On 27 and 28 March, AI4CCAM took part in HyCHA’24 – Hybrid Artificial Intelligence Workshop: From knowledge and human integration to model explanation (Knowledge, Human and Machine Learning Hybridization), held in Gif-sur-Yvette (France), co-organised among others by SystemX, AI4CCAM partner.
AI4CCAM was represented by Arnaud Gotlieb, Simula, AI4CCAM coordinator who participated in the event with a poster.
Hybrid Artificial Intelligence (AI) broadly encompasses all approaches that combine several AI methods, whether symbolic or numerical, qualitative, semi-qualitative or quantitative. It is a field that is currently enjoying renewed interest, particularly because the hybridization is a way of addressing the respective weaknesses of different approaches and tackling certain current AI challenges such as trust and explainability.
In Automated Driving, scene understanding is a crucial issue. However, today’s explanations are based on quantitative analysis: Heatmaps, importance map, occlusion analysis, etc. But there is still no understanding of events chain and consequences. Given an action taken by an ego-car in a scene, the goal is to automatically explain the interactions between the actors that led to such action Decision Making Explanation Action.
AI4CCAM is based on Qualitative Reasoning.
Given a scene, we apply Qualitative Constraint Acquisition on each frame Then, we create the Qualitative eXplainable Graph (QXG) from sensor data.
Given a set of QXGs and actions (cruising, accelerating, etc.), we train one-class classifiers to give action-labels to each object relation chains. These action labels can be helpful to automatically document the driving scene.
Considering action-labels with relation chains, our method automatically proposes possible ego-car action explanations, by considering interactions between the car and each object.
User feedback on possible explanations allows us to strengthen the result and properly document automated driving scenes.
For further information on the event, click here
Check the project library for the Poster!
AI4CCAM will be joining the 8th edition of the AUTONOMY Mobility World Expo, held in Paris on 20 and 21 March.
The event is among the world’s largest annual gathering of international policymakers, institutions, NGOs, corporations, companies, and start-ups focused on sustainable urban mobility solutions welcoming 250+ exhibitors, 400+ speakers, and 7,000+ participants every year in March.
AUTONOMY MOBILITY WORLD EXPO (AMWE) combines exhibition stands with a Startup Village, 5 conference stages, test tracks, demos, B2B & B2G meetings, and also the Startup Challenge and Innovation Awards ceremonies.
AI4CCAM will be represented by Arnaud Gotlieb, Simula, the project coordinator, who will be joining the session on “Steering Tomorrow: The Roadmap to Autonomous Vehicle Integration in Cities”.
This session explores the transformative impact of Autonomous Vehicles (AVs) on urban mobility. We’ll discuss advancements in AV technology, regulatory challenges, and integration strategies. Experts will share insights on improving urban transportation efficiency, ensuring safety, and the benefits of AV deployment in city landscapes.
For further information on the event, click here
On 11 March, the AI4CCAM Ethical and Scientific Advisory Board (ESAB) held its second meeting, in Paris.
The ESAB is led by Prof. Jean-Gabriel Ganascia, Professor of Computer Science at Sorbonne University. The ESAB also involves Margriet Van Schijndel-de Nooij, Program director Responsible Mobility bij Technische Universiteit Eindhoven; and Jerome Perrin, Recherche scientifique et développement industriel dans les domaines énergie, environnement, mobilité – Associations caritatives – Ethique et Théologie.
The scope of the Advisory Board is to maintain an accurate body of knowledge on AI in CCAM operations, enabling the continuous consideration and analysis of ethical aspects and also facilitating interactions with relevant associations and policy-makers.
The meeting started with a presentation of the follow-up actions on the ESAB initial recommendations, then the main focus was on AI4CCAM methodology and its application to the three use cases of the project. A first presentation was given by Simula, the project coordinator, then followed by three more presentations by project members on the AI4CCAM methodology for Trustworthy AI in CCAM, Use Case Experimentation and Evaluation and AI ethical risks in CCAM.
