BEGIN:VCALENDAR
VERSION:2.0
PRODID:icalendar-ruby
CALSCALE:GREGORIAN
X-WR-CALNAME:Bay Area Tech Economics Seminar with Avi Feller\, UC Berkeley
X-WR-TIMEZONE:Pacific Time (US & Canada)
BEGIN:VEVENT
DTSTAMP:20260714T132606Z
UID:tag:localist.com\,2008:EventInstance_53064140966372
DTSTART:20260605T013000Z
DTEND:20260605T033000Z
DESCRIPTION:Talk Title: Classical Statistics in the Age of AI\n\nAbstract: 
 Researchers increasingly use generative AI to create “digital twins” a
 nd conduct synthetic experiments. Recognizing that LLMs often fail to capt
 ure complex real-world behavior\, a growing literature has proposed novel 
 methods for combining synthetic and ground-truth data. This talk illustrat
 es the continued relevance of classical statistics for this challenge thro
 ugh two projects. In the first project\, we argue that off-the-shelf linea
 r regression is a natural approach for incorporating AI predictions into e
 xperiments\, building on the randomization inference framework dating back
  to Fisher and Neyman. Unlike many recent proposals\, standard linear regr
 ession inherits a “do no harm” property in that the adjusted estimator
  automatically reverts to the unadjusted difference in means when AI predi
 ctions are uninformative. In the second project\, we examine how to improv
 e the AI predictions themselves using “activation steering\,” a techni
 que from mechanistic interpretability that modifies internal LLM activatio
 ns to shift behavior toward a target concept. We show that many steering m
 ethods implicitly estimate the average gradient of an outcome regression\,
  a quantity with a long history in causal inference and econometrics. This
  connection immediately enables more flexible models\, including Neyman-or
 thogonal estimators\, which in turn lead to improved AI predictions for th
 e first project’s regression framework. Together\, these results show ho
 w classical statistical ideas continue to provide both conceptual clarity 
 and empirical gains at the intersection of causal inference and generative
  AI.\n\nSpeaker: Avi Feller\, Associate Professor in the Goldman School of
  Public Policy and the Department of Statistics\, UC Berkeley\n\nSpeaker B
 io: Avi Feller is an associate professor in the Goldman School of Public P
 olicy and the Department of Statistics at UC Berkeley\, working at the int
 erface of data science and the social sciences. His research focuses on de
 veloping practical\, transparent methods that can be applied at scale\, an
 d on deploying these tools in a range of policy domains and industry appli
 cations. His research has appeared in top methodological journals (such as
  Journal of the American Statistical Association\, Journal of the Royal St
 atistical Society\, and Econometrica)\, in leading machine learning confer
 ences (such as NeurIPS and ICML)\, and in interdisciplinary journals like 
 the Proceedings of the National Academy of Sciences.\n\nOutside of academi
 a\, Feller is a research consultant at Adobe and was previously a visiting
  researcher at Google. He also co-founded EveryDay Labs\, an edtech compan
 y focused on reducing student absenteeism (acquired in 2026).\n\nFeller ha
 s received multiple awards for his work\, including the COPSS Emerging Lea
 der Award\, the SREE Early Career Award\, the American Statistical Associa
 tion’s Outstanding Statistical Application Award\, and the Mid-Career Aw
 ard from the ASA Social Statistics Section. He received a PhD in statistic
 s from Harvard University\, MSc in Applied Statistics from Oxford Universi
 ty as a Rhodes Scholar\, and BA from Yale University. Prior to his doctora
 l studies\, he worked as a policy official in the White House Office of Ma
 nagement and Budget.\n\nThe logistics details for each session will be pro
 vided to registered participants.  Sign up here to be included on the mail
 ing list and learn about future events.\n\nThis talk is co-sponsored by US
 F's Master's in Applied Economics and the Stanford Causal Science Center. 
 For additional information and abstracts from past talks\, please click he
 re.
GEO:37.430043;-122.171561
LOCATION:Simonyi Conference Center\, CoDa
SUMMARY:Bay Area Tech Economics Seminar with Avi Feller\, UC Berkeley
URL;VALUE=URI:https://events.stanford.edu/event/bay-area-tech-economics-sem
 inar-with-avi-feller-uc-berkeley
CATEGORIES:Class/Seminar
END:VEVENT
END:VCALENDAR