Recherche avancée

Ingela Alger et Donald Cox

novembre 2019, 32 pages

Jean-François Bonnefon, Fatimah Ishowo-Oloko, Zakariyah Soroye, Jacob W. Crandall, Iyad Rahwan et Tahal Rahwan

vol. 1, novembre 2019, p. 517–521

Recent advances in artificial intelligence and deep learning have made it possible for bots to pass as humans, as is the case with the recent Google Duplex—an automated voice assistant capable of generating realistic speech that can fool humans into thinking they are talking to another human. Such...

Luke Glowacki, Samuel Mehr, Manvir Singh, Dean Knox, Daniel Ketter, Daniel Pickens-Jones, S. Atwood, Christopher Lucas, Nori Jacoby, Alena Egner, Erin Hopkins, Rhea Howard, Joshua Hartshorne, Mariela Jennings, Jan Simson, Constance Bainbridge, Steven Pinker, Timothy O'Donnell et Max Krasnow

vol. 366, n° 6468, novembre 2019

Thierry Magnac et Sébastien Roux

n° 19-1041, octobre 2019, révision mars 2021

Using panel data from a single cohort of French male wage earners observed over a long span of 30 years starting at their entry in the labor market, we estimate parameters of a human capital investment model by random and fixed effect methods. Individual wage proles are described by their...

Milo Bianchi et Philippe Jehiel

n° 19-1042, octobre 2019

We study banks incentive to pool assets of heterogeneous quality when investors evaluate pools by extrapolating from limited sampling. Pooling assets of heterogeneous quality induces dispersion in investors valuations without affecting their average. Prices are determined by market clearing...

Jérôme Bolte, Camille Castera, Edouard Pauwels et Cédric Févotte

n° 19-1043, octobre 2019

We devise a learning algorithm for possibly nonsmooth deep neural networks featuring inertia and Newtonian directional intelligence only by means of a backpropagation oracle. Our algorithm, called INDIAN, has an appealing mechanical interpretation, making the role of its two hyperparameters...

Jérôme Bolte et Edouard Pauwels

n° 19-1044, octobre 2019

Modern problems in AI or in numerical analysis require nonsmooth approaches with a exible calculus. We introduce generalized derivatives called conservative fields for which we develop a calculus and provide representation formulas. Functions having a conservative field are called path...

Sereeparp Anantavrasilp, Abe De Jong, Douglas V. DeJong et Ulrich Hege

n° 19-1045, octobre 2019

This paper focuses on dominant owners’ use of leverage to finance their blockholdings and its relationship to dividend policy. We postulate that blockholder leverage may impact payout policy, in particular when earnings are hit by a negative shock. We use panel data for France where blockholders...

Francesc Dilmé et Daniel F. Garrett

vol. 17, n° 5, octobre 2019, p. 1654–1686

Successes of law enforcement in apprehending offenders are often publicized events. Such events have been found to result in temporary reductions in of- fending, or “residual deterrence”. We provide a theory of residual deterrence which accounts for the incentives of both enforcement officials and...

José-Raimundo Carvalho, Thierry Magnac et Qizhou Xiong

vol. 10, octobre 2019, p. 1233–1277

We use rich microeconomic data on performance and choices of students at college entry to analyze interactions between the selection mechanism, eliciting college preferences through exams, and the allocation mechanism. We set up a framework in which success probabilities and student preferences are...