After my last blog post, I was contacted by a journalist who wanted to discuss the effects of the Netflix prize. It seems that now that the competition is winding to an end, one of the real questions that emerges is whether it was worth it. Below, I’m pasting part of my side of the dialogue; other blogs are posting similar discussions, and I’m curious as to what any of you fellow researchers may have to say.
Archive for June, 2009
Cybercrime is rapidly spreading on Facebook as fraudsters prey on users who think the world’s top social networking site is a safe haven on the Internet…
A combined team of the leaders of the Netflix prize (teams BellKor, Pragmatic Theory & BigChaos – together know as BellKor’s Pragmatic Chaos) has recently submitted an entry to the Netflix prize which surpasses Cinematch by 10.05%.
According to the competition rules, an announcement should follow soon saying that 30 days remain for competing teams to submit their final entries. A quick look at the leaderboard, however, shows that 5 of the top 10 leaders are members of one (or combined) of the groups who look like they are soon to win $1,000,000.
Interestingly, this submission appears just as researchers were discussing whether a 10% improvement was at all possible, while others rallied together to try and surpass the magic barrier. Also, funnily enough, I tweeted about the potential for this to happen a while back – and nearly got the team names right.
I hope that the competition ends soon; and I also wonder:
- Will the winning algorithm ever be used by Netflix?
- Now that the accuracy goal has been reached, what will researchers care about? (how many of them will stop caring?)
Recommender system research has definitely benefited from the competition (thanks to Netflix). But we all know that there is more to a recommendation that an intelligent, accurate forecast of what star rating I’m going to give something. Many things come to mind – but for now, congratulations to team BellKor’s Pragmatic Chaos.
Update: The notification email has been sent out; participating teams have until July 26, 2009 18:42:37 UTC to send in their submissions!
1) “In an amusingly titled WWW 2009 paper, “Computers and iPhones and Mobile Phones, oh my!” , a quartet of Googlers offer some thoughts on where mobile search may be going. In particular, based on log analysis of iPhone searches, they claim search on mobile devices is not likely to differ from normal web search once people upgrade to the latest phones. They go on to predict that an important future feature for mobile search will be providing history and personalization synchronized across all of a person’s computers and mobile devices.” (here)
2) Enhancing Mobile Recommender Systems with Activity Inference by Kurt Partridge and Bob Price “ This paper describes how to infer a user’s high-level activity automatically to improve recommendations. Activity is determined by interpreting a combination of current sensor data, models generated from historical sensor data, and priors from a large time-use study. We present an initial user study that shows an increase in prediction accuracy from 62% to over 77%, and discuss the challenges of integrating activity representations into a user model.
3) Potential for personalization. To appear in ACM:Transaction on Computer Human Interaction 09
An interesting conference at CASA (UCL) of few months ago. Few titles of interest:
- GMapCreator and MapTube: Web-Based Mapping for Sharing and Visualising Geographic Information (pdf)
- Public Engagement: The London Profiler, Public Profiler and the E-Society Classification (pdf)
- Mapping Peoples Mood: Crowdsourcing Spatial Surveys (pdf)
- Understanding Crowdsourced Geographical Information: An Analysis of OpenStreetMap(slideshare)
- Cellular Census: Explorations in Urban Data Collection (pdf)
Ushahidi (blog) is an open source platform for collecting, visualizing, and distributing information related to a crisis or ongoing public problem, such as swine flu, election fraud, and political violence:
- 4-minute TED talk
- Explaining Swift River – video on developing more sophisticated news filtering
- API – if you like to play with data
Plus, there is also the OMC – it is all about open source mobile phone software, with a focus on humanitarian needs.
- FlexRecs: Expressing and Combining Flexible Recommendations
- Building Community-Centric Information Exploration Applications on Social Content Sites
- ELMR: Lightweight Mobile Health Records
- MobileMiner: A Real World Case Study of Data Mining in Mobile Communication
- CourseRank: A Social System for Course Planning
- DataLens: Making a Good First Impression (could this be applied to recsys?)
- Keynote Talk: Transforming Data Access Through Public Visualization (IBM)
The primary goal of this project is to explore novel and effective ways to search geo-spatial data and leverage multi-lingual technologies within maps.