用小模型实现高效多模态搜索推荐，性能接近大模型。
Scaling Multimodal Search and Recommendation with Small Language Models via Upside-Down Reinforcement Learning
- 用逆向强化学习+合成数据训练1亿参数小模型生成多任务提示。
- 小模型在相关性和多样性上仅比大模型低6%，延迟和内存大幅降低。
- 适合资源受限的实时应用，如媒体推荐、创意内容生成。
本文研究如何将小型语言模型（SLMs）扩展至支持多模态搜索与推荐任务，同时保持实时、资源受限环境下的高效性。我们提出一个框架，结合逆向强化学习与来自大语言模型（Llama-3）的合成数据蒸馏，训练一个1亿参数的GPT-2模型以实现多任务提示生成。尽管该SLM比当前最先进的大语言模型（LLMs）小多达80倍，其在相关性和多样性指标上仍达到Llama-3 8B、Qwen3 8B和Ministral 8B等基线模型的94%水平。结果表明，小型模型可有效承担多模态搜索与推荐任务，显著降低推理延迟与内存开销。本研究凸显了轻量级模型在可扩展多模态发现中的潜力，弥合前沿研究与实际应用（如媒体推荐、创意内容生成）之间的差距。
原文摘要 · Abstract (English)
In this work, we investigate how small language models (SLMs) can be scaled to support multimodal search and recommendation use cases while remaining efficient enough for real-time, resource-constrained deployments. We present a framework that combines upside-down reinforcement learning with synthetic data distillation from a large language model (Llama-3) to train a 100M-parameter GPT-2 model for multitask prompt generation. Despite being up to 80 times smaller than state-of-the-art large language models (LLMs), our SLM achieves relevance and diversity scores within 6% of competitive baselines such as Llama-3 8B, Qwen3 8B, and Ministral 8B. These results demonstrate that SLMs can effectively handle multimodal search and recommendation tasks, while dramatically reducing inference latency and memory overhead. Our study highlights the potential of lightweight models as practical engines for scalable multimodal discovery, bridging the gap between cutting-edge research and real-world multimodal applications such as media recommendations and creative content generation.
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。