Automatic Code Assignment to Medical Text (2009)
Koby Crammer, Mark Dredze, Kuzman Ganchev, Partha Pratim Talukdar, Steven Carroll
Code assignment is important for handling large amounts of electronic medical data in the modern hospital. However, only expert annotators with extensive training can assign codes. We present a...
DRASO: Declaratively Regularized Alternating Structural Optimization (2009)
Partha Pratim Talukdar, Ted Sandler, Mark Dredze, Koby Crammer, John Blitzer, Fernando Pereira
Recent work has shown that Alternating Structural Optimization (ASO) can improve supervised learners by learning feature representations from unlabeled data. However, there is no natural way to...
A Rate-Distortion One-Class Model and its Applications to Clustering (2009)
Koby Crammer, Partha Pratim Talukdar
In one-class classification we seek a rule to find a coherent subset of instances similar to a few positive examples in a large pool of instances. The problem can be formulated and analyzed naturally...
Feature Design for Transfer Learning (2008)
Mark Dredze, John Blitzer, Koby Crammer, Fernando Pereira
Discriminative learning methods for classification perform well when training and test data are drawn from the same distribution and labeled using the same function. However, often we have labeled...
Penn/UMass/CHOP Biocreative II systems 1 Penn/UMass/CHOP Biocreative II systems (2008)
Kuzman Ganchev, Koby Crammer, Fernando Pereira, Gideon Mann, Kedar Bellare, Andrew Mccallum, ...
Our team participated in the entity tagging and normalization tasks of Biocreative II. For the entity tagging task, we used a k-best MIRA learning algorithm with lexicons and automatically derived...
Reranking candidate gene models with cross-species comparison for improved gene prediction (2008)
Liu, Qian, Crammer, Koby, Pereira, Fernando CN, Roos, David S
Abstract Background Most gene finders score candidate gene models with state-based methods, typically HMMs, by combining local properties (coding potential, splice donor and acceptor patterns, etc)....
Reranking candidate gene models with cross-species comparison for improved gene prediction (2008)
Liu, Qian, Crammer, Koby, Pereira, Fernando CN, Roos, David S
Background: Most gene finders score candidate gene models with state-based methods, typically HMMs, by combining local properties (coding potential, splice donor and acceptor patterns, etc)....
A Rate-Distortion One-Class Model and its Applications to Clustering ∗ (2008)
Koby Crammer, Partha Pratim, Talukdar Fern, O Pereira
We study the problem of one-class classification, in which we seek a rule to separate a coherent subset of instances similar to a few positive examples from a large pool of instances. For instance,...
Batch Performance for an Online Price (2008)
Koby Crammer, Mark Dredze, John Blitzer, O Pereira
Batch learning techniques achieve good performance, but at the cost of many (sometimes even hundreds) of passes over the data. For many tasks, such as web-scale ranking of machine translation...
Global Inference and Learning Algorithms for Multi-Lingual Dependency Parsing (2008)
Ryan Mcdonald, Koby Crammer, Fernando Pereira, Kevin Lerman
This paper gives an overview of the work of McDonald et al. (McDonald et al. 2005a, 2005b; McDonald and Pereira 2006; McDonald et al. 2006) on global inference and learning algorithms for data-driven...
Learning from Multiple Sources (2008)
Crammer, Koby, Kearns, Michael, Wortman, Jennifer
We consider the problem of learning accurate models from multiple sources of "nearby" data. Given distinct samples from multiple data sources and estimates of the dissimilarities between these...
Spanning Tree Methods for Discriminative Training of Dependency Parsers (2008)
Ryan Mcdonald, Koby Crammer, Fernando Pereira
Untyped dependency parsing can be viewed as the problem of finding maximum spanning trees (MSTs) in directed graphs. Using this representation, the Eisner (1996) parsing algorithm is sufficient for...
A Conservative Aggressive Subspace Tracker (2008)
The need to track a subspace accurately describing well a stream of points arises in many signal processing applications. In this work, we present a very efficient algorithm using a machine learning...
Discriminative Learning via Semidefinite Probabilistic Models (2008)
Discriminative linear models are a popular tool in machine learning. These can be generally divided into two types: linear classifiers, such as support vector machines (SVMs), which are well studied...
Automatic Code Assignment to Medical Text (2008)
Koby Crammer, Mark Dredze, Kuzman Ganchev, Partha Pratim Talukdar, Steven Carroll
Code assignment is important for handling large amounts of electronic medical data in the modern hospital. However, only expert annotators with extensive training can assign codes. We present a...
