• Publications
  • Influence
Gradient-based learning applied to document recognition
This paper reviews various methods applied to handwritten character recognition and compares them on a standard handwritten digit recognition task, and Convolutional neural networks are shown to outperform all other techniques. Expand
Support vector machines for histogram-based image classification
It is observed that a simple remapping of the input x(i)-->x(i)(a) improves the performance of linear SVM's to such an extend that it makes them, for this problem, a valid alternative to RBF kernels. Expand
Object Recognition with Gradient-Based Learning
This paper attempts to show that for recognizing simple objects with high shape variability such as handwritten characters, it is possible, and even advantageous, to feed the system directly with minimally processed images and to rely on learning to extract the right set of features. Expand
ACAS: automated construction of application signatures
This paper applies three statistical machine learning algorithms to automatically identify signatures for a range of applications and finds that this approach is highly accurate and scales to allow online application identification on high speed links. Expand
High quality document image compression with "DjVu"
A new image compression technique called DjVu is presented that enables fast transmission of document images over low-speed connections, while faithfully reproducing the visual aspect of the document, including color, fonts, pictures, and paper texture. Expand
Rational Kernels: Theory and Algorithms
A general family of kernels based on weighted transducers or rational relations, rational kernels, that extend kernel methods to the analysis of variable-length sequences or more generally weighted automata and show that rational kernels are easy to design and implement and lead to substantial improvements of the classification accuracy. Expand
SVMs for Histogram Based Image Classification
Traditional classification approaches generalize poorly on image classification tasks, because of the high dimensionality of the feature space. This paper shows that Support Vector Machines (SVM) canExpand
Optimizing SVMs for complex call classification
A global optimization process based on an optimal channel communication model that allows a combination of possibly heterogeneous binary classifiers to decrease the call-type classification error rate for AT&T's How May I Help You (HMIHY/sup (sm)/) natural dialog system by 50 % is proposed. Expand
Scaling large margin classifiers for spoken language understanding
  • P. Haffner
  • Computer Science
  • Speech Commun.
  • 1 March 2006
This paper provides an original and unified presentation of these algorithms within the framework of regularized and large margin linear classifiers, reviews some available optimization techniques, and offers practical solutions to scaling issues. Expand
A Modular Machine Learning System for Flow-Level Traffic Classification in Large Networks
A novel two-step model, which seamlessly integrates these collective traffic statistics into the existing traffic classification system is proposed, which easily scales to classify traffic on 10Gbps links and displays performance improvement on all traffic classes and an overall error rate reduction. Expand