ETSI, Sophia Antipolis, France The ETSI Artificial Intelligence (AI) Conference - Status, Implementation and Way Forward of AI Standardization, took place physically on 5-7 February 2024 in ETSI, Sophia Antipolis, France.
Artificial Intelligence / Machine Learning (AI/ML) technology has made major advancements in recent years and is becoming a key enabler for new disruptive features and applications in any digital product.
This face-to-face event offered attendees the unique opportunity to come together to exchange with experts, network with peers, visit the demos and posters around the subject of Machine Learning with a specific focus on Information and Communications Technology (ICT).
AI/ML will fundamentally shift the Information and Communications Technology (ICT) industry, as technologies such as Natural Language Processing, Deep Learning, and others will enhance the performance of communications, applications, content, and digital commerce.
Artificial Intelligence / Machine Learning are already part of the networks and services, and are present across multiple sectors such as finance, manufacturing, medical, telecommunications and transportation. AI Chatbots and AI voice assistants are widely deployed in society. Still this field is quickly evolving and the event will provide an opportunity for experts to have insights into the latest developments.
Europe and other regions are drafting policy and legislation which will impact the deployment of artificial intelligence. In the context of the draft AI Act, the EU puts forward objectives including:
- “ensure that AI systems placed on the Union market and used are safe and respect existing law on fundamental rights and Union values";
- "facilitate the development of a single market for lawful, safe and trustworthy AI applications and prevent market fragmentation”.
Attendees learned about the latest status and had the opportunity to meet and discuss with experts in this field.
The event presented the regulations in the different regions of the world and provided the opportunity to get involved in standardization that will play an important role in the context of European and global regulation. The topics also addressed the corresponding new market access requirements, conformity assessment, etc. The event was also an opportunity for research organizations (academia and industrial research) to discuss new trends and features in the field of AI.
The conference presented available work in the domain of AI. In this context, ETSI’s experts presented published and ongoing work on cybersecurity, testing of AI, trustworthy AI, sustainable AI, applications of AI for verticals and many more.
Demonstrations and Posters on the latest advances in the field of AI/ML applied to ICT systems and applications complemented the conference programme.
Demos and posters complemented the conference agenda. Event attendees had the opportunity to visit the demos and posters during the conference coffee and lunch breaks.
- Day 1: Monday 5 February
11:00 - 12:50 SESSION 1: Opening Address, Setting the Scene
Chaired by Markus Mueck, ETSI OCG-AI Chair, ETSI Board Chair, Programme Committee Chair- 11:00 Opening Address
- Luis Jorge Romero, ETSI Director General
- Markus Mueck, Programme Committee Chair
- 11:20 Keynote: Generative AI – Is It Even Possible To Do That Responsibly?
Hans Petter Dalen, IBM - 11:50 Keynote: Generative AI, from Vision to Language
Pietro Michiardi, Eurecom - 12:20 Keynote: Standardization Work in Support of the AI Act
Antoine-Alexandre Andre, DG CNECT, European Commission
12:50 - 14:15 Lunch & Networking Break - Demos and Posters Visits
14:15 - 15:55 SESSION 2: AI Regulation on a Global Level/Across regions
Chaired by Gabrielle Owen, Dutch Authority for Digital InfrastructureThis session gives an overview of the approaches in different regions with regard to AI regulation.
- 14:15 China AI Policies and Standardization
Betty Xu, SESEC - 14:35 Artificial Intelligence: Status Update from India
Dinesh Chand Sharma, EU Project SESEI - 14:55 The UK Government’s approach to the cyber security of AI
James Deacon, DSIT - 15:15 The NIST AI Risk Management Framework
Jesse Dunietz, NIST - 15:35 Overview of the Japanese position on AI
François Ortolan, NEC Labs Europe
15:55 - 16:25 Coffee & Networking Break - Demos and Posters Visits
16:25 - 17:40 SESSION 3: Standardization in Support of the Upcoming AI Act and the Legal Frameworks Outside EU
Chaired by Francois Ortolan, ETSI OCG-AI, NEC Labs Europe LtdThis session focuses on international and European standards supporting the locally applicable regulation of AI.
