Weightless Neural Networks (WNN) are good candidates for Federated Learning scenarios due to their robustness and computational lightness.In this work, we show that it is possible to aggregate the knowledge of multiple WNNs using more compact data structures, such as Bloom Filters, to reduce the amount of data transferred between devices.Finally, we explore variations of Bloom Filters and found that a particular data-structure, the Count-Min Sketch (CMS), is a good candidate for aggregation.Costing at most 3% of accuracy, CMS can be up to 3x smaller when compared to previous approaches, specially for large datasets.

The outsourcing of machine learning classification and data mining tasks can be an effective solution for those parties that need machine learning services, but lack the appropriate resources, knowledge and/or tools to carry them out, in their own premises. This solution, however, raises major privacy concerns, in particular, when irrevocable biometric data such as speech is involved. In this work, we focus on the development of privacy-preserving schemes in a speech emotion recognition task, as a proof of concept that could be extended to other speech analytics tasks. Our aim is to prove that the implementation of privacy-preserving speech mining schemes in challenging tasks involving paralinguistic features is not only feasible, but also accurate. Using distance-preserving hashing techniques in a first approach, and homomorphic encryption in a second approach, we successfully protect sensitive data with little degradation costs regarding the accuracy of the predictive models.

For many applications it is important to be able to detect what a human is currently doing. This ability is useful for applications such as surveillance, human computer interfaces, games and health-care. In order to recognize a human action, the typical approach is to use manually labeled data to perform supervised training. This paper aims to compare the performance of several supervised classifiers trained with manually labeled data versus the same classifiers trained with data automatically labeled. In this paper we propose a framework capable of recognizing human actions using supervised classifiers trained with automatically labeled data in RGB-D videos.

In our daily activities we perform prediction or anticipation when interacting with other humans or with objects. Prediction of human activity made by computers has several potential applications: surveillance systems, human computer interfaces, sports video analysis, human-robot-collaboration, games and health-care. We propose a system capable of recognizing and predicting human actions using supervised classifiers trained with automatically labeled data evaluated in our human activity RGB-D dataset (recorded with a Kinect sensor) and using only the position of the main skeleton joints to extract features. Using conditional random fields (CRFs) to model the sequential nature of actions in a sequence has been used before, but where other approaches try to predict an outcome or anticipate ahead in time (seconds), we try to predict what will be the next action of a subject. Our results show an activity prediction accuracy of 89.9% using an automatically labeled dataset.

Gaze information has the potential to benefit Human-Computer Interaction (HCI) tasks, particularly when combined with speech. Gaze can improve our understanding of the user intention, as a secondary input modality, or it can be used as the main input modality by users with some level of permanent or temporary impairments. In this paper we describe a multimodal HCI system prototype which supports speech, gaze and the combination of both. The system has been developed for Active Assisted Living scenarios.

The goal of any automatic process is to run without human intervention. However, it is not uncommon that unexpected issues recur, requiring repetitive human intervention in order to processes execute successfully. Such situations indicate that those processes were not designed with the necessary quality requirements to achieve the automation goal, and quality needs to be improved. This paper describes a case study of the quality improvements of several automatic ETL processes that required a growing amount of effort to ensure its successful execution. Key to the improvement process was the use of the ISO/IEC 25010 Standard for Systems and Software Quality Requirements and Evaluation to categorize all issues found. This approach proved beneficial in two important ways: first, it created a clear understanding of the overall ETL processes quality problems, second, it made obvious which issues deserved uttermost attention.

The elderly population is quickly increasing in the developed countries. However, in European Portuguese (EP) no studies have examined the impact of age-related structural changes in speech acoustics. The purpose of this paper is to analyse the effect of age ([60-70], [71-80] and [81-90]), gender and type of vowel in the acoustic characteristics (fundamental frequency (F0), first formant (F1), second formant (F2) and duration) of the EP vowels. A sample of 78 speakers was selected from the database of elderly speech collected by Microsoft Language Development Center (MLDC) within the Living Usability Lab (LUL) project. It was observed that duration is the only parameter that significantly changes with ageing, being the highest value found in the [81-90] group. Moreover, F0 decreases in females and increases in males with ageing. In general, F1 and F2 decreases with ageing, mainly in females. Comparing the data obtained with the results of previous studies with adult speakers, a trend towards the centralization of vowels with ageing is observed. This investigation is the starting point for a broader study which will allow to analyse the changes in vowels acoustics from childhood to old age in EP.

