Katrien Verbert, profile picture
Uploaded byKatrien Verbert
356 views

Human-centered AI: towards the next generation of interactive and adaptive explanation methods

The document describes research on human-centered AI and interactive explanation methods. It discusses explainable AI and the goals of explaining model outcomes to increase user trust and acceptance, and enabling users to interact with the explanation process to improve models. It then provides an overview of the Augment/HCI research group at KU Leuven and its work on explanation methods, recommendation techniques, and evaluating explanations through user studies.

Embed presentation

Downloaded 12 times
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
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
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
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.
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
Collaborative filtering – Content-based filtering Knowledge-based filtering - Hybrid Recommendation techniques
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.
Prediction models 9
Overview 10 Application domains Algoritmic foundation
Overview 11 Application domains Algoritmic foundation
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.
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
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
Results 15 The interaction effect between NFC (divided into 4 quartiles Q1-Q4) and interfaces in terms of confidence
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
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
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.
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.
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
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]
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.
Overview 23 Application domains Algoritmic foundation
Learning analytics Src: Steve Schoettler
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
User-centred design of explanations: 3 iterations & think-alouds Tutorial for full transparency Single-screen explanation Final explanation interface
Why? Justification Comparison with others Real explanation Placebo explanation No explanation
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
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
Learning analytics Src: Steve Schoettler
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
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)
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
Overview 35 Application domains Algoritmic foundation
Precision agriculture 36
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.
AHMoSe Visual Encodings 38 Model Explanations (SHAP) Model + Knowledge Summary
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]
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%
Qualitative Evaluation ¤ 10 open ended questions ¤ 5 viticulture experts and 4 ML experts. ¤ Thematic Analysis: potential use cases, trust, usability, and understandability.
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
Overview 43 Application domains Algoritmic foundation
Designing for interacting with predictions for finding jobs 44
Predicting duration to find a job 45 Key Issues: Missing data, prediction trust issues, job seeker motivation, lack of control.
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
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)
Overview 49 Application domains Algoritmic foundation
50 Healthcare
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
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
Recommender systems for food 53
54 https://augment.cs.kuleuven.be/demos
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).
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.
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
Ongoing work: PERSFO
59 https://www.imec-int.com/en/what-we-offer/research-portfolio/discrete
RECOMMENDER ALGORITHMS MACHINE LEARNING INTERACTIVE DASHBOARDS SMART ALERTS RICH CARE PLANS OPEN IoT ARCHITECTURE
User centered design approach 61
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.
Evaluation ¤ 12 nurses used the app for three months ¤ Data collection ¤ Interaction logs ¤ Resque questions ¤ Semi-structured interviews 63
¤ 12 nurses during 3 months 64
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
Motivation Context 60% of Belgian employees has pain complaints 33% of extended sick leave is due to pain complaints Solution 67
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
6 different designs 69 text inline reply tags
70 Word cloud Feature importance Feature importance+ % 6 different designs
User study 71
Results Feature Importance Bars (+%) favoured by most users 72
Keywords / themes as to why users (dis)like the explanation type 73
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
Next steps: health dashboard 75
76
77
Petal-X ¤ Explaining cardiovascular disease (CVD) risk predictions to patients.
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
Peter Brusliovsky Nava Tintarev Cristina Conati Denis Parra Collaborations Jurgen Ziegler Gregor Stiglic
Questions? katrien.verbert@cs.kuleuven.be @katrien_v Thank you! https://augment.cs.kuleuven.be/

More Related Content

PDF
Tutorial on User Profiling with Graph Neural Networks and Related Beyond-Acc...
byErasmo Purificato
 
PPTX
Collaborative Filtering Recommendation System
byMilind Gokhale
 
PPTX
Introduction to Deep learning
byleopauly
 
PPTX
Cloud Computing
byInternational School Of Management Excellence
 
PPTX
Soft computing01
byuniversity of sargodha
 
PPTX
Ensemble learning
byHaris Jamil
 
PDF
Resnet
byashwinjoseph95
 
PDF
A tutorial on GreenCloud
byHabibur Rahman
 
Tutorial on User Profiling with Graph Neural Networks and Related Beyond-Acc...
byErasmo Purificato
 