- 16:25 Status of JTC21 Activities in Response of the Standardization Request on AI
Kim Skov Hilding - Patrick Bezombes - Sebastian Hallensleben - CEN CENELEC JTC 21 - 16:40 Cooperation between ETSI and JTC21
Markus Mueck, ETSI OCG-AI Chair, ETSI Board Chair - 16:55 AI Overview of ISO/IEC SC42
Wael Diab, ISO/IEC SC42 Chair - 17:10 Q&A and Panel Discussion with the session speakers
17:40 Presentation from StandICT.eu and HSBooster Projects
Maria Giuffrida, StandICT.eu & HSBooster17:45 Networking Cocktail and End of Day 1
- 11:00 Opening Address
- Day 2: Tuesday 6 February
09:00 - 10:30 SESSION 4: Cyber Security in the Context of AI
Chaired by Ray Forbes, ETSI OCG AI, Programme Committee MemberWith AI becoming an ubiquitous tool in software deployment, this session will discuss the related cyber security challenges.
- 09:00 Current and Future Activities of ETSI TC SAI (Securing Artificial Intelligence)
Scott Cadzow, ETSI TC SAI Chair - 09:20 How to Deal with Cyber AI. Overlaps of AI with the standardization work in ETSI TC CYBER
Samim Ahmadi - 09:40 European Approach to AI Cybersecurity
Corina Pascu, ENISA - 10:00 Q&A / Panel discussion with the session' speakers
10:30 - 11:00 Coffee & Networking Break - Demos and Posters Visits
11:00 – 12:45 SESSION 5: Applications of AI in Cellular and Communications Networks
Chaired by Andreas Maeder, NokiaThis session discusses the applications of AI from the perspective of network manufacturers and operators.
- 11:00 Trustworthy AI in a Telecom Network
Andreas Maeder, Nokia - 11:20 Mobile Network Operator View on AI
Ahmet Faruk Coskun, Turkcell - 11:40 Overview of AI Specification in Experiential Networked Intelligence (ENI)
Ray Forbes, ETSI ISG ENI Chair (Experiential Networked Intelligence) - 12:00 The role of AI in Network Automation - The ZSM Case
Diego Lopez, ETSI ISG ZSM Chair (Zero Touch Network and Service Management) - 12:20 Q&A / Panel discussion with the session' speakers
12:45 - 14:15 Lunch & Networking Break - Demos and Posters Visits
14:15 – 16:00 SESSION 6: Applications of AI in Various Sectors
Chaired by Suno Wood, ETSI TC eHealth ChairThis session discusses the applications of AI in the health, industrial automation, automotive and aerospace sectors.
- 14:15 eHEALTH and AI - New Horizons 2024
Suno Wood, ETSI TC eHealth Chair - 14:35 Managing the European Airspace with AI
Benjamin Cramet, EUROCONTROL - 14:55 AI in Intelligent Transport
Niels Skov Andersen, ETSI TC ITS Chair (Intelligent Transport Systems) - 15:15 Potentials and Challenges of Industrial AI
Marco Ulrich, ABB - 15:35 Q&A / Panel discussion with the session' speakers
16:00 - 16:30 Coffee & Networking Break - Demos and Posters Visits
16:30 - 18:00 SESSION 7: Novel AI Features, and Related Conformity Assessment
Chaired by Dorian Knoblauch, ETSI MTS AI Working Group ChairThis session's objective is understanding how AI related requirements can be tested and validated in the context of conformity assessment.
