Preferencebased search with Suggestions

1
Preference-based search with Suggestions

• Paolo Viappiani
• Artificial Intelligence Lab (LIA)
• Ecole Polytechnique Federale Lausanne (EPFL)

2
Overview

• Preference-based search
• Example-critiquing with Suggestions
• Experimental results
• Scalability and Implementation
• Adaptive strategies

3
Traditional commerce
Electronic commerce

• Human interactions
• General outlook of possibilities
• Shop assistants
• ?
• Increase customers awareness
• Serendipitous discoveries
• Require long time and physical displacement

• Electronic commerce
• Human-computer interactions
• User interfaces
• ?
• Saved time
• Fixed interaction
• No third dimension

4
User knowledge
Database query
Mixed-initiative systems
Implicit recommender systems
User involvement
5
Form-filling

• Example actual scenario with travel website
  (July 5th, 2006)
• User wants to travel from Geneva to Dublin
• Return flight
• Preferences
• Outbound flight, arrive by 5pm
• Inbound flight, arrive by 3pm
• (Cheapest)

6
(No Transcript)
7
Example-based tools

• Several proposed systems
• Findme (Burke et al. 97)
• Smartclient (PuFaltings00)
• Expertclerk (Shimazu01)
• User expresses the preferences as critiques on
  displayed examples
• Feedback directs the next search cycle
• Motivation users preferences are often
  constructed when considering specific examples
• (Payne et al. 93 Slovic95)

Initial preference
The user critiques the examples
The system shows k examples
The user picks the final choice
8
Prominence effect
9
Anchoring effect

• Users are biased to what is shown to them
  (Tversky1974)
• Example
• Three laptops that all weigh around 3-4 kg
• The user might never consider a lighter model
• Metaphor local optimum
• When all options look similar, motivation to
  state additional preference is low

10
Example-critiquing with Suggestions
11
Suggestions

• Others have also recognized the need to help
  users consider potentially neglected attributes
• Show extreme examples (Linden97)
• Show diverse examples (Smyth McGinty03,
  McSherry02)
• Problems
• Extremes might be unrealistic
• Too many to choose from
• Diversity does not mean interesting
• Might introduce a even worse bias

12
Model-based Suggestions

• We show suggestions
• Based on the current preference model and
  possible extensions
• Optimally stimulate preference expression
• Metaphor of Active Learning

13
The lookahead principle

• Suggestions should not be optimal under the
  current preference model, but should provide a
  high likelihood of optimality when an additional
  preference is added
• Implemented with Pareto-optimality
• Avoid sensitivity to numerical errors
• ? Display options that have high probability of
  becoming Pareto-optimal

14
The model
Discrete domain
Continuous domains
I prefer attribute less than ?
penalty
?
Attribute values

• Preferences are order relations
• Distribution over possible missing preferences
  for suggestions
• Effective suggestions even with uniform
  distribution (user studies)

15
Ogt

• Probability of optimality Popt
• Pareto dominance ? partial order
• New preference ? option has to be better than all
  dominators

o better than o
For all dominators Ogt
To become optimal, the black option has to be
better than all dominators w.r.t a new preference
Integrate over possible preferences
H if c(?,o2) gt c(?,o1) then 1 else 0
16
Example
Preferences PRICElt500, DIST_UNIVlt10
CANDIDATES
SUGGESTIONS
17
User has to select a flight among a set of
options. 4 attributes fare, arrival time,
departure airport, airline.