Collaborative Filtering Recommendation System
byMilind Gokhale
 
Introduction to Deep learning
byleopauly
 
Cloud Computing
byInternational School Of Management Excellence
 
Soft computing01
byuniversity of sargodha
 
Ensemble learning
byHaris Jamil
 
Resnet
byashwinjoseph95
 
A tutorial on GreenCloud
byHabibur Rahman
 

What's hot

PDF
Graph Neural Networks for Recommendations
byWQ Fan
 
PPTX
ppt on machine learning to deep learning (1).pptx
byAnweshaGarima
 
PPTX
Computational Intelligence: concepts and applications using Athena
byPedro Almir
 
PPTX
Machine Learning Training Bootcamp
byTonex
 
PPTX
Intro to machine learning(with animations)
byRishiSingh189
 
PDF
NLP with Deep Learning
byfmguler
 
PPT
Content based recommendation systems
byAravindharamanan S
 
PPTX
Soft computing
byNabarun Paul
 
PPT
Recognition as Graph Matching
byVishakha Agarwal
 
PPT
Designing Distributed Systems: Google Cas Study
byMeysam Javadi
 
PPTX
Federated learning in brief
byShashi Perera
 
PDF
Data Science - Part XI - Text Analytics
byDerek Kane
 
PPTX
Regularization_BY_MOHAMED_ESSAM.pptx
byMohamed Essam
 
PPT
Rule Based System
bySuresh Sambandam
 
PPTX
Chapter 5 - Fuzzy Logic
byAshique Rasool
 
PDF
Topic 2: Cloud Computing Paradigms
byZubair Nabi
 
PPTX
Firewall log and network security management - Mumbai Seminar
byManageEngine, Zoho Corporation
 
PDF
Expert system neural fuzzy system
byJamil S. Alagha
 
PPT
Dynamic Resource Allocation Using Virtual Machines for Cloud Computing Enviro...
bySaikiranReddy Sama
 
PPTX
Activation function
byAstha Jain
 
Graph Neural Networks for Recommendations
byWQ Fan
 
ppt on machine learning to deep learning (1).pptx
byAnweshaGarima
 
Computational Intelligence: concepts and applications using Athena
byPedro Almir
 
Machine Learning Training Bootcamp
byTonex
 
Intro to machine learning(with animations)
byRishiSingh189
 
NLP with Deep Learning
byfmguler
 
Content based recommendation systems
byAravindharamanan S
 
Soft computing
byNabarun Paul
 
Recognition as Graph Matching
byVishakha Agarwal
 
Designing Distributed Systems: Google Cas Study
byMeysam Javadi
 
Federated learning in brief
byShashi Perera
 
Data Science - Part XI - Text Analytics
byDerek Kane
 
Regularization_BY_MOHAMED_ESSAM.pptx
byMohamed Essam
 
Rule Based System
bySuresh Sambandam
 
Chapter 5 - Fuzzy Logic
byAshique Rasool
 
Topic 2: Cloud Computing Paradigms
byZubair Nabi
 
Firewall log and network security management - Mumbai Seminar
byManageEngine, Zoho Corporation
 
Expert system neural fuzzy system
byJamil S. Alagha
 
Dynamic Resource Allocation Using Virtual Machines for Cloud Computing Enviro...
bySaikiranReddy Sama
 
Activation function
byAstha Jain
 

Similar to Human-centered AI: towards the next generation of interactive and adaptive explanation methods

PPTX
Explainable AI for non-expert users
byKatrien Verbert
 
POTX
Explaining recommendations: design implications and lessons learned
byKatrien Verbert
 
PPTX
Human-centered AI: how can we support end-users to interact with AI?
byKatrien Verbert
 
PDF
Human-centered AI: how can we support lay users to understand AI?
byKatrien Verbert
 
PPTX
Explainable AI is not yet Understandable AI
byepsilon_tud
 
PPTX
Mixed-initiative recommender systems
byKatrien Verbert
 
PDF
Interactive Recommender Systems
byKatrien Verbert
 
PDF
Mixed-initiative recommender systems: towards a next generation of recommende...
byKatrien Verbert
 
PPTX
Workshop on Designing Human-Centric MIR Systems
byepsilon_tud
 
PDF
Human-centered AI: how can we support end-users to interact with AI?
byKatrien Verbert
 
PDF
IUI Step-hai 2025 workshop Keynote by Martijn Willemsen
byEindhoven University of Technology / JADS
 
