Portail HAL de l'Enssib
Not a member yet
956 research outputs found
Sort by
What is the meaning of "bias" in artificial intelligence
No full textInternational audienceIn recent years, the word “bias” has frequently been used in the field of artificial intelligence (AI), and more precisely machine learning. It is an “umbrella term” which can refer to a great variety of issues: for example the lack of diversity among computer scientists, the non-representativity of data bases, or the existence of implicit stereotypes in statistical models. However, the concept of “bias”, its potential causes, and even consequences are not always clearly defined.Recent reviews have proposed classifications based on the sources of bias and how they impact the development and use of AI systems (Hovy, Prabhumoye 2021; Mehrabi et al. 2022). In this workshop, we wish to shed a new light on bias by discussing the underlying implicit norms. Our hypothesis is that in the field of AI, three different visions of bias are actually confounded: bias as a divergence from scientific standards (methodological bias), bias as a divergence from the “rational” decision (cognitive bias), and bias as a divergence from an ideal society (socio-historical bias). We will also question the ambition of “unbiasing” AI.Methodological bias concerns epistemological issues. In machine learning, it mostly depends on the quality of the data used to train the AI models. Taken seriously, these issues raise several non-trivial questions related to objectivity and “scientific truth”. However, a naive understanding of science may lead AI designers and users to think that data speaks for themselves. In fact, data-driven sciences require very careful methodologies for collecting and interpretating data, which are often lacking in AI applications. The field of critical data studies emerged in order to denounce the fact that data are never “neutral” or “raw”, but always “cooked” (Iliadis, Russo 2016). From this standpoint, “unbiasing” AI will always be a continuous and never-ending process (as well as scientific objectivity).Cognitive bias is a concept initially proposed in psychology (Tversky, Kahneman 1974), before gaining great popularity in economy. It is defined as a systematic error of judgement, attributed to a priori knowledge and mental shortcuts. The standard for “unbiased” or “rational” decisions are those of “Homo economicus”, who seeks to maximize their own interests. Contextual and qualitative information (like emotions of the subject or the framing of sentences) is considered irrelevant (Gigerenzer 2018). In the field of AI, the reference to cognitive bias is ambiguous, because it can be used to describe the limits of AI systems (most of them unable to emulate a reasoning and relying on statistical models), but also as an argument to replace human decisions by algorithmic computation (Kahneman et al. 2016). “Unbiasing” AI may therefore be seen as a technocratic project, aiming at avoiding conflicts and depoliticizing institutional decisions.Socio-historical bias is frequently mentioned when discussing the social impacts of AI systems. This issue goes beyond data quality or representativity issues: it is also question of spurious and unwanted correlations. Famous examples of unfairness and algorithmic discriminations have contributed to the rise of AI ethics, but the way to tackle these issues remains controversial. If several definitions of fairness and ethical frameworks have been proposed, many authors suggested that they are not a technical issue, but a political one (Binns 2018; D’Ignazio, Klein 2020; Green 2021). In this case, instead of “unbiasing” AI, we should talk of “rebiasing” AI in a way that is considered more ethical at a given time. Because of the high cost of training and testing AI, this “rebiased” AI will by definition be conservative.References:BINNS, Reuben, 2018. Fairness in Machine Learning: Lessons from Political Philosophy. In : Proceedings of the 1st Conference on Fairness, Accountability and Transparency [en ligne]. PMLR. 21 janvier 2018. pp. 149‑159. [Consulté le 22 août 2022]. Disponible à l’adresse : https://proceedings.mlr.press/v81/binns18a.htmlD’IGNAZIO, Catherine et KLEIN, Lauren F., 2020. Data Feminism. Cambridge, MA, USA : MIT Press. Strong Ideas. ISBN 978-0-262-04400-4.GIGERENZER, Gerd, 2018. The bias bias in behavioral economics. Review of Behavioral Economics. 2018. Vol. 5, n° 3‑4, pp. 303‑336.GREEN, Ben, 2021. The Contestation of Tech Ethics: A Sociotechnical Approach to Technology Ethics in Practice. Journal of Social Computing. septembre 2021. Vol. 2, n° 3, pp. 209‑225. DOI 10.23919/JSC.2021.0018.HOVY, Dirk et PRABHUMOYE, Shrimai, 2021. Five sources of bias in natural language processing. Language and Linguistics Compass. 