Intent-Driven Semantic ID Generation for Grounded Conversational News Recommendation
Not on the map. Why?
Abstract (arXiv)
Conversational news recommendation requires grounding each suggestion in a rapidly evolving article corpus while addressing implicit user intents that lack explicit retrievable keywords. To characterize this scenario, we identify 6 intent types from production dialogues: five are implicit and pose fundamental challenges to standard RAG pipelines, forming a critical retrieve-first bottleneck. To address these issues, we introduce intent-driven Semantic ID (SID) generation under a Generate-then-Match paradigm. With two-stage training that consists of multi-task SID alignment and GPT-4 Chain-of-Thought distillation, an LLM maps diverse intents to hierarchical SID prefixes, which are then fuzzy-matched to the current news pool to guarantee fully grounded recommendations. Profile-Aware Dual-Signal Reasoning (PADR) further enables cold-start users to obtain valid recommendations using only profiles. On a mainstream Chinese news platform, our 7B model achieves 0% hallucination and 12.4% L1 match in the 152K open-generation SID space (4x random baseline). It matches GPT-4+Hybrid RAG on L1 while surpassing it on finer-grained metrics (L2 2x, Category +1.2pp) at ~100x lower cost. Cold-start users, where existing baselines score 0%, achieve 18.0% L1 (6x random), the highest among all user groups.
Key facts. No citing papers in the map yet (data through 7 Oct 2026). Builds on 19 of its 33 references that are in the map. First posted on arXiv 8 May 2026.
Builds on
19 of 33 references.
References · 19
- 2020Retrieval-Augmented Generation for Knowledge-Intensive NLP Taskscited by 9,160
- 2018Self-Attentive Sequential Recommendationcited by 1,169
- 2019BERT4Rec: Sequential Recommendation with Bidirectional Encoder Representations from Transformercited by 952
- 2023C-Pack: Packed Resources For General Chinese Embeddingscited by 556
- 2022Teaching Small Language Models to Reasoncited by 279
- 2023Chat-REC: Towards Interactive and Explainable LLMs-Augmented Recommender Systemcited by 241
- 2017Lexically Constrained Decoding for Sequence Generation Using Grid Beam Searchcited by 234
- 2018Towards Deep Conversational Recommendationscited by 174
- 2023Adapting Large Language Models by Integrating Collaborative Semantics for Recommendationcited by 171
- 2020A Survey on Conversational Recommender Systemscited by 142
- 2023Large Language Models as Zero-Shot Conversational Recommenderscited by 126
- 2021Advances and Challenges in Conversational Recommender Systems: A Surveycited by 118
7 more
- 2019Towards Knowledge-Based Recommender Dialog Systemcited by 112
- 2022Towards Unified Conversational Recommender Systems via Knowledge-Enhanced Prompt Learningcited by 85
- 2025OneRec-Think: In-Text Reasoning for Generative Recommendationcited by 47
- 2020CR-Walker: Tree-Structured Graph Reasoning and Dialog Acts for Conversational Recommendationcited by 31
- 2025ThinkRec: Thinking-based recommendation via LLMcited by 11
- 2025Graph Retrieval-Augmented LLM for Conversational Recommendation Systemscited by 10
- 2025DiscRec: Disentangled Semantic-Collaborative Modeling for Generative Recommendationcited by 8
Built on by
Not cited yet.
Similar, not cited
Close in wording, no citation either way
- 2026Deep Interest Mining for Intent-Enriched Semantic IDs in Multimodal Generative Recommendation
- 2024Generative News Recommendation
- 2026Reasoning over Semantic IDs Enhances Generative Recommendation
- 2026The Disconnect Between Better Descriptive Reasoning Trace Quality and Recommendation Effectiveness
- 2026S2GR: Stepwise Semantic-Guided Reasoning in Latent Space for Generative Recommendation
- 2026Implicit Reasoning for Large Language Model-based Generative Recommendation
- 2025Refine-POI: Reinforcement Fine-Tuned Large Language Models for Next Point-of-Interest Recommendation
- 2026Generate to Explore, Select to Exploit: Aligning LLM-based Headline Generation with Personalized Recommendation