neuralnetwork   2056

« earlier    

Weight Agnostic Neural Networks (WANN)
Not all neural network architectures are created equal, some perform much better than others for certain tasks. But how important are the weight parameters of a neural network compared to its architecture? In this work, we question to what extent neural network architectures alone, without learning any weight parameters, can encode solutions for a given task. We propose a search method for neural network architectures that can already perform a task without any explicit weight training. To evaluate these networks, we populate the connections with a single shared weight parameter sampled from a uniform random distribution, and measure the expected performance. We demonstrate that our method can find minimal neural network architectures that can perform several reinforcement learning tasks without weight training. On supervised learning domain, we find architectures that can achieve much higher than chance accuracy on MNIST using random weights.
machine  learning  ai  neuralnetwork  paper 
10 days ago by dogrover
DiffEqFlux.jl – A Julia Library for Neural Differential Equations
the first toolbox to combine a fully-featured differential equations solver library and neural networks seamlessly together

New research as of early 2019 seamlessly combines differential equations along with neural networks.
machinelearning  deeplearning  neuralnetwork  differential-equations  math 
14 days ago by tobym
Teaching a Neural Network How to Drive a Car

I think one of the reason I find neural network training so fascinating is that you can observe, in a very simple and understandable way, the basic method by which all life on Earth evolved the ability to do things like move, see, swim, digest food, echolocate, grasp objects, and use tools. (via dunstan)
software  MachineLearning  evolution  video  NeuralNetwork 
17 days ago by jefframnani
alibaba/MNN: MNN is a lightweight deep neural network inference engine.
MNN is a lightweight deep neural network inference engine. - alibaba/MNN
neuralnetwork 
5 weeks ago by geetarista

« earlier    

related tags

2018  3d  ai  alexa  algorithms  amazon  anomaly-detection  anomaly  api  architecture  art  article  artificialinteligence  artificialintelligence  assistant  audio  automation  automl  autonomous  bias  book  brain  c#  can  cg  cinema  classifier  cli  clojure  cnn  code  commandline  computergraphics  computervision  contral  control  course  crossplatform  cs  darknet  data-mining  datascience  deathmetal  deep-learning  deeplearning  deeplerning  demo  design  development  diagram  diagrams  differential-equations  discussion  distributed  driving  education  embedding  evolution  example  explainableai  explanation  face  fashion  fatml  figure  film  game  gaming  gan  gcn  gdc19  generator  github  graph  hackaday  healthcare  hyperparameter  illustration  image  imageprocessing  imagerecognition  instagram  interaction  interesting  interface  interference  interpretability  intro  javascript  js  keras  labeling  latex  learning  library  linear-algebra  list  livestream  logic  lstm  machine-learning  machine  machinelearning  math  mathematics  medicine  metal  ml  mri  music  networks  neural-networks  neural-style  neural  neural_network  neuralnet  neuralnets  neuralnetworks  neuron  nlp  nn  opensource  optimization  painting  paper  parasite  perception  photo  photogrametery  politics  presentation  principle  privacy  programming  python  r  raspberrypi  research  resources  rlanguage  rnn  signal  simple  smarthome  snake  software  sse  stats  streaming  surveillance  synthetic  technical  technology  tensorflow  tesla  text  tool  training  transferlearning  trends  tutorial  tweetit  uber  user  video  vision  visual  visualization  voice  vr  youtube 

Copy this bookmark:



description:


tags: