Jaan Lı 李 PhD's Projects
American Community Survey data on people and households
A Python implementation of global optimization with gaussian processes.
Calculating CitiBike personal statistics for the community leaderboard 📈 🚴
CoFactor: Regularizing Matrix Factorization with Item Co-occurrence
Instead of classifying activity, this app does one thing: tells you how much you sit based on accelerometer data.
User-artist-song Poisson Factorization
Automated expansions of Hamiltonian cumulants for analyzing Monte Carlo simulations.
Deep exponential family models in MXNet/Gluon. Layers o' latents 💤
Dotfiles - main development environment is vim + tmux or emacs
Using the Census Bureau's American Community Survey data with `dbt` (data build tool) for creating compressed parquet files for exploratory data analysis and downstream applications.
implementation of fair dummies
Library for fast text representation and classification.
:hamburger:
Deep Latent Gamma Model / Gamma VAE
Gaussian mixture model implementation in C++ with black box variational inference and control variates
Gradient fuzzing for probabilistic programming
PyTorch implementation of graph ConvNets, NIPS’16
Hierarchical variational models for physics.
A Retina-ready Jekyll-powered blog with responsiveness, SEO, etc.; up at https://jaan.io
Python library for fast, interactive geospatial vector data visualization in Jupyter.
Modified Dvorak Model01 layout. Goal is to reduce chording via OneShot and other plugins
:goat: Lightweight configuration trees with command line flags :goat:
static site generator for maps + static site generator for data apps
This code accompanies the proximity variational inference paper.
PyTorch implementations of algorithms for density estimation
RankFromSets - SDSS submission code for reproducibility.
word2vec with a context based on sentences.