1. Human-centered AI: towards the next generation
of interactive and adaptive explanation methods
IHM 2022 – 8 April 2022
Katrien Verbert
Augment/HCI – Department of Computer Science - KU Leuven
@katrien_v

2. Human-Computer Interaction group
Explainable AI - recommender systems – visualization – intelligent user interfaces
Learning analytics &
human resources
Media
consumption
Precision agriculture
Healthcare
Augment Katrien Verbert
ARIA Adalberto Simeone
Computer
Graphics
Phil Dutré
LIIR Sien Moens
E-media
Vero Vanden Abeele
Luc Geurts
Kathrin Gerling

3. Augment/HCI team
Robin De Croon
Postdoc researcher
Katrien Verbert
Associate Professor
Francisco Gutiérrez
Postdoc researcher
Tom Broos
PhD researcher
Nyi Nyi Htun
Postdoc researcher
Houda Lamqaddam
PhD researcher
Oscar Alvarado
Postdoc researcher
https://augment.cs.kuleuven.be/
Diego Rojo Carcia
PhD researcher
Maxwell Szymanski
PhD researcher
Arno Vanneste
PhD researcher
Jeroen Ooge
PhD researcher
Aditya Bhattacharya
PhD researcher
Ivania Donoso Guzmán
PhD researcher

4. Explainable Artificial Intelligence (XAI)
“Given an audience, an explainable artificial
intelligence is one that produces details or reasons
to make its functioning clear or easy to understand.”
[Arr20]
4
[Arr20] Arrieta, A. B., Díaz-Rodríguez, N., Del Ser, J., Bennetot, A., Tabik, S., Barbado, A., ... & Herrera, F. (2020). Explainable Artificial
Intelligence (XAI): Concepts, taxonomies, opportunities and challenges toward responsible AI. Information fusion, 58, 82-115.

5. q Explaining model outcomes to increase user trust and acceptance
q Enable users to interact with the explanation process to improve the model
Research objectives
Models

6. 6

7. Collaborative filtering – Content-based filtering
Knowledge-based filtering - Hybrid
Recommendation techniques

8. Example: TasteWeights
8
Bostandjiev,
S.,
O'Donovan,
J.
and
Höllerer,
T.
TasteWeights:
a
visual
interactive
hybrid
recommender
system.
In
Proceedings
of
the
sixth
ACM
conference
on
Recommender
systems
(RecSys
'12).
ACM,
New
York,
NY,
USA
(2012),
35-42.

9. Prediction models
9

10. Overview
10
Application domains
Algoritmic foundation

11. Overview
11
Application domains
Algoritmic foundation

12. Explanations
12
Millecamp, M., Htun, N. N., Conati, C., & Verbert, K. (2019, March). To explain or not to explain:
the effects of personal characteristics when explaining music recommendations. In
Proceedings of the 2019 Conference on Intelligent User Interface (pp. 397-407). ACM.

13. Personal characteristics
Need for cognition
•Measurement of the tendency for an individual to engage in, and enjoy, effortful cognitive
activities
•Measured by test of Cacioppo et al. [1984]
Visualisation literacy
•Measurement of the ability to interpret and make meaning from information presented in the form
of images and graphs
•Measured by test of Boy et al. [2014]
Locus of control (LOC)
•Measurement of the extent to which people believe they have power over events in their lives
•Measured by test of Rotter et al. [1966]
Visual working memory
•Measurement of the ability to recall visual patterns [Tintarev and Mastoff, 2016]
•Measured by Corsi block-tapping test
Musical experience
•Measurement of the ability to engage with music in a flexible, effective and nuanced way
[Müllensiefen et al., 2014]
•Measured using the Goldsmiths Musical Sophistication Index (Gold-MSI)
Tech savviness
•Measured by confidence in trying out new technology 13

14. User study
¤ Within-subjects design: 105 participants recruited with Amazon Mechanical Turk
¤ Baseline version (without explanations) compared with explanation interface
¤ Pre-study questionnaire for all personal characteristics
¤ Task: Based on a chosen scenario for creating a play-list, explore songs and
rate all songs in the final playlist
¤ Post-study questionnaire:
¤ Recommender effectiveness
¤ Trust
¤ Good understanding
¤ Use intentions
¤ Novelty
¤ Satisfaction
¤ Confidence

