Person
Pranay Dighe
This person is no longer with EPFL
Related publications (11)
On quantifying the quality of acoustic models in hybrid DNN-HMM ASR
Hervé Bourlard, Afsaneh Asaei, Pranay Dighe
We propose an information theoretic framework for quantitative assessment of acoustic models used in hidden Markov model (HMM) based automatic speech recognition (ASR). The HMM backend expects that (i) the acoustic model yields accurate state conditional e ...
ELSEVIER2020Sparse and Low-rank Modeling for Automatic Speech Recognition
This thesis deals with exploiting the low-dimensional multi-subspace structure of speech towards the goal of improving acoustic modeling for automatic speech recognition (ASR). Leveraging the parsimonious hierarchical nature of speech, we hypothesize that ...
EPFL2019Low-rank and sparse subspace modeling of speech for DNN based acoustic modeling
Hervé Bourlard, Afsaneh Asaei, Pranay Dighe
Towards the goal of improving acoustic modeling for automatic speech recognition (ASR), this work investigates the modeling of senone subspaces in deep neural network (DNN) posteriors using low-rank and sparse modeling approaches. While DNN posteriors are ...
ELSEVIER SCIENCE BV2019Far-field ASR Using Low-rank and Sparse Soft Targets from Parallel Data
Hervé Bourlard, Afsaneh Asaei, Pranay Dighe
Far-field automatic speech recognition (ASR) of conversational speech is often considered to be a very challenging task due to the poor quality of alignments available for training the DNN acoustic models. A common way to alleviate this problem is to use c ...
IEEE2018Low-rank and Sparse Soft Targets to Learn Better DNN Acoustic Models
Hervé Bourlard, Afsaneh Asaei, Pranay Dighe
Conventional deep neural networks (DNN) for speech acoustic modeling rely on Gaussian mixture models (GMM) and hidden Markov model (HMM) to obtain binary class labels as the targets for DNN training. Subword classes in speech recognition systems correspond ...
Ieee2017Sparse Modeling of Neural Network Posterior Probabilities for Exemplar-based Speech Recognition
Hervé Bourlard, Afsaneh Asaei, Pranay Dighe
In this paper, a compressive sensing (CS) perspective to exemplar-based speech processing is proposed. Relying on an analytical relationship between CS formulation and statistical speech recognition (Hidden Markov Models HMM), the automatic speech recognit ...
2016Exploiting Low-dimensional Structures to Enhance DNN based Acoustic Modeling in Speech Recognition
Hervé Bourlard, Afsaneh Asaei, Pranay Dighe
We propose to model the acoustic space of deep neural network (DNN) class-conditional posterior probabilities as a union of low- dimensional subspaces. To that end, the training posteriors are used for dictionary learning and sparse coding. Sparse represen ...
IEEE2016Exploiting Low-dimensional Structures to Enhance DNN based Acoustic Modeling in Speech Recognition
Hervé Bourlard, Afsaneh Asaei, Pranay Dighe
We propose to model the acoustic space of deep neural network (DNN) class-conditional posterior probabilities as a union of lowdimensional subspaces. To that end, the training posteriors are used for dictionary learning and sparse coding. Sparse representa ...
ICASSP2016Sparse Hidden Markov Models for Exemplar-based Speech Recognition Using Deep Neural Network Posterior Features
Hervé Bourlard, Afsaneh Asaei, Pranay Dighe
Statistical speech recognition has been cast as a natural realization of the compressive sensing problem in this work. The compressed acoustic observations are sub-word posterior probabilities obtained from a deep neural network. Dictionary learning and sp ...
Idiap2016Low-Rank Representation of Nearest Neighbor Phone Posterior Probabilities to Enhance DNN Acoustic Modeling
Hervé Bourlard, Afsaneh Asaei, Pranay Dighe
We hypothesize that optimal deep neural networks (DNN) class-conditional posterior probabilities live in a union of low-dimensional subspaces. In real test conditions, DNN posteriors encode uncertainties which can be regarded as a superposition of unstruct ...
Idiap2016