Entering Year 2019, I plan to start a post series discussing what I have learned in statistics, machine learning, big data, computer science, and neuroscience (always!). I name this series “Connect the Dots”, as in the puzzle game “connect the dots“.
This series idea was long inspired by my conversation and discussion with friends, classmates, and coworkers, and reflected by the shared struggle on the intimidating breadth, challenging depth, and seemingly infinite unknown territory of knowledge and skills in data science. Let alone keeping up with latest research discovery and industry trend.
Table of Contents:
- Set up the supervised learning problem as function fitting
- Restrict function search space by assumptions
- Goal setting for function fitting (regression)
- Goal setting for function fitting (classification)
- Everything you need to know about matrix in machine learning (I): Solve Ax = b
- Everything you need to know about matrix in machine learning (II): eigendecomposition and singular value decomposition
- Numerical optimization in machine learning (I): the basics
- Numerical optimization in machine learning (II): unconstrained optimization
- Numerical optimization in machine learning (III): Constrained optimization
- Learn to learn: Hyperparameter Tuning and Bayesian Optimization
- How does Bayesian Optimization work?
What this series is NOT:
- It is NOT a checklist for data scientists.
- It is NOT a systematic and general introduction to statistics or computer science. If you have zero knowledge in such field, you may consider taking some 101 courses first.
- It is certainly NOT a textbook and does not intend to be thorough and complete.
What this series intends to include:
- Study note and reflection
- Code implementation
- Algorithm discussion
- Theoretical derivation of formula
- Expert interview
- Challenges in projects
The goal of this series is not only to recap and organize my personal learning experience, but also to share with others for further discussion.