Publications de Milos Cernak | EPFL Graph Search

Serab: A Multi-Lingual Benchmark For Speech Emotion Recognition

Recent developments in speech emotion recognition (SER) often leverage deep neural networks (DNNs). Comparing and benchmarking different DNN models can often be tedious due to the use of different datasets and evaluation protocols. To facilitate the proces ...

IEEE2022

Word-level Embeddings for Cross-Task Transfer Learning in Speech Processing

Pierre Anton Beckmann, Milos Cernak

Recent breakthroughs in deep learning often rely on representation learning and knowledge transfer. In recent years, unsupervised and self-supervised techniques for learning speech representation were developed to foster automatic speech recognition. Up to ...

EUROPEAN ASSOC SIGNAL SPEECH & IMAGE PROCESSING-EURASIP2021

Non-Intrusive Speech Quality Assessment with Transfer Learning and Subject-specific Scaling

Paolo Prandoni, Milos Cernak, Natalia Nessler

In communication systems, it is crucial to estimate the perceived quality of audio and speech. The industrial standards for many years have been PESQ, 3QUEST, and POLQA, which are intrusive methods. This restricts the possibilities of using these metrics i ...

ISCA-INT SPEECH COMMUNICATION ASSOC2021

Spiking Neural Networks Trained With Backpropagation For Low Power Neuromorphic Implementation Of Voice Activity Detection

Milos Cernak, Giorgia Dellaferrera

Recent advances in Voice Activity Detection (VAD) are driven by artificial and Recurrent Neural Networks (RNNs), however, using a VAD system in battery-operated devices requires further power efficiency. This can be achieved by neuromorphic hardware, which ...

IEEE2020

A Bin Encoding Training Of A Spiking Neural Network Based Voice Activity Detection

Milos Cernak, Giorgia Dellaferrera

Advances of deep learning for Artificial Neural Networks (ANNs) have led to significant improvements in the performance of digital signal processing systems implemented on digital chips. Although recent progress in low-power chips is remarkable, neuromorph ...

IEEE2020

Open-Vocabulary Keyword Spotting with Audio and Text Embeddings

Milos Cernak, Niccolò Sacchi, Martin Jaggi

Keyword Spotting (KWS) systems allow detecting a set of spoken (pre-defined) keywords. Open-vocabulary KWS systems search for the keywords in the set of word hypotheses generated by an automatic speech recognition (ASR) system which is computationally expe ...

2019

NeuroSpeech: An open-source software for Parkinson's speech analysis

Milos Cernak

A new software for modeling pathological speech signals is presented in this paper. The software is called NeuroSpeech. This software enables the analysis of pathological speech signals considering different speech dimensions: phonation, articulation, pros ...

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

Speech vocoding for laboratory phonology

Milos Cernak

2017

Characterisation of voice quality of Parkinson's disease using differential phonological posterior features

Milos Cernak

2017

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

Afsaneh Asaei, Milos Cernak

IEEE2017

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

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

Afsaneh Asaei, Philip Neil Garner, Milos Cernak

2016

Incremental Syllable-Context Phonetic Vocoding

Petr Motlicek, Philip Neil Garner, Milos Cernak

Current very low bit rate speech coders are, due to complexity limitations, designed to work off-line. This paper investigates incremental speech coding that operates real-time and incrementally (i.e., encoded speech depends only on already-uttered speech ...

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

A Simple Continuous Pitch Estimation Algorithm

Petr Motlicek, Philip Neil Garner, Milos Cernak

Recent work in text to speech synthesis has pointed to the benefit of using a continuous pitch estimate; that is, one that records pitch even when voicing is not present. Such an approach typically requires interpolation. The purpose of this paper is to sh ...

IEEE Institute of Electrical and Electronics Engineers2013