Archive for October, 2008

Pervasive Social Computing Seminar

Wednesday, October 29th, 2008

I gave a seminar yesterday in the Mobisys weekly seminars at UCL and I would be pleased to share my slides with you. It was about Pervasive Social Computing, or how to support social networks in pervasive computing environments in order to enable social interactivity between mobile users.

Any feedback is welcome and if you have any question I will pleased to answer.

Pervasive Social Computing

Mobisys Seminar 28/10/08

From: poline_sonia, 7 minutes ago

Mobisys Seminar 28/10/08

View SlideShare presentation or Upload your own. (tags: pervasive social)

SlideShare Link

Papers at RecSys 2008

Monday, October 27th, 2008

A lot of interesting papers have been discussed during RecSys conference. Here are a few list of interesting ones.


Tutorials at RecSys 2008

Friday, October 24th, 2008

Yesterday was the first day of RecSys 2008, and was dedicated to three very interesting tutorials:

1. Robust Recommender Systems. Robin Burke introduced the wide range of attacks that typical collaborative filtering algorithms are vulnerable to; scenarios that arise when people attempt to force, rather than express, opinions. An attack was strictly defined as a set of profiles intending to obtain excessive influence on others, which can be aimed at pushing (making recommendation more likely) or nuking (i.e. recommendation less likely) items. His talk was an interesting blend of attack strategies, knowledge that attackers need to have, and a high-level description of approaches aiming at preventing or fixing the system when attacked. Of course, there are strong overlaps between this work and work in other areas (p2p trust, adversarial information retrieval, search engine spam..); I particularly like this area as pushes the point that recommender systems are about people/dynamic datasets, and not just prediction.

2. Recent Progress in Collaborative Filtering. Yehuda Koren (who has recently moved from AT&T to Yahoo! Research) gave a tutorial about the leading approaches in the Netflix prize competition. The techniques he described blend matrix factorisation and neighbourhood models, and include a number of other factors (such as user biases and time) that result in techniques that have multiple-billions of parameters (and the resulting ranking of team BellKor in the competition). His work is remarkable and worthy of the progress prizes he has been awarded thus far. He also explored alternative techniques of evaluating recommender systems, explaining his take on evaluating top-N recommendation lists.

3. Context-Aware Recommendations. Gedas Adomavicius and Alex Tuzhilin introduced their work on incorporating context into recommender systems, including pre-, post-, and hybrid-filtering of recommendation algorithm results based on user context. A running example that was repeated throughout the tutorial was going to the theatre with your girlfriend on the weekend: if you always watch comedy, then your recommendations can be filtered to match what you did in previous instances of the same context (i.e. you can be recommended comedy). They have done a lot of cool stuff on multi-dimensional recommenders, extending the common rating scales into cubes of ratings, and stressed more than once that this is virgin territory. Their work is also impressive, but raised a few questions. For example, should context be described by a well-enumerated taxonomy? Moreover, if you always watch comedy at the theatre with your girlfriend on weekends, then why should you need a recommender system in the first place (especially a collaborative one- what happened to serendipity or diversity)? They have a number of papers that are worth reading before trying to answer these questions!

Selected start-ups at Mobile 2.0

Monday, October 20th, 2008

[Cool video on the aka-aki website]

  • aka-aki (Germany) – focuses on Proximity Networking, as in mobile social networking with Bluetooth-sensing capabilities.
  • Dial2Do (Ireland) – Dial2Do lets you do common tasks by just calling a number and speaking.
  • Nimbuzz (Netherlands) – Mobile IM and Text Message Service.
  • Rummble (UK) – a location based discovery tool and social search platform.
  • Seesmic (USA) – a video service mimicking and aggregating your favorite web products.
  • Zipipop (Finland) – a start-up that is developing Zipiko, a mobile service for sorting your social life on the go.
  • Wubud (UK) – Wubud allows you to take your social network in your back pocket everywhere you go.

MobiSys: UCL and Birkbeck Research

Tuesday, October 14th, 2008

The mobisys seminar series had its first presentation of the (academic) year today. We decided to slightly change the format. Instead of having a single speaker present for 45 minutes, we would have short research-pitches; that way, the various old and new group members could introduce themselves and their work to each other.  We were also very happy to have Birkbeck researchers present, since a lot of their work has strong overlaps with what is going on in the mobisys group.

By last night, I had about 19 speakers ready to present: way too many! So I decided to break the minute-madness presentations into two sessions. The first was today, centred around the theme of pervasive computing. The second session, broadly categorised as social computing, will be in the coming weeks.