PPTX
Towards the next generation of interactive and adaptive explanation methods
byKatrien Verbert
 
PDF
Mixed-initiative recommender systems
byKatrien Verbert
 
PPTX
Interactive recommender systems: opening up the “black box”
byKatrien Verbert
 
PDF
Interactive Recommender Systems
byKatrien Verbert
 
PDF
To explain or not to explain
byMartijn Millecamp
 
PDF
master_thesis.pdf
byEL MAJJODI Ayoub
 
PDF
Explaining job recommendations: a human-centred perspective
byKatrien Verbert
 
PDF
ifib Lunchbag: CHI2018 Highlights - Algorithms in (Social) Practice and more
byhen_drik
 
PPTX
Why one-size-fits all does not work in Explainable Artificial Intelligence!
byvoginip
 
Explainable AI for non-expert users
byKatrien Verbert
 
Explaining recommendations: design implications and lessons learned
byKatrien Verbert
 
Human-centered AI: how can we support end-users to interact with AI?
byKatrien Verbert
 
Human-centered AI: how can we support lay users to understand AI?
byKatrien Verbert
 
Explainable AI is not yet Understandable AI
byepsilon_tud
 
Mixed-initiative recommender systems
byKatrien Verbert
 
Interactive Recommender Systems
byKatrien Verbert
 
Mixed-initiative recommender systems: towards a next generation of recommende...
byKatrien Verbert
 
Workshop on Designing Human-Centric MIR Systems
byepsilon_tud
 
Human-centered AI: how can we support end-users to interact with AI?
byKatrien Verbert
 
IUI Step-hai 2025 workshop Keynote by Martijn Willemsen
byEindhoven University of Technology / JADS
 
Towards the next generation of interactive and adaptive explanation methods
byKatrien Verbert
 
Mixed-initiative recommender systems
byKatrien Verbert
 
Interactive recommender systems: opening up the “black box”
byKatrien Verbert
 
Interactive Recommender Systems
byKatrien Verbert
 
To explain or not to explain
byMartijn Millecamp
 
master_thesis.pdf
byEL MAJJODI Ayoub
 
Explaining job recommendations: a human-centred perspective
byKatrien Verbert
 
ifib Lunchbag: CHI2018 Highlights - Algorithms in (Social) Practice and more
byhen_drik
 
Why one-size-fits all does not work in Explainable Artificial Intelligence!
byvoginip
 

More from Katrien Verbert

PDF
Explainability methods
byKatrien Verbert
 
PDF
Visual analytics
byKatrien Verbert
 
PDF
Designing Learning Analytics Dashboards: Lessons Learned
byKatrien Verbert
 
PDF
Scalable Exploration of Relevance Prospects to Support Decision Making
byKatrien Verbert
 
PDF
Learning analytics dashboards
byKatrien Verbert
 
PDF
Student-facing Learning dashboards
byKatrien Verbert
 
PDF
Explaining and Exploring Job Recommendations: a User-driven Approach for Inte...
byKatrien Verbert
 
PDF
Information Visualisation: perception and principles
byKatrien Verbert
 
PDF
Information Visualisation: Introduction
byKatrien Verbert
 
PDF
Learning analytics for feedback at scale
byKatrien Verbert
 
PDF
Web Information Systems Lecture 1: Introduction
byKatrien Verbert
 
PDF
Web Information Systems Lecture 2: HTML
byKatrien Verbert
 
PDF
Interactive recommender systems and dashboards for learning
byKatrien Verbert
 
PPT
Agents vs Users: Visual Recommendation of Research Talks with Multiple Dimens...
byKatrien Verbert
 
PDF
Personalized food recommendations: combining recommendation, visualization an...
byKatrien Verbert
 
PDF
EC-TEL 2016 Opening
byKatrien Verbert
 
PPTX
Open science in the digital humanities
byKatrien Verbert
 
Explainability methods
byKatrien Verbert
 
Visual analytics
byKatrien Verbert
 
Designing Learning Analytics Dashboards: Lessons Learned
byKatrien Verbert
 
Scalable Exploration of Relevance Prospects to Support Decision Making
byKatrien Verbert
 
Learning analytics dashboards
byKatrien Verbert
 
Student-facing Learning dashboards
byKatrien Verbert
 
Explaining and Exploring Job Recommendations: a User-driven Approach for Inte...
byKatrien Verbert
 
