BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//Talks.cam//talks.cam.ac.uk//
X-WR-CALNAME:Talks.cam
BEGIN:VEVENT
SUMMARY:On Estimation of Unnormalized Density Models - Song Liu — Univer
 sity of Bristol
DTSTART:20191025T130000Z
DTEND:20191025T140000Z
UID:TALK130054@talks.cam.ac.uk
CONTACT:Dr Sergio Bacallado
DESCRIPTION:In many machine learning applications\, complicated parametric
  density models such as Deep Neural Networks are favoured due to their fle
 xibility and expressiveness. However\, unlike classic density models\, the
 se density functions cannot be easily normalised. Thus\, Maximum Likelihoo
 d Estimation (MLE) cannot be easily applied for model parameter estimation
 . If the dimensionality of the input variable is high\, MCMC based MLE may
  also fail. In the machine learning community\, some Stein Identity based 
 methods have risen in popularity due to their ability to measure the "good
 ness of fit" of unnormalisable models. In the first part of the talk [1]\,
  we study the performance of a parameter estimator using Stein's identity.
  Particularly\, we construct a Stein density ratio estimator\, which estim
 ates the ratio function between a data distribution and a model distributi
 on. Then we minimise the fitted likelihood ratio to estimate model paramet
 ers. In the second part of the talk [2]\, I discuss a specific type of unn
 ormalised density model: Truncated densities. We show how an augmented sco
 re matching estimator can be applied to estimate parameters of density mod
 els with a complex truncation domain (such as a polytope in R^2). \n\n[1] 
 Liu\, S.\, Kanamori\, T.\, Jitkrittum\, W.\, Chen\, Y\; Fisher Efficient I
 nference of Intractable Models\; arxiv:1805.07454\, 2019.\n[2] Liu\, S.\, 
 Kanamori\, T.\, Estimating Density Models with Complex Truncation Boundari
 es\; arXiv:1910.03834\, 2019. 
LOCATION:MR12
END:VEVENT
END:VCALENDAR