- 16:30 MTS AI Group's Efforts in Testing, Auditing, and Documentation Schemes
Dorian Knoblauch, ETSI MTS AI WG Chair, ETSI TS 104 008 Rapporteur - 16:50 3GPP (WI on AI/ML in the Network)
Puneet Jain, 3GPP SA Chair - 17:10 Future Conformity Assessment for AI Safety and Security
Xavier Valero, DEKRA - 17:30 Q&A / Panel Discussion with the session' speakers
17:45 Networking Drinks and End of Day 2
- 09:00 Current and Future Activities of ETSI TC SAI (Securing Artificial Intelligence)
- Day 3: Wednesday 7 February
09:00 – 10:45 SESSION 8: Deep Dive on Selected Topics
Chaired by Jochen Friedrich, Programme Committee Member, ETSI Board MemberIn ETSI’s work to date several complex peculiarities of AI have been uncovered. The intent in this session is to look in some detail at these topics and address trustworthiness, eco impact, ethics and similar to open debate on the relevance of standardisation.
- 09:00 An Enterprise Software Framework to Manage AI Governance at Scale
Martijn Wiertz, IBM EMEA Technical Sales Lead, AI governance - 09:20 Sustainable AI: Adding a Conscience
Greg Ainslie-Malik, SPLUNK - 09:40 AI and Moral Decision-Making
Bernard Sinclair-Desgagné, Skema Business School and Université Côte d’Azur (GREDEG) - 10:00 New Technological Trends and Standardization Activities at ITU on AI
Frederic Werner, ITU - 10:20 Q&A / Panel Discussion with the session' speakers
10:45 - 11:15 Coffee & Networking Break - Demos and Posters Visits
11:15 – 12:45 SESSION 9 - PANEL DISCUSSION: New ETSI Initiatives in the Field of AI
Moderated by Markus Mueck, ETSI OCG-AI Chair, ETSI Board Chair, Programme Committee ChairThe session will give an overview of the currently ongoing and planned ETSI activities in the field of AI. It is an opportunity for new stakeholders to better understand the current ETSI Standardization landscape and get engaged.
Panelists include:
- Scott Cadzow, ETSI TC SAI Chair (Securing Artificial Intelligence), Generative AI
- Ray Forbes, ETSI ISG ENI Chair
- Dorian Knoblauch, ETSI MTS AI Working Group (Methods for Testing & Specification - AI)
- Franck Le Gall, ETSI ISG CIM Chair (cross-cutting Context Information Management(CIM) Layer), Knowledge enabled AI
- Diego Lopez, ETSI ISG ZSM Chair (Zero-touch network and Service Management)
- Alain Mourad, ETSI ISG ISAC Chair (Integrated Sensing And Communications)
- Arman Shojaeifard, ETSI ISG RIS Chair (Reconfigurable Intelligent Surfaces), Smart surfaces/smart elements
12:45 ETSI AI Conference Wrap-up and Closing Remarks
Presented by Markus Mueck, Programme Committee Chair13:00 End of ETSI AI Conference
- 09:00 An Enterprise Software Framework to Manage AI Governance at Scale
Venue:
ETSI Headquarters
650 Route des Lucioles
06560 Valbonne - Sophia Antipolis
FRANCE
Tel: +33 4 92 94 42 00
Information on venue, travel, visa, local accommodation is available at www.etsi.org/events/find-us
The Conference attendees had the opportunity to visit the following demos and posters during the networking coffee and lunch breaks.
DEMOS:
- Generative AI based enterprise knowledge hub - SriLanka Telecom Mobitel PLC (Anil Pradeep Kumara and Navindu Madanayaka)
- Querio AI is a Generative AI based enterprise knowledge hub. It is a centralized platform that serves as a repository of information and resources. This innovative solution is developed with Generative AI and LLM to introduce the capacity for data-driven decision-making through AI-driven analysis of substantial datasets.It's designed to facilitate knowledge sharing, collaboration, and access to information within a particular organization, community, or field. This solution can take information in the form of documents, Web URLs,..etc depending on their purpose and the context. Most of the time information generated within the organization is huge and accessibility of those is a challenge. This is the solution for any organization where their most of the knowledge is scattered and accessibility takes time. This Generative AI based enterprise knowledge hub is quickly joining the workplace and it boosts in productivity and innovation. Querio AI can be in many forms such as sales assistant, call center assistant, field assistant, financial assistant, .etc depending on the information fed and context that employed.
