Blog

Not everything is meant to be. But everything is worth a try.

MetaStyle: Trading Off Speed, Flexibility, and Quality in Neural Style Transfer

This post briefly summarizes our work on trading off speed, flexibility, and quality in neural style transfer. For further details, please refer to our AAAI 2019 paper. Figure 1. Stylization ...

Mirroring without Overimitation

To learn more about the approach we take to teach a robot to open medicine bottles, please checkout our paper. 1. Introduction A hallmark of machine intelligence is the capability to adapt to...

Training a Quadcopter to Autonomously Learn to Track

This a summary of our IJCAI 2018 paper in training a quadcopter to learn to track. 1. Introduction We’ve witnessed the advent of a new era for robotics recently due to advances in control met...

Towards Reducing Repetitiveness in Community-Based Question Answering

This a summary of our Neurocomputing paper in solving the repetitiveness in Community-Based Question Answering. 1. Introduction Community-Based Question Answering (CQA) platforms, such as Quo...

Release of SceneSet-50 and Code for Batch Automatic Transformation of 3D Models

This blog post summarizes the interior scene dataset built when I was a research intern at VCLA@UCLA together with code for batch transformation of 3D models. 1. Introduction Computer Vision ...

Towards Real-Time Automatic Number Plate Detection: Dots in the Search Space

Tech-report on ANPR

A PDF version of this tech report could be found here. Abstract: Automatic Number Plate Detection [1], or ANPR, has long been a traditional problem and is already widely applicable in police fo...

A Bite of Computational Art

Convolutional and Neural Art

This blog post depicts my personal experience in playing with the computational art, or to be more precise, algorithms that populate pictures with elements of aesthetics. It really is great fun ...

Demystify DeepWalk and Crack Its Code

Brief Overview of DeepWalk

1. Introduction The algorithm of DeepWalk [1] was first introduced by Bryan Perozzi, Rami Al-Rfou and Steven Skiena from Stony Brook University. In a word, DeepWalk tries to learn good latent rep...

The Monty Hall Problem: A Bayesian Approach

How I Attacked a Probabilistic Puzzle

1. How I Came Across Monty Hall It is actually pretty straightforward: a similar problem showed up in Problem Set 1 of the course Introduction to Data Mining. It all came as follows: There are...