BEGIN:VCALENDAR
VERSION:2.0
PRODID:Linklings LLC
BEGIN:VTIMEZONE
TZID:Australia/Melbourne
X-LIC-LOCATION:Australia/Melbourne
BEGIN:DAYLIGHT
TZOFFSETFROM:+1000
TZOFFSETTO:+1100
TZNAME:AEDT
DTSTART:19721003T020000
RRULE:FREQ=YEARLY;BYMONTH=4;BYDAY=1SU
END:DAYLIGHT
BEGIN:STANDARD
DTSTART:19721003T020000
TZOFFSETFROM:+1100
TZOFFSETTO:+1000
TZNAME:AEST
RRULE:FREQ=YEARLY;BYMONTH=10;BYDAY=1SU
END:STANDARD
END:VTIMEZONE
BEGIN:VEVENT
DTSTAMP:20260114T163719Z
LOCATION:Meeting Room C4.11\, Level 4 (Convention Centre)
DTSTART;TZID=Australia/Melbourne:20231215T101500
DTEND;TZID=Australia/Melbourne:20231215T102500
UID:siggraphasia_SIGGRAPH Asia 2023_sess135_papers_964@linklings.com
SUMMARY:Face0: Instantaneously Conditioning a Text-to-Image Model on a Fac
 e
DESCRIPTION:Dani Valevski, Danny Lumen, Yossi Matias, and Yaniv Leviathan 
 (Google Research)\n\nWe present Face0, a novel way to instantaneously cond
 ition a text-to-image generation model on a face, in sample time, without 
 any optimization procedures such as fine-tuning or inversions. We augment 
 a dataset of annotated images with embeddings of the included faces and tr
 ain an image generation model, on the augmented dataset. Once trained, our
  system is practically identical at inference time to the underlying base 
 model, and is therefore able to generate images, given a user-supplied fac
 e image and a prompt, in just a couple of seconds. Our method achieves ple
 asing results, is remarkably simple, extremely fast, and equips the underl
 ying model with new capabilities, like controlling the generated images bo
 th via text or via direct manipulation of the input face embeddings. In ad
 dition, when using a fixed random vector instead of a face embedding from 
 a user supplied image, our method essentially solves the problem of consis
 tent character generation across images. Finally, while requiring further 
 research, we hope that our method, which decouples the model’s textual bia
 ses from its biases on faces, might be a step towards some mitigation of b
 iases in future text-to-image models.\n\nRegistration Category: Full Acces
 s\n\nSession Chair: Chongyang Ma (ByteDance)\n\n
URL:https://asia.siggraph.org/2023/full-program?id=papers_964&sess=sess135
END:VEVENT
END:VCALENDAR