Machine Learning

Paradigms

• Supervised Learning

• Unsupervised Learning

• Reinforcement Learning

Supervised Learning

Supervised Learning

• Data : (X, y)
• Model p( y | X )
• Unavailability of labelled data
• Data Tagging

Unsupervised Learning

Unsupervised Learning

• Exploit inherent latent structure in data
• Generative Modeling
• Model p(x)
• Generate new samples from model

Reinforcement Learning

Reinforcement Learning

• Unknown Environment
• Not-so-well-defined objective
• Learn through Self-play

Deep Learning

Deep Learning

Deep Learning

Applications

• Computer Vision
• Speech Recognition and Synthesis
• Language Understanding
  • Machine Translation
  • Dialogue Systems
  • Question Answering
• Recommendation Systems
• Robotics

Convolutional

Neural Networks

Convolution

CNN

CNN

Recurrent Neural Networks

RNN

Learning

Linear Regression

Linear Regression

Gradient Descent

Gradient Descent

Multi-layer Perceptron

Backprop

Research

Genomics

DecaNLP

Interpretability

Interpretability

Interpretability

Question Answering

Bayesian Learning

Bayesian Learning

Bayesian Learning

Bayesian Learning

Bayesian Learning

Hyperopt

Hyperopt