currently learning about ml

Resources

Systems Programming
- OSTEP - Operating Systems: Three Easy Pieces:
  The best textbook I have used in CS, and the best introduction to operating systems. Doesn't require much background past beginner knowledge of C. Very digestible, and lays out content in a manner that motivates the solutions presented.
Machine Learning
- Introduction to Statistical Learning
  Pretty good foray into classical ML techniques, doesn't expect knowledge of linear algebra, provides a decent mathematical base for everything while also providing practical labs. Also has an accompanying set of videos.

Log

week of 5/19/24
- Finished up the final chapter of Introduction to Statistical Learning on unsupervised learning
- Found more reasoning behind the use of PCA
- Started reading a few Kaggle competition notebooks to see how the theory translates / how much of what I learned is actually relevant in (Kaggle) practice.
- Finished Introduction to Statistical Learning
- Happy to be done with ISL, hoping to learn more about DL and GPU programmming now, while also applying the theory I learned
week of 5/12/24
- Started a brief introduction to neural networks
- Learned fundamentals of decision trees, bagging, boosting, and random forests.
- Learned about SVMs and their motivations
- continued working through introduction to statistical learning, covered:
  - generalized linear models
  - bayesian perspectives
  - more statistics knowledge
  - deep dive into regularization techniques
- finished learning about linear models, starting to move onto non-parametric models
week of 5/5/24
- interview cram
- continued learning about some basic statistical tests
- logistic regression
- generalized linear models helped relate a lot of concepts in my head
- resampling methods
- strengths of bayes classifiers
- learning about motivations behind AI developments
  - scaling is all we need?
  - field seems much less intimidating then when I started. it changes fast but the core remains the same
week of 4/28/24
- final exams
- deep dive into linear regression
- review of statistics
week of 4/21/24
- learned more about inference scheduling constraints
- math behind backpropagation
- attended YC startup school to learn more about entrepreneurship
  - cold emails
  - cold emails
  - cold emails
week of 4/14/24
- learned more about the internals of file systems in the OS
- the historical structure of neural networks
- coursework on fpga programming
week of 4/7/24
- interviewed with a company working on the application of reinforcement learning and finance
- studied stats and interview prep material
- worked primarily on coursework and non-ai computer science related material
- read about federation learning and attempted to grok SuperFed
- brushed up on algorithms
- found a great meta-resource by Sam Finlayson for machine learning
week of 4/1/24
- unresolved questions
  - what are bandits? getting conflicted answers
  - how do we deal with covariate shift and label shift?
- Realized my calculus understanding is pretty poor, was having trouble deriving the analytic solution to linear regression. Going to address this next week. Possibly working through the [MIT OCW course on single-variable calculus](https://ocw.mit.edu/courses/18-01sc-single-variable-calculus-fall-2010/).
- d2l: linear regression and gradient descent
- islp: k-nearest neighbors
- d2l: softmax and basics of information theory
  - cross-entropy loss
  - understanding the purpose of log
- 3B1B intro to transformers
- restarted use of anki