CO-LOCATED EVENTS

Hotel Metropol, Belgrade

18th-22nd November 2024

DSC Europe 24 SCHEDULE

Lorem ipsum dolor sit amet, consectetur adipiscing elit. Ut elit tellus, luctus nec ullamcorper mattis, pulvinar dapibus leo.

 Aristotle

Workshop: Think You Can Handle Football Player Data? Join the Club’s Data Lab

Workshop
Step into the world of football performance analysis with our hands-on Data Lab workshop! In this session, you’ll learn how to load and clean GPS data, visualize key player metrics, and calculate Exponential Weighted Moving Averages (EWMA) to track player load throughout the season. Dive into the winter conditioning phase to assess team workload and gain insights, and create actionable player progress reports that help coaches make data-driven decisions. By the end of the workshop, you’ll be equipped with the skills to turn raw data into meaningful insights that boost performance analysis.
11:00

 –

12:00

Workshop: Think You Can Handle Football Player Data? Join the Club’s Data Lab

Arian Skoki
Data & Growth Strategist @ Morplo

AI in Music

Breakout Session
Description:   AI is orchestrating a new era in music, blending creativity with technology in ways we’ve never imagined. In this session, we’ll begin by exploring how AI-generated content is evolving, and how close we are to making it indistinguishable from human-created music. As we look deeper, we’ll see how AI is not just enhancing creativity but also transforming the entire music industry—from marketing strategies to understanding audience behavior and predicting the next big hit.

The spotlight will then turn to the technical side of AI’s influence on music, diving into the creation of custom neural network architectures for challenges like music source separation. This session will also reveal how the technology behind autonomous trucks is quietly reshaping the music industry in the logistics of concerts and live events.

As we move from theory to application, we’ll uncover how AI is not just composing music, but revolutionizing the entire business model behind it, helping artists and businesses alike thrive in an era where data and creativity intersect. Join us for an engaging exploration of the future of music, where AI is the new maestro, conducting a symphony of innovation.



12:00 – 12:15   Fool me once, shame on you; fool me twice, shame on AI

Bojan Banjac
Assistant Professor @ Faculty of Technical Sciences

We are constantly exposed to images, music and text generated by AI. Yet, they are often easy to spot as not real. Let us talk about how to make them even more convincing. Methods vary from data preparation techniques to post-processing and even some new ideas for future directions. Part of lecture is determining how well we actually detect generated content. After all, who wouldn’t like to be fooled by smart AI.



12:15 – 12:30   Creating specialized AI tools for music production & education

Stipe Kabic
Machine Learning Engineer @ Atomic Intelligence

Applications of AI in music are a hot topic of research, with a lot of attention being given to large generative AI models able to create realistic music. Another kind of research is being done in smaller labs and companies, focusing on creating useful AI tools for musicians and producers, striving to help them rather than replace them. In such research lots of focus is on model efficiency and speed, as well as creating a quality user experience for music practicioners. In this talk we will discuss the process of creating such models with a focus on music source separation, and demonstrate our AI platform for music learning and production.



12:30 – 12:45   The Symphony of AI and Music: Composing Tomorrow’s Hits with Data-Driven Creativity

Suren Movsisyan
Founder, CTO @ Aigorithm | CMO @ Revolve Games | Tech Lead @ Bitcoin Center NYC

What happens when artificial intelligence becomes the composer? In this engaging keynote, Dr. Suren Movsisyan—an innovator at the crossroads of AI, blockchain, and web3 metaverse technologies—explores the convergence of data science, NFTs, and music creation. From AI-generated albums to fully immersive metaverse concerts, learn how data, machine learning, blockchain, and algorithms are rewriting the rules of creativity. Get insights into how artists and businesses can leverage these cutting-edge tools to create personalized musical experiences, unlock new revenue streams, and redefine the future of the music industry. This is not just the next chapter in music, it’s a symphony of innovation, powered by AI.



