Machine learning

VOT

Automatic detection of voiced onset time (VOT) for assessment of pathological speech

Automatic recognition of emotions from speech

Human emotions detection considering speech signals is a field that has attracted the attention of the research community since the last years. Several situations where the human integrity and security is at risk have been addressed; particularly the analysis of speech in emergency calls or in call-centers, are an interesting scenario. This project aimed to develop a methodology to classify different types of emotions such as anger, anxiety, disgust, and desperation, in scenarios where the speech signal is contaminated with noise or is coded by telephone channels.

Development of machine learning methods to analyze and to characterize the energy consumption of Oil & Gas Colombian companies

A significant amount of energy is wasted in industrial operations due to several factors such as electricity theft, fraud, billing errors, and fault devices, among others. It is important to monitor the behavior of the electricity consumption in order to improve energy usage efficiency. Smart meters have been developed in order to collect information about the electricity consumption behaviors and lifestyles of the consumers. One of the most important applications to monitor energy consumption corresponds to anomaly detection. The main challenge in this application is the lack of fully labeled datasets with annotated information about the presence of anomalies in the time-series. The prpject proposes an unsupervised approach where a deep learning strategy that combines autoencoders and recurrent neural networks is used to detect anomalies in time-series of energy consumption.

2016 Third Frederick Jelinek Memorial Summer Workshop

Remote Monitoring of Neurodegeneration through Speech

Multimodal PD

Asynchronous Non-Intrusive Multi-Modal Analysis of Bio-Signals for the Automatic evaluation of the Neurological State of People With Parkinson's Disease

TAPAS

Training Network on Automatic Processing of PAthological Speech

Natural Language Analysis to Detect Parkinson's Disease

Parkinson’s disease is a neuro-degenerative disorder characterized by different motor symptoms, including several gait impairParkinson’s disease (PD) is a neuro-degenerative disorder that produces motor and non-motor impairments. Non-motor …

Feature Space Visualization with Spatial Similarity Maps for Pathological Speech Data

The feature vectors of a data set encode information about relations between speaker groups, clusters and outliers. Based on the assumption that these relations are conserved within the spatial properties of feature vectors, we introduce similarity …

Analysis and evaluation of handwriting in patients with Parkinson’s disease using kinematic, geometrical, and non-linear features

Background and objectives: Parkinson’s disease is a neurological disorder that affects the motor system producing lack of coordination, resting tremor, and rigidity. Impairments in handwriting are among the main symptoms of the disease. Handwriting …

Automated Cross-language Intelligibility Analysis of Parkinson’s Disease Patients Using Speech Recognition Technologies

Speech deficits are common symptoms amongParkinson’s Disease (PD) patients. The automatic assessment of speech signals is promising for the evaluation of the neurological state and the speech quality of the patients. Recently, progress has been made …