RecoReward: Recommender-Guided Multimodal Description Generation for Recommendation

• URL: http://arxiv.org/abs/2607.25901v2
• Date: Sun, 02 Aug 2026 06:00:57 GMT
• Title: RecoReward: Recommender-Guided Multimodal Description Generation for Recommendation
• Abstract summary: We introduce RecoReward, which uses behavior-derived rewards during training and preserves content-only inference.<n>We show that RecoReward trains the MLLM to produce item features that benefit downstream recommendation while retaining content-only serving.
• Score: 41.239756064223776
• License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
• Abstract: Multimodal large language models (MLLMs) can convert multimodal item content into structured descriptions used as semantic features for recommendation. Conventional content-only generation, however, cannot use downstream user signals to determine which semantics should be emphasized. Recent user-conditioned methods incorporate these signals through user histories or profiles, but they require user information at inference and make generation user-dependent. In this paper, we introduce RecoReward, which instead uses behavior-derived rewards during training and preserves content-only inference. To instantiate this idea in live-stream recommendation, we treat historically engaged users as a proxy for future target users and use observational non-target users to estimate affinity shared broadly across users. The Recommender Affinity Score (RAS) contrasts these signals to provide user-selective feedback for reinforcement learning, allowing the learned policy to generate a single shared description without user inputs. In our offline benchmark, RecoReward-9B outperforms its Qwen3.5-9B baseline and all other evaluated models across seven recall metrics. Online A/B testing also shows performance gains. These results show that RecoReward trains the MLLM to produce item features that benefit downstream recommendation while retaining content-only serving.

Related papers

• Agents as Knowledge Integrator and Utilizer in Multimodal Recommendation [32.21401177935487]
  multimodal content should be interpreted together with user behavior before it is used to construct recommendation graphs or adjust rankings.<n>We propose AgentMMRec, an agent-based multimodal recommendation framework with two coordinated roles.<n> Experiments on three Amazon multimodal recommendation datasets show that AgentMMRec consistently improves Recall and NDCG over recent multimodal baselines.
  arXiv  Detail & Related papers  (2026-08-29T19:25:18Z)
• RECAP: Feedback-Driven Streaming Semantic User Profiles for Short-Video Recommendation [74.86562505934961]
  Language-based user profiles convert long behavioral histories into explicit semantic representations for recommendation.<n>We study this problem in real-world short-video recommendation, where user behaviors continuously arrive as streams and profiles must be incrementally updated under limited capacity.<n>We propose RECAP, an offline closed-loop framework for optimizing streaming structured semantic profiles with historical recommendation feedback.
  arXiv  Detail & Related papers  (2026-07-17T08:10:37Z)
• Toward User Preference Alignment in LLM Recommendation via Explicit Context Feedback [46.88440233575785]
  We advocate for prioritizing explicit context feedback in the next generation of Large Language Models (LLMs)-based RecSys.<n>We aim to foster more personalized, transparent, and explainable RecSys online platforms.
  arXiv  Detail & Related papers  (2026-05-27T22:10:33Z)
• RecoWorld: Building Simulated Environments for Agentic Recommender Systems [55.979427290369216]
  We present RecoWorld, a blueprint for building simulated environments tailored to agentic recommender systems.<n>A user simulator reviews recommended items, updates its mindset, and when sensing potential user disengagement, generates reflective instructions.<n>The agentic recommender adapts its recommendations by incorporating these user instructions and reasoning traces, creating a dynamic feedback loop.
  arXiv  Detail & Related papers  (2025-09-12T16:44:34Z)
• Multi-agents based User Values Mining for Recommendation [52.26100802380767]
  We propose a zero-shot multi-LLM collaborative framework for effective and accurate user value extraction.<n>We apply text summarization techniques to condense item content while preserving essential meaning.<n>To mitigate hallucinations, we introduce two specialized agent roles: evaluators and supervisors.
  arXiv  Detail & Related papers  (2025-05-02T04:01:31Z)
• SymCERE: Symmetric Contrastive Learning for Robust Review-Enhanced Recommendation [2.087411180679868]
  We propose SymCERE, a contrastive learning method that addresses false negatives in recommendation.<n>Experiments on 15 datasets from three platforms demonstrate that SymCERE outperforms several strong baselines.
  arXiv  Detail & Related papers  (2025-04-03T00:40:09Z)
• LLM-based User Profile Management for Recommender System [15.854727020186408]
  PURE builds and maintains evolving user profiles by systematically extracting and summarizing key information from user reviews.<n>We introduce a continuous sequential recommendation task that reflects real-world scenarios by adding reviews over time and updating predictions incrementally.<n>Our experimental results on Amazon datasets demonstrate that PURE outperforms existing LLM-based methods.
  arXiv  Detail & Related papers  (2025-02-20T13:20:19Z)
• Modeling User Retention through Generative Flow Networks [34.74982897470852]
  Flow-based modeling technique can back-propagate the retention reward towards each recommended item in the user session. We show that the flow combined with traditional learning-to-rank objectives eventually optimized a non-discounted cumulative reward for both immediate user feedback and user retention.
  arXiv  Detail & Related papers  (2024-06-10T06:22:18Z)
• Lusifer: LLM-based User SImulated Feedback Environment for online Recommender systems [0.0]
  Reinforcement learning (RL) recommender systems often rely on static datasets that fail to capture the fluid, ever changing nature of user preferences in real-world scenarios.<n>We introduce Lusifer, an LLM-based simulation environment designed to generate dynamic, realistic user feedback for RL-based recommender training.
  arXiv  Detail & Related papers  (2024-05-22T05:43:15Z)
• Latent User Intent Modeling for Sequential Recommenders [92.66888409973495]
  Sequential recommender models learn to predict the next items a user is likely to interact with based on his/her interaction history on the platform. Most sequential recommenders however lack a higher-level understanding of user intents, which often drive user behaviors online. Intent modeling is thus critical for understanding users and optimizing long-term user experience.
  arXiv  Detail & Related papers  (2022-11-17T19:00:24Z)
• ELIXIR: Learning from User Feedback on Explanations to Improve Recommender Models [26.11434743591804]
  We devise a human-in-the-loop framework, called ELIXIR, where user feedback on explanations is leveraged for pairwise learning of user preferences. ELIXIR leverages feedback on pairs of recommendations and explanations to learn user-specific latent preference vectors. Our framework is instantiated using generalized graph recommendation via Random Walk with Restart.
  arXiv  Detail & Related papers  (2021-02-15T13:43:49Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.