BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//wp-events-plugin.com//7.2.3.1//EN
TZID:America/New_York
X-WR-TIMEZONE:America/New_York
BEGIN:VEVENT
UID:906@fds.yale.edu
DTSTART;TZID=America/New_York:20260121T113000
DTEND;TZID=America/New_York:20260121T130000
DTSTAMP:20260120T193046Z
URL:https://fds.yale.edu/events/fds-colloquium-alexander-lew-yale/
SUMMARY:FDS Colloquium: Alexander Lew (Yale)\, "Automatic Integration and D
 ifferentiation of Probabilistic Programs"
DESCRIPTION:\nAbstract: By automating the error-prone math behind deep lear
 ning\, systems such as TensorFlow and PyTorch have supercharged machine le
 arning research\, empowering hundreds of thousands of practitioners to rap
 idly explore the design space of neural network architectures and training
  algorithms. This talk will show how new programming language techniques
 —particularly generalizations of automatic differentiation—make it pos
 sible to generalize and extend such systems to support probabilistic model
 s. Our tools can automate the computation of expected values\, probabilit
 y densities\, and their gradients\, as well as help users derive fast\, lo
 w-variance\, unbiased estimators of these quantities when they are too exp
 ensive to compute exactly\, enabling orders-of-magnitude speedups in downs
 tream optimization and inference problems. To illustrate the value of thes
 e techniques\, I’ll show how they have helped us build systems for (1) a
 uditable reasoning and learning in relational domains\, enabling the detec
 tion of thousands of errors across millions of Medicare records\, and (2) 
 probabilistic inference over large language models\, enabling small open m
 odels to outperform frontier models on several constrained generation benc
 hmarks.\n\n\n\nSpeaker Bio: Alex's research aims to automate and scale up 
 principled probabilistic reasoning\, drawing on techniques from programmin
 g languages\, machine learning\, Bayesian statistics\, and cognitive scien
 ce. Alex is especially interested in the theory and practice of probabilis
 tic and differentiable programming languages.\n\n\n\nAlex is also a member
  of the&nbsp\;GenLM consortium\, a multi-university partnership aiming to 
 better control\, compose\, and understand language models using the probab
 ilistic programming and Bayesian inference toolkits.\n\n\n\n\n
CATEGORIES:FDS Events,Colloquium
LOCATION:Yale Institute for Foundations of Data Science & Webcast\, 219 Pro
 spect Street\, 13th Floor\, New Haven\, CT\, 06511\, United States
X-APPLE-STRUCTURED-LOCATION;VALUE=URI;X-ADDRESS=219 Prospect Street\, 13th 
 Floor\, New Haven\, CT\, 06511\, United States;X-APPLE-RADIUS=100;X-TITLE=
 Yale Institute for Foundations of Data Science & Webcast:geo:0,0
END:VEVENT
BEGIN:VTIMEZONE
TZID:America/New_York
X-LIC-LOCATION:America/New_York
BEGIN:STANDARD
DTSTART:20251102T010000
TZOFFSETFROM:-0400
TZOFFSETTO:-0500
TZNAME:EST
END:STANDARD
END:VTIMEZONE
END:VCALENDAR