Lightning Research Introduction Seminars Series 3 (Fall 2015)

5 October 2015 (Monday)

Seminar: Lightning Research Introduction Seminar (3)

Date: 5 October 2015 (Monday)

Time: 4-5pm (light refreshment will be arranged outside the LTF after the seminar)

Venue: LTF (near lifts 25/26), HKUST

Host: Prof. Huamin QU (PG Coordinator)

Speakers: CSE Faculty Members

  1. Ke YI
  2. Nevin ZHANG
  3. Xiaojuan MA
  4. Brian MAK
  5. Kai CHEN
  6. Charles ZHANG
  7. Pan HUI
  8. Cunsheng DING

Schedule

Time Speaker Talk Title Abstract Slides Video
4:00-4:05pm Dr. Ke YI Small Summaries for Big Data Abstract Slides Playback
4:05-4:10pm Prof. Nevin Zhang Latent Tree Models and Applications Abstract Slides Playback
4:10-4:15pm Dr. Xiaojuan MA Multimodal Interaction @ Social-Digital Media Abstract Slides Playback
4:15-4:20pm Dr. Brian MAK Automatic Speech Recognition Abstract Slides Playback
4:20-4:25pm Dr. Kai CHEN Networked Systems for Big Data and Clouds Abstract Slides Playback
4:25-4:30pm Dr. Charles ZHANG Stop the Bleeding from the Heart Abstract Slides Playback
4:30-4:35pm Dr. Pan HUI From Connecting People to Connecting People with Machines Abstract Slides Playback
4:35-4:40pm Prof. Cunsheng DING My Past and Current Research Abstract Slides Playback
4:40-5:00pm Q&A Session Playback

Full Video Playback

Abstracts

Talk 1

Title: "Small Summaries for Big Data"

Speaker: Dr. Ke YI

Abstract:

The main approach for dealing with big data is to scale up / out 
computation. My research, however, takes the complimentary approach of 
scaling down data. In this talk, I will discuss a few applications and 
projects on how data summarization techniques can be used to tackle the big 
data challenge.

Talk 2

Title: "Latent Tree Models and Applications"

Speaker: Prof. Nevin ZHANG

Abstract:

Latent tree models are a class of probabilistic models with multiple latent 
variables. They are a tool for latent structure discovery, co-occurrence 
detection, and multidimensional clustering. They have been used to detect 
topics in text data, and to analyze survey data from marketing research and 
social sciences. They have also been used to solve some fundamental 
problems in traditional Chinese medicine.

Talk 3

Title: "Multimodal Interaction @ Social-Digital Media"

Speaker: Dr. Xiaojuan MA

Abstract:

I will give a brief overview on the three directions of my current research 
focus - multimodal interaction with and through social-digital media: 1) 
social interaction through digital media; 2) interaction in social media 
with rich digital contents; and 3) interaction with integrated 
social-digital media.

Talk 4

Title: "Automatic Speech Recognition"

Speaker: Dr. Brian MAK

Abstract:

I will give a 5-minute crash tutorial on acoustic modeling for automatic 
speech recognition.

Talk 5

Title: "Networked Systems for Big Data and Clouds"

Speaker: Dr. Kai CHEN

Abstract:

In this talk, I will introduce the networked systems research conducted in 
our group, with an emphasis on a real infrastructure platform we built, and 
a simple yet practical and effective flow scheduling framework we 
implemented on top of it (http://sing.cse.ust.hk/projects/PIAS).

Talk 6

Title: "Stop the Bleeding from the Heart"

Speaker: Dr. Charles ZHANG

Abstract:

Buffer-overflow, one of the first known computer security problems, is 
essentially caused by programming mistakes. Unfortunately, in the past 30 
years since its first appearance, buffer overflow bug remains as one of the 
deadliest security bugs, e.g., the recent ssl heartbleed fiasco. In the 
talk, I will provide an anatomy of the SSL heartbleed bug and highlight the 
challenges we need to overcome to automatically detect such problems.

Talk 7

Title: "From Connecting People to Connecting People with Machines"

Speaker: Dr. Pan HUI

Abstract:

In this talk I will briefly introduce my research journey in networking and 
human computer interaction.

Talk 8

Title: "My Past and Current Research"

Speaker: Prof. Cunsheng DING

Abstract:

In this five-minute talk, I will outline my past and current research in 
layman's terms, so that the audience will get a quick understanding of my 
research topics.