artificial-intelligence   1164

« earlier    

Squeezing Deep Learning Into Mobile Phones
This is a great deck on the things you can to to strip down trained models for running on mobile platforms.
neuralnetworks  frameworks  mobile  ai  ml  machine-learning  artificial-intelligence 
5 days ago by mcroydon
Building [a neural network using homomorphic encryption] - i am trask
<< With the approach above, you could train a regular, decrypted neural network for a while, encrypt it, send it to Party A with a public key (who trains it for a while on their own data... which remains in their possession). Then, you could get the network back, decrypt it, re-encrypt it with a different key and send it to Party B who does some training on their data. Since the network itself is what's enrypted, you get total control over the intelligence that you're capturing along the way. Party A and Party B would have no way of knowing that they each received the same network, and this all happens without them ever being able to see or use the network on their own data. You, the company, retain control over the IP in the neural network, and each user retains control over their own data. >>

I suspect that this isn't actually going to work but bookmarking for future digestion.
artificial-intelligence  neural-networks  cryptography 
6 days ago by absfac
Alan Jacobs | Analogy
For many years now Douglas Hofstadter, drawing on the work of the mathematician Stanislaw Ulam, has been convinced that the secret to creating artificial intelligence lies in teaching machines to recognize analogies. (Ulam says somewhere that it’s all about “as”: we see marks on a piece of wood pulp as a portrait of a beloved child, a cairn of stones as a monument to a dead chieftain.) Similar principles underlie the methods of Google Translate, which collects an enormous corpus of sentences and then tries to match your input to something in that corpus, and Apple’s “digital personal assistant,” Siri. Siri can’t parse what you say to her unless she can connect to the network, which undertakes a comparison of your utterance to other utterances on record. All this might be called brute-force analogizing, but it seems to me that my own understanding develops as I pursue the same method, though with far less force and (I hope) less brutishness.
douglas-hofstadter  alan-jacobs  artificial-intelligence  metaphor  language 
8 days ago by xianoforange
[1605.07133] Towards Multi-Agent Communication-Based Language Learning
We propose an interactive multimodal framework for language learning. Instead of being passively exposed to large amounts of natural text, our learners (implemented as feed-forward neural networks) engage in cooperative referential games starting from a tabula rasa setup, and thus develop their own language from the need to communicate in order to succeed at the game. Preliminary experiments provide promising results, but also suggest that it is important to ensure that agents trained in this way do not develop an adhoc communication code only effective for the game they are playing
artificial-intelligence  collective-intelligence  machine-learning  distributed-processing  to-write-about  philosophers'-conference 
18 days ago by Vaguery
[1606.00652] Death and Suicide in Universal Artificial Intelligence
Reinforcement learning (RL) is a general paradigm for studying intelligent behaviour, with applications ranging from artificial intelligence to psychology and economics. AIXI is a universal solution to the RL problem; it can learn any computable environment. A technical subtlety of AIXI is that it is defined using a mixture over semimeasures that need not sum to 1, rather than over proper probability measures. In this work we argue that the shortfall of a semimeasure can naturally be interpreted as the agent's estimate of the probability of its death. We formally define death for generally intelligent agents like AIXI, and prove a number of related theorems about their behaviour. Notable discoveries include that agent behaviour can change radically under positive linear transformations of the reward signal (from suicidal to dogmatically self-preserving), and that the agent's posterior belief that it will survive increases over time.
wtf  to-read  artificial-intelligence 
19 days ago by Vaguery

« earlier    

related tags

##pvlpo-project  @share  academic_lab  ads  aggregation  ai  alan-jacobs  algorithm  algorithms  allan-newell  analytics  api  app  approximation  archaeology  archives  art  artificial-neural-network  artificialintelligence  augmented-human-intelligence  basic-income  bias  black-mirrory  book  books  catalog  chess  citations  cognitive  collective-intelligence  complexity  computational-labor  computer-science  computer-vision  consider:looking-to-see  consider:performance-space-analysis  consider:representation  course  cpp  creative-coding  cryptography  curated  cyborgs  data  decision-making  deep-learning  deep_learning  dev  discovery  distributed-processing  douglas-hofstadter  economics  elixir  email  emoji  encryption  erlang  ethics  evolution  facebook  forecasting  frameworks  futurism  game-dev  game  games  gaming  generation  germany  github  google-analytics  google  grand-theft-auto  health  herbert-simon  history  hivemind  hn  homomorphic  homomorphic_encryption  ideas  ilpozzo  image-analysis  image-processing  intelligence  javascript  job-market  jonh-mccarthy  journalism  laar  labor  langley  language  learning  legal  libraries  lilly-irani  list  machine-learning  machine_learning  math  mechatronics  metaphor  microsoft  mit  ml  mobile  music  nasa  natural-language-processing  neural-nets  neural-networks  neural  neuralnets  neuralnetworks  noosphere  nudge-targets  nytimes  ocw  ontologies  openai  opensource  opinion  people  perception  philosophers'-conference  philosophy-of-engineering  philosophy  plane-geometry  poker  policy  politics  portugal  programming  psychology  public-knowledge  python  r-language  raspberry-pi  rather-interesting  reading  records_management  reference  reinforcement-learning  representation  researchers  resources  robotics  robots  science  search-engine  security  self-driving  semantics  singularity  sound  speech-recognition  stanford  statistics  talk  teaching-computer  tech  technology  tensorflow  texas-holdem  textbook  thoughtful  til  time  to-read  to-write-about  traditional-methods  transfer-learning  tutorial  ucsd  watson  webservice  westworld  wicked-problems  wired  work  wtf 

Copy this bookmark: