Reading

fixing a bug in google chrome as a first-time contributor

@ameobia10· completed

this was a cool read that detailed a first-time open source contributors pr process. but more than that @ameobia10 shared the thought process behind finding an issue, grabbing the codebase, navigating the codebase, reproducing the issue, writing code and tests, and going through the cl (chromium pr) process. then the rewarding parts of seeing it in chrome canary 128 which he actively uses.

a method for stochastic optimization

Diederik P. Kingma, et al.· soon

...

an exploration of clustering algorithms for customer segmentation in the uk retail market

Jean Mary John, et al.· in-progress

related to research work @ traderspost

customer segmentation in user behavior analysis: a comparative study of clustering algorithms

Yingze Liu· in-progress

related to research work @ traderspost

customer churn prediction in the software by subscription models it business using machine learning methods

Anna Kolomiiets, et al.· in-progress

related to research work @ traderspost

a survey on customer churn prediction using machine learning techniques

Saran Kumar, et al.· in-progress

related to research work @ traderspost

variational inference: a review for statisticians

David M. Blei, et al.· completed

getting into the weeds of variational inference (vi). the book does a good job explaining that vi is not just a bayesian tool but also has many other applications like reinforcement learning. the concepts were mostly new to me as i mainly studied mcmc during classes. vi is clearly a powerful tool for approximating posterior distributions in a optimization framework. it would be cool to go and replicate something similar to the nytimes probabilistic topic modeling example.

auto-encoding variational bayes

Diederik P. Kingma, et al.· completed

starting to see some patterns. estimate the variational lower bound (stochastic gradient variational bayes in this case). then apply an efficient algorithm for inference (auto-encoding variational bayes). this paper had good derivations and clear experiment procedures for comparing wake sleep, mcem, and aevb algorithms.

black box variational inference

Rajesh Ranganath, et al.· soon

...

a primer on probabilistic inference

Thomas L. Griffiths· completed

wow. great read. it smoothly goes from describing bayesian inference to using it in practice when dealing with complex distributions and problem types. also mentioned dynamic programming and graphs which made me happy since im doing a lot of dsa practice right now.

theory-based bayesian models of inductive learning and reasoning

Joshua T. Tenenbaum, et al.· completed

theory-based bayesian thinking behind human cognitive science. paper highlights interactions of sophisticated inference processes or sophisticated konwldge representations. a mix of tree-structured priors, casual graphs, and the tendancy for smaller more specific hypothesis to be preferred over larger generalized ones.

towards a future space-based, highly scalable ai infrastructure system design

Blaise Agüera y Arcas, et al.· completed

super early system design for a future space-based ai infrastructure. covers physical solar panel design, data transmission methods, cost estimatation based on spacex launch history, and tpu radiation testing. cool read.

core views on AI safety

Anthropic· completed

using llms every day is cool, but not being aware of the risks beyond a couple 'extremist' reels is not. i decided to read this to understand what anthropics views on ai safety and the future impact of ai. its actually insane how much effort is being put into research at anthropic. i personally find the ideas of mechanistic interpretability interesting, but i think i should align myself (no joke) with the alignment capabiliteis and alignment science research first since that is the core of the safety effort right now.

quantitative trading

Ernest P. Chan· completed

this book is high level for a good reason. it covers the practical implications of trading automated systems and scaling a portfolio. from data quality to risk management. when to start or end a business like this from someone who has done it in both a professional desk and home office.

structured computer organization

Andrew S. Tanenbaum· completed

a bottom-up walkthrough from digital logic to the OS layer. great visuals on microarchitecture. i enjoyed reading through the whole book in my free time but it would also be a good choice for referencing individual chapters.

a practical guide to quantitative finance interviews

Xinfeng Zhou· completed

the book did a good job tieing together the math and practical aspects of finance trading ie markov chains, risk, etc. a good read if you want to get a quick breadth refresher imo.