Track papers across labs, topics, and trends.
Search papers, labs, and topics across Lattice.
This paper introduces a Deep Interest Network (DIN) architecture enhanced with real-time urgency features and temporal positional encodings to improve match recommendations in daily fantasy sports (DFS). The model dynamically weighs the recency of past user actions and is trained using a listwise neuralNDCG loss function. Validated on a large industrial dataset, the system achieves a 9% lift in nDCG@1 compared to a heavily optimized LightGBM baseline, demonstrating its effectiveness for time-sensitive recommendations.
Injecting urgency and recency directly into a Deep Interest Network boosts recommendation quality by 9% in daily fantasy sports, proving that time-awareness is a game-changer for live event engagement.
In daily fantasy sports (DFS), match participation is highly time-sensitive. Users must act within a narrow window before a game begins, making match recommendation a time-critical task to prevent missed engagement and revenue loss. Existing recommender systems, typically designed for static item catalogs, are ill-equipped to handle the hard temporal deadlines inherent in these live events. To address this, we designed and deployed a recommendation engine using the Deep Interest Network (DIN) architecture. We adapt the DIN architecture by injecting temporality at two levels: first, through real-time urgency features for each candidate match (e.g., time-to-round-lock), and second, via temporal positional encodings that represent the time-gap between each historical interaction and the current recommendation request, allowing the model to dynamically weigh the recency of past actions. This approach, combined with a listwise neuralNDCG loss function, produces highly relevant and urgency-aware rankings. To support this at industrial scale, we developed a multi-node, multi-GPU training architecture on Ray and PyTorch. Our system, validated on a massive industrial dataset with over 650k users and over 100B interactions, achieves a +9% lift in nDCG@1 over a heavily optimized LightGBM baseline with handcrafted features. The strong offline performance of this model establishes its viability as a core component for our planned on-device (edge) recommendation system, where online A/B testing will be conducted.