Group items tagged

Introduction à la journée d'étude « Vox Machines », université Toulouse - Jean Jaurès / ESA Pyrénées, 10 décembre 2018

"Automatic Speech Recognition (ASR) struggles in the absence of an extensive volume of training data. We present SpecAugment, a new approach to augmenting audio data that treats it as a visual problem rather than an audio one."

"Une trentaine de startups et de laboratoires de recherche ont créé le Voice Lab, une place de marché qui a pour but de fédérer les acteurs Français et Européens de la Voix pour faire face aux géants chinois et américains. L'objectif du Voice Lab ? Offrir une alternative aux propositions existantes, principalement américaines et chinoises. Parmi ses membres figurent le moteur de recherche Qwant, la startup Snips, le hub France IA ou encore CandyVoice."

"Amazon.com Inc. employs thousands of people around the world to help improve the Alexa digital assistant powering its line of Echo speakers. The team listens to voice recordings captured in Echo owners' homes and offices. The recordings are transcribed, annotated and then fed back into the software as part of an effort to eliminate gaps in Alexa's understanding of human speech and help it better respond to commands. " "In marketing materials Amazon says Alexa "lives in the cloud and is always getting smarter." But like many software tools built to learn from experience, humans are doing some of the teaching." "In Alexa's privacy settings, the company gives users the option of disabling the use of their voice recordings for the development of new features. A screenshot reviewed by Bloomberg shows that the recordings sent to the Alexa auditors don't provide a user's full name and address but are associated with an account number, as well as the user's first name and the device's serial number." "Apple's Siri also has human helpers, who work to gauge whether the digital assistant's interpretation of requests lines up with what the person said. The recordings they review lack personally identifiable information and are stored for six months tied to a random identifier, according to an Apple security white paper. After that, the data is stripped of its random identification information but may be stored for longer periods to improve Siri's voice recognition. At Google, some reviewers can access some audio snippets from its Assistant to help train and improve the product, but it's not associated with any personally identifiable information and the audio is distorted, the company says. "

8 feminist personal intelligent assistant prototypes Designing a Feminist Alexa Seminar UAL Creative Computing Institute and Feminist Internet > https://www.youtube.com/watch?v=hQyKdC37M20&t=876s

Une alternative aux Gafam / Batx ?

"Users of Siri, Alexa, and Google Assistant conceptualize them in one of 3 ways: an interface, a personal assistant, or a brain. Frequent users are less likely to push the interaction limits of these AI systems than new users."

"Do-it-yourself artificial intelligence" by Google

carte interactive des 24 émotions : https://s3-us-west-1.amazonaws.com/vocs/map.html#

https://golem.ai/ NLP classique vs NLP Machine learning

"In a blog post this morning, Amazon details acoustic fingerprinting, a technique Alexa AI research scientists at the Seattle company use to "teach" Alexa what individual instances of its name sound like, so that the assistant can ignore them. The work is applied on the fly to detect when multiple Alexa-enabled devices around the globe hear the same command at roughly the same time - which Mike Rodehorst, machine learning scientist at Amazon's Alexa Speech division, says is needed to prevent Alexa from responding to pranks, references to people named Alexa, and other TV mentions Amazon isn't made aware of in advance."

"Manufacturers are already looking at ways to protect voice-activated devices. For example, ultrasound attacks could be stymied through detecting frequency alterations in received signals. It would be a nice idea to train all smart devices to recognize their owner's voice, although having already tested this on its own system, Google warns that such security can be fooled by a voice recording or a decent impersonation."

"[...] actuellement, Common Voice contiendrait 1 400 heures d'échantillons vocaux provenant de 42 000 contributeurs dans le monde, et couvrant 19 langues, dont l'anglais, le français, l'allemand, le farsi, le basque, le mandarin et le gaélique. Pour entraîner les IA, ces enregistrements sont accompagnés de métadonnées."

"Leon est un script open source (licence MIT) qui une fois installé et configuré selon vos besoins, saura répondre vocalement ou textuellement à vos questions. Leon fonctionne à l'aide de modules à paramétrer pour lui faire faire ce que vous voulez. Modules que vous pourrez ensuite partager avec la communauté."

"Dans chaque cabine, un petit boitier rond haute technologie avec micro et haut-parleur, à qui les passagers pourront parler et demander tout ou presque sur leurs vacances." "Zoé pourra répondre à 800 questions posées fréquemment par les passagers. 400 personnes, notamment chez Harman, la maison mère de la célèbre marque JBL, ont travaillé de longs mois pour le créer. Selon eux ,comme en radio, la voix c'est très simple à utiliser pour tout le monde : une question, une réponse. "' "Tous les nouveaux bateaux de MSC seront équipés de cette technologie notamment ceux fabriqués à Saint-Nazaire. Reste maintenant à savoir si les premiers passagers seront séduits après leur semaine à bord."

"[...] Google Assistant s'est enrichi d' une nouvelle fonction baptisée « conversation continue »." "Une « conversation continue » déjà sur Alexa"

Le laboratoire d'innovation numérique de la CNIL (LINC) publie son 6ème cahier Innovation et prospective, La Forme des choix - Données personnelles, design et frictions désirables : une exploration des enjeux du design dans la conception des services numériques, au prisme de la protection des données et des libertés.

"A glimmer of hope at CES 2019 came in the form of a new open-source virtual assistant called Mycroft that promises never to collect your personal data. CNet notes that it's still in early development, and it may never challenge Google Assistant or Alexa in general intelligence. But at least the device shows that some in the tech world are starting to regard privacy as a core feature of virtual assistants, rather than an afterthought."