Patrick Haffner

Bio: Patrick Haffner is an academic researcher from AT&T Labs. The author has contributed to research in topics: Support vector machine & Speaker recognition. The author has an hindex of 32, co-authored 97 publications receiving 42604 citations. Previous affiliations of Patrick Haffner include Nuance Communications & Carnegie Mellon University.

A Machine Learning Framework for Spoken-Dialog Classification

Abstract: One of the key tasks in the design of large-scale dialog systems is classification. This consists of assigning, out of a finite set, a specific category to each spoken utterance, based on the output of a speech recognizer. Classification in general is a standard machine-learning problem, but the objects to classify in this particular case are word lattices, or weighted automata, and not the fixed-size vectors for which learning algorithms were originally designed. This chapter presents a general kernel-based learning framework for the design of classification algorithms for weighted automata. It introduces a family of kernels, rational kernels, that combined with support vector machines form powerful techniques for spoken-dialog classification and other classification tasks in text and speech processing. It describes efficient algorithms for their computation and reports the results of their use in several difficult spoken-dialog classification tasks based on deployed systems. Our results show that rational kernels are easy to design and implement, and lead to substantial improvements of the classification accuracy. The chapter also provides some theoretical results helpful for the design of rational kernels.

Accelerated Parallel Optimization Methods for Large Scale Machine Learning.

Abstract: The growing amount of high dimensional data in different machine learning applications requires more efficient and scalable optimization algorithms. In this work, we consider combining two techniques, parallelism and Nesterov's acceleration, to design faster algorithms for L1-regularized loss. We first simplify BOOM, a variant of gradient descent, and study it in a unified framework, which allows us to not only propose a refined measurement of sparsity to improve BOOM, but also show that BOOM is provably slower than FISTA. Moving on to parallel coordinate descent methods, we then propose an efficient accelerated version of Shotgun, improving the convergence rate from $O(1/t)$ to $O(1/t^2)$. Our algorithm enjoys a concise form and analysis compared to previous work, and also allows one to study several connected work in a unified way.

Conversion of digital documents to multilayer raster formats

Abstract: How can we turn the description of a digital (i.e. electronically produced) document into something that is efficient for multi-layer raster formats? It is first shown that a foreground/background segmentation without overlapping foreground components can be more efficient for viewing or printing. Then, a new algorithm that prevents overlaps between foreground components while optimizing both the document quality and compression ratio is derived from the minimum description length (MDL) criterion. This algorithm makes the DjVu compression format significantly, more efficient on electronically produced documents. Comparisons with other formats are provided.

AT&T Research at 2007.

Abstract: AT&T participated in two tasks at TRECVID 2007: shot boundary detection (SBD) and rushes summarization. The SBD system developed for TRECVID 2006 was enhanced for robustness and efficiency. New visual features are extracted for cut, dissolve, and fast dissolve detectors, and SVM based verification method is used to boost the accuracy. The speed is improved by a more streamlined processing with on-the-fly result fusion. We submitted 10 runs for SBD evaluation task. The best result (TT05) was achieved with the following configuration: SVM based verification method; more training data that includes 2004, 2005, and 2006 SBD data; no SVM boundary adjustment; training SVM with high generalization capability (e.g., a smaller value of C). As a pilot task, rushes summarization aims to show the main objects and events in the raw material with least redundancy while maximizing the usability. We proposed a multimodal rushes summarization method that relies on both face and speech information. Evaluation results show that the new SBD system is highly effective and the human centric rushes summarization approach is concise and easy to understand.

GMM/SVM N-best speaker identification under mismatch channel conditions

Abstract: Under severe channel mismatch conditions, such as training with far-field speech and testing with telephone data, performance of speaker identification (SID) degrades significantly, often below practical use. But for many SID tasks, it is sufficient to recognize an N-best list of speakers for further human analysis. We investigate N-best SID accuracy for matched (telephone/telephone) and mismatched (far-field/telephone) train/test channel conditions. Using an SVM-GMM supervector (GSV), pitch and formant frequency histograms (PFH) and cross-channel adaptation using cohorts, we reduced matched channel error rate by over 25% relative to the baseline (GMM-UBM), for top-1, and achieved mismatched N-best accuracy comparable to the baseline.

Deep learning

Abstract: Deep learning allows computational models that are composed of multiple processing layers to learn representations of data with multiple levels of abstraction. These methods have dramatically improved the state-of-the-art in speech recognition, visual object recognition, object detection and many other domains such as drug discovery and genomics. Deep learning discovers intricate structure in large data sets by using the backpropagation algorithm to indicate how a machine should change its internal parameters that are used to compute the representation in each layer from the representation in the previous layer. Deep convolutional nets have brought about breakthroughs in processing images, video, speech and audio, whereas recurrent nets have shone light on sequential data such as text and speech.

Gradient-based learning applied to document recognition

Abstract: Multilayer neural networks trained with the back-propagation algorithm constitute the best example of a successful gradient based learning technique. Given an appropriate network architecture, gradient-based learning algorithms can be used to synthesize a complex decision surface that can classify high-dimensional patterns, such as handwritten characters, with minimal preprocessing. This paper reviews various methods applied to handwritten character recognition and compares them on a standard handwritten digit recognition task. Convolutional neural networks, which are specifically designed to deal with the variability of 2D shapes, are shown to outperform all other techniques. Real-life document recognition systems are composed of multiple modules including field extraction, segmentation recognition, and language modeling. A new learning paradigm, called graph transformer networks (GTN), allows such multimodule systems to be trained globally using gradient-based methods so as to minimize an overall performance measure. Two systems for online handwriting recognition are described. Experiments demonstrate the advantage of global training, and the flexibility of graph transformer networks. A graph transformer network for reading a bank cheque is also described. It uses convolutional neural network character recognizers combined with global training techniques to provide record accuracy on business and personal cheques. It is deployed commercially and reads several million cheques per day.

Going deeper with convolutions

Abstract: We propose a deep convolutional neural network architecture codenamed Inception that achieves the new state of the art for classification and detection in the ImageNet Large-Scale Visual Recognition Challenge 2014 (ILSVRC14). The main hallmark of this architecture is the improved utilization of the computing resources inside the network. By a carefully crafted design, we increased the depth and width of the network while keeping the computational budget constant. To optimize quality, the architectural decisions were based on the Hebbian principle and the intuition of multi-scale processing. One particular incarnation used in our submission for ILSVRC14 is called GoogLeNet, a 22 layers deep network, the quality of which is assessed in the context of classification and detection.

The Nature of Statistical Learning Theory

Abstract: Setting of the learning problem consistency of learning processes bounds on the rate of convergence of learning processes controlling the generalization ability of learning processes constructing learning algorithms what is important in learning theory?.

Generative Adversarial Nets

Abstract: We propose a new framework for estimating generative models via an adversarial process, in which we simultaneously train two models: a generative model G that captures the data distribution, and a discriminative model D that estimates the probability that a sample came from the training data rather than G. The training procedure for G is to maximize the probability of D making a mistake. This framework corresponds to a minimax two-player game. In the space of arbitrary functions G and D, a unique solution exists, with G recovering the training data distribution and D equal to ½ everywhere. In the case where G and D are defined by multilayer perceptrons, the entire system can be trained with backpropagation. There is no need for any Markov chains or unrolled approximate inference networks during either training or generation of samples. Experiments demonstrate the potential of the framework through qualitative and quantitative evaluation of the generated samples.