Transparent, Reliable and Unbiased Smart Tool for AI

Informations projet

TRUST-AI

N° de convention de subvention: 952060

DOI

10.3030/952060

Projet clôturé

Date de signature de la CE 4 Juin 2020

Date de début 1 Octobre 2020

Date de fin 31 Mars 2025

Financé au titre de

• EXCELLENT SCIENCE - Future and Emerging Technologies (FET)

Coût total

€ 3 996 418,75

Contribution de l’UE

€ 3 996 418,75

Coordonné par

INESC TEC - INSTITUTO DE ENGENHARIADE SISTEMAS E COMPUTADORES, TECNOLOGIA E CIENCIA
Portugal

CORDIS fournit des liens vers les livrables publics et les publications des projets HORIZON.

Les liens vers les livrables et les publications des projets du 7e PC, ainsi que les liens vers certains types de résultats spécifiques tels que les jeux de données et les logiciels, sont récupérés dynamiquement sur OpenAIRE .

Livrables

Documents, reports (20)

Management Report

This deliverable concerns a status report on the technical achievements of TRUST in the first nine months of the projects A brief description of the development of each task will be provided including documentation of procedures screenshots preliminary results and identified risks

Data management plan V2
Final validation of the learned explainable AI models (online retail)

In this deliverable, a final validation of the proposed approach will be provided. The results and final conclusions of the online retail use case will be described together with lessons learned and recommendations for future developments.

Final validation of the learned explainable AI models for the energy use case

In this deliverable, a final validation of the proposed approach will be provided. The results and final conclusions of the energy use case will be described together with lessons learned and recommendations for future developments.

Saliency measures for identifying causally variables of explanations

This report will present the saliency measures and code for identifying causally relevant variables of humanlike explanations The relevant variables are those to be considered during the discovery and communication of causal explanations These variables will be formalised and measurable in terms of their specificity insensitivity proximity and other characteristics known to be preferred by humans

User studies on the realization of explanations

The deliverable reports the result of the qualitative studies on the best way to present explanation content produced in WP3 and provides recommendations for Task 22

Communication & Dissemination report

This deliverable will present a final status on the achievement of the project objectives in terms of communication and dissemination. The report will refer to the KPIs proposed on Deliverable D8.1.

Evaluation by human observers of different explainability formats

This deliverable reports the results of the evaluation of explainability performed in Task2.3.

Final validation of the learned explainable AI models (healthcare)

In this deliverable, a final validation of the proposed approach will be provided. The results and final conclusions of the healthcare use case will be described together with lessons learned and recommendations for future developments.

Data management plan V3
Dialog WP4-WP3

In this deliverable, the results of the interaction between WP4 and WP3 will be presented. The final outcome is the generation of explainable expressions by iterated dialog with the user in the proposed toy problems.

Exploitation Plan

This report described the future exploration of the framework, detailed in Task 8.3.Partners will consolidate all relevant findings, identify risks and evaluate the potential applicability of TRUST components in different sectors, covering many aspects of AI.

Communication & Dissemination update

This deliverable will present an updated status of the CDP, reporting the achievements obtained for each KPI and possible necessary changes to the plan.

Framework requirements document

This deliverable will describe the functional and nonfunctional requirements of the framework as well as the interactions and dependencies between the building blocks The use case needs will also be detailed and reported in this document ensuring that the framework is adequate to different problems and sectors

Communication & Dissemination plan

This report will present the Communication Dissemination Plan of TRUST where the strategy to raise public awareness about the project outcomes will be detailed and scheduled In addition to academic publications and conferences the plan includes events promotion participation in working groups and online forums and educational content creation such as courseware and webinars The plan will include KPIs and their target values as well as the Partner responsible for each communicationdissemination method

Evaluation with healthcare experts of learned models

This deliverable concerns a formal validation of the AI models developed for the first simplified version of the healthcare problem These models will be designed by NWOI and validated by medical experts from LUMC The report will present the first insights on the models results and suggestions for modifications

Initial validation of the explainable AI models from business experts

This deliverable concerns a formal validation of the AI models developed for the first simplified version of the online retail problem These models will be designed by LTP and validated by practitioners from Sonae INESC will coordinate the development and validation process

Data management plan

This deliverable presents the Data Management Plan of TRUSTAI detailing the types of data generatedcollected how it will be exploited protected and the standards to be considered

Framework validation

This deliverable consolidates the final conclusions from the use cases and describes the primary outcomes of TRUST framework. Recommendations for future extensions will also be included.

Initial validation of the explainable AI models from energy experts

This deliverable concerns a formal validation of the AI models developed for the first simplified version of the energy problem These models will be designed by POLIS21 and validated by practitioners from the industry

Other (2)

Automated medical image analysis learning techniques

This deliverable concerns the automated image analysis learning techniques integrated with TRUST blocks.

