multimedia

ICSI in Geekosystem

When you’re trying to figure out statistics about something as nebulous and pervasive as “the Internet,” you’re going to have to performs some interesting mental contortions. After all, how many things out there are connected to the Internet? How many of those things would you consider to be actively “using” it, how much power is there in the world? All valid and difficult questions, valid and difficult questions that Justin Ma and Barath Raghavan, of UC Berkeley and the International Computer Science Institute respectively, were determined to tackle.

ICSI in The Daily Californian

"Internet Accounts for Almost 2 Percent of the World’s Total Energy Consumption"
October 30, 2011  |  Paras Shah, The Daily Californian

UC Berkeley researchers have estimated that it takes between 1 and 2 percent of the world’s energy to construct, run and maintain the Internet. According to the research, which will be presented Nov. 14 at the Workshop on Hot Topics in Networks in Cambridge, Mass., the Internet uses between 170 and 307 gigawatts — about 1.1 to 1.9 percent of the 16 terrawatts of energy consumed by humanity.

ICSI in InfoWorld

"Nowhere to Hide: Video Location Tech Has Arrived"
February 21, 2013  |  Bill Snyder, Tech's Bottom Line, InfoWorld

Researchers at ICSI are currently building a video database by analyzing videos downloaded from Flickr, says Gerald Friedland, who leads ICSI’s multimedia efforts. Data from videos taken at known locations is used to develop profiles of the respective locations.

BEARS Open House This Thursday

Tuesday, February 12, 2013

Bernhard Amann will demonstrate the SSL Trust TreeEvery year, we host an open house in conjunction with the Berkeley EECS Annual Research Symposium (BEARS). This year, BEARS is on Thursday, February 14. Group directors will be giving overviews of their groups' work, and senior researchers will be on hand throughout to talk about and demonstrate their research. The open house starts at 2 p.m. - we hope you can make it!

The abstracts for all our demos are on the events page, but here are a few previews:

Internet Privacy Misunderstandings, Part 4

Friday, December 7, 2012

privacy misunderstanding #4Even if you are careful not to post private information, you might not realize how much metadata becomes available whenever you post something online.

Metadata is information that is contained in files such as photos and status updates on social networks, but is not visible when looking at the content, so many people are not aware that it exists.

Qualifying Mechanical Turk Users for Skilled Tasks

Wednesday, December 5, 2012

geolocation of videoMechanical Turk, Amazon's popular crowdsourcing platform, is used by many people and organizations who need repetitive tasks completed. For rapid completion of unskilled (easy) tasks online, crowdsourcing can be an easy and affordable solution. But what if you need people to complete a skilled task? Researchers at ICSI decided to find out.

Drinking from a Multimedia Firehose: How to Deal with Billions of Daily Video Posts

Monday, October 29, 2012

At the ICSI Research Review, Gerald Friedland spoke about the challenges of performing empirical research on the astronomically large data set that is consumer-produced video. Here is Dr. Friedland's abstract:

Consumer-produced videos are the fastest-growing type of content on the Internet. YouTube claims that 72 hours of video are uploaded to its Web site alone every minute. Because the videos capture parts of the world, they are potentially useful for qualitative and quantitative empirical research on a larger scale than has ever been possible before. A major prerequisite to making social media videos usable for "field studies" is efficient and unbiased (e.g., keyword-independent) retrieval. More importantly, retrieval needs to go beyond simply finding objects to detecting more abstract concepts, such as “taking care of a car” or “winning a game that is not a card game.” Research on such a large corpus requires the creation of methods that exploit as many cues as possible from different modalities. ICSI has begun using novel acoustic methods to complement computer vision approaches. This talk summarizes ICSI’s research and progress in this area.