Publications de Afsaneh Asaei | EPFL Graph Search

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 ...

ELSEVIER2020

Low-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 BV2019

Phonetic Subspace Features for Improved Query by Example Spoken Term Detection

Hervé Bourlard, Afsaneh Asaei, Dhananjay Ram

This paper addresses the problem of detecting speech utterances from a large audio archive using a simple spoken query, hence referring to this problem as "Query by Example Spoken Term Detection" (QbE-STD). This still open pattern matching problem has been ...

2018

Far-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 ...

IEEE2018

Phonological Posterior Hashing for Query by Example Spoken Term Detection

Hervé Bourlard, Afsaneh Asaei, Dhananjay Ram

State of the art query by example spoken term detection (QbE-STD) systems in zero-resource conditions rely on representation of speech in terms of sequences of class-conditional posterior probabilities estimated by deep neural network (DNN). The posteriors ...

ISCA-INT SPEECH COMMUNICATION ASSOC2018

Sparse Subspace Modeling for Query by Example Spoken Term Detection

Hervé Bourlard, Afsaneh Asaei, Dhananjay Ram

2018

Perceptual Information Loss due to Impaired Speech Production

Hervé Bourlard, Afsaneh Asaei, Milos Cernak

Phonological classes define articulatory-free and articulatory-bound phone attributes. Deep neural network is used to estimate the probability of phonological classes from the speech signal. In theory, a unique combination of phone attributes form a phonem ...

2017

Low-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 ...

Ieee2017

Cognitive Speech Coding Examining the impact of cognitive speech processing on speech compression

Afsaneh Asaei, Milos Cernak

IEEE2017

Exploiting 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 ...

IEEE2016

Composition of Deep and Spiking Neural Networks for Very Low Bit Rate Speech Coding

Afsaneh Asaei, Philip Neil Garner, Milos Cernak

2016

On Structured Sparsity of Phonological Posteriors for Linguistic Parsing

Hervé Bourlard, Afsaneh Asaei, Milos Cernak

The speech signal conveys information on different time scales from short (20--40 ms) time scale or segmental, associated to phonological and phonetic information to long (150--250 ms) time scale or supra segmental, associated to syllabic and prosodic info ...

2016

Subspace Detection of DNN Posterior Probabilities via Sparse Representation for Query by Example Spoken Term Detection

Hervé Bourlard, Afsaneh Asaei, Dhananjay Ram

We cast the query by example spoken term detection (QbE-STD) problem as subspace detection where query and background subspaces are modeled as union of low-dimensional subspaces. The speech exemplars used for subspace modeling are class-conditional posteri ...

Idiap2016

Sparse 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 ...

2016

Binary Sparse Coding of Convolutive Mixtures for Sound Localization and Separation via Spatialization

Hervé Bourlard, Afsaneh Asaei, Mohammadjavad Taghizadeh, Volkan Cevher, Saeid Haghighatshoar

We propose a sparse coding approach to address the problem of source-sensor localization and speech reconstruction. This approach relies on designing a dictionary of spatialized signals by projecting the microphone array recordings into the array manifolds ...

Institute of Electrical and Electronics Engineers2016

Ad Hoc Microphone Array Calibration: Euclidean Distance Matrix Completion Algorithm and Theoretical Guarantees

Hervé Bourlard, Reza Parhizkar, Afsaneh Asaei, Mohammadjavad Taghizadeh, Philip Neil Garner

This paper addresses the problem of ad hoc microphone array calibration where only partial information about the distances between the microphones is available. We construct a matrix consisting of the pairwise distances and propose to estimate the missing ...

2015

On Application Of Non-Negative Matrix Factorization for Ad Hoc Microphone Array Calibration from Incomplete Noisy Distances

Hervé Bourlard, Afsaneh Asaei, Mohammadjavad Taghizadeh

We propose to use non-negative matrix factorization (NMF) to estimate the unknown pairwise distances and reconstruct a distance matrix for microphone array position calibration. We develop new multiplicative update rules for NMF with incomplete input matri ...

2015

Robust Microphone Placement for Source Localization from Noisy Distance Measurements

Hervé Bourlard, Afsaneh Asaei, Mohammadjavad Taghizadeh, Philip Neil Garner, Saeid Haghighatshoar

We propose a novel algorithm to design an optimum array geometry for source localization inside an enclosure. We assume a square-law decay propagation model for the sound acquisition so that the additive noise on the measured source-microphone distances is ...

2015

Novel GCC-PHAT Model in Diffuse Sound Field for Microphone Array Pairwise Distance Based Calibration

Hervé Bourlard, Afsaneh Asaei, Mohammadjavad Taghizadeh

We propose a novel formulation of the generalized cross correlation with phase transform (GCC-PHAT) for a pair of microphones in diffuse sound field. This formulation elucidates the links between the microphone distances and the GCC-PHAT output. Hence, it ...

2015

On Compressibility of Neural Network phonological Features for Low Bit Rate Speech Coding

Hervé Bourlard, Afsaneh Asaei, Milos Cernak

Phonological features extracted by neural network have shown interesting potential for low bit rate speech vocoding. The span of phonological features is wider than the span of phonetic features, and thus fewer frames need to be transmitted. Moreover, the ...

2015