- Querio AI is a Generative AI based enterprise knowledge hub. It is a centralized platform that serves as a repository of information and resources. This innovative solution is developed with Generative AI and LLM to introduce the capacity for data-driven decision-making through AI-driven analysis of substantial datasets.It's designed to facilitate knowledge sharing, collaboration, and access to information within a particular organization, community, or field. This solution can take information in the form of documents, Web URLs,..etc depending on their purpose and the context. Most of the time information generated within the organization is huge and accessibility of those is a challenge. This is the solution for any organization where their most of the knowledge is scattered and accessibility takes time. This Generative AI based enterprise knowledge hub is quickly joining the workplace and it boosts in productivity and innovation. Querio AI can be in many forms such as sales assistant, call center assistant, field assistant, financial assistant, .etc depending on the information fed and context that employed.
- AI-enabled Cyberattack Detection in Open-Radio Access Network - University of Western Macedonia (Dimitris Pliatsios):
- The open radio access network (O-RAN) is an industrial standard for RANs that specifies the interconnections and interfaces among various components towards increasing network flexibility and enhancing interoperability between equipment. Additionally, O-RAN facilitates the integration of network intelligence and programmability, by leveraging artificial intelligence (AI) and network softwarization approaches, enabling autonomous RAN orchestration. O-RAN is a critical component of Beyond-5th and 6th generation (B5G/6G) mobile networks as it is the main interface between the user equipment (UE) and the mobile network. Consequently, it is a primary target for adversaries seeking to disrupt communications or compromise data integrity. The utilization of AI techniques is a promising asset in detecting and addressing the impacts of potential cyberattacks against critical infrastructure, such as a telecommunications network. As a result, the combination of AI and the network status data obtained through O-RAN can lead to the development of novel countermeasures against cyberattacks.
- The open radio access network (O-RAN) is an industrial standard for RANs that specifies the interconnections and interfaces among various components towards increasing network flexibility and enhancing interoperability between equipment. Additionally, O-RAN facilitates the integration of network intelligence and programmability, by leveraging artificial intelligence (AI) and network softwarization approaches, enabling autonomous RAN orchestration. O-RAN is a critical component of Beyond-5th and 6th generation (B5G/6G) mobile networks as it is the main interface between the user equipment (UE) and the mobile network. Consequently, it is a primary target for adversaries seeking to disrupt communications or compromise data integrity. The utilization of AI techniques is a promising asset in detecting and addressing the impacts of potential cyberattacks against critical infrastructure, such as a telecommunications network. As a result, the combination of AI and the network status data obtained through O-RAN can lead to the development of novel countermeasures against cyberattacks.
- Deceiving Post-hoc Explainable AI (XAI) Methods in Network Intrusion Detection - University College Dublin (Thulitha Millawa Gamage):
- Artificial Intelligence used in future networks is vulnerable to biases, misclassifications, and security threats, which seeds constant scrutiny in accountability. Explainable AI (XAI) methods bridge this gap in identifying unaccounted biases in black-box AI/ML models. However, scaffolding attacks would hide the internal biases of the model from XAI methods, jeopardizing any auditory or monitoring processes, service provisions, security systems, regulators, auditors, and end-users in future networking paradigms, including Intent-Based Networking (IBN). For the first time ever, we formalize and demonstrate a framework on how an attacker would adopt scaffoldings to deceive the security operators in Network Intrusion Detection Systems (NIDS). Furthermore, we propose a detection method that auditors can use to detect the attack efficiently. We rigorously test the attack and detection methods using the NSL-KDD. We then simulate the attack on 5G network data. Our simulation illustrates that the attack adoption method is successful, and the detection method can identify an affected model with extremely high confidence.
