REACT Cluster Webinar Series – Episode 4: Synthetic Data Generation – Producing Data in Simulation, Training AI on Small and Larger Datasets

📅 May 20th, 2025
🕒 10:00–12:30 CET
💻 Online, Teams

➡️ Mandatory registration here.

 Join the REACT Cluster’s 4th webinar on synthetic data generation – producing data in simulation, training AI on small and larger datasets!

This webinar – organised by the GRINNER Project – focuses on the challenge of data scarcity when it comes to training models in industrial settings. Most of the time comprehensive datasets are either impossible or very expensive to collect. Even when collected, data might be commercially sensitive or severely imbalanced with certain categories represented only by a handful of data.

Involving partner projects from the cluster such as DARROW, ALCHIMIA, and RECLAIM we will share challenges and solutions for synthetic data, transfer learning, few-shot/meta learning, and anomaly detection.

AGENDA

11:00-11:10 Introduction of the REACT Cluster
11:10-11:20 Data Scarcity in Training Models in Industrial Settings | GRINNER | Antonis Porichis from University of Essex
11:20-11:25 Questions
11:25-11:35 Data Scarcity and the Use of Synthetic Data | ALCHIMIA | Marcos Varveris and James Lloyd from EXUS
11:35-11:40 Questions
11:40-11:50 Expanding AI’s Reach: Calibrated Mechanistic Models as a Solution to Data Scarcity | DARROW | Aitor Domec Paz from Ceit
11:50-11:55 Questions
11:55-12:05 Generating Synthetic Images of Waste Streams in Material Recovery Processes | RECLAIM | Michalis Maniadakis from FORTH
12:10-12:15 Questions
12:15-12:25 iBot4CRMs | Antonio M. Oritz from Norce Research
12:25-12:30 Questions