IntentRec: Predicting User Session Intent with Hierarchical Multi-Task Learning

This post summarizes the paper IntentRec: Predicting User Session Intent with Hierarchical Multi-Task Learning by Sejoon Oh, Moumita Bhattacharya, Yesu Feng, and Sudarshan Lamkhede, published on arXiv (2024-07-25): https://arxiv.org/abs/2408.05353.

Motivation

Recommender systems play a critical role in diverse digital services such as e-commerce, streaming media, and social networks. If a system knows what a user's intent is within a given session (e.g., do they want to watch short videos, a movie, or play games; are they shopping for a camping trip), it becomes much easier to deliver high-quality recommendations. Large-scale recommendation systems have long faced challenges around efficiency, scalability, and understanding user intent, especially as users increasingly expect natural-language interaction and real-time responsiveness.

What IntentRec Does

IntentRec is a novel recommendation framework based on a hierarchical multi-task neural network architecture. Its core idea:

1. Intent estimation: Estimate a user's latent intent using their short-term and long-term implicit signals as proxies. 2. Next-item prediction: Use the predicted intent directly to predict the next item the user is likely to engage with. 3. Personalization: By directly leveraging intent prediction, offer accurate and personalized recommendations.

The method follows a typical pipeline: encoding user context into representations, a hierarchical multi-task core that jointly models intent and next-item engagement, and an inference stage that serves recommendations under latency and budget constraints.

Evaluation

The authors run comprehensive experiments on Netflix user engagement data, showing that IntentRec outperforms state-of-the-art next-item and next-intent predictors. They also share several findings and downstream applications of IntentRec. Quantitative results should be verified against the original PDF tables.

Insights for Search / Rec / Personalization

1. Architecture: Cascaded retrieve–rank–generate remains mainstream, but agentic paradigms are making retrieval strategy itself a learnable object. 2. Data: High-quality instruction data and click/session logs are both critical; synthetic data must guard against knowledge leakage and distribution shift. 3. Evaluation: The gap between offline metrics and online satisfaction is widening; LLM-as-judge should be cross-validated with human evaluation. 4. Product: Latency, cost, interpretability, and safety are hard constraints for industrial deployment—do not optimize only academic benchmarks.

Engineering Checklist

| Item | Question | Suggestion | |------|----------|------------| | Data | Does training/index data contain PII? Versioning? | Partitioned indexes, anonymization, rollback-capable embedding versions | | Latency | What is the p99 budget? How many retrieval steps? | Cascade + early stopping, cache popular queries, async re-ranking | | Quality | Do offline gains translate to online CTR/satisfaction? | Interleaving experiments, human audits, citation checks | | Safety | Does open retrieval introduce poisoning/bias? | Source allowlists, adversarial detection, output filtering | | Cost | Token and GPU cost per query? | Route to smaller models, distillation, hybrid sparse+dense retrieval |

Limitations and Future Directions

Possible limitations include experiment scale constrained by compute budgets, benchmarks that may not match real user distributions, and unknown cross-lingual generalization. Future work may explore more efficient test-time compute allocation, deeper integration with knowledge graphs and structured databases, and causal/fairness constraints for recommender systems.

Related Reading

• A Survey of Personalization: From RAG to Agent (arXiv:2504.10147)
• Bridging Personalization and Control in Scientific Personalized Search (arXiv:2411.02790)
• Can Large Language Models Understand Preferences in Personalized Recommendation? (arXiv:2501.13391)
• Unified Embedding Based Personalized Retrieval in Etsy Search (arXiv:2306.04833)
• User-LLM: Efficient LLM Contextualization with User Embeddings (WWW, DOI: 10.1145/3701716.3715463)

Original Abstract

> Recommender systems have played a critical role in diverse digital services such as e-commerce, streaming media, social networks, etc. If we know what a user's intent is in a given session (e.g. do they want to watch short videos or a movie or play games; are they shopping for a camping trip), it becomes easier to provide high-quality recommendations. In this paper, we introduce IntentRec, a novel recommendation framework based on hierarchical multi-task neural network architecture that tries to estimate a user's latent intent using their short- and long-term implicit signals as proxies and uses the intent prediction to predict the next item user is likely to engage with. By directly leveraging the intent prediction, we can offer accurate and personalized recommendations to users. Our comprehensive experiments on Netflix user engagement data show that IntentRec outperforms the state-of-the-art next-item and next-intent predictors. We also share several findings and downstream applications of IntentRec.