2021. Vol. 15, n° 8, pp. e12432. DOI 10.1111/lnc3.12432.ILIADIS, Andrew et RUSSO, Federica, 2016. Critical data studies: An introduction. Big Data & Society. 1 décembre 2016. Vol. 3, n° 2, pp. 2053951716674238. DOI 10.1177/2053951716674238.KAHNEMAN, Daniel, ROSENFIELD, Andrew M, GANDHI, Linnea et BLASER, Tom, 2016. Noise: How to Overcome the High, Hidden Cost of Inconsistent Decision Making. Harvard Business Review. octobre 2016. pp. 9.MEHRABI, Ninareh, MORSTATTER, Fred, SAXENA, Nripsuta, LERMAN, Kristina et GALSTYAN, Aram, 2022. A Survey on Bias and Fairness in Machine Learning [en ligne]. 25 janvier 2022. arXiv. arXiv:1908.09635. [Consulté le 1 juillet 2022]. Disponible à l’adresse : http://arxiv.org/abs/1908.09635TVERSKY, Amos et KAHNEMAN, Daniel, 1974. Judgment under Uncertainty: Heuristics and Biases: Biases in judgments reveal some heuristics of thinking under uncertainty. science. 1974. Vol. 185, n° 4157, pp. 1124‑1131
Eulalie: a documentary system for the collaborative preservation of electroacoustic music based on the Doremus ontology
No full textInternational audienceEulalie is an operational, open-source, and free information system dedicated to electroacoustic music to document and preserve works. It is based on the Doremus ontology. We propose adapting this ontology to the specific field of electroacoustic music and its implementation in a database (Heurist, in dialogue with Nakala to host the documents). The project was born in the context of Art Zoyd Studios, a center for musical creation. One of the missions is to preserve the repertoire of the XXth and XXIst century pieces involving electronics. The principles of interoperability and Open Data have guided the different choices made throughout the project. Thus, it is not only possible to make the Art Zoyd Studios information database's content interact with other music resources, but the system itself, Eulalie, can be used by other institutions facing problems similar to those of Art Zoyd Studios
30 ans après le Web… Big bang dans les institutions de savoirs ? Introduction
No full textInternational audienceLe dossier thématique du sixième numéro de Balisages vise à interroger l’évolution des institutions de savoirs au regard des mutations de l’industrie de l’information dans le monde post-web. Née dans les années 50 et structurée dans les années 70-80, la société de l'inforrmation a été profondément restructurée par l’arrivée du web entrainant dans son sillage une remise en cause profonde du positionnement des institutions dans leur rapport aux savoirs et leur vocation de transmission.En 1989, lorsque Tim Berners Lee développe le Web pour partager les connaissances entre scientifiques, il ne mesurait peut-être pas la révolution des savoirs qu’il allait en grande partie initier. Le Web s’est progressivement hybridé à d’autres innovations (applications mobiles, réseaux sociaux…), à de nouveaux paradigmes (la participation, les plateformes, le web sémantique, l’open access, la science des données…) bousculant nos facultés de penser, de concevoir, d’agir, et de transmettre.Un peu plus de trente ans après, que s’est-il passé ? Comment comparer dans une lignée diachronique, la production-circulation des savoirs à l’ère post-web, à l’heure où ce sont également des transformations profondes d’élaboration des savoirs qui se jouent ?Ce numéro propose de porter son attention sur deux axes de réflexion, l'économie politique du Web d'une part et les enjeux d'innovation et de design de services de l'autre, pour répondre à ces questionnements de recherch
La fabrique de la bande dessinée, socio-logies et socio-sémiotiques de la bande dessinée
No full textInternational audienc
Ethics of usage for AI technologies
No full textInternational audiencePrésentation préparée pour la table ronde “Ethical Issues in Generative Speech technologies”, organisée lors du Speech Synthesis Worksho
Théorie, technique et pratique archivistique en environnement numérique
No full textInternational audienceLe principe de respect des fonds et le concept des trois âges des archivesse sont imposés comme norme et fondent encore aujourd’hui l’ archivistique contemporaine, en France. Le premier a été formulé au XIXe sièclepar Natalis de Wailly historien-archiviste et le second, en 1961, par Yves Pérotin,figure tutélaire du métier d’archiviste. L’un et l’autre ont ouvert la voie à unearchivistique nouvelle en formulant des concepts et des méthodes documentaires detraitement des archives en rupture avec ceux élaborés précédemment. Dans cetarticle, nous discutons la permanence de ces deux principes alors que les nouvellespratiques informationnelles et communicationnelles des organisations pourproduire, enregistrer, conserver et diffuser les documents sont radicalementdifférentes des pratiques en vigueur lorsque ces concepts ont été formulés