We describe a new family of topic-ranking algorithms for multi-labeled documents. The motivation for the algorithms stems from recent advances in online learning algorithms. The algorithms we present...
Koby Crammer, Ran Gilad-bachrach, Naftali Tishby, Amir Navot
One of the earliest and most powerful machine learning methods is the Learning Vector Quantization (LVQ) algorithm, introduced by Kohonen about 20 years ago. Still, despite its popularity, the...
A Temporal Kernel-Based Model for TrackingHand-Movements from Neural Activities (2008)
Lavi Shpigelman, Koby Crammer, Rony Paz, Eilon Vaadia, Yoram Singer
Koby Crammer, Jaz Kandola, Yoram Singer
Online algorithms for classification often require vast amounts of memory and computation time when employed in conjunction with kernel functions. In this paper we describe and analyze a simple...
Koby Crammer, Ran Gilad-bachrach, Naftali Tishby, Amir Navot
Prototypes based algorithms are commonly used to reduce the computational complexity of Nearest-Neighbour (NN) classifiers. In this paper we discuss theoretical and algorithmical aspects of such...
Koby Crammer, Jaz Kandola, Yoram Singer
Online algorithms for classification often require vast amounts of memory and computation time when employed in conjunction with kernel functions. In this paper we describe and analyze a simple...
Learning bounds for domain adaptation (2008)
John Blitzer, Koby Crammer, Alex Kulesza, O Pereira, Jennifer Wortman
Empirical risk minimization offers well-known learning guarantees when training and test data come from the same domain. In the real world, though, we often wish to adapt a classifier from a source...
Learning bounds for domain adaptation (2008)
John Blitzer, Koby Crammer, Alex Kulesza, O Pereira, Jennifer Wortman
Empirical risk minimization offers well-known learning guarantees when training and test data come from the same domain. In the real world, though, we often wish to adapt a classifier from a source...
Learning bounds for domain adaptation (2008)
John Blitzer, Koby Crammer, Alex Kulesza, O Pereira, Jennifer Wortman
Empirical risk minimization offers well-known learning guarantees when training and test data come from the same domain. In the real world, though, we often wish to adapt a classifier from a source...
Learning to Create Data-Integrating Queries (2008)
Partha Pratim Talukdar, Marie Jacob, Muhammad Salman Mehmood, Koby Crammer, Zachary G. Ives, Fernando Pereira, ...
The number of potentially-related data resources available for querying — databases, data warehouses, virtual integrated schemas — continues to grow rapidly. Perhaps no area has seen this problem...
BMC Bioinformatics BioMed Central Methodology article (2008)
Qian Liu, Koby Crammer, O Cn Pereira, David S Roos
Reranking candidate gene models with cross-species comparison for improved gene prediction
Global Discriminative Learning for Higher-Accuracy Computational Gene Prediction (2007)
Bernal, Axel, Crammer, Koby, Hatzigeorgiou, Artemis, Pereira, Fernando C.N.
Most ab initio gene predictors use a probabilistic sequence model, typically a hidden Markov model, to combine separately trained models of genomic signals and content. By combining separate models...
Global Discriminative Learning for Higher-Accuracy Computational Gene Prediction (2007)
Axel Bernal, Koby Crammer, Artemis Hatzigeorgiou, Fernando Pereira
Most ab initio gene predictors use a probabilistic sequence model, typically a hidden Markov model, to combine separately trained models of genomic signals and content. By combining separate models...
Global Discriminative Training for Higher-Accuracy Computational Gene Prediction (2007)
Axel E Bernal, Koby Crammer, Artemis Hatzigeorgiou, Fernando CN Pereira
Most ab initio gene predictors use a probabilistic sequence model, typically a hidden Markov model, to combine separately-trained models of genomic signals and content. By combining separate models...
Analysis of representations for domain adaptation (2007)
Shai Ben-david, John Blitzer, Koby Crammer, Presented Marina Sokolova
Domain is a distribution D on an instance set X Domain adaptation of a classifier A classification task Source domain (DS)
Learning from multiple sources (2007)
Koby Crammer, Michael Kearns, Jennifer Wortman
We consider the problem of learning accurate models from multiple sources of “nearby ” data. Given distinct samples from multiple data sources and estimates of the dissimilarities between these...