Project website

This deliverable will present the specification, organization and features of TRUST-AI website. The DNS, URL to access and screenshots on each page will also be presented.

Publications

Conference proceedings (16)

Coefficient mutation in the gene-pool optimal mixing evolutionary algorithm for symbolic regression

Auteurs: M. Virgolin and P.A.N. Bosman
Publié dans: GECCO '22: Proceedings of the Genetic and Evolutionary Computation Conference Companion, 2022, Page(s) 2289–2297
Éditeur: ACM
DOI: 10.1145/3520304.3534036

Symbolic Pricing Policies for Attended Home Delivery - the Case of a European Retailer

Auteurs: Miguel Lunet, Daniela Fernandes, Fábio Neves-Moreira, Pedro Amorim
Publié dans: GECCO '25: Genetic and Evolutionary Computation Conference, 2025
Éditeur: Digital Library

Multi-objective Genetic Programming for Explainable Reinforcement Learning

Auteurs: Videau, Mathurin; Ferreira Leite, Alessandro; Teytaud, Olivier; Schoenauer, Marc
Publié dans: EUROGP - 25th European Conference on Genetic Programming, part of EvoStar 2022, Numéro 25, 2022, Page(s) pp.278-293, ISBN 978-3-031-02055-1
Éditeur: Springer Verlag
DOI: 10.1007/978-3-031-02056-8_18

Exploring Commonalities in Explanation Frameworks: A Multi-Domain Survey Analysis

Auteurs: Eduard Barbu, Marharytha Domnich, Raul Vicente, Nikos Sakkas, André Morim
Publié dans: The 2nd World Conference on eXplainable Artificial Intelligence (xAI 2024), July 17-19, 2024 - Valletta, Malta, 2024
Éditeur: Springer
DOI: 10.48550/arxiv.2405.11958

Multi-modal multi-objective model-based genetic programming to find multiple diverse high-quality models

Auteurs: Sijben, Evi; Alderliesten, Tanja; Bosman, Peter
Publié dans: GECCO '22: Genetic and Evolutionary Computation Conference, 2022, ISBN 978-1-4503-9237-2
Éditeur: Association for Computing Machinery, New York, NY, United States
DOI: 10.48550/arxiv.2203.13347

Memetic Semantic Genetic Programming for Symbolic Regression

Auteurs: Alessandro Leite and Marc Schoenauer
Publié dans: 26th EuroGP - Part of EvoStar 2023, Numéro 26, 2023, Page(s) 198–212, ISBN 978-3-031-29572-0
Éditeur: Springer Verlag LNCS-13986
DOI: 10.1007/978-3-031-29573-7_13

A Guide for Practical Use of ADMG Causal Data Augmentation

Auteurs: Poinsot, Audrey; Leite, Alessandro
Publié dans: Workshop on the pitfalls of limited data and computation for Trustworthy ML, ICLR 2023, 2023
Éditeur: OpenReview
DOI: 10.48550/arxiv.2304.01237

Explanatory World Models via Look Ahead Attention for Credit Assignment

Auteurs: Oriol Corcoll and Raul Vicente
Publié dans: Numéro 26403498, 2022, ISSN 2640-3498
Éditeur: Proceedings of Machine Learning Research

Function Class Learning with Genetic Programming: Towards Explainable Meta Learning for Tumor Growth Functionals

Auteurs: Evi Sijben, Jeroen Jansen, Peter Bosman, Tanja Alderliesten
Publié dans: Proceedings of the Genetic and Evolutionary Computation Conference, 2024, Page(s) 1354-1362
Éditeur: ACM
DOI: 10.1145/3638529.3654145

Emergence of Adaptive Circadian Rhythms in Deep Reinforcement Learning

Auteurs: Labash, Aqeel; Fletzer, Florian; Majoral, Daniel; Vicente, Raul
Publié dans: ICML'23: Proceedings of the 40th International Conference on Machine Learning, Numéro 18, 2023
Éditeur: JMLR.org
DOI: 10.48550/arxiv.2307.12143

COIN: Counterfactual inpainting for weakly supervised semantic segmentation for medical images

Auteurs: Dmytro Shvetsov, Joonas Ariva, Marharyta Domnich, Raul Vicente, Dmytro Fishman
Publié dans: The 2nd World Conference on eXplainable Artificial Intelligence (xAI 2024), July 17-19, 2024 - Valletta, Malta, 2024
Éditeur: Springer
DOI: 10.48550/arxiv.2404.12832

Learning symbolic expressions to solve multi-period time slot pricing vehicle routing problems

