Categories

Workshop on Context-aware Retrieval and Recommendation

Earlier today we got the acceptance notification of our workshop on Context-aware Retrieval and Recommendation which we will be organizing at next years (2011) Intelligent User Interfaces (IUI) conference which will be held at Stanford University in February next year.

This workshop is somewhat related to our current one, the Challenge on Context-aware Movie Recommendation at this years RecSys conference. I’ve written about the challenge before here and here.

This the topic is a bit more general, not only focusing particular datasets, but rather a more general approach to the context concept.

We’re co-organizing the workshop with Matthias Böhmer from TU-Münster and Florian Michahelles from ETHZ who have been working with location based contexts, giving the workshop the needed contextual broadness as we’ve mostly been working with “abstract” contexts (mood, time, etc).

The questions we’re aiming at for the workshop are

  • What is context?
  • Is context-awareness in retrieval and recommendation necessary?
  • Which benefits come from context-aware retrieval and recommendation systems?
  • How do user interfaces handle context?
  • In what ways can context improve HCI?
  • How can we combine general- and user-centric context-aware technologies?
  • How should context affect the way information is presented?

which will hopefully attract authors to submit their work to the workshop.

Context-aware Movie Recommendation

As I’ve written earlier, at this year’s RecSys we’re organizing a challenge on context-aware movie recommendation. The challenge is part of a joint event together with the workshop on Context-Aware Recommender Systems (CARS) and I’m anticipating both events to be very interesting.

We’re done with the reviewing and currently waiting for the camera ready versions to come in, as well as onepagers describing each paper’s algorithm (these will be evaluated by an expert panel consisting of very distinguished researchers).

Organizing this challenge so far has been an extremely valuable experience. From the first abstract idea to creating dataset to answering all the different questions and inquiries from the participants. Even if it has taken up a quite large portion of my time for the last couple of months, it definitely has been worth it and I would recommend any PhD student to try to get involved in organizing at least one of these as there quite a lot of insight to gain.

We’ll have 10 papers presentations at the workshop, 5 long and 5 short, and I must say I’m very happy with the quality of the accepted papers and really looking forward to the presentations and discussions at the day of the workshop.

RecSys 2010 Workshop

It’s been official for a couple of days now, we got our workshop proposal for RecSys 2010 accepted.

The idea is to have a context-aware movie recommendation challenge. The challenge will be held as  joint event with the Context-aware Recommender Systems workshop,which was held for the first time at last years’ RecSys. We’re co-orinating the event currently, but what’s sure is that it’s a two day event, one for the workshop part (CARS), and one for the challenge part (CAMRa).

For the challenge we will be releasing two real-world datasets from two different movie recommendation websites, both containing a different set of features. The idea is for the participants to predict recommendations for different periods of time, for different users. The full set of details about the dataset, the challenge and all things relate will be posted on the challenge website. It’s still very much under construction, but will be updated on a (more or less) daily basis for the next couple of weeks.

What is context?

Recently we’ve started a new project on context aware-multimedia recommendation. The projects is still in the startup phase, nevertheless we’re collecting and going through relevant litterature.

There are huge amounts of research being conducted in the field of recommender systems, the amount of research on the topic of context-aware recommender systems is considerably smaller. One of the very few events that have been held on this topic was last year’s Workshop on Context-aware Recommender Systems (CARS-2009) at the ACM Recommender Systems 2009 (RecSys09) conference. I had the opportunity to attend the conference, but was sadly unable to attend the workshop. The papers presented at the workshop, all available on the website, all stress the importance of context-awareness.

All recommender systems are context-dependent, why nobody seems to be aware of that (or at least choses to ignore it) is a mystery to me. What does the rest of the RecSys community do, what are the motivation behind not using context? As the papers from CARS show, context-awareness does not have to be complicated, simple contextual assumptions can increase recommendation results considerably, still, looking at recsys tracks and papers at other conferences, you’re more likely not to find anything on context-awareness than the other way around.

I do believe this is not due to ignorance or disinterest from the community. The reason, I think, is that this concept has been overlooked, we’re all blinded by all the different collaborative filtering approach and their extensions. Often, when there is a paper on context-awareness, the authors will not even have considered their work to be related to the context-aware topic, instead they use terms such as hybrid, higher-order, temporal, etc. Making a simple search for any of those terms on either the ACM portal, Google Scholar or any related library you’ll find hundreds of papers, none of them actually stating the simple fact that they deal with context-awareness.

Zanzibar

Friends of mine decided to move their company to Zanzibar for three months (November 2009 – January 2010). This did of course mean that, being the good friend that I am, I had to visit.

So, on January first, not very much after I had gotten back from the New Year’s celebration I was at, me and M boarded a plane (in -15 degrees C) that first took us to Amsterdam, then Nairobi (+30) and finally, in something than more resembled a bus than a plane, we landed at the airport just outside of Stone Town on the beatiful island of Zanzibar (+35).

