ICT role in 21st century education and it's challenges.
Self-Adaptation of Online Recommender Systems via Feed-Forward Controllers
1. Self-Adaptation of
Online Recommender Systems
via Feed-Forward Controllers
Licia Capra
University College London
Workshop on Self-Awareness in Computing
June 27th, 2010
11. DOMAIN PROBLEM ANALYSIS
Method of Assessment
Time
User
Tag
Item
Period
T1
6%
17%
2%
T2
6%
1%
11%
T3
6%
2%
3%
Ti
1500
tests
Training
Ti(1)
Training
Ti(2)
12. DOMAIN PROBLEM ANALYSIS
Method of Assessment
Time
User
Tag
Item
Accuracy
Loss
Accuracy
Loss
Accuracy
Loss
Period
(25-‐75)
(50-‐50)
(75-‐25)
T1
6%
17%
2%
24%
32%
45%
T2
6%
1%
11%
12%
20%
23%
T3
6%
2%
3%
3%
10%
14%
Ti
1500
tests
Training
Ti(1)
Training
Ti(2)
13. DOMAIN PROBLEM ANALYSIS
Method of Assessment
Time
User
Tag
Item
Accuracy
Loss
Accuracy
Loss
Accuracy
Loss
Period
(25-‐75)
(50-‐50)
(75-‐25)
T1
6%
17%
2%
24%
32%
45%
T2
6%
1%
11%
12%
20%
23%
T3
6%
2%
3%
3%
10%
14%
Ti
1500
tests
Training
Ti(1)
Training
Ti(2)
15. DYNAMIC UPDATE METHODOLOGY
Recommender Systems as Self-Adaptive
Systems
x
[users,items,tags]
[Recommender
] y
System
[recommendaFon
list]
u
[update
frequency]
Feed-‐Back
16. DYNAMIC UPDATE METHODOLOGY
Feed-Forward Controller for Dynamic
Updating of Recommender Systems
x
[users,items,tags]
[Recommender
] y
[recommendaFon
u
System
list]
[update
frequency]
Feed-‐Forward
22. CONCLUSIONS
Accuracy vs Cost Tradeoff may
lead to suboptimal choices
Recommender Systems as Self-
Adaptive Systems
Feed-Forward Control Theory for
Unobservable Outputs
23. … & FUTURE WORK
Automation of Empirical Evaluation
Beyond Accuracy and Cost (diversity,
surprise, serendipity)
24. On self-adaptation
• B.H. Cheng, et al. Software Engineering for Self-Adaptive Systems: A Research Roadmap.
In Software Engineering for Self-Adaptive Systems, pages 1-26, 2009. Springer-Verlag
• Y. Brun, et al. Engineering Self-Adaptive Systems through Feedback Loops. In Software
Engineering for Self-Adaptive Systems, pages 48-70, 2009. Springer-Verlag.
On recommender-systems
• J. Herlocker, et al. An Algorithmic Framework for Performing Collaborative Filtering. In
Proc. of the 22nd Annual International Conference on Research and Development in
Information Retrieval, pages 230-237, New York, NY, USA, 1999. ACM.
• G. Adomavicius and A. Tuzhilin. Context-Aware Recommender Systems. In Proc. of the ACM
Conference on Recommender Systems, 2008.
From my group
• V. Zanardi and L. Capra. Social Ranking: Uncovering Relevant Content using Tag-based
Recommender Systems. In Proc. of the Conference on Recommender Systems, pages
51-58, 2008. ACM.
• V. Zanardi and L. Capra. "Dynamic Updating of Online Recommender Systems via Feed-
Forward Controllers". In 6th Intl. Symposium on Software Engineering for Adaptive and
Self-Managing Systems (SEAMS 2011). Waikiki, Honolulu, Hawaii, USA. May 2011
THANK YOU!