Welcome address by Wen-Chung Kao, a chair professor with the Department of Electrical Engineering, National Taiwan Normal University, New Taipei, Taiwan.
Responsible and Trustworthy Artificial Intelligence (AI) has become a critical topic in policy, industry, and the sciences, as it is crucial for the adoption and acceptance of AI systems in various applications such as healthcare, business, government, and education. A trustworthy AI requires a range of properties such as safety, justice, explainability, human-centeredness, beneficence, autonomy, robustness, fairness, transparency, non-discrimination, promotion of social and environmental well-being, non-maleficence, and sustainability. Global initiatives put the aspect of responsible AI development at the focus of their efforts, and Dubravko will provide an overview of these initiatives and Serbia's role in AI.
"AI is about to change how you use computers completely", said Bill Gates on November 09, 2023. Today, we are witnessing that such change is near. We are confident that the Retrieval Augmented Generation (RAG) method will advance this discourse greatly. The integration of retrieval mechanisms with generative capacities is a significant step toward the mentioned change, amplifying AI's ability to produce responses that are relevant to the context. Having said such strong averment, we will explore the principles of RAG in a discussion aimed at clarifying the complexities of developing RAG-based applications. The talk will debunk these applications, offering insights into the use of both proprietary and open-source large language models, and highlighting areas where RAG technology could play a transformative role. It will also detail the development lifecycle, identify industry areas suited for RAG application, and share success stories from our innovation hub. The talk will also address ethical considerations in RAG application engineering, focusing on the ethical aspects of AI and dissecting relevant regulations, such as the notorious EU AI Act. This talk is designed as a synthesis of practical experience, academic insights, and industry knowledge, aiming to connect these domains. It is ideal for those interested in AI development and the societal impact of AI applications. Attendees will gain a comprehensive understanding of RAG's role in GenAI, exploring the intersection of industry practices and ethical frameworks. So, has AI changed how we use computers in just as little time as three months? Yes. Will it continue to do so henceforward? Yes. And we need to (light)talk about it.
We will discuss the challenges of delivering an OS to an ecosystem of OEMs including ensuring the quality for 400+ commercial brands spanning across 10+ manufacturers. From code management, individual customization to final certifications.
We live in a world being redefined by software. In the past, connecting endpoints over a distance was the big challenge. The telephone system is the high point of that generation. The endpoints were just microphones and speakers, which are dependent upon the network to preserve the waveform.
Today we live in a very different world in which our devices are smart, and we no longer depend on a network provider's smarts. In fact, those smarts create problems such as buffer bloat. We now implement systems as a collection of devices and are no longer confined to a single box. The components can be scattered across the world and can be part of multiple systems.
This new way of thinking is at odds with an infrastructure that puts barriers in the path (subscriptions) and second-guesses innovation by trying to "add value". This is even more the case as AI has accelerated the pace of innovation.
The Matter protocol, formerly known as Project CHIP (Connected Home over IP), represents a significant advancement in smart home technology standardization. By unifying various smart devices under a common communication standard, Matter aims to simplify device setup, enhance cross-compatibility, and bolster data privacy for consumers.
This presentation aims to provide a comprehensive overview of this emerging standard and its implications for the future of smart home technology (and possibly wider). We will explore the key technical details of the Matter protocol, various communication standards it uses, standardized device profiles, and security mechanisms. The potential benefits of Matter, such as seamless device integration across different ecosystems, improved user experience, and increased market adoption of smart home devices will be tackled. However, alongside its promising features, the Matter protocol also faces challenges that need to be addressed, like user experience, maintaining backward compatibility with existing smart home ecosystems, and ensuring the maturity and evolution of the Matter ecosystem.
SAE J1939 is the backbone of heavy-duty vehicles, such as trucks, construction equipment, and other off-road machinery. It offers a transmission ranging from 250 Kbps up to 1 Mbps of highly structured data about all aspects of the machinery.
SAE J1939 is a high-layer protocol modelled on the ISO/OSI network protocol stack with the physical layer functionalities being realized by the CAN protocol. CAN protocol, which relies on the shielded twisted-pair wiring has brought in the reduced wiring combined with the prevention of message collision (no data lost). The SAE J1939 keeps building atop it.
Thanks to its structure, the SAE J1939 protocol is easy to integrate with various diagnostic tools.
The protocol has already proven itself on many levels within the machines' inter-communications, as well as off-duty diagnostics. But, in the era of the Internet of Things, what are the potential fruits of this protocol propagated to the Cloud? The already known fact is that the protocol is about smart communication - enabling filtering methods such as source address, parameter group filters, and message ID, leading to efficient data traffic. But, what are the benefits of the remote diagnostics in industrial IoT that SAE J1939 protocol brings in?
The presentation will delve into the challenges faced by the European automotive industry, highlighting the lack of software expertise and complex supplier interdependencies. It will propose continuous software integration and verification (CX) as a solution to improve software KPIs and enhance competitiveness in the Software Defined Vehicle (SDV) race. The insights are drawn from Pickhard's extensive practical experience rather than academic theory.
Software container technologies, such as containers and Kubernetes, are revolutionizing the way applications are developed, deployed, and managed. This presentation provides a concise introduction to containerization, discussing its benefits, such as increased portability, scalability, and security. We will explore the key concepts of containerization, including container images, registries, and orchestration tools. Moreover, this presentation will delve into the practical applications of containers within the domain of Industrial IoT. From edge computing devices to the cloud, containers play a pivotal role in enabling seamless deployment and management of IoT applications across diverse environments. We will discuss how containers facilitate the encapsulation of IoT software components into lightweight, portable units, ensuring consistency and reliability across the entire IoT ecosystem. In addition to discussing the benefits of containers in Industrial IoT, we will also address the unique challenges and considerations associated with containerizing IoT applications. From resource-constrained edge devices to stringent security requirements, we will explore strategies for overcoming these challenges and optimizing the deployment of container-based IoT solutions.
Keynote expert address by Dr Sven Laemmermann, Functional Algorithm and Software Development Group Lead in ZF Serbia.
In recent years, car manufacturers have been refining their in-vehicle infotainment systems to offer unique user experiences. Traditionally, these systems relied heavily on Linux. However, with the rise of Android's popularity mostly because of its rich ecosystem of apps, there's a growing interest in incorporating Android into these systems. There are three primary approaches to integrating Android into in-vehicle infotainment: Complete replacement of Linux with Android, Hypervisor solutions, and Containerized approach. The pros and cons of the approaches mentioned will be discussed along with the Company's experiences with each of them. Some of the metrics would be time to market, resource consumption, architecture, etc. Additionally, the most common implementation technologies would be given. What do You think? Which solution will prevail in the market?
OTT technology is slowly taking over and becoming massively used for streaming on various multimedia platforms. In today's modern era streaming is possible on almost every device which has internet connection (e.g. mobile phones, TV/STB, personal computers, cars equipped with infotainment systems).
The presentation aims to explain challenges in the industry with latency while watching live OTT content. It highlights complexity of the issue to reduce live latency and presents one of the solutions how OTT player was extended to solve the problem. Technology can make significant impact on every day's life in some specific fields / areas. Examples are live events broadcasting (e.g. sport events, concerts), online betting. However, there are many challenges handling the issue, especially in times where still different transportation medias (RF signal, IPTV and OTT) are used for same content distribution. Complexity of the solution requires improvements that need to be made in every node of an AV pipe - from source, encoders, etc., down to user's device and player on it.