Standing on the tarmac it finally hit us that we were there. Next step for me was to get my visa, M had already gotten his back in Sweden. Surprisingly (or perhaps not so), security was minimal. Haven gotten my visa we went for our luggage only to find it in the hands of “helpful” cab drivers as the baggage retrieval hall was on the other side of (the unmanned) passport controls. In the taxi, M realized his passport had not been checked, and stamped by officials. So back we went, through the non-existen security controls, into the arrival hall and to the visa personnel to get a proper entry stamp. At this point, we could have just as easily walked back out to the tarmac, as back to the arrival hall.

But, back to the taxi we went. The funny thing about Zanzibar: no addresses. Or as M2 put it when we were asking for the SLX-crews address: “Zanzibar hasn’t started with addresses and other modern stuff”. We simply asked to be taken to the police station in Malindi outside of which we found M2 waiting for us and a couple of minutes later we were in the Stone Town house they were renting.

As an added bonus, we were told that the electricity (which had been cut off for the last 5-6 weeks) had still no been restored. Actually, all of Zanzibar was without electricity for some months from the end of 2009 to the beginning of 2010. No electricity of course meant no water. In 40+ degree heat. This made the Africa experience a little more genuine.

Having changed into clothes a little more suitable for 40+ degrees we explored Stone Town, went swimming and spent the evening on the roof top of the Malindi house looking at the skyline.

Day 2 we walked around in Stone Town a bit more as M had some work to do before we’d leave the city. Stone Town is a vibrant, colorful, chaotic and very nice town. In parts, especially around Malindi, it reminds of Venice, with all the narrow alleys and shops everywhere. Although it does have a completely different atmosphere.

We had not intended to stay in Stone Town particularly long, so on day 3 we (me, M, M2, and H) took a Dalladalla to Jambiani on the east coast of Zanzibar. We found a couple of bungalows to rent, left our bags and headed out to the ocean. I’ve had coffee colder than the water we swam in. We went in just as the high tide was coming in, thus all the shallow water up to the reef that had been heated up during the day was flowing in.

We spent the next couple of days hopping between villages on the east coast, mingling with the locals and tourists, where there were any. Heading north we spent nights in Jambiani, Paje, Bweju, Matemwe and Nungwi, heading back to Stone Town once as the only (currently working) ATM’s on the island were there.

Whilst in Zanzibar, eating anything else than seafood is just a waste. However, when uttering a comment like “The food’s great, could have been a bit warmer though” when having a grilled tuna you should rethink if you really want to finish your food. As M found out, it wasn’t. So while me, M2, and H where having the time of our lives in Nungwi and the Kendwa Rocks party, M was having the most non-cooperative stomach of his life. In hindsight, he probably felt better than we did the day after the party as we spent amounts of money on Konyagi drinks comparable to what a night out in Sweden costs.

Next, we were heading for a safari in Mikumi National Park, which meant we had to get ourselves to the mainland, i.e. Dar es-Salaam, but, knowing that Stone Town did not offer neither electricity nor water, we opted for staying one last night in our bungalow (which from time to time had diesel powered electricity and water in the pipes). This in turn lead to us missing our desired overning ferry to Dar, which lead to us taking a ferry at ungodly 5 in the morning the following day. Surprisingly, security checks before boarding the boats were more rigorous than those at Zanzibar Airport (i.e. people actually asked to see our papers).

Post ferry trip and breakfast in Dar, we had another 5 hours to Mikumi in our private safari minivan whith a driver/guide. Mainland Tanzania is a bit different from the Hakuna Matata mentality everyone on Zanzibar has. Still beautiful.

After a couple of hours of driving, we passed into Mikumi and the animals started to appear, elephants, zebras, giraffes and baboons crossing the street as if we weren’t there, quite a sight. And this wasn’t even the cotrnolled part of the park. Another hour or so, we drove off of the main road and into the maintained part of the park. We’d booked an all-inclusive bungalow at the Mikumi Wild Life Camp lodge, which turned out to be fantastic. From our porch we were looking at African buffalos, baboons, elephants and giraffes grazing close by.

The two following days, we drove around in the park looking for animals. This turned out to be quite easy as on our first day we saw all four of the big five they have in the park, i.e. lions, leopards, elephants, and african buffalos. Since Mikumi does not have any rhinos we did not get to see any. More than that we saw wildebeests, zebras, hippos, etc. Truly fantastic experience, would recommend to everyone. Day two we started out early morning and went to the farther parts of the park, needless to say, the scenes were beautiful. Sometimes we’d see big herds of giraffes (40-50) walking past graciously as if they’d walk in slow motion. Five minutes later we’d drive by a herd of impalas or elands.

All in all, the safari was an experience that suspassed the expectations I had for it.

Zanzibar day 1 Zanzibar day 2 Zanzibar day 3 Zanzibar day 4 Zanzibar day 5 Zanzibar day 6






Zanzibar day 7 Zanzibar day 8 Zanzibar day 9 Last night in Z Mikumi day 1 Mikumi day 2
Get Adobe Flash playerPlugin by wpburn.com wordpress themes