Towards an Automated and exPlainable ATM System

Risultati finali

Report including the final publishable summary report and all the technical activities performed during the whole project It includes an assessment of the project achievements towards the RD initial goals An initial version will be available at M22

TAPAS Validation Plan

This document develops a clear plan of the validation activities that will be conducted within the project for each one of the target domains CDR and ATFM including the scope of the validation approach and methodology relevant criteria to be assessed and humanspecific measures that need to be gathered for the defined scenarios

Consolidated Requirements and Functional Roadmap

This document will deliver a merge of contextual requirements from both the operational and technical use cases descriptions, as well as a roadmap of delegation of functionalities between human actors and the machine.

Use Cases Transparency Requirements

This document will consolidate the transparency requirements derived from the human machine interactions for each level of automation in the ATFM and CDR operational use cases, and that will need to be accomplished during the implementation activities. A first draft will be available at M6 for the ATFM transparency requirements.

TAPAS Validation Report

This document will consolidate the results of the analysis performed based on the feedback and data gathered during the execution of the ATFM and CDR experiments including objectives achieved further research and recommendations for derivation of transparency criteria A first draft will be available at M14 containing the preliminary results of the ATFM experiment

TAPAS Integrated Prototype

Document that contains the description of the ATFM and CDR prototype including functionalities and improvements derived from the validation activities conducted A first draft will be available at M11 describing the ATFM prototype

Principles for Transparency in AI/ML automation in ATM

This document constitutes the final identification of the requirements that will ensure that different transparencyexplainability criteria are fulfilled when AIMLbased models are implemented This document will take as input the results from the validation activities conducted along the project and will combine analysis and expertise to specify the requirements A first draft will be available at M14

Reference of XAI Methods

This document will provide an identification of promising state of the art explainable reinforcement learning methods and deliver a reference of the XAI techniques for detecting problems and prescribing solutions in the ATFM and CDR operational cases A first draft will be available at M12 delivering a reference for the ATFM prototype

Visualizations and Visual Analytics methods

Description of the Visualizations and Visual Analytics methods implemented to improve the explainability and understandability of XAI ATFM and CDR Operational Cases A first draft will be available at M12 containing the description of the visualizations and visual analytics methods implemented for the ATFM use case

Exploitation and Dissemination Plan

This deliverable comprises the planning of the dissemination exploitation and communication activities

TAPAS Use Cases Description

This document will describe in detail the use cases to be developed under TAPAS project: (a) from the operational point of view; and (b) from the technological perspective, providing also additional requirements and feasibility analysis of the solutions proposed.

Exploitation and Dissemination Report

This deliverable comprises the results of the implementation of the dissemination exploitation and communication activities according to the Exploitation and Dissemination Plan A first draft will be available at M24

Pubblicazioni

Automating the resolution of flight conflicts: Deep reinforcement learning in service of air traffic controllers

Autori: George Vouros, George Papadopoulos, Alevizos Bastas, Jose Manuel Cordero, Ruben Rodriguez Rodriguez
Pubblicato in: Volume 351: PAIS 2022, 2022
Editore: PAIS (Prestigious Applications of Intelligent Systems)
DOI: 10.48550/arxiv.2206.07403

Supporting Visual Exploration of Iterative Job Scheduling

Autori: Gennady Andrienko, Natalia Andrienko, Jose Manuel Cordero Garcia, Dirk Hecker, George A. Vouros
Pubblicato in: IEEE Computer Graphics and Applications, Numero 02721716, 2022, ISSN 0272-1716
Editore: Institute of Electrical and Electronics Engineers
DOI: 10.1109/mcg.2022.3163437

Explaining deep reinforcement learning decisions in complex multiagent settings: towards enabling automation in air traffic flow management

Autori: Theocharis Kravaris, Konstantinos Lentzos, Georgios Santipantakis, George A. Vouros, Gennady Andrienko, Natalia Andrienko, Ian Crook, Jose Manuel Cordero Garcia & Enrique Iglesias Martinez
Pubblicato in: Applied Intelligence, Numero 0924669X, 2022, ISSN 0924-669X
Editore: Kluwer Academic Publishers
DOI: 10.1007/s10489-022-03605-1

Visual Analytics for Human-Centered Machine Learning

Autori: N. Andrienko, G. Andrienko, L. Adilova and S. Wrobel
Pubblicato in: IEEE Computer Graphics and Applications, Numero 02721716, 2022, ISSN 0272-1716
Editore: Institute of Electrical and Electronics Engineers
DOI: 10.1109/mcg.2021.3130314

Explainable Deep Reinforcement Learning: State of the Art and Challenges

Autori: George A. Vouros
Pubblicato in: ACM Computing Surveys, Numero 03600300, 2022, ISSN 0360-0300
Editore: Association for Computing Machinary, Inc.
DOI: 10.1145/3527448

Deep Multiagent Reinforcement Learning Methods Addressing the Scalability Challenge

Autori: Theocharis Kravaris and George A. Vouros
Pubblicato in: Multi-Agent Technologies and Machine Learning, 2022
Editore: IntechOpen
DOI: 10.5772/intechopen.105627

Risultati finali

Pubblicazioni

Condividi questa pagina

Scarica