Welcome to Linh's home page, a place for me to organize and also share my programming projects, independent films, short stories, photos, and ramblings.
- TensorFlow JS Example using the TensorFlow JS machine learning library
- Pyodide Run and edit a neural network written in Python 3 and Numpy directly in your browser!
- KIC 8462852 My first co-authored scientific paper on a 1574-day periodicity of transits orbiting KIC 8462852 also known as Tabby's Star or the Alien Megastructure Star
- Exo-planet transit explorer Create planets, asteroids, megastructures, and shapes to discover how a transiting exoplanet or alien megastructure might affect a star's light curve as seen by Kepler and other observatories. NEW: mobile friendly!
- C# Roslyn I embedded the Roslyn C# compiler/interpreter that lets you edit and run C# or run pre-formatted code snippets via your browser.
I'm working on a C#/.NET re-imagining of Hadoop/Spark. Let me know if you'd like to help! The design goals are:
- Zero-configuration. It just works. Add a new node and it is automatically accepted into the swarm.
- Each node is independent and smart. No need for master or control nodes.
- Memory-based computing (a la Spark) with hard disk for back ups.
- Supports streaming and batch including all C#/.NET features including TPL, Rx, channels, etc.
- Simple distributed in-memory cache with parallel/async support with plug-in support for alternatives like Redis.
- Simple distributed message bus with plug-in support for alternatives like MassTransit/RabbitMQ.
- Online dashboard that lets you test new code in realtime. Type and run! No build, copy, and run/test cycle needed.
- Fully redundant and resilient: The network organically recovers from node or connection loss.
- Beyond Spark and current paradigms, this could be used for: in-memory distributed databases and a general distributed computing platform or as an alternative to microservices architecture.
- Deep Dream Run Google's incredible Deep Dream neural network algorithm right in your browser with parallel worker tasks.
- Image Painting From ConvNetJS, train a neural network to paint a photo you upload. Fascinating watching the learning happening.
- Deep Q Learning Train a ConvNetJS creature to eat berries and avoid poison. My plan is to add procedural level generation, additional agents, etc.
- Recurrent Neural Network Train a network to predict the next words in the style of a particular author. Nifty.
- 2D Machine Learning My experiments with a 2D physics engine and deep learning.
- Python & Turtle Graphics Run Python right in your browser with turtle graphics snippet.
- Leaflet Example using the LeafletJS mapping library
AI & CNNs
This section will list my code and thoughts on AI.
This section lists my independent film projects.
This section will show highlights from my years in photojouralism and artistic photography.