Auteurs: Fábio Neves-Moreira, Daniela Fernandes, Miguel Lunet, Pedro Amorim
Publié dans: IJCAI 2024 - International Joint Conference on Artificial Intelligence, Jeju, South Korea, 2024
Éditeur: IJCAI

Deep learning-based auto-segmentation of paraganglioma for growth monitoring

Auteurs: E.M.C. Sijben, J.C. Jansen, P.A.N. Bosman (Peter), and T. Alderliesten
Publié dans: Proceedings Volume 12929, Medical Imaging 2024: Image Perception, Observer Performance, and Technology Assessment, 2024, Page(s) 1292916
Éditeur: SPIE
DOI: 10.1117/12.3006413

Enhancing Counterfactual Explanation Search with Diffusion Distance and Directional Coherence

Auteurs: Marharyta Domnich, Raul Vicente
Publié dans: The 2nd World Conference on eXplainable Artificial Intelligence (xAI 2024), July 17-19, 2024 - Valletta, Malta, 2024
Éditeur: Springer
DOI: 10.48550/arxiv.2404.12810

Real time Data and Application Sharing and Collaboration for the Building Energy Domain

Auteurs: N. Sakkas, M. Papadopoulou, D. Sakkas
Publié dans: WDBE 2021, 2021
Éditeur: World of Digital Built Environment WDBE 2021

Evolvability degeneration in multi-objective genetic programming for symbolic regression

Auteurs: Dazhuang Liu, Marco Virgolin, Tanja Alderliesten, Peter A. N. Bosman
Publié dans: GECCO '22: Genetic and Evolutionary Computation Conference, 2022
Éditeur: Association for Computing Machinery, New York, NY, United States
DOI: 10.1145/3512290.3528787

Peer reviewed articles (17)

Improving asset management in capital-intensive industries: Case study of a Portuguese water utility

Auteurs: Mariana Casalta, Flávia Barbosa, Luciana Yamada, Lígia B. Ramos
Publié dans: Utilities Policy, Numéro 91, 2024, Page(s) 101822, ISSN 0957-1787
Éditeur: Pergamon Press Ltd.
DOI: 10.1016/j.jup.2024.101822

Mind the gap: challenges of deep learning approaches to Theory of Mind

Auteurs: Aru, Jaan; Labash, Aqeel; Corcoll, Oriol; Vicente, Raul
Publié dans: Artificial Ingtelligence Review, Numéro 3, 2023, ISSN 0269-2821
Éditeur: Kluwer Academic Publishers
DOI: 10.1007/s10462-023-10401-x

Learning efficient in-store picking strategies to reduce customer encounters in omnichannel retail

Auteurs: Fábio Neves-Moreira, Pedro Amorim
Publié dans: International Journal of Production Economics, 2024, ISSN 0925-5273
Éditeur: Elsevier BV
DOI: 10.1016/j.ijpe.2023.109074

Scheduling wagons to unload in bulk cargo ports with uncertain processing times

Auteurs: Cristiane Ferreira, Gonçalo Figueira, Pedro Amorim, Alexandre Pigatti
Publié dans: Computers & Operations Research, 2023, ISSN 0305-0548
Éditeur: Pergamon Press Ltd.
DOI: 10.1016/j.cor.2023.106364

Open data or open access? The case of building data.

Auteurs: Sakkas, N., Yfanti, S
Publié dans: Academia Letters, 2021, ISSN 2771-9359
Éditeur: Academia.edu
DOI: 10.20935/al3629

Towards Unifying Evaluation of Counterfactual Explanations: Leveraging Large Language Models for Human-Centric Assessments

Auteurs: Marharyta Domnich, Julius Välja, Rasmus Moorits Veski, Giacomo Magnifico, Kadi Tulver, Eduard Barbu, Raul Vicente
Publié dans: Proceedings of the AAAI Conference on Artificial Intelligence, Numéro 39, 2025, Page(s) 16308-16316, ISSN 2374-3468
Éditeur: Association for the Advancement of Artificial Intelligence (AAAI)
DOI: 10.1609/aaai.v39i15.33791

Memetic semantic boosting for symbolic regression

Auteurs: Alessandro Leite, Marc Schoenauer
Publié dans: Genetic Programming and Evolvable Machines, Numéro 26, 2025, ISSN 1389-2576
Éditeur: Kluwer Academic Publishers
DOI: 10.1007/s10710-024-09506-1

Explainable Approaches for Forecasting Building Electricity Consumption

Auteurs: Nikos Sakkas, Sofia Yfanti,Pooja Shah, Nikitas Sakkas, Christina Chaniotakis, Costas Daskalakis, Eduard Barbu and Marharyta Domnich
Publié dans: Energies, 2023, ISSN 1996-1073
Éditeur: Multidisciplinary Digital Publishing Institute (MDPI)
DOI: 10.3390/en16207210

