personalization in the context of relevance-based visualization
TRANSCRIPT
Personalization in the Context of Relevance-Based VisualizationPeter Brusilovsky Jae-Wook AhnDenis ParraKatrien Verbert
University of PittsburghPUC ChileIBMUniversity of Leuven
Outline• Problem• History
– InfoCrystall, VIBE, TileBars
• A quest to Adaptive VIBE– KS-VIBE, QuizVIBE, Adaptive VIBE
• Combining social and adaptive relevance prospects in Conference Navigator– TalkExplorer– SetFusion– Intersection Explorer
2
Why Relevance Visualization?• Items might be relevant for a query for
different reasons– I.e., match different keywords
• Ranked list fuses and hides different relevance aspects– Not transparent, not controllable
• Focus on relevant items while keeping relevance dimensions recognizable?
3
4
InfoCrystal (Spoerri 1993)
From Venn Diagram to IC
6
More "InfoCrystals"
7
VIBE (Korfhage, 1991)
8
TileBars (Hearst, 1995)
9
Towards Adaptive VIBE
• Adaptive nodes– Social: KS-VIBE
– Knowledge-based: QuizVIBE
• Adaptive topology– Keyword-based: Adaptive VIBE– Concept-based: Adaptive VIBE+NE
10
KS-VIBE (Ahn et al. 2006)
11
12Ahn, J.-w., Farzan, R., and Brusilovsky, P. (2006) A two-level adaptive visualization for information access to open-corpus educational resources. In: Proc of Workshop on the Social Navigation and Community-Based Adaptation Technologies at the 4th International Conference on Adaptive Hypermedia and Adaptive Web-Based Systems, Dublin, Ireland, June 20th, 2006, 497-505.
QuizVIBE (2006, Ahn et al.)
Ahn, J.-w., Brusilovsky, P., and Sosnovsky, S. (2006) QuizVIBE: Accessing Educational Objects with Adaptive Relevance-Based Visualization. In: Proc. of World Conference on E-Learning, E-Learn 2006, Honolulu, HI, USA, October 13-17, 2006, AACE, pp. 2707-2714.
• User control in personalized Filtering in ROSETTA project– Users choose to ranks search results according to user profile, query, or both
• α * user profile + (1–α) * user query (α = 0, 0.5, 1)
• Users wanted more control
The motivation for Adaptive VIBE
PersonalizedIR system
Ranked list : User Profile
Ranked List : User
Query
FusedSearc
hResult
Adaptive VIBE Idea: Query and UM for Document Space Separation
15https://www.youtube.com/watch?v=Yt1fMEFlLVA&index=2&list=PLyCV9FE42dl7JG_i7m_kvwuYRpfwwJ4iY
VIBE based query-profile fusion
User Profile Terms
Query Terms
Documents
Mixing user profile and query terms as VIBE POI
VIBE based fusion
Document-queryrelevance
Profile-queryrelevance
VIBE based fusion (cont’d)
More aboutN. Koreannuclear weaponMore about
GenericNuclear weapon
VIBE POI presets
“Circular” preset
“Parallel” preset
• User profile is added on the same playfield as user query
• Topology is adaptive• Mediate between profile (green POI) and query
(red POI) terms• Browse documents free with control on profile
and query terms
Adaptive topology in VIBE
Adaptive VIBE+NE
22
Some Study Results• A sequence of user studies
– Search vs. VIBE vs. VIBE+NE
• Search -> VIBE -> VIBE+NE offers:– Better visual separation of relevant documents
(system)– Supports better opening relevant documents (user)
• VIBE+NE supports more meanigful interaction– No degradation found even with active visual UM
manipulation– While over performance retained or increased
Ahn, J., Brusilovsky, P., and Han, S. (2015) Personalized Search: Reconsidering the Value of Open User Models. In: Proceedings of Proceedings of the 20th International Conference on Intelligent User Interfaces, Atlanta, Georgia, USA, March 29-April 1, 2015, ACM, pp. 202-212
Fusing Multiple Relevance Prospects in Conference Navigator
• Conference Navigator• TalkExplorer• SetFusion• Intersection Explorer
24
Relevance in Conference Navigator
• Classic content-based relevance prospects– Items that has a specific keyword
• Social relevance prospects– Items bookmarked by a specific user
• Tag relevance prospects (content+community)– Items tagged by a specific tag
• Personal relevance prospects– Several different recommender engines– Each engine offer one relevance prospect
25Brusilovsky, P., Oh, J. S., López, C., Parra, D., and Jeng, W. (2017) Linking information and people in a social system for academic conferences. New Review of Hypermedia and Multimedia.
Social Prospect: Details of a talk in CN3
26
Social prospect: User schedules
27
Tag relevance: Tag page in CN3
28
Multiple Recommender Engines
29
Challenge• Idea: Fuse traditional, social, personal
relevance prospects• Approach: fuse several relevance lists
– Several recommendation approaches– Items bookmarked by valuable users– Items tagged by interesting tags
• Challenge: How to make it transparent and keep users in control– i.e., allowing to focus on a subset of
relevance prospects30
John O'Donovan, Barry Smyth, Brynjar Gretarsson, Svetlin Bostandjiev, and Tobias Höllerer. 2008. PeerChooser: visual interactive recommendation. CHI '08
Related work: PeerChooser
31
Related work: TasteWeights
32Bos
tand
jiev,
S.,
O'D
onov
an, J
. and
Höl
lere
r, T.
