Apprendre à aimer l’opacité : les arbres de décision et l’histoire de la boîte noire algorithmique

Le Médialab de Sciences Po, en partenariat avec le Centre Alexandre-Koyré, accueille Matthew L. Jones (Columbia University) pour présenter son travail sur la genèse et le développement d’un des principaux types d’algorithmes pour l’apprentissage supervisé : les arbres de décision.

La présentation aura lieu en anglais.

Séminaire ouvert à tou.te.s, dans la limite des places disponibles.

Inscription préalable fortement recommandée à cette adresse.

Présentation

A series of researchers, each slightly askew to the dominant practices and epistemic virtues of their fields, came obliquely to trees in the 1970s: a data-driven statistician, a machine learning expert focused on large data sets, social scientists unhappy with multivariate statistics, a physicist interested mostly in computers who eventually was tenured in a statistics department. In case after case, the creators of different forms of trees deployed “applied” philosophies of science in critiquing contemporary practices, epistemic criteria and even promotion practices in academic disciplines. Faced with increasing amounts of high-dimensional data, these authors time and again advocated a data-focused positivism. The history of trees does not cleanly divide into a theoretical and an applied stage; an academic and a commercial phase; a statistical and a computational stage; or even an algorithm design and an implementation stage.  This history is iterative: the implementation of algorithms on actually existing computers with various limitations drives the development and transformation of the techniques. Before the very recent renaissance and current triumph of neural networks, decision trees were central to the transformation of artificial intelligence and machine learning of recent years: the shift in the central goal to a focus on prediction at the expense of concerns with human intelligibility, and of a shift from symbolic interpretation to potent but inscrutable black-boxes. Trees exploded in the late 1980s and 1980s as paragons of interpretable algorithms but developed in the late 1990s into a key example of powerful but opaque ensemble models, predictive but almost unknowable. We need to explain, rather than take as given, the shift in values to prediction—to an instrumentalism—central to the ethos and practice of the contemporary data sciences. Opacity needs its history—just as transparency does.

Matthew L. Jones specializes in the history of science and technology, focused on early modern Europe and on recent information technologies. He was a Guggenheim Fellow in 2012-13 and a Mellon New Directions fellow in 2012-15. He is finishing two books, Great Exploitations: Data Mining, Legal Modernization, and the NSA and  Data Mining: The Critique of Artificial Reason, 1963-2005, a study of “big data” and its growth as a new form of technical expertise in business and scientific research.  Reckoning with Matter: Calculating Machines, Innovation, and Thinking about Thinking from Pascal to Babbage is appearing this fall from the University of Chicago Press. His first book The Good Life in the Scientific Revolution (University of Chicago Press, 2006) focused on the mathematical innovations of Descartes, Pascal, and Leibniz.

► Le blog du Médialab