12:45 – 13:00   The Beat of Data: AI-Driven Innovation in Music Production

Maksim Kariagin
Ex Director of Research and Analysis @ Warner Music Group

Discover how AI and advanced analytics are transforming decision-making in the music industry! This presentation dives into the practical use of AI to enhance strategic planning and optimize marketing efforts. Learn how data-driven insights are shaping decisions that allow music labels to stay competitive and creative in a fast-evolving environment. Through real-world examples and a case study on AI-powered planning, you’ll see how technology is driving innovation and unlocking new opportunities in the industry.



13:00 – 14:00   Panel duscussion

12:00

 –

14:00

AI in Music

Bojan Banjac & Stipe Kabic & Suren Movsisyan & Maksim Kariagin
Assistant Professor @ Faculty of Technical Sciences & Machine Learning Engineer @ Atomic Intelligence & Founder, CTO @ Aigorithm | CMO @ Revolve Games | Tech Lead @ Bitcoin Center NYC & Ex Director of Research and Analysis @ Warner Music Group

Autonomous Driving

Breakout Session
Description:   Autonomous driving is not just about cars learning to drive themselves; it’s about AI reshaping the future of transportation and how we interact with it. In this session, we’ll begin by exploring the broader challenges and opportunities that AI faces in autonomous driving—from interpreting human behavior to navigating complex environments. These obstacles also open doors for innovations that will transform the automotive industry and society as a whole.

Next, we’ll dive into the critical role of edge computing in enabling real-time decision-making in autonomous vehicles. When split-second choices are the difference between safety and danger, understanding how data engineering works at the edge is crucial.

We’ll then shift focus to the less-discussed, but rapidly advancing, sector of autonomous trucking. These trucks are quietly leading the charge in automation, reshaping logistics and the transport industry in ways passenger cars have yet to achieve.

Join us for a journey through the AI-driven future of autonomous driving, where each step brings us closer to safer, more efficient roads and a transformed transportation landscape.



15:00 – 15:15   Challenges and opportunities for AI in autonomous driving

Velibor Ilic
Senior AI/ML Research Scientist @ The Institute for Artificial Intelligence Research and Development of Serbia

Artificial Intelligence (AI) is revolutionizing the automotive industry by enhancing autonomous driving technologies, but it faces significant challenges. These include interpreting complex human behaviors, navigating diverse weather conditions, and adapting to various global road infrastructures. Ethical and legal considerations also pose significant challenges, particularly in terms of decision-making during unforeseen events and the potential impact on employment within the industry. Despite these obstacles, AI presents vast opportunities such as dramatically improving road safety by reducing human error and influencing new business models in the automotive sector. The integration of artificial intelligence into autonomous vehicles not only promises to transform the way vehicles operate, but also has a major impact on the wider social and economic landscape.



15:15 – 15:30   Edge Computing and Data Engineering in Autonomous Driving

Dejan Djekic
CEO @ Functor

In autonomous driving, real-time decision-making is often too latency-sensitive to rely solely on external (cloud) infrastructure. For instance, when a vehicle encounters an obstacle, it must decide in milliseconds whether to stop, steer, or continue. This decision-making must happen as close to the vehicle as possible to reduce the risk of latency-related accidents. This talk explores the topic a bit more in-depth.



15:30 – 15:45   Decision outsourcing – Data take the wheel

Uros Urosevic
Data Engineer @ TeleSign

This talk delves into the evolving role of data as the foundation of decision-making in autonomous driving. We’ll examine how the sheer volume, diversity, and depth of data—from sensor streams to environmental variables—enable autonomous vehicles to make nuanced, context-aware choices. By exploring the intricacies of data-driven decisions, we’ll uncover how autonomous systems transform raw information into safe, reliable actions on the road, redefining traditional notions of control and adaptation.



15:45 – 16:00   The Silent Revolution: How Autonomous Trucks are Outpacing Passenger Cars

Ninoslav Jovanov
Software Group Leader @ Continental Automotive Serbia

The landscape of autonomous vehicles is often dominated by discussions of self-driving cars, but the real revolution is unfolding on the highways and in industrial settings. This topic explores the rapid advancements in autonomous trucking technology and its far-reaching implications for the logistics and transportation industries. As the autonomous vehicle debate continues, it’s the heavy-duty trucks that are quietly leading the charge toward more automated future.