Information Visualisation: perception and principles
byKatrien Verbert
 
Information Visualisation: Introduction
byKatrien Verbert
 
Learning analytics for feedback at scale
byKatrien Verbert
 
Web Information Systems Lecture 1: Introduction
byKatrien Verbert
 
Web Information Systems Lecture 2: HTML
byKatrien Verbert
 
Interactive recommender systems and dashboards for learning
byKatrien Verbert
 
Agents vs Users: Visual Recommendation of Research Talks with Multiple Dimens...
byKatrien Verbert
 
Personalized food recommendations: combining recommendation, visualization an...
byKatrien Verbert
 
EC-TEL 2016 Opening
byKatrien Verbert
 
Open science in the digital humanities
byKatrien Verbert
 

Recently uploaded

PPTX
answer the question my lord dear lord.pptx
bykumailjavaid21
 
PDF
A Graphic Designing Project on "Skeuomorphism"
byRed Apple Learning
 
PPTX
Blue Illustrative Social Media Engagement Presentation_20251110_210350_0000.pptx
byabiharini55555
 
PPTX
Top Background Removal Benefits for Stunning and Professional Photos
bySam Smith
 
PPTX
Lec_5 Components of Architecture A&TP.pptx
byGhulamYaseen44
 
PDF
A NEW GAMMA MINING PROCEDURE CASE STUDY: BREAST CANCER TUMOR DIAGNOSIS
byamsjournal1
 
PDF
ColorTheory 2 -sensation and perception of colors 2025.pdf
byabeerartdesign
 
PPTX
IGCSE Examination dates for examination school.pptx
byRoyArjuna
 
PPTX
This presentation is about basic designing to website using basic html and css
byChristopherPurcia1
 
PPTX
CRAYON OFFICE PPT.pptx which is very use
byruchitakatre17
 
PPTX
digital pop up book prototype to show ho
byyourlyricsy
 
PPTX
Interior Design Type with Discription.pptx
byarsabbir65
 
DOCX
Personal Rapid Transit (PRT) System in Bandra Kurla Complex (BKC), Mumbai
byMayureshKumbhar4
 
PPTX
Windows 12.pptxghfghndf bhgfhghrbghjbdhgbcfyhg
bynguyenanhkietvn2014
 
PDF
A Graphic Designing Project on "Flat Designing"
byRed Apple Learning
 
PPTX
1cmeb02hj_9853rsdfsdfsdfsdfsdfsdf29.pptx
byLAWRENCEJEREMYBRIONE
 
PPTX
Bring the Beauty of Fall Indoors with Felt and Yarn’s Autumn Collection.pptx
bydavidbrown179291
 
PDF
Dyspnea approach with cases.pdf. J
by2986ishw
 
PPTX
Wooden_Furniture_Presentation.pptx......
byfaizaranjha959
 
PDF
mela design | Architectural design |urban design
byJnaneshPreethan
 
answer the question my lord dear lord.pptx
bykumailjavaid21
 
A Graphic Designing Project on "Skeuomorphism"
byRed Apple Learning
 
Blue Illustrative Social Media Engagement Presentation_20251110_210350_0000.pptx
byabiharini55555
 
Top Background Removal Benefits for Stunning and Professional Photos
bySam Smith
 
Lec_5 Components of Architecture A&TP.pptx
byGhulamYaseen44
 
A NEW GAMMA MINING PROCEDURE CASE STUDY: BREAST CANCER TUMOR DIAGNOSIS
byamsjournal1
 
ColorTheory 2 -sensation and perception of colors 2025.pdf
byabeerartdesign
 
IGCSE Examination dates for examination school.pptx
byRoyArjuna
 
This presentation is about basic designing to website using basic html and css
byChristopherPurcia1
 
CRAYON OFFICE PPT.pptx which is very use
byruchitakatre17
 
digital pop up book prototype to show ho
byyourlyricsy
 
Interior Design Type with Discription.pptx
byarsabbir65
 
Personal Rapid Transit (PRT) System in Bandra Kurla Complex (BKC), Mumbai
byMayureshKumbhar4
 
Windows 12.pptxghfghndf bhgfhghrbghjbdhgbcfyhg
bynguyenanhkietvn2014
 
A Graphic Designing Project on "Flat Designing"
byRed Apple Learning
 
1cmeb02hj_9853rsdfsdfsdfsdfsdfsdf29.pptx
byLAWRENCEJEREMYBRIONE
 
Bring the Beauty of Fall Indoors with Felt and Yarn’s Autumn Collection.pptx
bydavidbrown179291
 
Dyspnea approach with cases.pdf. J
by2986ishw
 
Wooden_Furniture_Presentation.pptx......
byfaizaranjha959
 
mela design | Architectural design |urban design
byJnaneshPreethan
 
In this document
Powered by AI

Introduction to human-centered AI with focus on interactive and adaptive explanation methods.