Each speaker had 4 minutes to introduce and overview their work (and were promptly interrupted by my annoying alarm clock sound if they went over): the slides are below. The whole idea of these sessions is to highlight how many people are working on similar problems and to foster discussion (perhaps in an effort to combat “phd depression?”)- unfortunately we did not have much time for the latter today, and I hope that we will in future events.

We rounded up the session with an overview of current research projects going on in the respective departments by Steve Hailes and George Roussos.

MobiSys Group Presentation
View SlideShare presentation or Upload your own. (tags: networks sensors)

The topology of dark networks

Monday, October 13th, 2008

In the current issue of CACM, there is a very interesting article on “The topology of dark networks“. This is the first study I see of the topological structure of dark networks, that is, networks hidden from view and yet with a potentially devastating impact on society. (more…)

Social network collaborative filtering

Monday, October 13th, 2008

Interestingly, “This paper demonstrates that “social network collaborative filtering” (SNCF), wherein user-selected like-minded alters are used to make predictions, can rival traditional user-to-user collaborative filtering (CF) in predictive accuracy. “

Informal Networks from Movements

Saturday, October 11th, 2008

I’m looking for  academics/practionaires interested in the following research  and for companies willing to give me access to relevant data ;-) My website. My email:

To facilitate the flow of information or to promote cultural change, companies often focus on formal organizational charts. Alas, those charts do not reveal the invisible networks that employees use to get things done. One way of identifying invisible networks is to keep track of how staff moves and, more specifically, “who talks to whom”. This can be done automatically by programming mobile phones to keep track of their owners’ location and proximity to other people.  By aggregating data on those phones, one  then produces “informal networks” and can harness them to:

  • Make change stick by identifying influential employees. If management can persuade influential people to be proponent of a big change, then the change is far more likely to succeed.
  • Focus on points in an informal network where relationships should be expanded or reduced. Imagine that, from its informal network, a company finds out that old personnel are extraordinary well-connected and central to collaboration, while many newcomers are stuck on the periphery. To fix the situation, the company may launch “mentoring programs” in which old personnel (central to collaboration) mentor newcomers. That is just one example of how the study of informal networks can break down barriers that hinder collaboration.
  • Measure the effectiveness of major initiatives. Informal networks are also a way for companies to measure the impact of changes. By measuring key network metrics (such as density, cohesion, and centrality) before and after a change, companies can asses whether the change has been a positive one. For example, if an informal network shows higher density after introducing a mentoring program, then the program has been a positive change in that it has reduced the number of steps for any individual to get in touch with a colleague.

KDD-08 video lectures

Tuesday, October 7th, 2008

KDD-08 videos include:


Creating Social Networks Will No Longer Be a Luxury

Thursday, October 2nd, 2008

In early August, the specialized publication Infoworld gave an award to Elgg, a platform for creating social networks, in the category of best open-source software for collaboration. At that time, hardly anyone other than specialists even knew what it was. That remains the case but things could change in the future. Specialized blogs have celebrated the recent release of version 1.0 of Elgg ( as evidence that open-source software is ready to have its voice be heard in the world of social network creation platforms. In short, Elgg provides a content-management solution that lets anyone create and manage their own social network. The fact that it uses the general public license means it can avoid the limitations of proprietary social networking sites, such as YouTube, Facebook and MySpace, for instance, which make their own rules for admission and content.


Analysis of Social Networks using Ucinet and Siena @ Ox

Thursday, October 2nd, 2008

A one-week workshop on Analysis of Social Networks using Ucinet and Siena, taught by Martin Everett and Tom Snijders, will be held in Oxford, December 15-19, 2008. Registration will be opened soon at the website of the Oxford Spring School. It will be possible to register independently for the Ucinet part (Monday-Tuesday, taught by Martin Everett) and for the Siena part (Wednesday-Friday, taught by Tom Snijders). The program is available at

Machine Learning Applications to Music

Wednesday, October 1st, 2008

Louis gave a very interesting talk about his research on applying machine learning to music. Interestingly, among other things, he discussed two issues of music retrievial:

1) Does the use tags improve retrieval algorithms? In my opinion, the answer is a qualified yes. Case in point: this month at Recsys, Licia and Valentina will present an effective  way of retrieving items (e.g., music files) . Their technique exploits two types of similarity  (item similarity and user similarity), both of which are computed only from (user-specified) tags.  Check section 3 of this paper (short description) and their RecSys paper (complete description).

2) Exsting retrievial algorithms learn your music taste and assume  that it does not change over time. What if you change your taste? That’s a question Neal will answer at, again, RecSys this month. Check his paper.