• Initial preference lowest price
• O1 is the highest ranked
• Other (hidden) preferences
• Arrive by 1200
• Leave from City airport
• gt O4 is the best compromise (TARGET option)

18
O1 is the best option w.r.t. the current model O2
and O3 are the best suggestions to stimulate
preference over the other attributes Extreme/diver
sity will select O5 or O6 O4 the real best
option, became highest ranked once the hidden
preferences are considered
Model based suggestion strategy ranks the options
according to P, the likelihood to become optimal
when new preferences are stated
19
Evaluation with Simulations
20
User Studies

• Two versions of the tools
• C showing only Candidates at each interaction
• CS showing Candidates and Suggestions
• Main objectives
• Decision accuracy the percentage of times the
  user succeeded in finding the target
• User effort the task time a user takes to make
  choice
• Between / Within groups experiments

21
Hypothesis

• Model-based suggestions ? more complete
  preference models
• Model-based suggestions ? more accurate decisions
• More complete preference models ? more accurate
  decisions (12)
• Question/answering ? incorrect preference models
  and inaccurate decisions
• Most preferences are stated when the user sees an
  additional opportunity,
• i.e. most critiques are positive reactions to the
  displayed options

22
(No Transcript)
23
H1. Model-based suggestions leads to more
complete preference models

• Interface CS (3 candidates, 3 suggestions) vs.
  interface C
• Users of the CS interface stated more
  preferences
• Incremental addition of preferences during the
  use
• Interface CS vs. Interface C showing 6
  candidates
• When suggestions are present, users state more
  preferences (5.8 versus 4.8)

24
Online user study

• FlatFinder was hosted on the laboratory server
  for one year
• Collected the results in log files, several
  hundreds users
• Filtered out incomplete interaction

25
Decision accuracy user effort
Decision accuracy
H2 Model-based suggestions leads to more accurate
decisions H4 Question/answering leads to
inaccurate decisions
26
H3. More complete preference models lead to more
accurate decisions

• Users that found their target stated more
  preferences (5.57) than users who did not (4.88)
• More preference revisions ? higher decision
  accuracy
• People who found their targets made more
  revisions
• 6.9 as opposed to 4.5, statistically significant
  (p0.0439)
• Mediation analysis
• Increase of accuracy not only because the
  preference model is more complete

Within-group experiments difference in the
number of preferences in the two use of the
interface and difference in accuracy.
27
H4. Question/answering leads to incorrect
preferences and inaccurate decisions

• Form-filling is not effective
• Only 25 decision accuracy
• Incorrect means objectives
• Average of 7.5 preferences
• Stated before having considered any of the
  available options
• Even after revisions, preferences were not
  retracted
• Example-critiquing
• Users begin with average of only 2.7 preferences
• Added average of 2.6 to reach 5.3
• 50 preferences were added during interaction
• Results suggest that volunteered preferences are
  more accurate

28
Other observations

• Price
• When suggestions are present, users were willing
  to pay 7 more
• Within group experiments
• Majority of times the user switched choice, the
  last choice was more expensive
• Weights (attribute importance levels)
• No correlation between number of changes and
  interface type

29
Subjective evaluation

• We asked questions at the end of the interaction
• Example-critiquing is easier to use, enjoyable
  and make the user more convinced about their
  choice than form-filling
• Results confirmed in the within group experiments
• Suggestions do not make example-critiquing
  significantly easier to use or more enjoyable

30
Preferred interface
Within group experiments
Example-critiquing with Suggestions versus
form-filling
Example-critiquing with/without suggestions
31
H5. Most preferences are stated when the user
sees an additional opportunity
32
Scalability and Implementation
33
Large databases

• Relaxation of the look-ahead principle
• Goal overcome quadratic complexity of matching
  each options with its dominators.
• Select suggestions from the top-k options
  top-suggestions
• Replace Pareto-optimality with Utility-dominance
  utilitarian-suggestions
• Assume dominating options are a fixed number at
  the top top-escape suggestions

34
Top suggestions

• Suggestions are not evenly distributed, but they
  are often at the top.
• The fraction required to guarantee that
  respectively 50 and 80 of the suggestions are
  in the top positions.
• O(n1.2) for the given k

35
Top-escape suggestions

• Consider few top options
• Maximize the probability Pesc of breaking the
  dominance with top options.
• Advantage constant number of comparisons for
  each option
• Problem the suggestions might not be good enough
  in existing preference
• High Popt ? high Pesc
• But not always the contrary