16:00 – 17:00   Panel duscussion

Moderator: Dejan Djekic
CEO @ Functor

15:00

 –

17:00

Autonomous Driving

Velibor Ilic & Dejan Djekic & Uros Urosevic & Ninoslav Jovanov
Senior AI/ML Research Scientist @ The Institute for Artificial Intelligence Research and Development of Serbia & CEO @ Functor & Data Engineer @ TeleSign & Software Group Leader @ Continental Automotive Serbia

 Plato

Prompt Engineering

Breakout Session
Description:   As AI becomes an integral part of modern solutions, prompt engineering has emerged as a crucial skill in shaping the performance and reliability of AI systems. In this session, we will begin by exploring the foundational principles of precision prompting, covering strategies to standardize prompts, improve reproducibility, and manage performance through modular designs. We’ll then move on to examining the current limitations of large language models (LLMs) and how smart prompt engineering can mitigate some of these challenges. Finally, we will address the practical obstacles of deploying generative AI, particularly LLMs, into production environments, with a focus on scalability, privacy, and ethical concerns. This session will provide both theoretical insights and real-world applications, equipping attendees with the tools and knowledge to effectively harness the power of AI through prompt engineering.



10:00 – 10:15   Precision Prompting: Building a Foundation for Reliable and Efficient AI Systems

Mikhail Rozhkov
Technical Product Manager @ Nebius

Discover how to leverage precision prompting to create more reliable and efficient AI solutions. This talk explores practical strategies for overcoming common challenges in prompt engineering, including reproducibility, versioning, and performance tracking. Learn how to standardize prompts, implement version control, and use modular design to build scalable AI systems. We’ll also cover the latest tools that streamline the prompt engineering process. Whether you’re an AI expert or newcomer, you’ll gain valuable insights to enhance your AI development workflow.



10:15 – 10:30   What LLMs are not (yet) good at?

Nikola Milosevic
Science Fellow @ Bayer

Large language models have revolutionized the field of artificial intelligence and we see more and more practical use cases where they can improve productivity and quality of work being done. This can range from document creation, to some areas of data analysis and evaluation, as well as assisting with obtaining crisp answers to complicated questions from multiple sources. However, there are still a number of areas where LLMs are not performing too well, from first order reasoning to cryptography. In this talk, we will cover some of the areas where LLMs are not performing well, and whether some smart prompt engineering can help.



10:30 – 10:45   Optimizing LLM Prompts: Techniques and Tools to Amplify Model Accuracy and Efficiency

Nikola Milojevic
Co-Founder & CEO @ Neural Factory

Unlock the full potential of Large Language Models (LLMs) with cutting-edge prompt optimization techniques. In this talk, we’ll explore strategies like gradient-based tuning, zero-shot and few-shot learning, and real-time prompt refinement. Discover tools like AdalFlow, DsPy, and Text-Grad that automate and enhance the prompt engineering process, boosting accuracy and efficiency. Join us to learn actionable techniques and tools that will transform how you interact with and deploy LLMs across sectors like finance, healthcare, and customer service. Perfect for data scientists ready to push the boundaries of AI!



10:45 – 11:00   Practical challenges in deploying GenAI to Production

Srdjan Jovanovic
Machine Learning Team Lead @ Loka

As generative AI, especially large language models (LLMs), becomes central to technological advancements, organizations face challenges moving from proof of concept to production. This session will explore practical obstacles such as scalability, reliability, data management, maintenance, privacy, and compliance, as well as issues related to model unpredictability, including hallucinations, prompt management, guardrails, security, and ethical concerns. We’ll also discuss how to effectively integrate LLMs into your system to maximize ROI. Using real-world examples, we aim to provide actionable insights for overcoming these challenges and successfully deploying generative AI.