15. Results
15
The interaction effect between NFC (divided into
4 quartiles Q1-Q4) and interfaces in terms of confidence

16. Design implications
¤ Explanations should be personalised for different groups of
end-users.
¤ Users should be able to choose whether or not they want to
see explanations.
¤ Explanation components should be flexible enough to present
varying levels of details depending on a user’s preference.
16

17. User control
Users tend to be more satisfied when they have control over
how recommender systems produce suggestions for them
Control recommendations
Douban FM
Control user profile
Spotify
Control algorithm parameters
TasteWeights

18. Controllability Cognitive load
Additional controls may increase cognitive load
(Andjelkovic et al. 2016)
Ivana Andjelkovic, Denis Parra, andJohn O’Donovan. 2016. Moodplay: Interactive mood-based
music discovery and recommendation. In Proc. of UMAP’16. ACM, 275–279.

19. Different levels of user control
19
Level
Recommender
components
Controls
low
Recommendations
(REC)
Rating, removing, and
sorting
medium User profile (PRO)
Select which user profile
data will be considered by
the recommender
high
Algorithm parameters
(PAR)
Modify the weight of
different parameters
Jin, Y., Tintarev, N., & Verbert, K. (2018, September). Effects of personal characteristics on music
recommender systems with different levels of controllability. In Proceedings of the 12th ACM Conference
on Recommender Systems (pp. 13-21). ACM.

20. User profile (PRO) Algorithm parameters (PAR) Recommendations (REC)
8 control settings
No control
REC
PAR
PRO
REC*PRO
REC*PAR
PRO*PAR
REC*PRO*PAR

21. Evaluation method
¤ Between-subjects – 240 participants recruited with AMT
¤ Independent variable: settings of user control
¤ 2x2x2 factorial design
¤ Dependent variables:
¤ Acceptance (ratings)
¤ Cognitive load (NASA-TLX), Musical Sophistication, Visual Memory
¤ Framework Knijnenburg et al. [2012]

22. Results
¤ Main effects: from REC to PRO to PAR → higher cognitive
load
¤ Two-way interaction: does not necessarily result in higher
cognitive load. Adding an additional control component
to PAR increases the acceptance. PRO*PAR has less
cognitive load than PRO and PAR
¤ High musical sophistication leads to higher quality, and
thereby result in higher acceptance
22
Jin, Y., Tintarev, N., & Verbert, K. (2018, September). Effects of personal characteristics on music
recommender systems with different levels of controllability. In Proceedings of the 12th ACM
Conference on Recommender Systems (pp. 13-21). ACM.

23. Overview
23
Application domains
Algoritmic foundation

24. Learning analytics
Src: Steve Schoettler

26. Explaining exercise recommendations
How to automatically
adapt the exercise
recommending on Wiski to
the level of students?
How do (placebo)
explanations affect initial
trust in Wiski for
recommending exercises?
Goals and research questions
Automatic
adaptation
Explanations & trust
Young target
audience
Middle and high school
students
Ooge, J., Kato, S., Verbert, K. (2022) Explaining Recommendations in E-Learning: Effects on
Adolescents' Initial Trust. Proceedings of the 27th IUI conference on Intelligent User Interfaces

27. User-centred design of explanations: 3
iterations & think-alouds
Tutorial for full transparency Single-screen explanation Final explanation interface

28. Why?
Justification
Comparison
with others
Real
explanation
Placebo
explanation
No
explanation

29. Results: Real explanations…
… did increase multidimensional initial trust
… did not increase one-dimensional initial trust
… led to accepting more recommended exercises
compared to both placebo and no explanations

30. Results: No explanations
Can be acceptable in low-stakes situations (e.g.,
drilling exercises):
indications of difficulty level might suffice
Personal level
indication: Easy,
Medium and Hard tags

31. Learning analytics
Src: Steve Schoettler

32. 32
uncertainty
Gutiérrez Hernández F., Seipp K., Ochoa X., Chiluiza K., De Laet T., Verbert K. (2018). LADA: A
learning analytics dashboard for academic advising. Computers in Human Behavior, pp 1-13. doi:
10.1016/j.chb.2018.12.004
LADA: a learning analytics dashboard
for study advisors

33. Evaluation method
33
Evaluation @KU Leuven Monitoraat
N = 12
6 Experts (4F, 2M)
6 Laymen (1F, 5M)
Evaluation @ESPOL (Ecuador)
N = 14
8 Experts (3F, 5M)
6 Laymen (6M)