Tas
teW
eigh
ts: a
vis
ual i
nter
activ
e hy
brid
re
com
men
der s
yste
m. I
n P
roce
edin
gs o
f the
six
th A
CM
con
fere
nce
on R
ecom
men
der
syst
ems
(Rec
Sys
'12)
. AC
M, N
ew Y
ork,
NY
, US
A (2
012)
, 35-
42.
The Approach• Using set relevance visualization
– One dimension of relevance = one set
• Agent metaphor to mix user- tag- and engine-based relevance– recommender systems are shown as agents
– in parallel to real users collecting talks
– tags are also perceived as agents collecting talks
– users can interrelate entities to find items
33
TalkExplorer
• Recommendation engines are shown as agents in parallel to users and tags• Uses Aduna clustermap library: http://www.aduna-software.com/
34
Entity selection
35
Canvas area
36
TalkExplorer
37
overview selected talks
Interrelations agents and users
38
Evaluation• Setup
– supervised user study– 21 participants at UMAP 2012 and ACM Hypertext 2012
conferences
• Results– The more aspects of relevance are fused, the more effective it
is for getting to relevant items. Especially effective are fusions across relevance dimensions
– The more relevance prospects are merged, the better is the yield, the easier is to find good items
– Dimensions of relevance are not equal– ADUNA approach is challenging for beyond fusion of 3 aspects 39
Verbert, K., Parra-Santander, D., and Brusilovsky, P. (2016) Agents Vs. Users: Visual Recommendation of Research Talks with Multiple Dimension of Relevance. ACM Transactions on Interactive Intelligent Systems 6 (2), Article No. 11
SetFusion• Using set relevance visualization
in the familiar Venn diagram form– One recommendation source = one
set
• Allow controlled ranking fusion
• Combine ranking with annotation showing source(s) of recommendation
40
Brief Results of Two Studies• SetFusion provides strong engaging effect
– Number of engaged users, bookmarked talks, explored talks doubled
– The effect is larger in UMAP “natural” settings
• SetFusion allows more efficient work– Increases yield of bookmarks in relation to
overhead actions
• But only 3 dimensions of relevance!47
Intersection Explorer• Based on ideas of
Talk Explorer• New approach for
scalable multi-set visualization
• Try it yourself at IUI2017 Conference Navigator
48
Verbert, K., Seipp, K., He, C., Parra, D., Wongchokprasitti, C., and Brusilovsky, P. (2016) Scalable Exploration of Relevance Prospects to Support Decision Making. In: P. Brusilovsky, et al. (eds.) Proceedings of Joint Workshop on Interfaces and Human Decision Making for Recommender Systems at 10th ACM Conference on Recommender Systems, Boston, MA, USA, September 16, 2016, pp. 28-35, also available at http://ceur-ws.org/Vol-1679/paper5.pdf.
Intersection Explorer at IUI2017
49http://halley.exp.sis.pitt.edu/cn3/iestudy3.php?conferenceID=148
Readings• Ahn, J.-w., Farzan, R., and Brusilovsky, P. (2006) A two-level adaptive visualization for information access to open-corpus educational
resources. Proceedings of Workshop on the Social Navigation and Community-Based Adaptation Technologies at the 4th International Conference on Adaptive Hypermedia and Adaptive Web-Based Systems, Dublin, Ireland, June 20th, 2006, pp. 497-505, also available at http://www.sis.pitt.edu/%7epaws/SNC_BAT06/crc/ahn.pdf.
• Ahn, J.-w., Brusilovsky, P., and Sosnovsky, S. (2006) QuizVIBE: Accessing Educational Objects with Adaptive Relevance-Based Visualization. In: T. C. Reeves and S. F. Yamashita (eds.) Proceedings of World Conference on E-Learning, E-Learn 2006, Honolulu, HI, USA, October 13-17, 2006, AACE, pp. 2707-2714.
• Ahn, J. and Brusilovsky, P. (2013) Adaptive visualization for exploratory information retrieval. Information Processing and Management 49 (5), 1139–1164.
• Ahn, J., Brusilovsky, P., and Han, S. (2015) Personalized Search: Reconsidering the Value of Open User Models. In: Proceedings of Proceedings of the 20th International Conference on Intelligent User Interfaces, Atlanta, Georgia, USA, March 29-April 1, 2015, ACM, pp. 202-212
• Verbert, K., Parra-Santander, D., and Brusilovsky, P. (2016) Agents Vs. Users: Visual Recommendation of Research Talks with Multiple Dimension of Relevance. ACM Transactions on Interactive Intelligent Systems 6 (2), Article No. 11
• Parra, D. and Brusilovsky, P. (2015) User-controllable personalization: A case study with SetFusion. International Journal of Human-Computer Studies 78, 43–67.
• Verbert, K., Seipp, K., He, C., Parra, D., Wongchokprasitti, C., and Brusilovsky, P. (2016) Scalable Exploration of Relevance Prospects to Support Decision Making. In: Proceedings of Joint Workshop on Interfaces and Human Decision Making for Recommender Systems at 10th ACM Conference on Recommender Systems, pp. 28-35, also available at http://ceur-ws.org/Vol-1679/paper5.pdf.
• Verbert, K., Parra-Santander, D., Brusilovsky, P., Cardoso, B., and Wongchokprasitti, C. (2017) Supporting Conference Attendees with Visual Decision Making Interfaces. In: Companion of the 22nd International Conference on Intelligent User Interfaces (IUI '17), Limassol, Cyprus, ACM.
50
