machine learning

author: John cron

Deep Learning vs machine mearning

  • Machine learning works on feature engineering. Ex finding the edges and recognizing faces
  • Deep learning is focussed on statistical engineering. Ex learning about faces..lots of them and then recognizing faces
  • As we got more GPU and processing power and data storage became cheap deep learning outperformed machine learning

Representation of language

  • language consists of

    • phonology -> the way a word sounds

    • morphology -> structure of word

    • words -> tokenization

    • syntax -> phrases etc

    • semantics -> meaning of words (this is very hard to infer as same sentences may have humor or rude meaning)

  • All these require lot of calculation in ML world, In DL we use vectors to do all this calculation.

  • In ML we need to create huge matrix for words , Its very hard to find something like similar words etc. In ML these words are located in N dimensional space and closer words are colocated

22 articles

Running code First neural network Jupiter cheat sheet networks supervised learning unsupervised learning model representation cost function matrix linear regression with multiple features gradient descent for multiple variables normal equation summary 1-4 + logistic regression octave neural networks cost fn for neural networks neural networks gradient back_propogation_intuition BackPropogation-unrolling parameters numerical gradient checking random initialization putting it together BP
Published with GitBook · Last edit January 24th 2019

results for ""

    No results matching ""