Programme 

Classic bots, such as chatbots, voicebots and smart assistants, are in vogue. For example, in call and service centers, companies automate answering questions, solving problems, or selling simple products. Generative AI promises to improve these solutions. But how does one deal with the sometimes-glaring errors? Foundation models play a crucial role in this. How can all three approaches be combined?

It's hard to open a social media application or news source without hearing about Generative AI. The hype is real, and growing! But how are we supposed to actually use Generative AI in a business, and is it even safe to use? This talk will cover evaluation techniques to determine the risks and opportunities for leveraging these new AI architectures at your organization, focusing on questions like data security and privacy, risk appetite, leveraging your own text data and tips to develop valuable and secure NLP applications.

Companies have been investing heavily in data teams and capabilities over the last few years. The hype around AI accelerates that even further. But intelligent solutions and fast, reliable decision-making need high-quality data. This session will share how introducing data product managers and “data as a product” thinking at HelloFresh helped to unlock the value of data faster.

In today's data-driven landscape, AI's transformative potential is undeniable. As AI systems steer critical decisions across industries, the imperative for trust in both their data foundation and outcomes grows. The presentation explores the interplay between data quality and AI reliability. Emphasizing high-quality, curated data's role in training AI models, it unveils risks tied to biased or incomplete information.

The development of driving functions for autonomous vehicles requires extensive testing and validation to ensure their reliability and safety. However, conducting tests in highly complex traffic situations can be difficult and time-consuming. For this reason, virtual assessment techniques have become increasingly important. By creating realistic virtual environments, developers can simulate complex traffic scenarios and evaluate the performance of autonomous driving functions. 

To do this, OEMs and suppliers need a data ecosystem that combines real-world traffic scenarios and accident scenarios. Automated processes as a One-Click-Application enable efficient testing, reducing human errors in data preparation and fusion. When evaluating the roadworthiness of driving functions, it is crucial to use comparable criticality criteria. This enables objective learning and testing of driver assistance systems and helps identify areas for improvement. By establishing standardized metrics for evaluating criticality, developers can more effectively measure and improve the performance and safety of driving functions. 

In addition, test tracks play a critical role in testing real-world driving scenarios of autonomous driving systems. A data ecosystem enables developers to bring systems developed in the virtual environment to the driving platforms with one click to recreate difficult situations in the real world and verify the functionality of driving functions. 

The audience will be introduced for the first time to the ability to easily apply, master and use complex procedures with TASC (Traffic Accident Scenario Community).

Since large language models broke into the public consciousness, lead-ers are increasingly seeking for ways to extend the benefits of generative AI ca-pabilities beyond the circle of experts, targeting deep domain experts who view these technologies as computational engines to drive breakthroughs in their field. Democratizing generative AI across the enterprise is doable. This is not just about putting new AI-based low/no-code tools and services into the hands of business users, but also about sharing vetted data and building data literacy across the en-terprise so they can understand the power of generative AI, develop the right use cases, and apply the results to drive business outcomes and insights. To prevent challenges and reduce the risks, smart investments are essential. In a highly demonstrative session, Alteryx will share its experience in helping business users build key applications on top of LLMs that were entirely finetuned on internal data.

As the new Generative AI technology grabs headlines and captures imaginations for businesses and consumers alike, organizations have a unique opportunity to harness the power of AI to drive value faster in their own businesses. The presentation provides an overview of the differentiated value of the Microsoft Cloud to help customers and the personas across their organization power their AI transformation, featuring practical examples from across Microsoft’s ecosystem.

AI driven applications brought immense benefits for both individual & society but also if not regulated or moderated, has potential to impact decisional autonomy and privacy that are core aspects of human right. The session will introduce the privacy threats from AI specific to personal information based use cases, how regulations are emerging globally, various controls needed to minimize harms, challenges in deploying the controls and approach to make it a positive sum game.

Successful risk mitigation requires more than just evaluating financial data. Since abrupt changes in the general environment can only be reflected in ratings with a delay, German Landesbanken also rely on the analysis of press data. In order to recognize credit risks at an early stage and avoid losses, early warning systems like Risk Guard are needed. This application calculates risk scores on a daily basis, forecasts significant changes in creditworthiness and generates warning signals with a lead time of up to one year for companies, banks, entire sectors, countries and regions. In this presentation we want to talk about the resilience and reliability of the developed news based model.

Schwarz Group is not only Europe's largest retailer (Lidl, Kaufland) but also highly vertically integrated thanks to its food manufacturing and waste management and recycling capabilities. Therefore, our supply chain can be considered end-to-end, which gives us the chance to follow a holistic approach that is both economically beneficial and sustainable in terms of a circular economy.

