adversarial   197

« earlier    

[1905.08233] Few-Shot Adversarial Learning of Realistic Neural Talking Head Models
Basically they made a talking head avatar from a single picture, and make it look good. Example, the Mono Lisa, now can talk like a millenial...
deep  machine  learning  deepfake  adversarial  network 
4 weeks ago by asteroza
This X Does Not Exist
Using generative adversarial networks (GAN), we can learn how to create realistic-looking fake versions of almost anything, as shown by this collection of sites that have sprung up in the past month.
gan  generative  adversarial  neural  network  compsci  ai  fake  photo  generation  collection  site 
march 2019 by vicchow
Which face is real?
Jevin West and Carl Bergstrom at the University of Washington:
<p>while we’ve learned to distrust user names and text more generally, pictures are different. You can't synthesize a picture out of nothing, we assume; a picture had to be of someone. Sure a scammer could appropriate someone else’s picture, but doing so is a risky strategy in a world with google reverse search and so forth. So we tend to trust pictures. A business profile with a picture obviously belongs to someone. A match on a dating site may turn out to be 10 pounds heavier or 10 years older than when a picture was taken, but if there’s a picture, the person obviously exists.

No longer. New adverserial machine learning algorithms allow people to rapidly generate synthetic 'photographs' of people who have never existed.

Computers are good, but your visual processing systems are even better. If you know what to look for, you can spot these fakes at a single glance — at least for the time being. The hardware and software used to generate them will continue to improve, and it may be only a few years until humans fall behind in the arms race between forgery and detection.

Our aim is to make you aware of the ease with which digital identities can be faked, and to help you spot these fakes at a single glance.</p>

So now we're using humans as the adversarial network (which calls out the generative network).
Machinelearning  gan  adversarial  internet 
february 2019 by charlesarthur
Lawyers of reddit around the globe, what countries do you think have the best/worst judicial system? Why so? : AskReddit
> I think the Commonwealth countries (particularly England, Australia, Canada and New Zealand) have the best judicial systems, for the following reasons:

In my view there's a lot to be said for the adversarial system. I think arguments can be really rigorously tested by independent advocates advancing each side of a controversy before an essentially passive judge. I prefer that approach to one in which the course and focus of inquiry is more directed by the judge. It seems more even-handed, and less susceptible to judicial whim than an inquisitorial system. This essentially narrows me down to the common law jurisdictions. I am probably biased on this point.

Of the great common law, adversarial judicial systems, I think the US system is deeply flawed because its judiciary is overtly politicised. From what I've seen this affects all levels of the judiciary, from Supreme Court appointees to local judges (some of whom are elected, which seems insane to me - you might as well select surgeons by popular vote).
law  global  adversarial 
december 2018 by porejide
“@lopp @aantonop adversarial thinking is counter-intuitive to most, and very hard to learn - like chess vs grandmaster: most people will never be able to beat a grandmaster even if they dedicated the rest of life to it. sounds elitist, but it's reality: schneier's law. best can do is seek review.” via Pocket
IFTTT  Pocket  adversarial  attack  opsec  security  thinking 
november 2018 by ChristopherA

