BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//Talks.cam//talks.cam.ac.uk//
X-WR-CALNAME:Talks.cam
BEGIN:VEVENT
SUMMARY:Accelerating Bayesian Inference and Data Acquisition via Amortizat
 ion - Daolang Huang (Aalto University)
DTSTART:20250624T101500Z
DTEND:20250624T111500Z
UID:TALK232213@talks.cam.ac.uk
DESCRIPTION:Many critical applications demand systems that can both strate
 gically acquire the most informative data and instantaneously perform infe
 rence based upon it. Bayesian inference and Bayesian experimental design o
 ffer principled mathematical means for reasoning under uncertainty and for
  strategically gathering data\, respectively. While foundational\, both me
 thods introduce notorious computational challenges. In recent years\, amor
 tized solutions have been proposed to address these issues by pre-training
  neural networks\, significantly reducing computational costs at deploymen
 t.\nIn this talk\, I will first introduce the Amortized Conditioning Engin
 e (ACE)\, a flexible amortized inference framework that affords conditioni
 ng on both observed data and interpretable latent variables\, the inclusio
 n of priors at runtime\, and outputs predictive distributions for both dis
 crete and continuous data and latents. I will then share our latest work\,
  the Amortized Active Learning and Inference Engine (ALINE). ALINE combine
 s the advantages of amortized inference and experimental design into a sin
 gle\, unified framework. It is capable of rapidly proposing valuable data 
 points while simultaneously performing fast\, flexible inference based on 
 the collected data\, thus seamlessly closing the loop between active data 
 acquisition and real-time reasoning.
LOCATION:Seminar Room 1\, Newton Institute
END:VEVENT
END:VCALENDAR