EMP: Enhanced Multi-modal Prediction for fashion sales using Fourier Mapping and ERP-based contrastive learning
Sanguk Park, Byungyoon Park, Seohyun Lee, Subin Shin, Wooju Kim · 2026
Read EMP: Enhanced Multi-modal Prediction for fashion sales using Fourier Mapping and ERP-based contrastive learning by Sanguk Park, Byungyoon Park, Seohyun Lee, Subin Shin, Wooju Kim (2026) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/plos-pone-0350107.txt.
Description
New fashion items have no sales history, so Sanguk Park and colleagues built a forecaster that borrows the past of similar items. Their 2026 PLOS One model predicts a product's first twelve weeks of sales from its photos, text, and look-alike items' records.
First, a distance measure that tolerates stretched timing finds true look-alikes and shapes a comparison space. Then a transformer network reads the item's features, with sales values converted through a wave-based mapping that preserves their size. Because everything comes from in-house data, no outside feeds are needed. The authors report sharper forecasts than earlier models on their tests.
Catalog record
| CREATOR | Sanguk Park, Byungyoon Park, Seohyun Lee, Subin Shin, Wooju Kim |
|---|---|
| YEAR | 2026 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 72419 |
| LICENSE | CC_BY |
| RIGHTS | CC BY |
| SOURCE | PLOS |