11:00 – 12:00   Panel duscussion

Moderator: Mikhail Rozhkov
Technical Product Manager @ Nebius

10:00

 –

12:00

Prompt Engineering

Mikhail Rozhkov & Nikola Milosevic & Nikola Milojevic & Srdjan Jovanovic
Technical Product Manager @ Nebius & Science Fellow @ Bayer & Co-Founder & CEO @ Neural Factory & Machine Learning Team Lead @ Loka

Human-Computer Interaction

Breakout Session
Description:   Human-Computer Interaction & AI explores how artificial intelligence is transforming the way humans interact with machines, creating more intuitive, adaptive, and personalized experiences. This session will examine advancements in AI-driven interfaces, voice recognition, and immersive technologies, while addressing challenges like user privacy, transparency, and the ethical implications of AI decision-making. Attendees will gain insights into how AI is shaping the future of user experience design, making technology more accessible and responsive to human needs.



12:00 – 12:15   Agentic AI: Towards AI Agent-Computer Interaction

Mladjan Jovanovic
Professor of Computer Science @ Singidunum University

There is a shift in how Language Models interact with digital environments, moving from passive text processors and straightforward responses to (pro)active agents capable of problem-solving and performing tasks autonomously. The agents can query complex data, perform computations, execute applications, and perform functions by operating within visual user interfaces. Constraining the agent’s action space enables them to perform clear actions and immediately adapt to feedback while maintaining accuracy and safety. Empowering AI agents to interact effectively in dynamic environments makes them reliable for various tasks with safe and controlled executions.



12:15 – 12:30   Challenges in teaching HCI, from we don’t need it to everybody knows it

Dragan Dinu
Associate professor @ University of Novi Sad, Faculty of Technical Sciences

Dragan has been doing research in Human Computer Interaction (HCI) and teaching HCI for more than 20 years. During that time he faced different obstacles and challenges in delivering HCI knowledge and experience to his students. Biggest obstacles that he had to overcome was attitude of his students towards HCI. He will discuss how this attitude changed over time and what he is using in changing that attitude while teaching best HCI practice.



12:30 – 13:30   Panel duscussion

Moderator: Tomislav Krizan
12:00

 –

13:30

Human-Computer Interaction

Mladjan Jovanovic & Dragan Dinu & Tomislav Krizan
Professor of Computer Science @ Singidunum University & Associate professor @ University of Novi Sad, Faculty of Technical Sciences

AI for a Greener Planet

Breakout Session
Description:   AI for a Greener Planet explores how artificial intelligence is driving sustainable solutions across various industries, from energy and water management to agriculture. This session will showcase innovative AI applications, such as optimizing resource consumption in the energy sector and enhancing water infrastructure with predictive maintenance. We will also discuss AI’s role in supporting environmental efforts like beekeeping, where technology is being integrated into complex, natural systems. Attendees will gain insights into how AI can address critical environmental challenges and contribute to a more sustainable and greener future.



15:00 – 15:15   AI in Water and Energy: Innovations and Use Cases

Boban Stojanovic
Founder and CEO @ Vodéna

This talk explores the innovative application of artificial intelligence in the water and energy sectors. By leveraging advanced AI algorithms, our company develops solutions that enhance water management, optimize energy consumption, and improve sustainability. We will discuss real-world use cases, including smart grid management, flood predictions, predictive maintenance for water infrastructure, and efficient resource allocation. Attendees will gain insights into how AI-driven approaches can address critical challenges, improve operational efficiency, and drive significant progress in these vital industries, ultimately contributing to a more sustainable future.



15:30 – 15:45   Beyond the buzz: bringing AI into the challenging world of beekeeping

Jelena Pejic
CTO @ Beehold

Using AI in beekeeping comes with its own set of challenges. Data collection is difficult due to the unpredictable nature of living environments and seasonal changes. Our team is developing unique hardware for precise, continuous data collection, but constant field testing is limited because critical hive processes occur only a few times per year. Additionally, beekeepers often resist adopting new technologies, complicating AI integration. Furthermore, the long return on investment for deep tech in this field—typically 5 to 10 years—makes securing funding difficult. We will explore these obstacles and discuss strategies to overcome them, emphasizing the importance of innovating not just in AI models, but in every aspect surrounding their application, to unlock full potential of AI to transform apiculture.



