neural   3155

« earlier    

BBC - Future - An effortless way to improve your memory
When trying to memorise new material, it’s easy to assume that the more work you put in, the better you will perform. Yet taking the occasional down time – to do literally nothing – may be exactly what you need. Just dim the lights, sit back, and enjoy 10-15 minutes of quiet contemplation, and you’ll find that your memory of the facts you have just learnt is far better than if you had attempted to use that moment more productively.
memory  brain  neural 
11 days ago by bouvin
How I Shipped a Neural Network on iOS with CoreML, PyTorch, and React Native - Stefano J. Attardi
My plan was to Google my way out the problem, as one does. I soon found the right keywords: segmented regression, and piecewise linear regression. Then I found one person who solved this exact problem using basic math. Jackpot!

Or not. That approach tries to split the trendline at every possible point and then decides which splits to keep based on how much they improve the mean squared error. Worth a shot, I guess.

No matter how I tweaked the parameters, the algorithm was either splitting too ...
coreml  ai  ml  regression  linear-regression  neural  neural-network  best 
12 days ago by hellsten

« earlier    

related tags

affordance  ai  algorithm  alphazero  analysis  ann  anomaly  approximation  architecture  art  artificial  assessment  attention  audio  autoencoder  automl  backpropagation  best  bike  biomimicry  birds  block  blog:  board  book  books  brain  calculus  capsule  carlo  categorical  chemistry  chemoinformatics  cnn  code  coloring  column  complex  compression  comps  computer  control  convolutional  coreml  cost  course  cs  culture  dataflow  deep-learning  deep  deeplearning  deepmind  dependent  detection  dev  differential  discussion  drone  education  elmo  embedding  embeddings  emulation  engines  episodic  equation  explainability  extraction  fake  first  frontend  function  funny  game  games  gaming  gated  general  git  google  gopro  gpu  graph  group  hackathon  hamiltonian  haskell  health  hidden  howto  idea  identity  ids  image  images  information  intelligence  interfaces  intro  introducing  ios  javascript  js  layers  lear  learn  learning  library  linear-regression  lisp  locative  logic  logistic  lstm  machine-learning  machine  machinelearning  mashup  math  mathematics  matrix  memnn  memory  message  meta  ml  mnist  mockups  models  monitoring  monte  music  native  netherlands  nets  network  networks  neural-network  neural-style-transfer  neuralnetworks  neuro  nids  nima:  nlp  nn  olfactory  online  openai  opensource  optimization  order  papers  partial  passing  perception  physics  pocket  poetry  porn  pornography  program  programming  pruning  psychology  python  pytorch  questions  react  reality  recnn  recursive  regression  regularization  reinforcement  relation  relational  renormalization  research  resources  review  rnn  robots  rule  scala  science  security  sentiment  sgd  simulation  sketch  software  sparse  spiking  spreadsheet  steering  stories  stream  style-transfer  survey  syncwalk  synthesis  technology  tensors  theory  top  traffic  tronto  turing  type  universal  variational  vision  word  worm  xor 

Copy this bookmark:



description:


tags: