2026-09-10 –, Main stage
Most ML practitioners working on ranking have had this experience: the model looks good on paper - AUC, precision are all healthy - but something feels off when the scores drive real decisions.
The underlying cause is often feature neglect. In supervised pointwise ranking, where a model predicts a relevance score for each item-context pair independently, learners frequently over-rely on contextual signals at the expense of item-specific features. This may yield decent classification metrics, but it leads to poorly calibrated scores - and in practice, we rarely stop at ranking. Model outputs feed downstream tasks: thresholds, priority queues, risk tiers, decision triggers. That's where miscalibration stops being a metric problem and becomes a decision-making problem.
In this talk, we present architectural interventions to address feature neglect directly: item-aware attention mechanisms, cross-feature interactions, and feature representation enhancements that explicitly strengthen item-context relationships. Through real-world experiments, we show these changes not only improve ranking performance but also produce better-calibrated scores - outputs you can trust beyond the leaderboard.
The talk is targeted at Data/ML Scientists and ML Engineers working on recommendation or ranking systems who want to move towards more calibrated and robust ranking architectures.
In this session, we address a common problem in deep ranking systems: When models prioritize contextual signals over item-specific attributes, they lose the ability to differentiate between items in the same context, leading to poor calibration and unreliable downstream decision-making.
Besides introducing different architectural interventions to address item-neglect in deep neural networks, we will share real-world results from productionized models. We will demonstrate that forcing the model to prioritize item features and specific interactions can improve ranking performance. To ensure these concepts are actionable, we will showcase practical implementations and code snippets that can be integrated into modern deep-learning pipelines.
Talk Outline:
Introduction & Problem Statement (3 min)
The Impact of Item Neglect (5 min)
Architectural Interventions (14 min)
Real-World Example and Results (3 min)
Q&A (5 min)
