ZODIAC—ZERO-INFLATED OVERSHOOT CONTROLLED DUAL-HEAD INTEGRATION FOR ASYMMETRIC CROSS-DOMAIN FORECASTING

Igor Yakushin, Sai Krishna Kiran Beathanabhotla, Dhruv Garg, Md Mahmudur Rahman

Presentation Attendance: No, we cannot present in-person

Keywords: forecast, cold start, zero-inflated demand, asymmetric costs

TL;DR: If a seller in market A starts selling in market B, forecast the sales based on the performance of an offer in market A, the performance of similar offers in market B, and static features of the product and the seller..

Abstract: Foundation models promise zero-shot forecasting across domains, yet their effectiveness for cold-start scenarios with zero-inflated distributions remains underexplored. We study cross-domain demand forecasting, predicting outcomes for items launching in new domains without historical data where a substantial fraction of launches ($\approx 30\%$) yield zero outcomes and overestimation carries asymmetric costs. We propose a specialized architecture---ZODIAC---combining: (1) dual-domain temporal integration via stacked recurrent layers processing source and target domain signals, (2) a dual-head design with classifier and regressor explicitly modeling zero-inflated distributions, and (3) an asymmetric loss function penalizing overestimation to align with domain-specific costs. We benchmark our approach against a pretrained in-context learner (TabPFN), an AutoML ensemble (AutoGluon), and a neural time-series model (Temporal Fusion Transformer) across six cross-domain forecasting tasks. Our model achieves 80\% WAPE, a 13\% relative improvement over TabPFN, 25\% over AutoGluon, and 26\% over TFT while reducing systematic overprediction from 66--87\% to under 41\%, a property unachievable with models lacking loss customization.

Track: Research Track (max 4 pages)

Submission Number: 28