Deep neural networks using a single neuron: folded-in-time architecture using feedback-modulated delay loops

Auteurs: Stelzer, Florian; Röhm, André; Vicente, Raul; Fischer, Ingo; Yanchuk, Serhiy
Publié dans: Nature Communications, Numéro 20411723, 2021, ISSN 2041-1723
Éditeur: Nature Publishing Group
DOI: 10.48550/arxiv.2011.10115

Technology Readiness Levels (TRLs) in the Era of Co-Creation

Auteurs: Sofia Yfanti, Nikos Sakkas
Publié dans: Applied System Innovation, 2024, ISSN 2571-5577
Éditeur: MDPI
DOI: 10.3390/asi7020032

A cooperative coevolutionary hyper-heuristic approach to solve lot-sizing and job shop scheduling problems using genetic programming

Auteurs: Yannik Zeiträg, José Rui Figueira, Gonçalo Figueira
Publié dans: International Journal of Production Research, 2024, ISSN 0925-5273
Éditeur: Elsevier BV
DOI: 10.1080/00207543.2023.2301044

Building data models and data sharing. Purpose, approaches, and a case study on explainable demand response

Auteurs: Nikos Sakkas, Christina Chaniotaki and Nikitas Sakkas
Publié dans: IOP Conference Series: Earth and Environmental Science, 2023, ISSN 1757-899X
Éditeur: IOP Science
DOI: 10.1088/1755-1315/1122/1/012066

Personalized choice model for forecasting demand under pricing scenarios with observational data—The case of attended home delivery

Auteurs: Özden Gür Ali, Pedro Amorim
Publié dans: International Journal of Forecasting, 2024, Page(s) 706-720, ISSN 0169-2070
Éditeur: Elsevier BV
DOI: 10.1016/j.ijforecast.2023.04.008

Quantifying Reinforcement-Learning Agent’s Autonomy, Reliance on Memory and Internalisation of the Environment

Auteurs: Anti Ingel, Abdullah Makkeh, Oriol Corcoll and Raul Vicente
Publié dans: Entropy, Numéro 10994300, 2022, ISSN 1099-4300
Éditeur: Multidisciplinary Digital Publishing Institute (MDPI)
DOI: 10.3390/e24030401

Interpretable Forecasting of Energy Demand in the Residential Sector

Auteurs: Nikos Sakkas; Sofia Yfanti; Costas Daskalakis; Eduard Barbu; Marharyta Domnich
Publié dans: Energies, Numéro 1, 2021, ISSN 1996-1073
Éditeur: Multidisciplinary Digital Publishing Institute (MDPI)
DOI: 10.3390/en14206568

Do Deep Reinforcement Learning Agents Model Intentions?

Auteurs: Tambet Matiisen; Aqeel Labash; Daniel Majoral; Jaan Aru; Raul Vicente
Publié dans: Stats, Vol 6, Iss 1, Pp 50-66 (2022), Numéro 5, 2022, ISSN 2571-905X
Éditeur: MDPI
DOI: 10.3390/stats6010004

Drivers of and counterfactuals for the final energy and electricity consumption in EU industry

Auteurs: Sakkas, N., Athanasiou, N.
Publié dans: Academia Letters, Numéro 27719359, 2021, ISSN 2771-9359
Éditeur: Academia.edu
DOI: 10.20935/al3451

Thesis and dissertations (4)

Genetic Programming Approaches for Solving Transportation Problems

Auteurs: Catarina Furtado Martins da Rocha Leite
Publié dans: 2022
Éditeur: University of Porto

Using Dimensionally Aware Genetic Programming to find interpretable Dispatching Rules for the Job Shop Scheduling Problem

Auteurs: Álvaro Manuel Festas Pereira da Silva
Publié dans: 2021
Éditeur: University of Porto

A generic scalable web platform for XAI algorithms

Auteurs: Luís Pedro Viana Ramos
Publié dans: 2022
Éditeur: University of Porto

Interface Design for Human-guided Explainable AI

Auteurs: João Rafael Gomes Varela
Publié dans: 2022
Éditeur: University of Porto

Book chapters (1)

Simultaneous Model-Based Evolution of Constants and Expression Structure in GP-GOMEA for Symbolic Regression

Auteurs: Johannes Koch, Tanja Alderliesten, Peter A. N. Bosman
Publié dans: Lecture Notes in Computer Science, Parallel Problem Solving from Nature – PPSN XVIII, 2024, Page(s) 238-255
Éditeur: Springer Nature Switzerland
DOI: 10.1007/978-3-031-70055-2_15