nhaliday + siggraph   28

How do these "neural network style transfer" tools work? - Julia Evans
When we put an image into the network, it starts out as a vector of numbers (the red/green/blue values for each pixel). At each layer of the network we get another intermediate vector of numbers. There’s no inherent meaning to any of these vectors.

But! If we want to, we could pick one of those vectors arbitrarily and declare “You know, I think that vector represents the content” of the image.

The basic idea is that the further down you get in the network (and the closer towards classifying objects in the network as a “cat” or “house” or whatever”), the more the vector represents the image’s “content”.

In this paper, they designate the “conv4_2” later as the “content” layer. This seems to be pretty arbitrary – it’s just a layer that’s pretty far down the network.

Defining “style” is a bit more complicated. If I understand correctly, the definition “style” is actually the major innovation of this paper – they don’t just pick a layer and say “this is the style layer”. Instead, they take all the “feature maps” at a layer (basically there are actually a whole bunch of vectors at the layer, one for each “feature”), and define the “Gram matrix” of all the pairwise inner products between those vectors. This Gram matrix is the style.
techtariat  bangbang  deep-learning  model-class  explanation  art  visuo  machine-learning  acm  SIGGRAPH  init  inner-product  nibble 
february 2017 by nhaliday
Spaceship Generator | Hacker News
some interesting discussion of the value of procedural generation in the comments
commentary  hn  graphics  games  programming  libraries  repo  oss  project  SIGGRAPH 
june 2016 by nhaliday
a search system for videos based on drawing
video  search  computer-vision  ai  worrydream  organization  software  skunkworks  SIGGRAPH 
march 2016 by nhaliday

bundles : techieworrydream

related tags

acm  acmtariat  adversarial  ai  analysis  announcement  app  applications  art  article  atoms  audio  bangbang  beauty  better-explained  books  bots  calculation  caltech  classification  comedy  commentary  composition-decomposition  computer-vision  concept  cool  counterexample  course  cs  deep-learning  differential  dirty-hands  dropbox  embodied  engineering  exocortex  explanation  exposition  extrema  features  finance  fintech  fluid  frontend  games  generative  geometry  gif  glitch  graphics  hacker  hmm  hn  identity  idk  init  inner-product  interdisciplinary  interpretability  ios  javascript  lecture-notes  let-me-see  letters  libraries  links  list  local-global  machine-learning  math  mechanics  mobile  model-class  models  multi  news  nibble  ocr  optimization  org:bleg  org:edu  org:junk  org:lite  org:mag  org:med  org:popup  organization  oss  papers  pdf  performance  perturbation  physics  pic  profile  programming  project  recommendations  reference  regularization  repo  research  robust  scaling-tech  search  security  SIGGRAPH  simulation  skunkworks  software  stanford  startups  stat-mech  structure  subculture  summary  symmetry  systems  techtariat  tools  topology  tumblr  tutoring  unit  video  virtu  visual-understanding  visualization  visuo  wiki  worrydream  writing 

Copy this bookmark: