2 days / 20+ talks
Find best practices, new exciting AI start-ups, ask experts or find partners that can help you understand how to apply AI
October 17th and 18th, 2024, Zagreb, Kraš Auditorium
Conference Schedule
Day 2 – Oct 18th
Engineering Track – Conference Room 2
Marko Velić, Google
Reasoning in LLMs and their limitations
09:20 – 10:00 | Conference Room 2
How is reasoning in LLMs achieved and do they reason at all? What are the limitations of the current models? Scaling laws, impact of model size on reasoning and some intuitions behind it….
Traditional Deep Learning Methods in the Era of Foundation Models
10:00 – 10:30 | Conference Room 2
he Large Language Model (or LLM) revolution has in recent years not only transformed how we process text, but it has also brought in a radical shift in how many new deep learning projects were structured…
The Journey of Innovation: Stories from Building AI products
10:30 – 11:00 | Conference Room 2
This presentation highlights our journey of AI innovation through accounts of our latest projects. We will discuss the development processes, the technical challenges we faced, and how we overcame them to deliver AI solutions. Attendees will gain insights into the complexities and successes of AI development, with anecdotes and practical lessons.
MLOps from Idea to production with Amazon SageMaker
11:30 -12:00 | Conference Room 2
An overview of the MLOps journey, from ideation to production, using Amazon SageMaker. Attendees will learn how SageMaker supports the full machine learning lifecycle, including data preparation, model training and tuning, and automated deployment and monitoring.
12:00 -12:30 | Conference Room 2
The C++ programming language has been around since the 1980s and is still the tool of choice for doing high-performance systems programming. In this talk, Matej will talk about the team’s decision to go with C++ as the main language for the Omnisearch engine
AI in Financial Services: Automated Company Valuation
12:30 – 13:00 | Conference Room
This talk presents an AI-powered digital solution for automating the valuation of Zagreb Stock Exchange Prime Market companies.
Using satellite imagery to track water quality – how did we implement it?
13:00 -13:30 | Conference Room 2
This presentation describes the implementation of a system using Sentinel-2 satellite imagery and advanced algorithms to monitor water quality parameters such as chlorophyll-a concentration and turbidity….
How to Build an In-House Automated Video Verification System
14:30 – 15:00 | Conference Room 2
Remote onboarding and video verification of new users have become hot topics in recent times. In this lecture, we’ll explore the development of an automated and robust in-house video verification system, focusing on various required AI models.
Deep Forecasting for Sequential Data in Finance with LSTMs & Transformers
15:00 – 15:30 | Conference Room 2
In this talk, the speaker will present the common challenges of applying machine learning in Finance. Focus will be on machine learning for sequential decision making with different types of neural networks.
Mastering the Impossible: Heuristic Solutions to Complex Scheduling Challenges
15:30 – 16:00 | Conference Room 2
In this insightful lecture, discover how the Sofascore AI team harnessed the power of heuristic optimization methods to tackle the multi-criteria challenges of scheduling, where multiple constraints and objectives collide.
Future of the future: creating an AI product beyond hype
10:00 – 10:30 | Conference Room 1
When a company creates one of a bunch of AI products, what should distinguish it from competitors? We know the initial AI hype is behind us, the real things are happening now and now everyone is asking: what`s next?
Urban Pfeifer i Ksenija Latkovic Kozarac, C1/PM & Podravka
Superfood Chefbot Case Study
11:30 – 11:50 | Conference Room 1
This presentation of Podravka’s case shows how integration of AI into web platforms can shaping the future of digital business transformation.
Ivar Scheers, Embassy of the Netherlands in Croatia
The Netherlands: Implementation of the AI Act
11:50 – 12:00 | Conference Room 1
TBD
Davor Aničić (Velebit.ai) / Ivan Rimac (Barrage) / Marijana Šarolić Robić (Attorney at Law)/ Mihael Mudrić (Associate Professor, Faculty of Law)
12:50 – 13:10 | Conference Room 1
The hype is there, people are playing with AI and piloting across organizations. This lecture will aim to share some learnings on how to make AI work in larger organizations in a secure and cost-effective way.
