David Donoho (Stanford)

Add to calendar:

Digital Media Transformation is Transforming the Purpose and Content and Value of Data Science Research

Media Ecology is the field that studies the effects of the media we use on the thoughts we have and on who we are. Media Ecology predicts that each media transition—say from print to radio to TV to smartphones will drastically change our mentalities. And we have recently seen plenty of such media effects in the general population since the advent of smartphones, which seem to include the death of traditional journalistic institutions and the rise of extremist niche influencers, for example, on podcasts. Mental effects include political polarization, an epidemic of mental health issues and social isolation, increasing adherence to utterly false worldviews like flat earth and holocaust denial, and a global rise in authoritarianism. These effects are so large and pervasive that one is forced to think about media ecology and its consequences in other settings.

Matan Gavish (CS at Hebrew University) and David Donoho have been discussing the effects of media on academic research in fields that are close to our interests. Their starting point is to recognize that a new mode of publication has arisen in recent years—the so-called "CS conference style"—and publications in this style, particularly those related to empirical machine learning, have rapidly climbed the "league tables" of citation impact. Twenty years ago, the top impact journals were the general interest science journals like Science, Nature, and PNAS, and the medical journals like NEJM, JAMA, BMJ, and Lancet. These have now been joined by 5 CS conferences: NeurIPS, ICLR, JCCV, PAMI, ICML. Never before were peri-data-science publications very impactful in this way.

This new (to us) type of academic publication involves short highly "templatable" papers in which the deliverable is often a table of numbers with some entries in boldface. Alongside this, authors engage in publicity operations gaining the attention of influencers and promoters on social media and podcasts. Separately, the conferences process submissions of this form through a historically novel open review process. 

In his talk, David Donoho will discuss these global trends from the viewpoint of a trained mathematical scientist; how the world differs today from the world he was trained in, our sense of the changing mentality of researchers, our perceptions of the pitfalls of this new model we already see trapping many researchers, and the opportunities the new situation presents, many of which are unseen by most, so far.

This is joint work with Matan Gavish, Hebrew University.

About the Seminar Series

The last few years have seen a substantial increase in the reported success of machine learning (ML), and generative artificial intelligence (AI). These impact practices in delivering services from financial institutions to entertainment and medicine. However scientific research also increasingly relies on large data sets, whose analysis leverages ML/AI. This seminar series aims to investigate if and how the paradigm for scientific research has changed or should change to incorporate these new tools and the possibilities they open.

A diverse group of scholars engaged in scientific research, method development, and historical and epistemological investigations will give a 50-minute presentation, followed by discussion.

The event is open to all. Stanford students and postdocs have the opportunity to engage more directly with speakers and topics by enrolling in the Canvas course here.