Exploring the Design Space for Adaptive Graphical User Interfaces

Transcription

1 Exploring the Design Space for Adaptive Graphical User Interfaces Krzysztof Gajos Mary Czerwinski Desney Tan Daniel S. Weld (University of Washington) (Microsoft Research) (Microsoft Research) (University of Washington)

2 Scope Graphical User Interfaces where the system automatically adapts the presentation of the functionality

3 Scope Graphical User Interfaces where the system automatically adapts the presentation of the functionality The Split Interface

4 Scope Graphical User Interfaces where the system automatically adapts the presentation of the functionality The Moving Interface

5 Scope Graphical User Interfaces where the system automatically adapts the presentation of the functionality The Visual Popout Interface

6 Scope Graphical User Interfaces where the system automatically adapts the presentation of the functionality

7 Motivation

8 Motivation

9 Motivation

10 Motivation They disorient the user!

11 Motivation They optimize the UI for the individual! They disorient the user!

12 Prior Work

13 Prior Work Greenberg and Witten [1985] Trevellyan and Browne [1987] Mitchell and Shneiderman [1989] Sears and Shneiderman [1994]? McGrenere, Baecker and Booth [2002] Findlater and McGrenere [2004] Tsandilas and shraefel [2005]

14 Commercial Deployments

15 Commercial Deployments

16 Our Goal Uncover the factors and relationships that influence users satisfaction and actual performance when using adaptive UIs

17 Road Map Introduce and motivate the problem Video Experiment 1: qualitative results Experiment 2: quantitative results Synthesis Conclusions

18

19

20 Potential Benefit Potential Disorientation

21 Potential Benefit Potential Disorientation The Split Interface Medium Low

22 Potential Benefit Potential Disorientation The Split Interface Medium Low The Moving Interface High Medium

23 Potential Benefit Potential Disorientation The Split Interface Medium Low The Moving Interface High Medium The Visual Popout Interface Low Low

24 Experiment 1 Goal: collect informative subjective data

25 Participants 26 volunteers (10 female) aged 25 to 55 (mean=46) moderate to high experience using computers (as indicated by a validated screener) intermediate to expert users of MS Office (as indicated by a validated screener) participants received software gratuity

26 Tasks Three classes of editing tasks: Flow chart edits Text edits Combined text and graphical edits

27 Procedures

28 Procedures Start Training

29 Procedures Start Training Flow Chart task Quotes task Poster task Questionnaire

30 Procedures Start Training Flow Chart task Change Interface Quotes task Poster task Questionnaire Done 4 conditions?

31 Procedures Start Training Flow Chart task Change Interface Quotes task Poster task Questionnaire Done 4 conditions? Final Questionnaire End

32 Results: Ranking Users ranked the Split Interface the highest (p<0.001)

33 Results: Ranking Users ranked the Split Interface the highest (p<0.001)

34 General 7 Satisfaction Ease of Use Satisfaction Unchanging Split Moving Visual Popout

35 General 7 Satisfaction Ease of Use Satisfaction Unchanging Split Moving Visual Popout

36 General 7 Satisfaction Ease of Use Ease of Use Satisfaction Satisfaction Unchanging Split Moving Visual Popout

37 General 7 Satisfaction Ease of Use Ease Ease of of Use Use Satisfaction Satisfaction Satisfaction Unchanging Split Moving Visual Popout

38 General 7 Satisfaction Ease Ease of of Use Use Satisfaction Satisfaction Unchanging Split Moving Visual Popout

39 Usability Discoverability Sense of Control Predictability of adaptation Unchanging Split Moving Visual Popout

40 Usability Discoverability Sense of Control Predictability of adaptation Unchanging Split Moving Visual Popout

41 Subjective Cost and Benefit

42 Subjective Cost and Benefit Subjective cost based on: Mental demand Physical Demand Frustration Confusion due to adaptation

43 Subjective Cost and Benefit Subjective cost based on: Mental demand Physical Demand Frustration Confusion due to adaptation Subjective benefit based on: Performance Efficiency due to adaptation

44 Subjective Cost and Benefit Subjective cost based on: Mental demand Physical Demand Frustration Confusion due to adaptation Subjective benefit based on: Performance Efficiency due to adaptation Subjective benefit Split Interface Non-adaptive baseline Moving Interface Visual Popout Interface Subjective cost

