Leveraging LLMs and Heterogeneous Knowledge Graphs for Persona-Driven Session-Based Recommendation

Muskan Gupta; Suraj Thapa; Jyotsana Khatri

Home/Archive/arXiv

arXiv · 2604.06928

Leveraging LLMs and Heterogeneous Knowledge Graphs for Persona-Driven Session-Based Recommendation

Muskan Gupta·Suraj Thapa·Jyotsana Khatri

View original source ↗Read source document ↗Export metadata ↓

Abstract

Session-based recommendation systems (SBRS) aim to capture user's short-term intent from interaction sequences. However, the common assumption of anonymous sessions limits personalization, particularly under sparse or cold-start conditions. Recent advances in LLM augmented recommendation have shown that LLMs can generate rich item representations, but modeling user personas with LLMs remains challenging due to anonymous sessions. In this work, we propose a persona driven SBRS framework that explicitly models latent user personas inferred from a heterogeneous knowledge graph (KG) and integrates them into a data-driven SBRS. Our framework adopts a two-stage architecture consisting of personalized information extraction and personalized information utilization. In the personalized information extraction stage, we construct a heterogeneous KG that integrates time-independent user-item interactions, item-item relations, item-feature associations, and external metadata from DBpedia. We then learn latent user personas in an unsupervised manner using a Heterogeneous Deep Graph Infomax (HDGI) objective over a KG initialized with LLM-derived item embeddings. In the personalized information utilization stage, the learned persona representations together with LLM-derived item embeddings are incorporated into a modified architecture of data-driven SBRS to generate a candidate set of relevant items, followed by reranking using the base sequential model to emphasize short-term session intent. Unlike prior approaches that rely solely on sequence modeling or text-based user representations, our method grounds user persona modeling in structured relational signals derived from a heterogeneous KG. Experiments on Amazon Books and Amazon Movies & TV demonstrate that our approach consistently improves over sequential models with user embeddings derived using session history.

Explore related subjects

cs.IR

Keep this discovery

Explore connections, maps & timelines

BibTeX RIS

Muskan Gupta, Suraj Thapa, Jyotsana Khatri. 2026-04-14. Leveraging LLMs and Heterogeneous Knowledge Graphs for Persona-Driven Session-Based Recommendation. https://arxiv.org/abs/2604.06928

Cite the original work for its findings. Save a collection to share your selection of sources.

KEEP EXPLORING

Related papers

arXiv2026

Do Large Language Models Favor Recent Content? A Study on Recency Bias in LLM-Based Reranking

Large language models (LLMs) are increasingly deployed in information systems, including being used as second-stage rerankers in information retrieval pipelines, yet their susceptibility to recency bias has received little attention. We investigate whether LLMs implicitly favour newer documents by prepending artificial publication dates to passages in the TREC Deep Learning passage retrieval collections in 2021 (DL21) and 2022 (DL22). Across seven models, GPT-3.5-turbo, GPT-4o, GPT-4, LLaMA-3 8B/70B, and Qwen-2.5 7B/72B, "fresh" passages are consistently promoted, shifting the Top-10's mean publication year forward by up to 4.78 years and moving individual items by as many as 95 ranks in our listwise reranking experiments. Although larger models attenuate the effect, none eliminate it. We also observe that the preference of LLMs between two passages with an identical relevance level can be reversed by up to 25% on average after date injection in our pairwise preference experiments. These findings provide quantitative evidence of a pervasive recency bias in LLMs and highlight the importance of effective bias-mitigation strategies.

cs.IR↗

arXiv2026

UniRec: Cross-stage Multi-Task Fusion with Preference Alignment for Cascaded Recommender Systems

Industrial recommender systems cascade stages with different objectives, feature spaces, and latency constraints. Optimizing pre-ranking and ranking separately induces cross-stage inconsistency: upstream models may filter out items preferred by downstream rankers, while independently tuned downstream fusion can offset upstream improvements. Most existing multi-task fusion methods target the ranking stage alone, and cross-stage methods often align with a downstream-derived score, leaving joint optimization of fusion modules across cascaded stages largely unexplored. We propose UniRec, a Unified Cross-stage Recommendation Fusion model. First, the two fusion agents partially share input embeddings in a single computation graph, allowing gradients from either stage to propagate through the shared representations. Second, a dual-axis preference alignment objective coordinates the two stages: horizontally, a compact aggregation term reorganizes dozens of pairwise objectives over heterogeneous prior signals into bidirectional preference evidence; vertically, a cross-stage consistency term transfers downstream pairwise preferences to the upstream fusion score. Third, we introduce attribute group-relative regularization, which computes relative advantages and normalizes policy updates within each attribute group, ensuring that uniformly promoting all items in a high-reward group provides no additional optimization gain. Offline experiments demonstrate UniRec consistently outperforms single-stage fusion and cross-stage coordination baselines; online A/B experiments show a 0.616% gain in app usage duration. UniRec has been fully deployed on the Kuaishou platform.

cs.IR↗

arXiv2026

Scaling Articulated Rationales for MLLM-based Recommendation

We presented SARA, an industrial framework that transforms sparse articulated user rationales into scalable recommendation signals. Its data engine curates questionnaire responses into SARA-HQ, providing explicit preference supervision for aligning SARA-7B through SFT and Quality-Refining DPO. This alignment extends rationale generation from $86{,}564$ questionnaire-covered authors to the full $10$M-author space. SARA-Ranker translates the generated positive and negative rationales into features for user--author interaction modeling and negative-feedback history modeling, connecting articulated reasons to production ranking. Evaluation on unseen authors demonstrates that SARA-7B generates more specific, relevant, and grounded rationales than the evaluated general-purpose MLLMs. On top of a strong industrial ranking baseline with multimodal features, separate online A/B tests show that positive-rationale integration increases watch time by $0.99\%$, while negative-rationale integration reduces Hate feedback by $8.16\%$. Daily refresh and more than $30$ days of production deployment further demonstrate the operational feasibility of the approach. These findings establish articulated rationales as a useful complement to behavioral and content signals, and demonstrate a practical role for MLLMs in scaling sparse human explanations into preference information that improves industrial recommendation.

cs.IR↗