We employ artificial intelligence at multiple stages of the supply chain, e.g., by enabling automatic disposition both on store and warehouse level, to optimize logistics, or for a better allocation of recyclables. We will discuss our demand forecasting project built in-house as an example of what it takes to build and operate a scalable (>10 000 stores) AI solution in terms of project management, team setup, solution architecture, and machine learning assets. Specific challenges include the reliable handling of big data, a huge amount of models, and analytical edge cases.

We also elaborate on how the highlighted project is just one out of several examples of data products that we have built and brought into production, which help us to achieve the group's ambitious sustainability goals.

Generative AI has become a widely available and used technology. Cybercriminals use their substitutes WormGPT or FraudGPT to leverage AI for criminal activities in the Cyberspace. This lowers the entry barrier for less sophisticated cybercriminals on the one hand and enables higher automation and autonomous attack scenarios for organized cybercrime groups on the other hand. These developments force defenders of public and corporate networks to reduce their mean times to response and leverage the very same technologies in their favor. The presentation will describe the latest of these developments and demonstrate how threat researchers of Unit 42 collect and process unprecedented amounts of telemetry data to investigate new tools, tactics and procedures of cybercriminals to prevent attacks from happening. It also concludes how latest generative AI can be used to support network defence more efficiently.

How does CrowdStrike Data Science leverage machine learning to massively increase cost for adversaries?

To answer this question we’ll be taking a closer look at how CrowdStrike employs machine learning to detect and stop both file-based and fileless cyberattacks at multiple layers.

We'll also be doing a deep dive into a new technology called “AI Powered Indicators of Attack” which detects attacks based on the behaviors they exhibit.

"History never repeats itself, but it does often rhyme.", Mark Twain

This talk explores the paradox of the exponential development of AI and the challenges organisations face in harnessing the technology. It emphasises the importance of proactive transformation to remain competitive in this fast-paced digital era. What can we learn from the past for a future with AI in the enterprise? Using experiences from the start-up world as a case study, it looks at what to consider when applying AI.

We present Fujitsu's approach and current activities involving the whole quantum computing technology stack. In particular, we focus on the recent developments of a quantum gate computing chip based on superconducting transmon qubits which is done in collaboration with the RIKEN Center for Quantum Computing.

Artificial intellgience, machine learning, big data and cloud computing are becomming indespensible technologies for companies and thus, the global amount of data to process is exploding. Three factors are most critical for software in the future: 1) high performance, 2) low data storage costs, 3) simple implementation processes for a short time-to-market. Today’s modern programming languages are ready for it. They are fast, let us build almost any use-case, and running plain programs in the cloud is simple. As soon as we have to store data, everything changes dramatically. To process and store data, we are used to use database systems as of course, because it works for us already for decades. However, compared to programming languages, traditional database systems cannot keep pace. They are 1000x slower, lead to high complexity of the architecture, gigantic implementation effort, and exploading cloud costs. 

Databaseless in-memory data processing is changing this. Companies that need uncompromising performance for their business such as online banking and online gaming, are using this approach already very successfully, but individually implementing is effortful and expensive. 

At the Eclipse foundation, we built a framework that enables all companies to use Databaseless in-memory data processing now to build high performance software applications that provide up to 1000x faster queries and save up to 98% database costs in the cloud. In addition to that, this approach will revolutionary simplify and accelerate your entire software development process. Companies such as Allianz use it in production successfully already. Learn, how your business can also benefit.

The road is long and complex: 

What will a sustainable corporate strategy bring me, and, above all, what will it cost me? 

How do I break down my goals?

IBM helps with proven technology in strategic or operational planning to quantify in real time the outcome of, for example, financial costs and CO2e emissions. A planner can simulate in "what-if" scenarios and ultimately decide which of the scenarios fits best. An optimization engine helps find the optimal path for sustainability versus cost savings.

Get your hands dirty in this hands-on, fast-paced workshop where you’ll create a chatbot persona (and a chatbot) from scratch. Together, we’ll delve into the world of prompt design, mastering the art of crafting natural and engaging conversational experiences with the assistance of state-of-the-art Large Language Models. 

Get ready to elevate your chatbot development skills and enhance user experiences. Join us for a workshop that’s all about fun, innovation and practical learning.

In his keynote “Artificial Intelligence – duel or duet of human and machine?”, Thomas Saueressig will talk about how to make the best out of Artificial Intelligence. He will share SAP’s experiences with AI and why SAP has a unique chance to win the race. Saueressig will also give insights on how SAP is making Artificial Intelligence relevant, reliable and responsible.