• Publications
  • Influence
The SuperSID project: exploiting high-level information for high-accuracy speaker recognition
This paper provides an overview of the structure, data, task, tools, and accomplishments of the SuperSID project. Expand
Short-time Gaussianization for robust speaker verification
In this paper, a novel approach for robust speaker verification, namely short-timeGaussianization, is proposed. Expand
Spoken language recognition-a step toward multilinguality in speech processing
  • Jirí Navrátil
  • Computer Science
  • IEEE Trans. Speech Audio Process.
  • 1 September 2001
We present systems for language identification as well as for unknown-language rejection based on phonotactic-acoustic features and presents a framework for experiments and discussion on these two tasks. Expand
A hybrid GMM/SVM approach to speaker identification
A hybrid system which appropriately combines the advantages of both the generative and discriminant model paradigms is described and experimentally evaluated. Expand
Using prosodic and conversational features for high-performance speaker recognition: report from JHU WS'02
We report here on work from the JHU 2002 Summer Workshop exploring a range of prosodic features, using as testbed the 2001 NIST Extended Data task. Expand
The MIT-LL/IBM 2006 Speaker Recognition System: High-Performance Reduced-Complexity Recognition
In 2006, MIT Lincoln Laboratory focused on a new system architecture which emphasized reduced complexity. Expand
Syntax Based Reordering with Automatically Derived Rules for Improved Statistical Machine Translation
We present a simple, automatic method to learn rules that reorder source sentences to more closely match the target language word order using only a source side parse tree and automatically generated alignments. Expand
A Word Reordering Model for Improved Machine Translation
We use the Lin-Kernighan heuristic to find the best source reordering efficiently during training and testing and show that it suffices to provide good quality reordering. Expand
Combining cross-stream and time dimensions in phonetic speaker recognition
We present an approach and results aimed at modeling the statistical pronunciation patterns across multiple phone streams, which we refer to as phonetic information in the cross-stream (cross-language) dimension. Expand
Confidence Scoring Using Whitebox Meta-models with Linear Classifier Probes
We propose a novel confidence scoring mechanism for deep neural networks based on a two-model paradigm involving a base model and a meta-model. Expand