- Artificial Intelligence used in future networks is vulnerable to biases, misclassifications, and security threats, which seeds constant scrutiny in accountability. Explainable AI (XAI) methods bridge this gap in identifying unaccounted biases in black-box AI/ML models. However, scaffolding attacks would hide the internal biases of the model from XAI methods, jeopardizing any auditory or monitoring processes, service provisions, security systems, regulators, auditors, and end-users in future networking paradigms, including Intent-Based Networking (IBN). For the first time ever, we formalize and demonstrate a framework on how an attacker would adopt scaffoldings to deceive the security operators in Network Intrusion Detection Systems (NIDS). Furthermore, we propose a detection method that auditors can use to detect the attack efficiently. We rigorously test the attack and detection methods using the NSL-KDD. We then simulate the attack on 5G network data. Our simulation illustrates that the attack adoption method is successful, and the detection method can identify an affected model with extremely high confidence.
- Explainable AI-based Data Poisoning Attacks Detection and Defence: Approaches for Better Accountability in Robust Algorithms for Federated Learning - University College Dublin (Chamara Sandeepa):
- Next-generation networks like Beyond 5G (B5G)/6G are envisioned to be fully AI-driven. Therefore, it can be expected to rely highly on up-to-date ML models, continuously trained by a distributed and privacy-preserved architecture like Federated Learning (FL). However, a surge of new threats to FL's security and privacy has been observed over recent years. Data poisoning is a critical attack aiming to manipulate client model updates from their training objective. This results in the degradation of model utility and increased vulnerability to privacy attacks. This demo aims to introduce a novel approach to detecting and defending against poisoning attacks via Explainable AI (XAI) using SHAP to identify anomalous clients and eliminate them from an FL system. We demonstrate the potential of adopting post-hoc explanations as a part of a robust aggregation algorithm to filter poisoning clients and enhance the explainability of decisions effectively.
- Next-generation networks like Beyond 5G (B5G)/6G are envisioned to be fully AI-driven. Therefore, it can be expected to rely highly on up-to-date ML models, continuously trained by a distributed and privacy-preserved architecture like Federated Learning (FL). However, a surge of new threats to FL's security and privacy has been observed over recent years. Data poisoning is a critical attack aiming to manipulate client model updates from their training objective. This results in the degradation of model utility and increased vulnerability to privacy attacks. This demo aims to introduce a novel approach to detecting and defending against poisoning attacks via Explainable AI (XAI) using SHAP to identify anomalous clients and eliminate them from an FL system. We demonstrate the potential of adopting post-hoc explanations as a part of a robust aggregation algorithm to filter poisoning clients and enhance the explainability of decisions effectively.
- Towards Safer Large Language Models - NEC Laboratories Europe (Kiril Gashteovski):
- Large Language Models (LLMs) are revolutionizing our world. They have impressive textual capabilities that will fundamentally change how human users can interact with intelligent systems. Nonetheless, they also still have a series of limitations that are important to keep in mind when working with LLMs. We explore how these limitations can be addressed from two different angles. First, we look at options that are currently already available, which includes (1) assessing the risk of a use case, (2) prompting a LLM to deliver explanations and (3) enacsing LLMs in a human-centered system design. Second, we look at technologies that we are currently developing, which will be able to (1) more accurately assess the quality of an LLM for a high-risk domain, (2) explain the generated LLM output by linking to the input and (3) fact check the generated LLM output against external trustworthy sources.
