multiclass problems that
combines large margin training, confidence weighting, and the capacity to handle non …
combines large margin training, confidence weighting, and the capacity to handle non …
Selective transfer learning for cross domain recommendation
Collaborative filtering (CF) aims to predict users' ratings on items according to historical user-
item preference data. In many real-world applications, preference data are usually sparse …
item preference data. In many real-world applications, preference data are usually sparse …
Reward Certification for Policy Smoothed Reinforcement Learning
Reinforcement Learning (RL) has achieved remarkable success in safety-critical areas, but it
can be weakened by adversarial attacks. Recent studies have introduced``smoothed …
can be weakened by adversarial attacks. Recent studies have introduced``smoothed …
[PDF][PDF] Confidence-Weighted Linear Classification for Text Categorization.
Confidence-weighted online learning is a generalization of margin-based learning of linear
classifiers in which the margin constraint is replaced by a probabilistic constraint based on a …
classifiers in which the margin constraint is replaced by a probabilistic constraint based on a …
Variance-based regularization with convex objectives
J Duchi, H Namkoong - arXiv preprint arXiv:1610.02581, 2016 - arxiv.org
We develop an approach to risk minimization and stochastic optimization that provides a
convex surrogate for variance, allowing near-optimal and computationally efficient trading …
convex surrogate for variance, allowing near-optimal and computationally efficient trading …
Maximum weighted loss discrepancy
Though machine learning algorithms excel at minimizing the average loss over a
population, this might lead to large discrepancies between the losses across groups within …
population, this might lead to large discrepancies between the losses across groups within …