« earlier    

related tags

2016-election  :/  a  ability  abstraction  academia  acm  acmtariat  adversarial-examples  adversarial-learning  adversarial_examples  adversarialattack  adversarialexample  adversarialnn  adversaries  adversary  adversity  age-generation  aggregation  ai-control  ai  aibusting  alexa  alien-character  ama  ambition  analysis  analytical-holistic  anglosphere  anthropology  antidemos  antisurveillance  apollonian-dionysian  art  article  articles  artificial  arxiv  asia  asr  attack  attacks  audio  authoritarianism  average-case  bad.advice  basics  being-right  big-peeps  big-yud  bird  birds  blog:  books  bottleneck  broad-econ  camouflage  capital  captialism  captioning  car  challenge  challenges  chart  china  classification  classifier  clever-rats  climate-change  clothing  coalitions  cog-psych  collection  command  commentary  comparison  competition  complement-substitute  complexity  composition-decomposition  comprehension  compsci  computation  computer-vision  computer  conquest-empire  contracts  convexity-curvature  coordination  counter-revolution  counterexample  cracker-econ  crime  crooked  crypto  cs  cui  cultural-dynamics  current-events  cv  data  dataset  decentralized  deep-learning  deep.learning  deep  deep_learning  deepfake  deeplearning  defence  democracy  descriptive  detail-architecture  devops  discovery  discussion  distance  distribution  divergence  dnn  domain  dropout  early-modern  economics  econotariat  elections  elegance  elon  embedded  ems  enhancement  environment  error  escapegoat  essence-existence  europe  evaluation  example  examples  expansionism  expert-experience  expert  explanans  exploit  fabric  fake  fake_facts  faq  features  flappy  flexibility  fooling  for  fourier  frequency  frontier  futurism  game  gan  generalization  generation  generative  geometry  github  global  gnosis-logos  goals  government  graphics  great-powers  growth-econ  guide  gwern  hack  hacker  hacking  hamming  hanson  hardware  hari-seldon  hash  history  hmm  hn  humanity  ideas  ideology  ifttt  illusions  image-recognition  image  images  impetus  incorrect  info-dynamics  info-foraging  insight  institutions  intelligence  internet  interpretability  intersection  intro  intuition  investing  iq  iron-age  iteration-recursion  judaism  language  later  law  leadership  learn  learning  left-wing  legacy  lens  library  lightweight  linear-algebra  links  lower-bounds  machiavelli  machine-learning  machine.learning  machine  machine_learning  machinelearning  magnitude  manipulation  marginal-rev  markets  martial  math  mediterranean  meta:war  metabuch  metameta  metta  microfoundations  military  ml-interpretability  ml  model-class  modeling  moments  monitoring  mood-affiliation  mostly-modern  multi  multiplicative  musk  n-factor  network  networks  neural-network  neural  neural_networks  neuralnet  neuralnetwork  neuralnetworks  neuro  neurons  nibble  nips  nn  no-go  noahpinion  nonlinearity  objects  occident  open-closed  openai  opensource  opsec  org:bleg  org:econlib  org:mat  org:med  orient  paper  papers  parable  patch  patches  pdf  peace-violence  pentesting  people  perception  perceptual  performance  perturbation  photo  picture  piracy  play  playing  poast  pocket  polanyi-marx  polisci  political-econ  politics  poll  power  powerlessness  prediction  preference-falsification  preprint  privacy  problem-solving  profile  programming  pseudoe  psychology  psychometrics  python  pytorch  q-n-a  qra  randy-ayndy  rationality  ratty  reading  realness  reason  recognition  recon  religion  rent-seeking  research-program  research  resistance  revolution  right-wing  rigidity  rigor  rigorous-crypto  risk  robot  robust  robustness  roots  s:*  sampling-bias  scapegoating  science  scitariat  securities  security  sensing  set  siggraph  simulation  singularity  sinosphere  site  smartspeaker  social-norms  social  software  sound  spatial  speculation  speech  speedometer  ssc  stalking  state-of-art  sticker  strategy  structure  study  systematic-ad-hoc  tcs  tech  technology  tensorflow  tesla  tfgan:  the-classics  the-great-west-whale  theory-practice  thick-thin  things  thinking  threat-modeling  threats  time-complexity  to_be_shot_after_a_fair_trial  tool  toread  track-record  tracking  training  trigger  trivia  trump  tutorial  twitter  tyrrany  uber  unaffiliated  uncertainty  universalism-particularism  usa  values  vision  voice  vui  war  wardriving  waves  wealth-of-nations  wealth  west-hunter  white  whitepaper  wifi  wonkish  wordlessness  zeitgeist  🎩  🔬  🤖 

Copy this bookmark: