• Publications
  • Influence
On the Algorithmic Implementation of Multiclass Kernel-based Vector Machines
tl;dr
In this paper we describe the algorithmic implementation of multiclass kernel-based vector machines. Expand
  • 2,065
  • 325
  • Open Access
Online Passive-Aggressive Algorithms
tl;dr
We present a unified view for online classification, regression, and uni-class problems through the same algorithmic prism. Expand
  • 1,608
  • 222
  • Open Access
Online Large-Margin Training of Dependency Parsers
tl;dr
We present an effective training algorithm for linearly-scored dependency parsers that implements online large-margin multi-class training (Crammer and Singer, 2003) on top of efficient parsing techniques for dependency trees (Eisner, 1996). Expand
  • 884
  • 190
  • Open Access
A theory of learning from different domains
tl;dr
We address the first question by bounding a classifier’s target error in terms of its source error and the divergence between the two domains. Expand
  • 1,125
  • 175
  • Open Access
Analysis of Representations for Domain Adaptation
tl;dr
We formalize the tradeoffs inherent in designing a representation for domain adaptation and give a new justification for a recently proposed model. Expand
  • 919
  • 124
  • Open Access
On the Learnability and Design of Output Codes for Multiclass Problems
tl;dr
Output coding is a general framework for solving multiclass categorization problems. Expand
  • 703
  • 69
  • Open Access
Adaptive regularization of weight vectors
tl;dr
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-separable data. Expand
  • 322
  • 62
  • Open Access
Ultraconservative online algorithms for multiclass problems
In this paper we study a paradigm to generalize online classification algorithms for binary classification problems to multiclass problems. The particular hypotheses we investigate maintain oneExpand
  • 413
  • 56
Pranking with Ranking
We discuss the problem of ranking instances. In our framework each instance is associated with a rank or a rating, which is an integer from 1 to k. Our goal is to find a rank-predict ion rule thatExpand
  • 612
  • 43
  • Open Access
Confidence-weighted linear classification
tl;dr
We introduce confidence-weighted linear classifiers, which add parameter confidence information to linear classifier. Expand
  • 382
  • 43
  • Open Access