- Large Language Models (LLMs) are revolutionizing our world. They have impressive textual capabilities that will fundamentally change how human users can interact with intelligent systems. Nonetheless, they also still have a series of limitations that are important to keep in mind when working with LLMs. We explore how these limitations can be addressed from two different angles. First, we look at options that are currently already available, which includes (1) assessing the risk of a use case, (2) prompting a LLM to deliver explanations and (3) enacsing LLMs in a human-centered system design. Second, we look at technologies that we are currently developing, which will be able to (1) more accurately assess the quality of an LLM for a high-risk domain, (2) explain the generated LLM output by linking to the input and (3) fact check the generated LLM output against external trustworthy sources.
- Enabling AI in the physical Layer for 6G: Learning channel estimation on a Sub-Terahertz Over-the-Air link - Keysight Technologies (Alan Anderson):
- Keysight Technologies present a 6G prototype which uses AI in the physical layer to handle the challenges of a complex channel in the untapped spectrum in the sub-THz frequency range. The setup includes an over-the-air link and an AI-powered receiver. The AI algorithms are able to cope with the complexity of the channel compensation problem by automatically learning about their surroundings and therefore is able to improve the link quality, which in turn boosts throughput and allows for power-aware optimisations. In the demo, we will illustrate how this method improves on the MMSE conventional algorithm baseline.
- Keysight Technologies present a 6G prototype which uses AI in the physical layer to handle the challenges of a complex channel in the untapped spectrum in the sub-THz frequency range. The setup includes an over-the-air link and an AI-powered receiver. The AI algorithms are able to cope with the complexity of the channel compensation problem by automatically learning about their surroundings and therefore is able to improve the link quality, which in turn boosts throughput and allows for power-aware optimisations. In the demo, we will illustrate how this method improves on the MMSE conventional algorithm baseline.
- Managing the European Aviation Network with AI - Eurocontrol (Benjamin Cramet):
- Eurocontrol, the European Organisation for the Safety of Air Navigation, plays a crucial role in predicting the Aviation Network and supporting the aviation industry in the development of AI models to enhance the efficiency of its operations. The organization provides valuable support to airlines, airports and air traffic controllers in the management of daily operations, using AI models to predict the impact of various factors such as weather conditions, delays, arrival times, taxi times, and passenger demand. These models are instrumental in ensuring smooh th and seamless operations, making Eurocontrol an indispensable partner for the aviation. https://www.eurocontrol.int/artificial-intelligence
- Eurocontrol, the European Organisation for the Safety of Air Navigation, plays a crucial role in predicting the Aviation Network and supporting the aviation industry in the development of AI models to enhance the efficiency of its operations. The organization provides valuable support to airlines, airports and air traffic controllers in the management of daily operations, using AI models to predict the impact of various factors such as weather conditions, delays, arrival times, taxi times, and passenger demand. These models are instrumental in ensuring smooh th and seamless operations, making Eurocontrol an indispensable partner for the aviation. https://www.eurocontrol.int/artificial-intelligence
- Resilient cobot-powered warehouse inventory/manufacturing management through AI/ML mechanisms - WINGS ICT Solutions (Christina Karousatou):
- The goal of this demo is to leverage Artificial Intelligence/Machine Learning (AI/ML) together with 6G capabilities, such as robotics and extended reality, to address the challenges of the rapidly growing landscape of warehouse inventory management and industrial manufacturing. The demo consists of collaborative robots (cobots), i.e., autonomous mobile robots (AMRs), as well as drones (UAVs) cooperating in conducting automated tasks in an industrial environment. The use case exhibited here, primarily concerns a Resource Allocation scenario, where the objective is to optimize the placement of a) the inventory management services, e.g., allocation of item scanning cobot role, and b) workloads requiring considerable computational resources (e.g., computer vision tasks), based on current workload, energy availability, hardware capabilities (e.g., ground/aerial node), as well as physical environment parameters, such as real-time proximity to the inventory locations.
