^B00:00:00
[ Music ]
>> We're always looking for better ways to organize our media, whether it's our music, our news, our information. At Techfest 2008 in Redmond, Washington, Microsoft researchers shared new ideas to help us do just that.
^M00:00:16
[ Music ]
^M00:00:20
This music steering project from a Microsoft Asia team allows you to better search through your own music library. It works similarly to Pandora's Music Genome Project, but on your own MP3 player and with your own songs.
>> Pandora actually acts as a professionals [inaudible] music. So we actually develop a technology to automatically detects the tags of each music.
>> So we've chosen Faith Hill, "The Lucky One" as a song that we like, and we want to find other songs similar?
>> Yes. Yes. Yes.
>> Okay.
^M00:00:52
[ Music ]
^M00:01:00
>> Whys they are similar, so we detect that automatic tags. So they have 40% [inaudible], and they also have 20% has soft rock.
>> If and when it's released to the public, users will also be able to search and purchase new music online based on their preferences, and create playlists built around an activity, like or running or studying.
>> So you can set a mood filter, for example, you want to listen to some soft music with slow tempos.
>> We're trying to put news stories in context by showing what blogs are talking about what stories.
>> That's the premise behind BLEWS, a political blog aggregator that sorts the news, left leaning or right leaning.
>> So it's everything from, like, the big name blogs all the way down to just, like, a mom in Illinois who may be blogging.
>> If the mom in Illinois is blogging, and we're able to find her on one of the several different feeds that we're talking about, and she's put herself on one of those big lists of Hi, I'm a lefty blog or Hi, I'm a righty blog, then yeah, we've got here on the list.
>> In coordination with Microsoft's Live Labs, the blogs are scanned for key words to gauge the author's political bent. And algorithms are deployed to judge the emotional response.
>> So, for example, over here, this story about Bill Buckley passing away. Here's the commentators on the right that are talking about it. Here's the two different blog entries that are on the left.
>> As of now, Microsoft won't disclose if or when BLEWS will be available to the public.
>> Translation websites are nothing new. But Microsoft researchers feel they have a way to improve the process with their Machine Translation Project.
>> All we do is choose the language, press the button, and you'll see a side-by-side translation of this page right here. You can do parallel scrolling, parallel highlighting. And if you wanted to share it with somebody, say a colleague who was Chinese, all you do is just choose the Chinese language and it changes as quick as that.
>> Researcher Heather Thorne explains that the statistical machine translations program learns from existing translated materials to produce a better translation.
>> Even though the translation quality will never be as good as human translation quality, the Machine Translator really allows us to break down our language barriers in various different scenarios, whether it be through translation of text on a web page, through search translation, or just communication with other people.
>> Besides translating web pages in real time, one of the most useful features is a live chat, where your instant messages can be immediately translated.
>> So imagine you have a colleague that's in another country and you want to be able to talk to them about a project that you're working on together, and you don't quite speak their language well enough to talk to them in their language, all you would have to do is go in and say, okay. I want to talk to my colleague who's in -- who speaks Italian. I would just choose Italian.
>> The program launched last year free to the public and can be found at translator.live.com. I'm Kara Tsuboi, CNET News.com
^M00:03:47.
[ Music ]