Analysis of representations for domain adaptation (2007)
Shai Ben-david, John Blitzer, Koby Crammer, O Pereira
Discriminative learning methods for classification perform well when training and test data are drawn from the same distribution. In many situations, though, we have labeled training data for a...
Learning from multiple sources (2007)
Koby Crammer, Michael Kearns, Information Science
Abstract We consider the problem of learning accurate models from multiple sources of"nearby " data. Given distinct samples from multiple data sources and estimates of the...
Learning from multiple sources (2007)
Koby Crammer, Michael Kearns, Information Science
Abstract We consider the problem of learning accurate models from multiple sources of"nearby " data. Given distinct samples from multiple data sources and estimates of the...
Learning from multiple sources (2007)
Koby Crammer, Michael Kearns, Jennifer Wortman
We consider the problem of learning accurate models from multiple sources of “nearby ” data. Given distinct samples from multiple data sources and estimates of the dissimilarities between these...
Analysis of representations for domain adaptation (2007)
Shai Ben-david, John Blitzer, Koby Crammer, O Pereira
Discriminative learning methods for classification perform well when training and test data are drawn from the same distribution. In many situations, though, we have labeled training data for a...
Room Impulse Response Estimation using Sparse Online Prediction and Absolute Loss (2006)
The need to accurately and efficiently estimate room impulse responses arises in many acoustic signal processing applications. In this work, we present a general family of algorithms which contain...
Online Passive-Aggressive Algorithms (2006)
Crammer, Koby, Dekel, Ofer, Keshet, Joseph, Shalev-Shwartz, Shai, Singer, Yoram
We present a family of margin based online learning algorithms for various prediction tasks. In particular we derive and analyze algorithms for binary and multiclass categorization, regression,...
Spanning Tree Methods for Discriminative Training of Dependency Parsers (2006)
McDonald, Ryan, Crammer, Koby, Pereira, Fernando C.N.
Untyped dependency parsing can be viewed as the problem of finding maximum spanning trees (MSTs) in directed graphs. Using this representation, the Eisner (1996) parsing algorithm is sufficient for...
Online Passive-Aggressive Algorithms (2006)
Crammer, Koby, Dekel, Ofer, Keshet, Joseph, Shalev-Shwartz, Shai, Singer, Yoram
We present a family of margin based online learning algorithms for various prediction tasks. In particular we derive and analyze algorithms for binary and multiclass categorization, regression,...
Online Passive-Aggressive Algorithms (2006)
Koby Crammer, Ofer Dekel, Joseph Keshet, Shai Shalev-Shwartz, Yoram Singer
We present a family of margin based online learning algorithms for various prediction tasks. In particular we derive and analyze algorithms for binary and multiclass categorization, regression,...
Online passive-aggressive algorithms (2006)
Koby Crammer, Ofer Dekel, Shai Shalev-shwartz, Yoram Singer
We present a unified view for online classification, regression, and uniclass problems. This view leads to a single algorithmic framework for the three problems. We prove worst case loss bounds for...
Tracking Hand Movements from Neuronal Activity with a Dynamic Kernel-Based Model (2005)
Shpigelman, Lavi, Crammer, Koby, Paz, Rony, Vaadia, Eilon, Singer, Yoram
It is well known that population activity in motor cortex can predict movement direction. This allows for development of brain machine interfaces (BMI) that read brain activity and produce movements....
Online Passive-Aggressive Algorithms (2005)
Crammer, Koby, Dekel, Ofer, Keshet, Joseph, Shalev-Shwartz, Shai, Singer, Yoram
We present a family of online learning, margin based, algorithms for various prediction tasks. In particular we derive and analyze algorithms for binary and multiclass categorization, regression,...
A temporal kernel-based model for tracking hand-movements from neural activities (2005)
Lavi Shpigelman, Koby Crammer, Rony Paz, Eilon Vaadia, Yoram Singer
We devise and experiment with a dynamical kernel-based system for tracking hand movements from neural activity. The state of the system corresponds to the hand location, velocity, and acceleration,...
Online large-margin training of dependency parsers (2005)
Ryan Mcdonald, Koby Crammer, Fernando Pereira
We present an effective training algorithm for linearly-scored dependency parsers that implements online largemargin multi-class training (Crammer and Singer, 2003; Crammer et al., 2003) on top of...
Flexible text segmentation with structured multilabel classification (2005)
Ryan Mcdonald, Koby Crammer, Fernando Pereira
Many language processing tasks can be reduced to breaking the text into segments with prescribed properties. Such tasks include sentence splitting, tokenization, named-entity extraction, and...