Automated assistant for logistics and supply chain management
13:10 – 13:30 | Conference Room 1
The focus of the lecture will be on demand forecasting, replenishment optimization, production planning, and inventory management.
15:20 – 16:05| Conference Room 1
Matej Ciglenečki, Tensorpix
Ozren Kobsa, Ulpian AI
Lucija Modrić, Rearm
Nenad Mijić, SRCE
High Performance Computing: Empowering the AI
16:05 – 16:25 | Conference Room 1
In this workshop, we will present benchmark results to demonstrate the tangible benefits of using our HPC system for AI workloads. Furthermore, we will detail the available HPC systems at SRCE and in Europe and how to access them.
Matija Piškorec, Research Associate at the RBI & Postdoc at the University of Zurich
AI and distributed ledger technologies: A match made in heaven (or hell)
10:00 – 10:30 | Conference Room 3
In this talk I will try to draw parallels between artificial intelligence (AI) and distributed ledger technologies (DLT, also more commonly known under less precise terms such as “blockchain” and “crypto”) by framing their development through the lens of digital revolution initiated by Shannon in the 1940’s.
Beyond the Labels: Handling Subjectivity in Textual Analysis
10:30 – 11:00 | Conference Room 3
Ana Barić and Laura will present their co-authored research papers on sentiment analysis in news headlines. Addressing a topic where neutrality is challenging, they will explore key principles and pitfalls of data annotation, the relationship between annotator subjectivity and model uncertainty, and compare two distinct NLP approaches: the conventional fine-tuning method and the emerging in-context learning paradigm.
Coffee Break
11:00 – 11:30 | Coffee Space
Viviane Potocnik, ETH Zurich
Optimizing Transformer Model Inference on a RISC-V Platform
11:30 – 12:00 | Conference Room 3
This presentation will cover the optimization techniques for deploying transformer-based foundation models, including Large Language Models (LLMs) and Vision Transformers (ViTs), on an open-source many-core RISC-V platform.
12:00 – 12:30 | Conference Room 3
This presentation is divided into three parts: an abstract and specific approach to reducing a problem to optimization, an explanation of the basic principles and examples behind PINNs, and taking an educated guess on the future development of this technology and its significance.
Enhancing Medical Research through Synthetic Data Generation
12:30 – 13:00 | Conference Room 3
This presentation discusses the application of machine learning in medicine, focusing on the challenges posed by limited and restricted data due to privacy concerns. It highlights the importance of synthetic data generation using generative models, reviews their application across various medical data types
Ema Puljak, Universitat Autonoma de Barcelona
Bridging the Gap: How Tensor Networks Connect Classical and Quantum Machine Learning
13:00 – 13:30 | Conference Room 3
This talk addresses the intriguing question of whether Machine Learning algorithms can benefit from the advancements in Quantum Computing.
Veridical benchmarks for testing the characteristics of deep learning models: Definitions Matter for GPT
14:30 – 15:00 | Conference Room 3
This talk explores the testing capabilities of large language models, focusing on their ability to generate definitions from examples and perform fine-grained zero-shot classification using class definitions. It introduces new methods for generating and evaluating definitions, demonstrating how improved class label definitions enhance downstream classification performance.
Bogdan Okreša Đurić, University of Zagreb Faculty of Organization and Informatics
From Ontology to Action: Streamlining Multiagent System Development with SPADE
15:00 – 15:30 | Conference Room 3
This talk introduces a framework for modelling multiagent systems using a specialized ontology that captures the key components and interactions within such systems…
It is not just the most important conference about Artificial Intelligence in Croatia, it is also a great place where you can meet and greet people that are having the same interests as you! So, you will have networking opportunity to start the discussion and Networking Zone, where you can sit and chat with them.
Meet the
Speakers
Some great speakers and moderators are coming your way! You will learn a ton from the presentation, but they are open to networking and your direct questions in the network zone.
Koristimo kolačiće (eng. „cookies“) za pružanje boljeg korisničkog iskustva. Nastavkom pregleda web-stranice slažete se s uvjetima korištenjaPrihvatiPročitajte više