- The goal of this demo is to leverage Artificial Intelligence/Machine Learning (AI/ML) together with 6G capabilities, such as robotics and extended reality, to address the challenges of the rapidly growing landscape of warehouse inventory management and industrial manufacturing. The demo consists of collaborative robots (cobots), i.e., autonomous mobile robots (AMRs), as well as drones (UAVs) cooperating in conducting automated tasks in an industrial environment. The use case exhibited here, primarily concerns a Resource Allocation scenario, where the objective is to optimize the placement of a) the inventory management services, e.g., allocation of item scanning cobot role, and b) workloads requiring considerable computational resources (e.g., computer vision tasks), based on current workload, energy availability, hardware capabilities (e.g., ground/aerial node), as well as physical environment parameters, such as real-time proximity to the inventory locations.
POSTERS:
- Cybersecurity for AI Applications - Cyber Governance Ltd (Rita Esen)
- Ethics-by-Design in AI-Driven Human-Robot Collaborative Manufacturing: Navigating Regulatory and Ethical Challenges - Eunomia Limited (Christina Nanou)
- Standardization: AI Act’s cornerstone - Eunomia Limited (Nikolaos Koulierakis)
- Multi-modal Latent Diffusion - Eurecom / Renault Software Labs (Mustapha Bounoua)
- AI Standards in Aviation - EUROCAE (Tri Thuc Nguyen)
- Cybersecurity Requirements for High Risk AI Systems: The Roadmap for an Holistic Approach - INESC TEC (Vasco Rosa Dias)
- 6G-XR - An evolvable experimental infrastructure of key B5G/6G candidate technologies - Intel Deutschland GmbH (Markus Mueck)
- Transforming ICT with Cutting-Edge AI Enablers: A 6G Perspective - Nextworks (Giacomo Bernini)
- Generative AI for Network AIOps - Omdia (James Crawshaw)
- Enhancing remote patient monitoring through Integrated AI and Open-Source platforms: A Detailed Examination of Complex Cardiovascular and COVID-19 Symptomatology - OpenNovations (Hans de Raad)
- EU AI Act’s High-Risk AI Applications and Harmonised Standards - ADAPT Centre, Trinity College Dublin (David Lewis)
- 6G Revolution in Healthcare: Bridging the Healthcare Data Chasm - Turkcell İletişim A.Ş. (Buse Bilgin)
- Resilient Mobile Networks under Earthquake: Next-Generation Disaster Management - TURKCELL (Buse Bilgin)
- End-to-end AI for Intelligent 6G Networks - University of Bristol (Juan Parra)
- AI-powered evolution towards open and secure edge architectures - UPC - Polytechnic University of Catalonia (Irene Vila Muñoz)
- Leveraging Machine Learning for Throughput Prediction in Cellular Networks - University College Cork (Cormac Sreenan)
- Preserving Privacy of Federated Learning Systems via Recommendation-based Defence - University College Dublin (Chamara Sandeepa)
- Deceiving LIME Explainable AI Method in Network Intrusion Detection - University College Dublin (Thulitha Millawa Gamage)
- Artificial intelligence for optical wireless communication - University of Strathclyde (H. Safi)
INFOSTAND
- Markus Mueck, Programme Commitee Chair, ETSI OCG-AI Chair, ETSI Board Chair, INTEL (Germany)
- Alexander Cadzow, ETSI TC SAI Secretary, Cadzow Communications Consulting Ltd (UK)
- Scott Cadzow, ETSI TC SAI Chair, Cadzow Communications Consulting Ltd (UK)
- Ray Forbes, ETSI OCG-AI Vice Chair, ETSI ISG ENI Chair, Huawei (UK)
- Jochen Friedrich, ETSI Board Member, IBM (Germany)
- Antoine Mouquet, ETSI OCG-AI and ETSI ISG SAI Technical Officer, ETSI (France)
- François Ortolan, ETSI OCG-AI Vice Chair, NEC (Germany)
- Helene Schmidt, ETSI OCG-AI Support Coordinator, ETSI (France)
Presentations made during the event are available in PDF format.