Latent-Lens：通过潜空间通信让小语言模型获得视觉感知

View PDF HTML (experimental)

Abstract:Bridging independent vision-language models (VLMs) and small language models (SLMs) in a modular multi-agent system typically forces them to exchange information as discrete text, which creates a lossy bottleneck for spatial and fine-grained details. This constraint is especially acute for SLMs, whose limited capacity makes them more reliant on the quality of the input signal. Recent latent-communication methods exchange states between language models. How a VLM can pass visual information to a separately pretrained text-only SLM though remains unexplored. We propose \textbf{Latent-Lens}, a lightweight codec that bridges the representation gap between a frozen VLM and a frozen SLM directly in latent space, enabling them to communicate purely through hidden states. The codec consists of a domain encoder that projects hidden states from the VLM into the SLM's embedding space, and an attention decoder formalised as a LoRA that teaches the SLM to attend to these visual prefixes, while training only ${\sim}1.6$M parameters (0.2\% of the combined model size). On visual question answering (CLEVR, GQA), Latent-Lens achieves $81.0\pm7.9$\% and $65.4\pm5.4$\%, compared with $58.3\pm5.7$\% and $61.8\pm1.0$\% for text communication with the same LoRA budget (+30.2--69.8pp relative to frozen text communication), while on image captioning (Flickr8k) it improves over the same baseline by +16.5 (+18.8 frozen). We also evaluate Latent-Lens with TinyLlama-1.1B and InternVL2-1B, which use different architectures, vocabularies, and hidden dimensions.

Comments:	Published at NeurIPS 2026
Subjects:	Machine Learning (cs.LG); Artificial Intelligence (cs.AI)
Cite as:	arXiv:2610.41751 [cs.LG]
	(or arXiv:2610.41751v1 [cs.LG] for this version)
	https://doi.org/10.48550/arXiv.2610.41751 arXiv-issued DOI via DataCite (pending registration)

Submission history

From: Mo Shahdloo [view email]
[v1] Fri, 30 Oct 2026 13:53:44 UTC (2,969 KB)