15:45 – 16:00   Precision agriculture with AI

Catalin Gheorghiu
Solutions Architect @ EPAM

Hook statement: You cannot do precision agriculture without IoT and AI. Or let’s see an AI/Custom Vision solutions from the real world, and see what makes them tick. And this opens a lot of other fields (pun intended 😊) of application. In this session I will share from the experience creating a couple of solutions in the field of precision agriculture. We will see what kind of problems are tackling a solution of this kind. And of course, how to solve them fastest. We will try even live demos and of course chocolates.



16:00 – 17:00   Panel duscussion

Moderator: Catalin Gheorghiu
Solutions Architect @ EPAM

15:00

 –

17:00

AI for a Greener Planet

Boban Stojanovic & Jelena Pejic & Catalin Gheorghiu
Founder and CEO @ Vodéna & CTO @ Beehold & Solutions Architect @ EPAM

Elytis

Sino-Serbian Academic AI Exchange

Breakout Session
Description:   The Sino-Serbian Academic AI Exchange session focuses on the exchange of knowledge and perspectives between leading AI experts from Serbia and China. This session will provide a platform for top academics and researchers to share insights on the latest advancements in AI, discuss emerging trends, and explore the future of artificial intelligence from the viewpoints of two distinct academic communities. Attendees will have the opportunity to engage with thought leaders and gain a deeper understanding of the state of AI research in both countries.



10:00 – 10:05   Opening Remarks

David Rao
Program Coordinator @ DSC



10:05 – 10:10   Opening Remarks

Dubravko Culibrk
Director of The Institute for Artificial Intelligence Research and Development of Serbia



10:10 – 10:15   Opening Remarks

Irina Karabulatova
Professor @ Heilongjiang University



10:15 – 10:20   Opening Remarks

Bojan Lalic
Director of Belt and Road Institute Belgrade



10:20 – 10:30   Ceremonial signing of MOU between Heilongjiang University and The Institute for Artificial Intelligence Research and Development of Serbia



10:30 – 10:45   TBA

Yang Xiang
Deputy Director of Cloud Computing Institute @ Peng Cheng Laboratory

TBA



10:45 – 11:00   TBA

Zhengyu Ma
Associate Professor @ Peng Cheng Laboratory

TBA



11:00 – 11:15   From chatbots toward multimodal and multilingual voice-bots

Vlado Delic
Full Professor @ Faculty of Technical Sciences, University of Novi Sad

The rapid advancement of Large Language Models (LLMs) has revolutionized the AI landscape. Large Multilingual Models facilitate swift development for low-resourced languages, ensuring not only impressive language models and high levels of human-machine communication but also the preservation of linguistic heritage and the reinforcement of historical and cultural national identities, which vary across different countries. Certain institutions, such as those in the medical, legal, and governmental sectors, require on-premise solutions to provide data security and maintain independence from foreign AI providers. Multimodal communication encompasses text, speech, and other modalities in human-machine interactions. Progress in speech technology is driven by AI on one side, and on the other side, it results in the ability to comprehend human speech, mood, and intent, as well as to generate appropriate multimodal responses—hallmarks of AI capabilities. This presentation will explore these topics, comparing the cases of Serbian and Chinese languages, addressing the challenges faced by speech technology companies and academic researchers, and discussing prospects for future collaboration.



