Garrett Bingham

About Me

I am a senior at Yale University majoring in Computer Science & Mathematics. I am interested in the automatic design of neural network architectures, and more broadly in techniques for automated machine learning. Parts of this site are outdated. For the most up-to-date information about my research and publications, please download my resume on the right.

Contact Info

Please ask by email



Resume (11.30.2018)


Random Subspace Two-dimensional LDA for Face Recognition

Submitted to Pattern Analysis and Applications

Presented at:

Abstract: In this paper, a novel technique named random subspace two-dimensional LDA (RS-2DLDA) is developed for computer face recognition. This approach yields a number of improvements over the random subspace two-dimensional PCA (RS-2DPCA) framework introduced by Nguyen et al. Firstly, RS-2DLDA is a supervised method. Because of this, the features learned by RS-2DLDA are generally more discriminative than those from RS-2DPCA. RS-2DLDA further offers the flexibility of choosing a distance metric to be used during classification. Some metrics are better suited to face recognition than others and can help to compensate for issues such as face images in different lighting conditions. Finally, RS-2DLDA is an ensemble method. Experiments are conducted to demonstrate ways to increase classifier diversity through parameter tuning. A weighting scheme is introduced which exploits this classifier diversity to further boost accuracy. A series of experiments on the MORPH-II and ORL datasets demonstrate the effectiveness of this approach.

MORPH-II: Inconsistencies and Cleaning Whitepaper

Published to Seahawk DOCKS

Abstract: This paper presents a detailed summary of the inconsistencies in the non-commercial release of the MORPH-II dataset and covers the steps and strategy taken to clean it. In addition, examples of prior research that made use of the uncleaned data are briefly introduced and the potential implications on their results are discussed.