Explores XAI concepts, user trust, and interactive model explanations in recommendation systems.

Investigates user characteristics affecting trust, satisfaction, and control in recommendation systems.

Examines the impact of user control levels on cognitive load and acceptance in recommendation systems.

Focuses on developing explanations for exercise recommendations in educational platforms and their impact on trust.

Evaluates LADA for academic advising, emphasizing the need for transparency to enhance decision-making.

Discusses the implementation of AHMoSe in agriculture for grape quality prediction and model transparency.

Design aspects of job finding dashboards, highlighting the significance of user customization and control.

Explores health recommender systems and user engagement in selecting healthier food products.

Presents user-centered designs of health recommendations, emphasizing the importance of simplicity and customization.

Wrap-up of the presentation and invitation for questions with contact details.

Human-centered AI: towards the next generation of interactive and adaptive explanation methods

  • 1.
    Human-centered AI: towardsthe 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 ExplainableAI - 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 DeCroon 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 modeloutcomes to increase user trust and acceptance q Enable users to interact with the explanation process to improve the model Research objectives Models
  • 6.
  • 7.
    Collaborative filtering –Content-based filtering Knowledge-based filtering - Hybrid Recommendation techniques
  • 8.
  • 9.
  • 10.
  • 11.
  • 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 forcognition •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-subjectsdesign: 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 effectbetween NFC (divided into 4 quartiles Q1-Q4) and interfaces in terms of confidence
  • 16.
    Design implications ¤ Explanationsshould 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 tendto 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 Additionalcontrols 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 ofuser 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.
  • 24.
  • 26.
    Explaining exercise recommendations Howto 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 ofexplanations: 3 iterations & think-alouds Tutorial for full transparency Single-screen explanation Final explanation interface
  • 28.
  • 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 Canbe acceptable in low-stakes situations (e.g., drilling exercises): indications of difficulty level might suffice Personal level indication: Easy, Medium and Hard tags
  • 31.
  • 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 @KULeuven 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 wasperceived 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.
  • 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 ModelExplanations (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 ¤ AHMoSevs 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 ¤ 10open 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.
  • 44.
    Designing for interactingwith predictions for finding jobs 44
  • 45.
    Predicting duration tofind a job 45 Key Issues: Missing data, prediction trust issues, job seeker motivation, lack of control.
  • 46.
    Methods ¤ A CustomerJourney 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.
  • 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.
  • 50.
  • 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.
  • 54.
  • 55.
    Design and evaluation 55 GutiérrezF., 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 allowsusers 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 tocultivate 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.
  • 59.
  • 60.
  • 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 nursesused the app for three months ¤ Data collection ¤ Interaction logs ¤ Resque questions ¤ Semi-structured interviews 63
  • 64.
    ¤ 12 nursesduring 3 months 64
  • 65.
    Results ¤ Iterative designprocess 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.
  • 67.
    Motivation Context 60% of Belgianemployees has pain complaints 33% of extended sick leave is due to pain complaints Solution 67
  • 68.
    Explaining health recommendations ¤ 6different 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 textinline reply tags
  • 70.
    70 Word cloud Featureimportance Feature importance+ % 6 different designs
  • 71.
  • 72.
    Results Feature Importance Bars(+%) favoured by most users 72
  • 73.
    Keywords / themesas to why users (dis)like the explanation type 73
  • 74.
    Results “Insight vs. informationoverload” ¤ 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: healthdashboard 75
  • 76.
  • 77.
  • 78.
    Petal-X ¤ Explaining cardiovasculardisease (CVD) risk predictions to patients.
  • 79.
    Take-away messages ¤ Involvementof 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 NavaTintarev Cristina Conati Denis Parra Collaborations Jurgen Ziegler Gregor Stiglic
  • 81.