Deep learning notes Link to heading

Deep Learning is a subset of Machine Learning and this one is a subset of Artificial Intelligence.

Required Math:

  • Linear Algebra
  • Calculus
  • Probability and statistics

Frameworks: Numpy, Pnadas, Matplotlib, tensorflow, Keras, pytorch, hugging Face

Keys:

  • Data Preprocessing: data cleaning, normalization
  • ML algos: linear regression, decision trees, random forest, SVM, K-NN
  • Evaluaction metrics: Accuracy, precision, Recall, F1 Score
  • Deeplearning: neural networks. CNNs, RNNs, GANs, Transformers
  • NPL: text mining, sentiment analysis
  • Reinforcement learning: q-learning, policy gradients
ML vs DL
― ML vs DL ―

LLMs: GPT (openai), BERT (google), Roberta (facebook)

Models:

  • Bert for undesratndinc a context in a text
  • GPTs (2, 3, etc) generates texts
  • T5 for translations and summarization