34. Results
¤ LADA was perceived as a valuable tool for more
accurate and efficient decision making.
¤ LADA enables expert advisers to evaluate significantly
more scenarios.
¤ More transparency in the prediction model is required in
order to increase trust.
34
Gutiérrez Hernández F., Seipp K., Ochoa X., Chiluiza K., De Laet T., Verbert K. (2018). LADA: A
learning analytics dashboard for academic advising. Computers in Human Behavior, pp 1-13. doi:
10.1016/j.chb.2018.12.004

35. Overview
35
Application domains
Algoritmic foundation

36. Precision agriculture
36

37. AHMoSe
Rojo, D., Htun, N. N., Parra, D., De Croon, R., & Verbert, K. (2021). AHMoSe: A knowledge-based visual
support system for selecting regression machine learning models. Computers and Electronics in
Agriculture, 187, 106183.

38. AHMoSe Visual Encodings
38
Model Explanations
(SHAP)
Model + Knowledge Summary

39. Case Study – Grape Quality Prediction
39
¤ Grape Quality Prediction Scenario
[Tag14]
¤ Data
¤ Years 2010, 2011 (train) 2012 (test)
¤ 48 cells (Central Greece)
¤ Knowledge-based rules
[Tag14] Tagarakis, A., et al. "A fuzzy inference system to model grape
quality in vineyards." Precision Agriculture 15.5 (2014): 555-578. Source: [Tag14]

40. Simulation Study
¤ AHMoSe vs full AutoML approach to support model
selection.
40
RMSE (AutoML) RMSE (AHMoSe) Difference %
Scenario A
Complete
Knowledge
0.430 0.403 ▼ 6.3%
Scenario B
Incomplete
Knowledge
0.458 0.385 ▼ 16.0%

41. Qualitative Evaluation
¤ 10 open ended questions
¤ 5 viticulture experts and 4 ML experts.
¤ Thematic Analysis: potential use cases, trust, usability,
and understandability.

42. Qualitative Evaluation - Trust
42
¤ Showing the dis/agreement of model outputs with
expert’s knowledge can promote trust.
“The thing that makes us trust the models is the fact that most of the
time, there is a good agreement between the values predicted by the
model and the ones obtained for the knowledge of the experts.”
– Viticulture Expert

43. Overview
43
Application domains
Algoritmic foundation

44. Designing for interacting with predictions for
finding jobs
44

45. Predicting duration to find a job
45
Key Issues: Missing data, prediction trust issues, job
seeker motivation, lack of control.

46. Methods
¤ A Customer Journey approach. (5 mediators).
¤ Hands-on time with the original dashboard (22 mediators).
¤ Observations of mediation sessions. (3 mediators, 6 job seekers).
¤ Questionnaire regarding perception of the dashboard and
prediction model (15 Mediators).
46
Charleer S., Gutiérrez Hernández F., Verbert K. (2018). Supporting job mediator and job seeker
through an actionable dashboard. In: Proceedings of the 24th IUI conference on Intelligent User
Interfaces Presented at the ACM IUI 2019, Los Angeles, USA.

47. 47

48. Take away messages
¤ Key difference between actionable and non-actionable
parameters
¤ Need for customization and contextualization.
¤ The human expert plays a crucial role when interpreting
and relaying in the predicted or recommended output.
48
Charleer S., Gutiérrez Hernández F., Verbert K. (2019). Supporting job mediator and job
seeker through an actionable dashboard. In: Proceedings of the 24th IUI conference on
Intelligent User Interfaces Presented at the ACM IUI 2019, Los Angeles, USA. (Core: A)

49. Overview
49
Application domains
Algoritmic foundation

50. 50
Healthcare

51. 51
Ooge, J., Stiglic, G., & Verbert, K. (2021). Explaining artificial intelligence with
visual analytics in healthcare. Wiley Interdisciplinary Reviews: Data Mining
and Knowledge Discovery, e1427. https://doi.org/10.1002/widm.1427

52. https://www.jmir.org/2021/6/e18035
Nutrition
Nutrition advice (7)
Diets (7)
Recipes (7)
Menus (2)
Fruit (1)
Restaurants (1)
Doctors (4)
Hospital (5)
Thread / fora (3)
Self-Diagnosis (3)
Healthcare information (5)
Similar users (2)
Advise for children (2)
General
health
information
Routes (2)
Physical activity (10)
Leisure activity (2)
Wellbeing motivation (2)
Behaviour (7)
Wearable devices (1)
Tailored messages (2)
Routes (2)
Physical activity (10)
Leisure activity (2)
Behaviour (7)
Lifestyle
Specific
health
conditions
Health
Recommender
Systems

53. Recommender systems for food
53

54. 54
https://augment.cs.kuleuven.be/demos

55. Design and evaluation
55
Gutiérrez F., Cardoso B., Verbert K. (2017). PHARA: a personal health augmented reality assistant to
support decision-making at grocery stores. In: Proceedings of the International Workshop on Health
Recommender Systems co-located with ACM RecSys 2017 (Paper No. 4) (10-13).