36
Configurable problems

• Configurable products consisting of many parts
• Constraint Satisfaction Problems (CSPs)
• The constraints represent the feasible
  assignments
• Preferences are soft constraints
• We need to generate
• Candidate solutions
• Branch and Bound techniques
• Suggestions
• Top-escape strategy
• ? Generate suggestions solving a single
  optimization problem

37
Preference Distribution
Unknown preferences Distribution of Soft
Constraints
The preferences that are known Soft constraints
Feasible configurations Hard constraints
Preference-based search in configurable catalogs
38
Top-escape for CSPs
Top options
SoftCSP CSP preferences
BB
stop
Aux WeightedCSP Variables, HC same
SoftConstraints For each variable vi d ?
prob(dgt vi(stop) ) in pi
Suggestions (top-escape)
39
Relation between Pesc and Popt
3 cases in which s1 escape s0
Current situation
New Preference
s0
s
s1
s1
s
s0
s1
s
s
s1
s0
s0
Dominated
P.O
P.O
40
Relation between Pesc and Popt
Top-escape suggestions have high probability Pesc
of escaping top options, but not necessarily
become Pareto-optimal
s0
We can express the probability of becoming
optimal with respect to the probability of
escaping the top solution.
s1 s2 s3
Popt Pesc(stop) P(no s in S sgts1
s1gtstop)
Approximation solve a WCSP that retrieve the S
that maximize the contribution
s1
41
Iterative strategy
s0

• Algorithm
• Top-escape suggestion s1
• Repeat
• Generate new set of dominators D by solving a
  new Auxiliary-WCSP
• The solutions of this problem are the dominators
  that most contribute in the formula of Popt
• Calculate approximate value for Popt considering
  D as dominators
• Consider options in D as possible suggestions
• Stop when Popt does not increase anymore

Aux Weighted-CSP
s1 s2 s3
Aux Weighted-CSP Constraints dominate s1 Soft
constraints Probability that is better than s1
when s1 is better thn s0
s1
42
Adaptive strategies
43
Learning by observing the user

• We want improve suggestions by considering
• Prior distribution about previous users
• Adaptation by learning from users response
• Adaptive question answering strategy
• Only ask questions that have impact
• Chajewska (2000) chooses questions to maximize
  VOI
• Boutilier (2002) considers the value of future
  questions

44
The system
Probability distribution of missing preferences
The user
Current preferences
preferences
Generation of Adaptive Suggestions
suggestions
Current preferences
preferences
Distribution update
Generation of Adaptive Suggestions
45
Bayesian update
IF NO REACTION ? probability of TransSubway is
decreased
Predicates State the user expresses the
preference Critique the user has this
preference ?
Depends on the displayed options. Assumption gt
Popt
46
Evaluation of Adaptive Suggestions
Preference discovery

• Simulations
• Number of preferences discovered according to the
  lookahead principle
• Adaptive model-based suggestions perform even
  better than simple model-based suggestions

Number of shown suggestions
47
Conclusions and Contributions

• Emergence of e-commerce
• ? Need of personalized technologies
• Preference-based search
• Inefficiency of current web tools
• Form filling achieves only 25 of accuracy due to
  means-objectives
• Example critiquing
• Incremental preference acquisition
• Interaction paradigm that avoid means-objectives
• Increase user awareness, preferences are
  constructed
• Any critiques, at any time ? The user states only
  the preferences they are sure about
• The need for Suggestions to avoid the anchoring
  effect

48
Contributions/2

• Look-ahead principle
• Preferences are stated when an opportunity is
  identified
• Model-based suggestions
• Metaphor of active learning
• Model-based suggestions effectively stimulate
  users to express accurate preferences
• Dramatic increase of decision accuracy up to
  70
• Scalability
• Large datasets relaxation of the look-ahead
  principle
• Configurable products
• Retrieve suggestions solving a single
  optimization problem
• Adaptive suggestions
• Inference about user behavior