Scalable large-margin online learning for structured classification (2005)
Koby Crammer, Ryan Mcdonald, Fernando Pereira
We investigate large-margin online learning algorithms for large-scale structured classification tasks, focusing on a structured-output extension of MIRA, the multi-class classification algorithm of...
Online large-margin training of dependency parsers (2005)
Ryan Mcdonald, Koby Crammer, Fernando Pereira
We present an effective training algorithm for linearly-scored dependency parsers that implements online largemargin multi-class training (Crammer and Singer, 2003; Crammer et al., 2003) on top of...
Flexible text segmentation with structured multilabel classification (2005)
Ryan Mcdonald, Koby Crammer, Fernando Pereira
Many language processing tasks can be reduced to breaking the text into segments with prescribed properties. Such tasks include sentence splitting, tokenization, named-entity extraction, and...
Loss bounds for online category ranking (2005)
Abstract. Category ranking is the task of ordering labels with respect to their relevance to an input instance. In this paper we describe and analyze several algorithms for online category ranking...
Loss bounds for online category ranking (2005)
Abstract. Category ranking is the task of ordering labels with respect to their relevance to an input instance. In this paper we describe and analyze several algorithms for online category ranking...
Loss bounds for online category ranking (2005)
Abstract. Category ranking is the task of ordering labels with respect to their relevance to an input instance. In this paper we describe and analyze several algorithms for online category ranking...
A temporal kernel-based model for tracking hand movements fron neural activities (2004)
Shpigelman, Lavi, Crammer, Koby, Paz, Rony, Vaadia, Eilon, Singer, Yoram
We devise and experiment with a dynamical kernel-based system for tracking hand movements from neural activity. The state of the system corresponds to the hand location, velocity, and acceleration,...
A needle in a haystack: Local one-class optimization (2004)
This paper addresses the problem of finding a small and coherent subset of points in a given data. This problem, sometimes referred to as one-class or set covering, requires to find a small-radius...
A needle in a haystack: Local one-class optimization (2004)
This paper addresses the problem of finding a small and coherent subset of points in a given data. This problem, sometimes referred to as one-class or set covering, requires to find a small-radius...
A needle in a haystack: Local one-class optimization (2004)
This paper addresses the problem of finding a small and coherent subset of points in a given data. This problem, sometimes referred to as one-class or set covering, requires to find a small-radius...
A needle in a haystack: Local one-class optimization (2004)
This paper addresses the problem of finding a small and coherent subset of points in a given data. This problem, sometimes referred to as one-class or set covering, requires to find a small-radius...
Online Passive-Aggressive Algorithms (2003)
Crammer, Koby, Dekel, Ofer, Keshet, Joseph, Shalev-Shwartz, Shai, Singer, Yoram
We present a unified view for {\em online} classification, regression, and uniclass problems. This view leads to a single algorithmic framework for the three problems. We prove worst case loss bounds...
IB for General Representations of Data:Bregman Divergence to the Rescue (2003)
The Information Bottleneck (IB) method [1] is an information-theoretic formulation for clus-tering problems. Given a weight distribution p(x) and a set of distributions over some space Y, p(y|x),...
Ultraconservative online algorithms for multiclass problems (2003)
Abstract. In this paper we study online classication algorithms for multiclass problems in the mistake bound model. The hypotheses we use maintain one prototype vector per class. Given an input...
A family of additive online algorithms for category ranking (2003)
Koby Crammer, Yoram Singer, Jaz K, Thomas Hofmann, Tomaso Poggio, John Shawe-taylor
We describe a new family of topic-ranking algorithms for multi-labeled documents. The motivation for the algorithms stem from recent advances in online learning algorithms. The algorithms are simple...
Learning Algorithms for Enclosing Points in Bregmanian Spheres (2003)
We discuss the problem of nding a generalized sphere that encloses points originating from a single source. The points contained in such a sphere are within a maximal divergence from a center point....
Ultraconservative Online Algorithms for Multiclass Problems (2003)
Koby Crammer, Yoram Singer, K. Warmuth
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 one...
Kernel design using boosting (2003)
Koby Crammer, Joseph Keshet, Yoram Singer
The focus of the paper is the problem of learning kernel operators from empirical data. We cast the kernel design problem as the construction of an accurate kernel from simple (and less accurate)...
Ultraconservative online algorithms for multiclass problems (2003)
Koby Crammer, Yoram Singer, K. Warmuth
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 one...