56. Results
¤ PHARA allows users to make informed decisions, and
resulted in selecting healthier food products.
¤ Stack layout performs better with HMD devices with a
limited field of view, like the HoloLens, at the cost of some
affordances.
¤ The grid and pie layouts performed better in handheld
devices, allowing to explore with more confidence,
enjoyability and less effort.
56
Gutiérrez Hernández F., Htun NN., Charleer S., De Croon R., Verbert K. (2018). Designing
augmented reality applications for personal health decision-making. In: Proceedings of the 2019
52nd Hawaii International Conference on System Sciences Presented at the HICSS, Hawaii, 07
Jan 2019-11 Jan 2019.

57. Biofortification info
Plants to cultivate
Ongoing work: PERNUG
¤ Increased access to more nutritious plants
¤ Improved iron and B12 intakes for vegan and vegetarian
subgroups
Consumer app with recipe
recommendations
Hydroponic system with
biofortified plants
https://www.eitfood.eu/projects/pernug

58. Ongoing work: PERSFO

59. 59
https://www.imec-int.com/en/what-we-offer/research-portfolio/discrete

60. RECOMMENDER
ALGORITHMS
MACHINE
LEARNING
INTERACTIVE DASHBOARDS
SMART ALERTS
RICH CARE PLANS
OPEN IoT
ARCHITECTURE

61. User centered design approach
61

62. 62
Gutiérrez Hernández, F.S., Htun, N.N., Vanden Abeele, V., De Croon, R., Verbert, K.
(2022). Explaining call recommendations in nursing homes: a user-centered design
approach for interacting with knowledge-based health decision support systems.
Proceedings of IUI 2022.

63. Evaluation
¤ 12 nurses used the app for three months
¤ Data collection
¤ Interaction logs
¤ Resque questions
¤ Semi-structured interviews
63

64. ¤ 12 nurses during 3 months
64

65. Results
¤ Iterative design process identified several important features, such as the pending
list, overview and the feedback shortcut to encourage feedback.
¤ Explanations seem to contribute well to better support the healthcare
professionals.
¤ Results indicate a better understanding of the call notifications by being able to
see the reasons of the calls.
¤ More trust in the recommendations and increased perceptions of transparency
and control
¤ Interaction patterns indicate that users engaged well with the interface, although
some users did not use all features to interact with the system.
¤ Need for further simplification and personalization.
65

66. 66

67. Motivation
Context
60% of Belgian employees has
pain complaints
33% of extended sick leave is
due to pain complaints
Solution
67

68. Explaining health
recommendations
¤ 6 different explanation designs
¤ Explain WHY users are given a
certain recommendation for
their (chronic) pain based on
their inputs
68
Maxwell Szymanski, Vero Vanden Abeele and Katrien Verbert Explaining
health recommendations to lay users: The dos and don’ts – Apex-IUI 2022

69. 6 different designs
69
text inline reply tags

70. 70
Word cloud Feature importance Feature importance+ %
6 different designs

71. User study
71

72. Results
Feature Importance Bars (+%) favoured by most users
72

73. Keywords / themes as to why users
(dis)like the explanation type
73

74. Results
“Insight vs. information overload”
¤ Most users prefer more information (holistic overview of inputs)
¤ However, some users experienced information overload
→ Future work - Do personal characteristics such as NFC
influence this?
74

75. Next steps: health dashboard
75

76. 76

77. 77

78. Petal-X
¤ Explaining cardiovascular disease (CVD) risk predictions
to patients.

79. Take-away messages
¤ Involvement of end-users has been key to come up with
interfaces tailored to the needs of non-expert users
¤ Actionable vs non-actionable parameters
¤ Domain expertise of users and need for cognition
important personal characteristics
¤ Need for personalisation and simplification
79

80. Peter Brusliovsky Nava Tintarev Cristina Conati
Denis Parra
Collaborations
Jurgen Ziegler
Gregor Stiglic

81. Questions?
katrien.verbert@cs.kuleuven.be
@katrien_v
Thank you!
https://augment.cs.kuleuven.be/