Latent-Aligned Reasoning for Multimodal Recommendation

2026-09-03

AI Summary

为解决多模态推荐中视觉和文本信号衰减的问题，提出LARK框架，通过两阶段潜对齐机制保持推理链中的感知细节。

Abstract

Multimodal Vision-Language Models (VLMs) have demonstrated remarkable capabilities in cross-modal understanding, yet a fundamental challenge persists when applying them to recommendation: as representations propagate through multi-step reasoning, both visual and textual signals progressively attenuate - a phenomenon we term cross-modal dilution. To address this, we propose LARK (Latent-Aligned Reasoning frameworK), a two-stage latent reasoning framework with complementary alignment mechanisms within a single VLM. In the first stage, learnable latent tokens are interleaved with multi-step chain-of-thought (CoT) reasoning and explicitly aligned with a frozen vision encoder, serving as visual checkpoints that preserve perceptual details throughout the reasoning chain. In the second stage, the latent representations are projected via a bridge MLP and trained with item-to-item contrastive learning; to prevent the reasoning semantics from fading, intermediate features are aligned with the CoT hidden states from the first stage, anchoring the final embeddings to the model's own reasoning output. Experiments on three public benchmarks and one industrial dataset show that LARK achieves state-of-the-art performance across multiple recommendation architectures, with controlled ablations confirming the distinct contribution of each component.

Problem

• Multimodal Vision-Language Models
• cross-modal dilution
• recommendation

Innovation

• Latent-Aligned Reasoning
• Cross-Modal Dilution
• Chain-of-Thought
• Contrastive Learning
• Multimodal Recommendation