11:15 – 12:00   Panel duscussion

Moderator: Ivan Lukovic
10:00

 –

12:00

Sino-Serbian Academic AI Exchange

David Rao & Dubravko Culibrk & Irina Karabulatova & Bojan Lalic & Yang Xiang & Zhengyu Ma & Vlado Delic & Ivan Lukovic
Program Coordinator @ DSC & Director of The Institute for Artificial Intelligence Research and Development of Serbia & Professor @ Heilongjiang University & Director of Belt and Road Institute Belgrade & Deputy Director of Cloud Computing Institute @ Peng Cheng Laboratory & Associate Professor @ Peng Cheng Laboratory & Full Professor @ Faculty of Technical Sciences, University of Novi Sad

Data & AI for Decision Making

Breakout Session
Description:   AI for Decision Making focuses on how artificial intelligence is transforming the decision-making processes in various industries, enabling faster, data-driven insights that improve outcomes. This session will explore the integration of AI tools for strategic planning, risk management, and operational efficiency. From predictive analytics to automated decision systems, AI is enhancing human judgment and helping organizations navigate complex environments. Attendees will gain insights into how AI can be leveraged to make more informed, accurate, and impactful decisions, driving innovation and competitiveness across sectors.



12:00 – 12:15   Dun & Bradstreet Data and analytics are key to making smart decisions

Aleksandar Golubovic
Business Development Advisor @ Dun & Bradstreet Europe

We need to Replace gut feeling with smart data and we need to Act on the right information and make informed decisions. The need to make fast, confident and effective decisions is becoming more urgent, and the risks associated with poor judgement are greater than ever. How can you be sure that you are acting on the right information? We surround ourselves with enormous volumes of information, and the amount of data continues to increase as digitalization progresses. The challenge is to refine, manage and analyze this data to produce tangible insights and accurate decision support.
At D&B, we are convinced that smart data already does and will continue to change the ways we make decisions/
We will focus on familiarization with our data and analytics and practical application in decision-making processes.



12:15 – 12:30   Transforming Recruitment with AI Speech Analysis: Insights from ZenHire

Dusan Pavlovic
Machine Learning Engineer @ ZenHire

In today’s HR landscape, AI-driven speech analysis is revolutionizing recruitment by enabling deeper insights into candidates. This talk highlights how speech and language characterization models evaluate fluency, vocabulary, and accent from audio interviews. We’ll discuss integrating these models with other assessments to create comprehensive candidate profiles, supporting data-driven decisions in large-scale recruitment systems. By combining structured speech evaluations with broader tools, AI enhances fairness, efficiency, and precision in hiring. Attendees will gain insight into how this approach fosters a more transparent and meritocratic recruitment process.



12:45 – 13:00   AI-Driven Decision Making in Fintech: A Case Study from TOM

Halil Yildirm
Head of Data @ TOM (Technology of Money)

In this session, we’ll explore the implementation of AI-driven decision-making in fintech based on a case study at TOM, leading digital bank in Turkey. We’ll focus on how AI was used to optimize risk assessment, customer segmentation, and personalized marketing. Attendees will learn strategies for deploying AI solutions in highly regulated industries, addressing key challenges, and enhancing operational efficiency, with actionable insights for improving data-driven decisions and customer engagement across sectors.



12:45 – 13:45   Panel duscussion

Moderator: Aleksandar Golubovic
Business Development Advisor @ Dun & Bradstreet Europe

12:00

 –

14:00

Data & AI for Decision Making

Aleksandar Golubovic & Dusan Pavlovic & Halil Yildirm
Business Development Advisor @ Dun & Bradstreet Europe & Machine Learning Engineer @ ZenHire & Head of Data @ TOM (Technology of Money)

Socrates

Black

12:00

 –

14:00

Black

Black

12:00

 –

14:00

Black

Workshop: Detecting Misinformation in the US 2024 Elections

Workshop
In the era of rapid information exchange, addressing misinformation is crucial, especially during the US 2024 elections. In this workshop, we’ll tackle this challenge by utilizing natural language processing techniques to identify and reduce the spread of false narratives in election discourse. Participants will observe live coding demonstrations to learn how to create effective detection systems and enhance their ability to identify misleading information
15:00

 –

16:00

Workshop: Detecting Misinformation in the US 2024 Elections

Vladimir Lukovic
Data Scientist @ Basiq

Socrates

* this program is not final and is subject to change, full schedule will be available soon