Center for AI Value PhD Candidate Wins Prestigious EHI Retail Science Award
We are thrilled to announce that our very own PhD candidate, Daniel Schoess, has been honored with the EHI Retail Science Award in the category of “Best Master’s Thesis.” The award ceremony, hosted by EHI Stiftung and GS1 Germany during EuroCIS in Düsseldorf, brought together around 250 industry leaders from the retail and consumer goods sectors.
About Daniel’s Award-Winning Work
Daniel’s thesis, conducted at ETH Zurich in collaboration with the Massachusetts Institute of Technology (MIT), focuses on multimodal deep-learning models for demand forecasting in stationary retail. His research stands out for its ability to capture cross-effects between products—relationships that go beyond mere substitutions, encompassing how items complement one another in a store setting. This modeling approach uses a wide range of product data, from textual descriptions to images, forming robust embeddings that help retailers:
- Predict sales of newly introduced items without historical data
- Identify product affinities more accurately for more efficient assortments
- Reduce inventory costs and minimize overproduction, thus supporting sustainability
In practical terms, Daniel’s innovation addresses one of the biggest challenges in the Fast-Fashion segment: the need to forecast demand for a constantly changing lineup of products. His model demonstrated that roughly 45% of frequently co-purchased products end up in very similar “embedding space” clusters—validating how closely related these products truly are. Impressively, the AI effectively recognized relationships even among items it had never seen before, showcasing exciting potential for dynamic, rapidly evolving retail contexts.
Significance for the Retail and Consumer Goods Sector
Multimodal AI, often hailed as the next big thing in machine learning, has not yet been widely adopted in mainstream retail research. Daniel’s work pushes this frontier forward by illustrating how transformer-based AI systems can integrate diverse data modalities—an approach that’s still relatively rare in real-world merchandising. For retailers:
- Increased Forecast Accuracy: More precise demand predictions and stock-level planning
- Flexible Product Launches: Quick adoption of brand-new products into forecasting tools
- Eco-Conscious Strategy: Lower waste through better alignment of inventory with actual demand
A Shared Achievement for the Center for AI Value
Daniel’s success also underscores the Center for AI Value’s core mission of marrying academic excellence with practical business impact. Our team investigates AI applications across multiple industries, with special attention to real-world impact. By focusing on customer-centric and value-driven AI projects, we aim to bridge the gap between cutting-edge AI research and day-to-day industry needs.
We extend our heartfelt congratulations to Daniel Schoess for this exceptional achievement and look forward to seeing how his innovative work will continue shaping the future of retail.