Michigan State University Skip navigation MSU Home    CSE & MSU    Embodied Intelligence Lab.   


[As of October 2013, I have graduated and moved to Microsoft Corp. This page will no longer be updated.]

Where-What Network (WWN) in C++. The Where-What Networks with Qt4 GUI used for various research projects. Download

Requirements: ANSI C++ compiler, CMake and Qt4.

MILN in C++. A simple version of our Multilayer In-place Learning Networks (MILN) prepared as a template for a class project. Download

Requirements: ANSI C++ compiler.

Stereo network modeled after the 6-layer architecture of cortex to develop binocular receptive fields in disparity-sensitive neurons.

Source code and data are available upon request.

Below is an example of the running of the program.

(Click on Change quality for better viewing.)

Locally-connected, temporal MILN used for visual object recognition. The program is GUI-enabled. Download.

Requirements: ANSI C++ compiler, Automake Tools, Qt 4.0 or higher.

(Click on Change quality for better viewing.)

Convolutional Deep Belief Networks are implemented in PyCuda (Cuda wrapper for Python). The implementation is much easier to read and faster (due to GPU) than the original Matlab implementation. Here is a techinical report for the software.

Source code is available upon request.

Last version of Aria2005 soccer simulator project that ranked 1st among more than 30 teams both the Robocup2005 world championship in Osaka, Japan and Robocup2004 in Lisbon, Portugal.

Source code is available upon request.

Requirements: ANSI C++ compiler, The Robocup Soccer Simulator.


Home | Contact | Publications | Software | Hobbies