RESEARCH
Computer Vision and Machine Learning
Conformal Training
End-to-end training of deep neural networks and conformal predictors to reduce confidence set size and optimizer application-specific objectives.
Conformal Training
Robust Generalization and Flatness
Robust generalization and overfitting linked to flatness of robust loss surface in weight space.
Robust Generalization and Flatness
Random and Adversarial Bit Robustness
Random and adversarial bit error robustness of DNNs for energy-efficient and secure DNN accelerators.
Random and Adversarial Bit Robustness Random Bit Error Robustness
Confidence-Calibrated Adversarial Training
Confidence calibration of adversarial training for "generalizable" robustness.
Confidence-Calibrated Adversarial Training
Adversarial Patch Training
Adversarial training on location-optimized adversarial patches.
Adversarial Patch Training
Adversarial Robustness and Generalization
Disentangling the relationship between adversarial robustness and generalization.
Adversarial Robustness and Generalization
Improved Shape Completion
Learning 3D shape completion under weak supervision; on ShapeNet, ModelNet, KITTI and Kinect data; published at CVPR and on ArXiv.
Improved Shape Completion Shape Completion
Superpixel Benchmark (2016)
A comprehensive comparison and evaluation of 28 superpixel algorithms on 5 different datasets; published in CVIU and GCPR.
Superpixel Benchmark (2016) Superpixel Benchmark (2014)
Superpixel Algorithms
Revised C++ implementations of two popular superpixel algorithms, SEEDS and FH, which are shown to outperform the original implementations.
Superpixel Algorithms
Video Segmentation
Efficient C++ implementation for hierarchical graph-based image segmentation.
Video Segmentation
Non-Convex Optimization
Efficient C++ implementation of iPiano, a proximal algorithm with inertial force for non-convex and non-smooth optimization; including applications to image segmentation.
Non-Convex OptimizationMISC
Educational
MatLab Two-Layer Perceptron
A two-layer perceptron for digit classification on MNIST implemented in MatLab.
MatLab Two-Layer Perceptron
PHP Matrix Decompositions
Common matrix decompositions, including LU, Cholesky and QR decompositions, implemented in PHP; also includes an interactive application to demonstrate and explain the material.
PHP Matrix Decompositions
Functional Dependencies
Relational databases can be described in terms of functional dependencies; this project implements functional dependencies as well as second and third normal form in PHP.
Functional DependenciesOPEN SOURCE
Web Development
Twitter Bootstrap Plugins
Some plugins for Twitter Bootstrap and jQuery; including a multiselect plugin and a password strength meter.
Twitter Bootstrap Plugins
Kohana Modules
Modules for the HMVC framework Kohana; including modules for user authentication, access control, form generation and validation, CSS and JS assets and more.
Kohana Modules
CMSimple Plugins and Themes
Plugins for CMSimple XH, a lightweight content management system; including, among others, news article, image gallery and YouTube plugins.
CMSimple Plugins and Themes
Wordpress Plugins and Themes
Experimental Wordpress plugins and themes; including this theme and plugins for author biographies and live data from GitHub.
Wordpress Plugins and Themes