The ESAB reported to be very satisfied with the project activities, including communication and dissemination towards the CCAM community. Before ending the meeting, the board also provided the coordination team with recommendation to connect with other relevant initiatives in the CCAM sphere.
AI4CCAM has a whole Work Package (WP4) dedicated to defining a validation process (including KPIs and methodological aspects from WP1 and WP3) that embraces the variety of CCAM Use Cases involving explainability and Trustworthy AI concepts.
Under this WP the AI4CCAM digital framework and AI models (from WP1 and WP2) will be evaluated in dedicated Use Cases representing simulated CCAM scenarios and collect all necessary evidence for validation. An impact assessment in line with the use case scenarios will be performed as well.
Three complementary use cases also covering user acceptance will be implemented:
- Human-vehicle AI-based interactions and ethical dilemmas
- Advanced AI models for CCAM situation awareness and VRUs behavior anticipation
- Car trajectory prediction.
This short video by Akkodis showcases a scenario where a VRU (bicyclist) is occluded by a larger vehicle (a bus) and the challenge is to identify and predict the risk involved to the VRU by the autonomous vehicle.
Co-organised by IRT SystemX, AI4CCAM partner, the Confiance.ai Day 2024 will be held on 7 March in Paris, acting as the meeting place for trusted AI. Several topics related to CCAM will be discussed, exchanging ideas and experiences with the French and international engineering and scientific communities involved in trusted AI:
- Trustworthiness as a pillar of AI adoption in industry
- Industrial AI toward 2030: trustworthiness beyond Confiance.ai
- Norms and standards: current status and future challenges
- The future challenges of trustworthy AI
- Hybrid AI, safety, cybersecurity, or ethics: spotlight on some of the upcoming challenges to address in developing trustworthy AI in industry.
Confiance.ai Day will be a new opportunity to meet AI4CCAM partners and find out more about the project!
For further information, click here
AI4CCAM has a whole Work Package (WP4) dedicated to defining a validation process (including KPIs and methodological aspects from WP1 and WP3) that embraces the variety of CCAM Use Cases involving explainability and Trustworthy AI concepts.
Under this WP the AI4CCAM digital framework and AI models (from WP1 and WP2) will be evaluated in dedicated Use Cases representing simulated CCAM scenarios and collect all necessary evidence for validation. An impact assessment in line with the use case scenarios will be performed as well.
Three complementary use cases also covering user acceptance will be implemented:
- Human-vehicle AI-based interactions and ethical dilemmas
- Advanced AI models for CCAM situation awareness and VRUs behavior anticipation
- Car trajectory prediction.
Skoda is elaborating its crosswalk scenario, which will be used to challenge and validate the Trustworthy AI.
AI4CCAM partners met on 25 January in Athens for the third General Assembly of the project, hosted by Inlecom.
A crowdy room to discuss project developments and upcoming achievements, especially considering that AI4CCAM is now one year old: the perfect moment for summing up progresses made so far.
The main focus of the meeting was on the three AI4CCAM Use Cases (UCs), discussing actions and on complementary of UCs, as well as common methodology, and user acceptance in UCs.
Of course, other relevant technical aspects of the project were discussed as well, such as
Ethical AI issues in autonomous driving, models and datasets, technical architecture. However, AI4CCAM is also looking into CCAM Partnership and internationalization activities that were discussed too, along with dissemination and communication activities.
Missed our insightful webinar on 29 November on building trust in AI for automated vehicles? The replay is now available!
This webinar aimed to delve into the challenges and opportunities of AI in the context of road safety, congestion reduction, and environmental impact. It primarily focused on what it takes to build trust and the conditions for adoption of AI-driven solutions. It gathered 58 attendees, notably from the industry, and also some researchers (from the CEA for instance), beyond of course AI4CCAM partners.
