Jiquan Ngiam

Co-Founder, CEO @ Lutra AI

Previously:
Senior Staff Software Engineer @ Google Brain
Founding Team @ Coursera
LinkedIn | Writings

Ph.D. Candidate Computer Science | Stanford University (advisor: Andrew Ng)
B.S. Computer Science | Carnegie Mellon University

jngiam

cs.stanford.edu

Publications (at Stanford)

J. Ngiam, P. Koh, Z. Chen, S. Bhaskar, A.Y. Ng.
Sparse filtering.
NIPS 2011. [PDF] [Supplementary]
Q.V. Le, A. Karpenko, J. Ngiam, A.Y. Ng.
ICA with reconstruction cost for efficient overcomplete feature learning.
NIPS 2011. [PDF] [Supplementary]
J. Nam, J. Ngiam, H. Lee, M. Slaney.
A classification-based polyphonic piano transcription approach using learned feature representations.
ISMIR 2011. [PDF]
J. Ngiam, Z. Chen, P. Koh, A.Y. Ng.
Learning deep energy models
ICML 2011. [PDF]
J. Ngiam, A. Khosla, M. Kim, J. Nam, H. Lee, A.Y. Ng.
Multimodal deep learning
ICML 2011. [PDF]
also appeared in the Deep Learning and Unsupervised Feature Learning Workshop (NIPS 2010). [PDF] Q.V. Le, J. Ngiam, A. Coates, A. Lahiri, B. Prochnow, A.Y. Ng.
On optimization methods for deep learning
ICML 2011. [PDF]
Q.V. Le, J. Ngiam, Z. Chen, D. Chia, P. Koh, A.Y. Ng.
Tiled convolutional neural networks.
NIPS 2010. [PDF] [Visualizations]

Code, Tutorials and Courses

Sparse Filtering
Matlab code that demonstrates how to run sparse filtering to train a two layer network.
http://github.com/jngiam/sparseFiltering
(Clone using

git clone --recursive git://github.com/jngiam/sparseFiltering.git

) Deep Learning Tutorial
Tutorial on deep learning, covering sparse autoencoders, whitening, softmax regrssion, deep neural networks, convolution and pooling.
http://deeplearning.stanford.edu/wiki/index.php/UFLDL_Tutorial Machine Learning Class
Machine learning course with a focus on applications; course covers supervised learning (classification, regression), unsupervised learning (clustering, dimension reduction) and recommender systems.
http://www.ml-class.org/ Score Matching with Independent Component Analysis (ICA)
Matlab code that learns overcomplete ICA bases using Score Matching and minFunc. Very fast!
http://github.com/jngiam/scoreMatching
(Clone using

git clone --recursive git://github.com/jngiam/scoreMatching.git

)