SPANER: Shared Prompt Aligner for Multimodal Semantic Representation
This work addresses the limitation of cross-modal generalization in multimodal learning, offering a scalable solution for aligning embedding structures, though it appears incremental as it builds on existing PEFT approaches.
The paper tackled the problem of isolated modality-specific representations in multimodal learning by introducing SPANER, a modality-agnostic PEFT framework that embeds inputs into a unified semantic space, achieving competitive few-shot retrieval performance on vision-language and audio-visual benchmarks.
Recent advances in multimodal Parameter-Efficient Fine-Tuning (PEFT) have significantly improved performance on downstream tasks such as few-shot retrieval. However, most existing approaches focus on task-specific gains while neglecting the structure of the multimodal embedding space. As a result, modality-specific representations often remain isolated, limiting cross-modal generalisation. In this work, we introduce Shared Prompt AligNER (SPANER), a modality-agnostic PEFT framework designed to embed inputs from diverse modalities into a unified semantic space. At its core, SPANER employs a shared prompt mechanism that acts as a conceptual anchor, enabling semantically related instances to converge spatially regardless of modality. This shared prompt design is inherently extensible, supporting the seamless integration of additional modalities, such as audio, without altering the core architecture. Through comprehensive experiments across vision-language and audio-visual benchmarks, SPANER demonstrates competitive few-shot retrieval performance while preserving high semantic coherence in the learned embedding space. Our results highlight the importance of aligning embedding structures, rather than merely tuning adapter weights, for scalable multimodal learning.