45 Subjective Cost and Benefit Subjective cost based on: Mental demand Physical Demand Frustration Confusion due to adaptation Subjective benefit based on: Performance Efficiency due to adaptation Subjective benefit Split Interface Non-adaptive baseline Moving Interface Visual Popout Interface Subjective cost

46 User Comments Split Interface Moving Interface Visual Popout Interface

47 User Comments Split Interface Moving Interface Visual Popout Interface - stability - semantic grouping

48 User Comments Split Interface Moving Interface Visual Popout Interface - stability - semantic - discoverability grouping

49 User Comments Split Interface Moving Interface Visual Popout Interface - stability - semantic - discoverability grouping

50 User Comments Split Interface Moving Interface Visual Popout Interface - stability - semantic - discoverability grouping - poor discoverability

51 User Comments Split Interface Moving Interface Visual Popout Interface - stability - semantic - discoverability grouping - poor discoverability - instability

52 User Comments Split Interface Moving Interface Visual Popout Interface - stability - semantic - discoverability grouping - poor discoverability - instability - anti-salience

53 Road Map Introduce and motivate the problem Video Experiment 1: qualitative results Experiment 2: quantitative results Synthesis Conclusions

54 Experiment 2 Goals: Collect accurate performance data Investigate how the accuracy of the adaptive algorithm affects how adaptation is used

55 Participants 8 research colleagues (2 female) aged 25 to 58 (mean=36) high experience using computers expert users of MS Office participants received two meal vouchers as gratuity

56 Tasks

57 Tasks

58 Tasks

59 Tasks

60 Tasks

61 Procedures Introduction and a brief training on a nonadaptive version of the interface Each participant used each of the three interfaces (Unchanging, Split and Moving) at two different accuracy levels (30% and 70%)

62 Performance Vs. Adaptation Type Completion time (seconds) None Split Moving

63 Performance Vs. Adaptation Type Completion time (seconds) None Split Moving

64 Performance Vs. Adaptation Type Participants were significantly faster using Split Interface than Nonadaptive baseline (p<0.003) Completion time (seconds) None Split Moving

65 Performance Vs. Adaptation Type Participants were significantly faster using Split Interface than Nonadaptive baseline (p<0.003) Completion time (seconds) None Split Moving

66 Performance Vs. Adaptation Type Participants were significantly faster using Split Interface than Nonadaptive baseline (p<0.003) Completion time (seconds) Participants were marginally faster using Moving Interface than Non-adaptive baseline (p<0.073) None Split Moving

67 Performance Vs. Adaptation Type Participants were significantly faster using Split Interface than Nonadaptive baseline (p<0.003) Completion time (seconds) Participants were marginally faster using Moving Interface than Non-adaptive baseline (p<0.073) None Split Moving

68 Performance Vs. Accuracy Both adaptive interfaces resulted in faster performance at the higher (70%) accuracy level than at the lower (30%) level (p<0.001) 70 30% 70% 30% 70% Split Moving

69 Frequency of Use Vs. Accuracy

70 Frequency of Use Vs. Accuracy

71 Frequency of Use Vs. Accuracy?

72 Frequency of Use Vs. Accuracy 19% 81% 30% accuracy

73 Frequency of Use Vs. Accuracy 7% 93% 70% accuracy 19% 81% 30% accuracy

74 User Comments Split Interface Moving Interface

75 User Comments Split Interface Moving Interface - discoverability

76 User Comments Split Interface Moving Interface - discoverability - poor discoverability

77 User Comments Split Interface Moving Interface - discoverability - poor discoverability - instability

78 Exploring the Design Space for Adaptive Graphical User Interfaces

79 Exploring the Design Space for Adaptive Graphical User Interfaces

80 Putting It All Together

81 Putting It All Together Interaction Mechanics stability locality

82 Putting It All Together Interaction Mechanics stability Algorithm Behavior frequency of adaptation locality accuracy predictability

83 Putting It All Together Interaction Mechanics stability locality Algorithm Behavior frequency of adaptation accuracy predictability Context interaction frequency task complexity

84 Interaction Mechanics Algorithm Behavior Context User Stability stability locality frequency of adaptation accuracy predictability interaction frequency task complexity satisfaction Split Interfaces Moving Interface Low stability High stability

85 Interaction Mechanics Algorithm Behavior Context User Stability stability locality frequency of adaptation accuracy predictability interaction frequency task complexity satisfaction Split Interfaces Moving Interface MS Smart Menus Low stability High stability

86 Interaction Mechanics Algorithm Behavior Context User Stability stability locality frequency of adaptation accuracy predictability interaction frequency task complexity satisfaction Split Interfaces Moving Interface MS Smart Menus Visual Popout Low stability High stability

87 Interaction Mechanics Algorithm Behavior Context Locality stability locality frequency of adaptation accuracy predictability interaction frequency task complexity User comments indicate that, especially for manual tasks, high locality improves discoverability of adaptation.