The webinar also aimed to shed light on the psychological and emotional factors that influence the adoption of AI in automated vehicles. By understanding these dynamics, we can contribute to an informed and well-rounded perspective on AI-driven mobility.
In this 1-hour webinar, join Marc Eynaud, Arnaud Gotlieb, Lucie Regereau, and Isabelle Vallet as they unveil user acceptance nuances, emotions, and barriers in automated mobility.
- Discover key insights from the European project AI4CCAM (AI for Connected Cooperative and Automated Mobility). Arnaud Gotlieb, the project coordinator, emphasizes the importance of trustworthy AI for autonomous driving.
- Explore Lucie Regereau’s methodology, delving into the emotions of daily car users in Paris, Berlin, and Warsaw. Understand why user acceptance is crucial for automated driving’s future.
- Isabelle Vallet shares fascinating results on traffic regulations, attitudes, and representations’ impact on automated vehicle adoption. See how these factors shape perceptions and trust.
AI4CCAM interviewed Nafsica Papakosta, Project Manager and Communications Specialist at INLECOM, leader of the WP5, which involves Communication, Dissemination and Exploitation for the AI and CCAM ecosystems.
In this interview, Nafsica addresses the methodology employed to identify and assess the outcomes of the AI4CCAM project with the greatest exploitation and innovation potential.
What is the purpose of the AI4CCAM’s Innovation & Exploitation Methodology?
As AI4CCAM’s Innovation and Exploitation Manager, INLECOM employs the Innovation & Exploitation Methodology (IEM), a well-established and validated framework to efficiently address the AI4CCAM project’s objectives in the areas of Exploitation, Innovation Registry, and Patent Filing. This methodology enables the documentation and evaluation of all project outcomes based on the owners’ input, with the purpose of producing a collection of outcomes and assessing their potential for innovation and exploitation. The AI4CCAM IEM is comprised of 3 phases and is being used to account for both exploitable and innovative project outcomes.
How does the Outcomes Registry phase of the IEM methodology work?
The initial phase of the IEM focuses on the completion of the Outcomes Registry via a web form that consists of 11 to 15 questions which help us pinpoint AI4CCAM’s innovative outcome(s) and key exploitative results filled out by project partners The Outcomes Registry is essentially a list of significant expected project outcomes together with preliminary descriptions of their characteristics. Partners submit their expected outcomes to the registry and assist in categorizing outcomes according to a certain exploitation path. The first phase of AI4CCAM’s IEM Methodology identified 11 expected Key Exploitable Results (KERs).
A feedback procedure will then be established to assist partners in better addressing difficulties, exploitation queries, and concerns. Based on the material collected, the initial Innovation & Exploitation assessment is performed, based on seven criteria: Solution Readiness, Anticipated market Interest, Anticipated size of the market in question, Level of innovation, Competitive landscape, Market readiness, and Solution reproducibility and re-usability.
What is the purpose of the AI4CCAM Innovation Registry and how will the decision on which patents to file be made?
On the innovation, IP, and IPR management front, INLECOM will use its experience as an Innovation Manager to help establish the AI4CCAM Innovation Registry, which will include novel AI-driven models used in safety-critical CCAM applications.
The methodology for cataloguing and supporting AI4CCAM outputs with high innovation potential will be provided in the third phase of the IEM. The AI4CCAM consortium partners will be able to fully explore and comprehend the innovation potential of their IP assets as a direct result of the innovation management initiatives guided by INLECOM, which will evaluate the outcomes’ innovation and commercial ambitions. A preliminary assessment of the outcomes’ innovative dimension will take place using the criteria of utility, novelty, and non-obviousness.
We will then be collaboratively evaluating, scoring, and prioritizing the areas where we will focus the IP related actions that have a planned commercial trajectory and/or strategic interest to AI4CCAM partners.
INLECOM intends to formally protect the AI4CCAM outputs through the submission and issuance of two patents. Our work will be guided by the EU’s legislative framework and the patent filing procedural requirements, as well as significant experience from past projects.