A family of additive online algorithms for category ranking (2003)
Koby Crammer, Yoram Singer, Jaz K, Thomas Hofmann, Tomaso Poggio, John Shawe-taylor
We describe a new family of topic-ranking algorithms for multi-labeled documents. The motivation for the algorithms stem from recent advances in online learning algorithms. The algorithms are simple...
Learning Algorithms for Enclosing Points in (2003)
Bregmanian Spheres Koby, Koby Crammer, Yoram Singer
We discuss the problem of finding a generalized sphere that encloses points originating from a single source. The points contained in such a sphere are within a maximal divergence from a center point.
Online Passive-Aggressive Algorithms (2003)
Koby Crammer, Ofer Dekel, Shai Shalev-Shwartz, Yoram Singer
We present a unified view for online classification, regression, and uniclass problems. This view leads to a single algorithmic framework for the three problems. We prove worst case loss bounds for...
Online Classification on a Budget (2003)
Koby Crammer, Jaz Kandola, Royal Holloway, Yoram Singer
Online algorithms for classification often require vast amounts of memory and computation time when employed in conjunction with kernel functions. In this paper we describe and analyze a simple...
Kernel design using boosting (2003)
Koby Crammer, Joseph Keshet, Yoram Singer
The focus of the paper is the problem of learning kernel operators from empirical data. We cast the kernel design problem as the construction of an accurate kernel from simple (and less accurate)...
A new family of online algorithms for category ranking (2002)
We describe a new family of topic-ranking algorithms for multi-labeled documents. The motivation for the algorithms stems from recent advances in online learning algorithms. The algorithms we present...
Margin Analysis of the LVQ Algorithm (2002)
Koby Crammer, Ran Gilad-bachrach, Amir Navot, Naftali Tishby
Prototypes based algorithms are commonly used to reduce the computational complexity of Nearest-Neighbour (NN) classifiers. In this paper we discuss theoretical and algorithmical aspects of such...
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 nd a rank-prediction rule that...
On the algorithmic implementation of multiclass kernel-based vector machines (2001)
Koby Crammer, Yoram Singer, Nello Cristianini, John Shawe-taylor, Bob Williamson
In this paper we describe the algorithmic implementation of multiclass kernel-based vector machines. Our starting point is a generalized notion of the margin to multiclass problems. Using this notion...
On the Algorithmic Implementation of Multiclass Kernel-based Vector Machines (2001)
Koby Crammer, Yoram Singer, Nello Cristianini, John Shawe-taylor, Bob Williamson
In this paper we describe the algorithmic implementation of multiclass kernel-based vector machines. Our starting point is a generalized notion of the margin to multiclass problems.
Improved output coding for classification using continuous relaxation (2001)
Output coding is a general method for solving multiclass problems by reducing them to multiple binary classification problems. Previous research on output coding has employed, almost solely,...
On the learnability and design of output codes for multiclass problems (2000)
Output coding is a general framework for solving multiclass categorization problems. Previous research on output codes has focused on building multiclass machines given predefined output codes. In...
On the learnability and design of output codes for multiclass problems (2000)
Output coding is a general framework for solving multiclass categorization problems. Previous research on output codes has focused on building multiclass machines given predefined output codes. In...
On the Learnability and Design of Output Codes for Multiclass Problems (2000)
Output coding is a general framework for solving multiclass categorization problems. Previous research on output codes has focused on building multiclass machines given predefined output codes. In...
Learning from data of variable quality (1995)
Koby Crammer, Michael Kearns, Jennifer Wortman
We initiate the study of learning from multiple sources of limited data, each of which may be corrupted at a different rate. We develop a complete theory of which data sources should be used for two...
Learning from data of variable quality (1995)
Koby Crammer, Michael Kearns, Jennifer Wortman
We initiate the study of learning from multiple sources of limited data, each of which may be corrupted at a different rate. We develop a complete theory of which data sources should be used for two...
Learning from data of variable quality (1995)
Koby Crammer, Michael Kearns, Jennifer Wortman
We initiate the study of learning from multiple sources of limited data, each of which may be corrupted at a different rate. We develop a complete theory of which data sources should be used for two...
Global Discriminative Learning for Higher-Accuracy Computational Gene Prediction
Bernal, Axel, Crammer, Koby, Hatzigeorgiou, Artemis, Pereira, Fernando
Most ab initio gene predictors use a probabilistic sequence model, typically a hidden Markov model, to combine separately trained models of genomic signals and content. By combining separate models...