In research on Silent Speech Interfaces (SSI), different sources of information (modalities) have been combined, aiming at obtaining better performance than the individual modalities. However, when combining these modalities, the dimensionality of the feature space rapidly increases, yielding the well-known "curse of dimensionality". As a consequence, in order to extract useful information from this data, one has to resort to feature selection (FS) techniques to lower the dimensionality of the learning space. In this paper, we assess the impact of FS techniques for silent speech data, in a dataset with 4 non-invasive and promising modalities, namely: video, depth, ultrasonic Doppler sensing, and surface electromyography. We consider two supervised (mutual information and Fisher's ratio) and two unsupervised (meanmedian and arithmetic mean geometric mean) FS filters. The evaluation was made by assessing the classification accuracy (word recognition error) of three well-known classifiers (knearest neighbors, support vector machines, and dynamic time warping). The key results of this study show that both unsupervised and supervised FS techniques improve on the classification accuracy on both individual and combined modalities. For instance, on the video component, we attain relative performance gains of 36.2% in error rates. FS is also useful as pre-processing for feature fusion

International audience

This paper presents a study of European Portuguese elderly speech, in which the acoustic characteristics of two groups of elderly speakers (aged 60-75 and over 75) are compared with those of young adult speakers (aged 19-30). The correlation between age and a set of 14 acoustic features was investigated, and decision trees were used to establish the relative importance of the features. A greater use of pauses characterized speakers aged 60 and over. For female speakers, speech rate also appeared to correlate with age. For male speakers, jitter distinguished between speakers aged 60-75 and older. The correlation between the features and speech recognition performance was also investigated. Word error rate correlated mostly with the use of pauses, speech rate, and the ratio of long phone realizations. Finally, by comparing the phone sequences used by the recognizer on the most frequent words, we observed that the young adult speakers reduced schwas more than the elderly speakers. This result seems to confirm the common idea that young speakers reduce articulation more than older speakers. Further investigation is needed to confirm this result by determining whether this is due to ageing or to the generation gap.

Immersive 3D Virtual Environments (VE) have become affordable for many research centers. However, a complete solution needs several integration steps to be fully operational. Some of these steps are difficult to accomplish and require an uncommon combination of different skills. This tutorial presents the most recent techniques developed to address this problem, from displays to software tools. The hardware in a typical VR installations combines projectors, screens, speakers, computers, tracking and I/O devices. The tutorial will discuss hardware options, explaining their advantages and disadvantages. We will cover design decisions from basic software and hardware design, through user tracking, multimodal human-computer interfaces and acoustic rendering, to how to administrate the whole solution. Additionally, we will provide an introduction to existing tracking technologies, explaining how the most common devices work, while focusing on infrared optical tracking. Finally, we briefly cover integration software and middleware developed for most VE settings.

In the last few years the number of systems and devices that use voice based interaction has grown significantly. For a continued use of these systems the interface must be reliable and pleasant in order to provide an optimal user experience. However there are currently very few studies that try to evaluate how good is a voice when the application is a speech based interface. In this paper we present a new automatic voice pleasantness classification system based on prosodic and acoustic patterns of voice preference. Our study is based on a multi-language database composed by female voices. In the objective performance evaluation the system achieved a 7.3% error rate

Abstract To achieve a full‐scale simulation of a pyrite mine, a highly immersive environment becomes necessary and this research has led to a complex system enabling users to walk through a virtual mine in real time, presenting all the behaviours present in such environment. Some of the problems encountered are the tunnels behaviours, including highly contrasted images due to the presence of the head light, narrow paths, elevators, sound reverberation and tunnels texture shades. The use of immersive virtual reality enables the generation of high‐quality simulations, because it is possible to control several feedback mechanisms such as the degree of luminance of produced imagery and spatial sound. In this research, a projection infrastructure and tracking system were specified and developed, aiming at producing the best results for this kind of simulation. To achieve our purposes, distributed algorithms were developed to run in a cluster solution that drives a four‐sided CAVE‐like environment. The complete production pipeline is presented, ranging from the developed authoring techniques, enabling fast production of new content for the simulation, to the tracking techniques produced for the improvement of the interaction.

In this paper, a rule-based automatic syllabifier for Danish is described using the Maximal Onset Principle. Prior success rates of rule-based methods applied to Portuguese and Catalan syllabification modules were on the basis of this work. The system was implemented and tested using a very small set of rules. The results gave rise to 96.9% and 98.7% of word accuracy rate, contrary to our initial expectations, being Danish a language with a complex syllabic structure and thus difficult to be rule-driven. Comparison with data-driven syllabification system using artificial neural networks showed a higher accuracy rate of the former system.

Stress and syllabification are essential attributes for several components in text-to speech (TTS) systems. They are responsible for improving grapheme-to-phoneme conversion rules and for enhancing the synthetic intelligibility, since stress and syllable are key units in prosody prediction. This paper presents three linguistically rule-based automatic algorithms for Catalan text-to-speech conversion: a word stress marker, an orthographic syllabification algorithm and a phonological syllabification algorithm. The systems were implemented and tested. The results gave rise to the following word accuracy rates: 100% for the stress marker algorithm, 99.7% for the orthographic syllabification algorithm and 99.8% for the phonological syllabification algorithm.