88 Adaptation Frequency Interaction Mechanics stability locality Algorithm Behavior frequency of adaptation accuracy predictability Context interaction frequency task complexity Two studies of Split Menus: Sears and Shneiderman [1994] Findlater and McGrenere [2004]

89 Adaptation Frequency Interaction Mechanics stability locality Algorithm Behavior frequency of adaptation accuracy predictability Context interaction frequency task complexity Two studies of Split Menus: Sears and Shneiderman [1994] adaptation once per user/session Findlater and McGrenere [2004] adaptation once per interaction

90 Interaction Mechanics Algorithm Behavior Context Accuracy stability locality frequency of adaptation accuracy predictability interaction frequency task complexity

91 Interaction Mechanics Algorithm Behavior Context Accuracy stability locality frequency of adaptation accuracy predictability interaction frequency task complexity Participants performed faster at higher accuracy levels (also in [ Tsandilas and schraefel CHI 05])

92 Interaction Mechanics Algorithm Behavior Context Accuracy stability locality frequency of adaptation accuracy predictability interaction frequency task complexity Participants performed faster at higher accuracy levels (also in [ Tsandilas and schraefel CHI 05]) Participants were more likely to take advantage of adaptation at higher accuracy levels

93 Interaction Mechanics Algorithm Behavior Context Accuracy stability locality frequency of adaptation accuracy predictability interaction frequency task complexity Participants performed faster at higher accuracy levels (also in [ Tsandilas and schraefel CHI 05]) Participants were more likely to take advantage of adaptation at higher accuracy levels More disorienting interfaces affected more by reduced accuracy [ Tsandilas and schraefel CHI 05]

94 Interaction Mechanics stability Predictability locality Algorithm Behavior frequency of adaptation accuracy predictability Context interaction frequency task complexity A study in progress!

95 Interaction Frequency Interaction Mechanics stability locality Algorithm Behavior frequency of adaptation accuracy predictability Context interaction frequency task complexity Two studies of adaptive deep hierarchical menus: Greenberg and Witten [1985] Trevellyan and Browne [1987]

96 Interaction Frequency Interaction Mechanics stability locality Algorithm Behavior frequency of adaptation accuracy predictability Context interaction frequency task complexity Two studies of adaptive deep hierarchical menus: Greenberg and Witten [1985] 30 interactions per trial Trevellyan and Browne [1987] 100 interactions per trial: -- first 30 positive -- last 30 neutral or negative

97 Interaction Mechanics Algorithm Behavior Context Task Complexity stability locality frequency of adaptation accuracy predictability interaction frequency task complexity Experiment 1 Experiment 2 Split Moving Split Moving Interface Interface Interface Interface - stability - semantic grouping - discoverability - discoverability - poor discoverability - instability - poor discoverability - instability

98 Interaction Mechanics Algorithm Behavior Context Task Complexity stability locality frequency of adaptation accuracy predictability interaction frequency task complexity Experiment 1 Experiment 2 Split Moving Split Moving Interface Interface Interface Interface - stability - semantic grouping - discoverability - discoverability - poor discoverability - instability - poor discoverability - instability

99 Conclusions

100 Conclusions Split Interface Moving Interface Visual Popout

101 Conclusions Split Interface Moving Interface Visual Popout Preferred [Experiment 1] Disliked

102 Conclusions Split Interface Moving Interface Visual Popout Preferred Disliked Faster [Experiment 2]

103 Conclusions Interaction Mechanics stability locality Algorithm Behavior frequency of adaptation accuracy predictability Context interaction frequency task complexity

104 Acknowledgments Andrea Bunt, Leah Findlater and Joanna McGrenere at UBC Members of the VIBE Group at MSR

105 Contact Information Krzysztof Gajos: Mary Czerwinski: Desney Tan: Daniel Weld: