Computational Culture

(Photo: Oscar at LSAS2006)

Oscar Celma has been working for a couple of years on music recommendations. I know him since ISMIR2004 and over the years we met regularly and exchanged ideas, sketches, prototypes and more “philosophically influenced thoughts about the future of music in general”. Finally he finished his Ph.D about this topic. The written dissertation gives insights into the top-notch state-of-the-art in  recommender systems fusing content-based, collaborative and innovative network-centric features.

“A must read before you die!”

Check his updated website here!

Download section: O.Celma:Music Recommendation and Discovery in the Long Tail

PS I am very happy to be part of the date of defense in BCN …

Tomorrow I have my talk at the University of Alberta. Randy Goebel was so kind to make this possible and I am really looking forward to have it as part of the Artificial Intelligence Seminar. So far I had excellent discussions with all the people here; including Sandra Zilles -our DFKI-colleague- being an incredible hostess during my stay here. We discussed a lot about the challenges for Incremental Machine Learning methods if applied to the emerging Mobile Socio-Semantic Web.

Tough challenges, save rocket science!

It is more fun to watch than to read! The C4DM guys dont hesitate to stream and record the seminar talks. What a great service for the audience being online in the net but “kind of offline” in the real world. Follow the link to see it all, including my talk (shameless selfpromotion done!).

UPDATE: Paul commented that it would be helpful to have the Slides.pdf.

I have been visiting some friends and colleagues in London. Giving talks, having discussions, etc.

First I met the guys at the Centre for Digital Music (C4DM) headed by Mark Sandler: They showed me really cool things around the semantic web for music, check the blog of Yves to find the stuff about querying a musicontology with SWI-Prolog. Mark showed me his social tag-based music browser for mood-oriented seeking … and the lunch offered excellent sushi bento. Wow, what a great day!

Next day I made a stop at Goldsmiths College. My old buddy Daniel from Germany is meanwhile there and joined the group working on Computational Creativity, including especially music in the digital, too. They have less DSP and more psychological hooks, also very interesting. And Geraint Wiggins and Tim Crawford showed up! I have seen ongoing stuff about the analysis of MySpace artist relations there, cool!

And finally in the afternoon Elias invited me to have an insight look on Last.fm. My god! Approx. 70 young and addicted people in one big room take care of the social music revolution, indeed!
Damn … the UK has BIG MOMENTUM! Maybe the situation with national research funding for these issues is not so bad overthere!

Next week I talk about my ideas towards highly personlized music recommendation at the Digital Music Centre in London. The announcement is on their Website. As already posted earlier I am truely convinced that there are good chances to combine MIR techniques with Web2.0 lifestreams, incremental machine learning and prior work in music psychology from Sloboda, Gabrielsson and Schönberger.

You may claim that this sounds too much like rocket science or blue sky research?

But I think that the time is right to cope with these challenges. When I started my Ph.D about perception of music similarity 5 years ago the technological situation was quite different. No lifestreams, MIR still at the beginning, etc. I am optimistic to find a setup of crowdsourced collections of goosebump sensations plus the latest state-of-the-art in MIR and digital lifestream aggregation (maybe even Google Open Social? )!
By accident I stumbled over a brandnew book of Oliver Sacks entitled Musicophilia. He deals with the very nature of music and human mankind by giving a series of medical case studies related to processes in the brain when performing, listening to music. I just received the book this morning! I have to read it …

The social network thing is not at all rocket science. Back in the good ol Friendster days I was seeking around this stuff for only a few weeks … and lost interest.

Some months later I joined Open BC -now XING- in order to find interesting business contacts. So far for the masterplan, but in the end after being for 1 year a frequent user I lost interest, I made no business within this community. And even worse I still receive many many request for friendship nowadays! If you look at my activity level you should no that I am not anymore interested.

In the meantime I used Flickr, YouTube and Last.fm for the pure reason of media management and feeding my 5 or so different wordpress blogs. This was efficient and FUN! But I did not really get deep into social involvement.

I missed MySpace completely because the interface was a complete desaster of design - for my very personal taste-. At least I grabbed user names for my bands. Just to be sure to be prepared for a record deal with a major label.

And now for facebook? I missed it in the beginning because I was so fed up with all this Web2.0 service overload. I still have invitations for Joost, Twitter, Pownce, etc. everything from “media-galore” to “teenage micro-blogging”. Never used them, I have no time, I have to work. BUT something happened 3 weeks ago. I got addicted again! It started when Facebook announced the great and damn smart “Open-API-Applet-Move”. I checked in with a pseudonym just to get a feeling for the technical stuff going on there. I discovered very good groups overthere by accident, joined in, pokin around … in addition to this I was part of real good event in firstlife: the summer school semantic web. I met talented and interesting people there who meet now in Facebook. The rest of the story is straightforward: I enjoyed so much the social patterns in this special interest group that I had to change from pseudonym to my real name. And then the heatwave started again. I received invitations of very very good friends, I became part of 3-4 interest groups, people gave me positive feedback on actions, walls, my photos, etc. Social gratification by people whom I trust for their own great work and passion! I explored actively further applet features, and finally I am IN! I am so facebooked!

Straightforward.

I had a talk at Re:publica about aspects of digital identity from a research point of view.

Download the pdf here: Identität im Netz.

Sunshine vs. rain.

Athens(LSAS2006.SAMT2006) vs. Leeds(VirtualGoods2006.AXMEDIS2006).

The beauty of music and technology vs. the future of -if any- DRM.

I had 2 very different experiences in a very short amount of time. Nevertheless I gave 2 talks sharing 1 hope.

#1: The Beauty of the Beast: Semantic Audio.
#2: BluetunA: Sharing your taste in music with people nearby.

I continued my previous nomadic research on phone spotting. Being in Paris I checked frequently my environment (1) while using the Metro, (2) being part of a technology-related symposium and (3) walking longer distances at urban sidewalks. Back home I checked the crowd (4) being at a live concert of Matthew Herbert. People listening to Matthew Herbert’s postmodern definition of Pop may be a very special “technophile” target group. The results are:
(1) Metro: ~ 10-15 phones with activated bluetooth
(2) Tec-symposium: >>20 (150 participants)
(3) Flaneur: ~ 3-5 (forget this in a rural place! remind: it was PARIS)
(4) Post-Pop Concert: ~ 8 (not in Paris, it was in MANNHEIM, 300 visitors)

If you are interest into this kind of research you should read e.g. the paper of O’Neill et al., Instrumenting the city: developing methods for observing and understanding the digital cityscape presented at UbiComp2006. MIT medialab is running a project on Reality Mining, too.