LUMA-RAG: Lifelong Multimodal Agents with Provably Stable Streaming Alignment

TL;DR

LUMA-RAG introduces a lifelong multimodal agent architecture that maintains stable, efficient streaming alignment across heterogeneous data streams, enabling reliable retrieval and cross-modal consistency in dynamic environments.

Contribution

The paper presents a novel streaming, multi-tier memory system and alignment method that ensure stable, efficient multimodal retrieval in continuous data streams, addressing key challenges in dynamic AI agents.

Findings

Achieves 94% recall in text-to-image retrieval

Maintains stable audio-to-image rankings with Safe@1=1.0

Demonstrates graceful degradation under quantization offloading

Abstract

Retrieval-Augmented Generation (RAG) has emerged as the dominant paradigm for grounding large language model outputs in verifiable evidence. However, as modern AI agents transition from static knowledge bases to continuous multimodal streams encompassing text, images, video, and audio, two critical challenges arise: maintaining index freshness without prohibitive re-indexing costs, and preserving cross-modal semantic consistency across heterogeneous embedding spaces. We present LUMA-RAG, a lifelong multimodal agent architecture featuring three key innovations: (i) a streaming, multi-tier memory system that dynamically spills embeddings from a hot HNSW tier to a compressed IVFPQ tier under strict memory budgets; (ii) a streaming CLAP->CLIP alignment bridge that maintains cross-modal consistency through incremental orthogonal Procrustes updates; and (iii) stability-aware retrieval telemetry providing Safe@k guarantees by jointly bounding alignment drift and quantization error. Experiments demonstrate robust text-to-image retrieval (Recall@10 = 0.94), graceful performance degradation under product quantization offloading, and provably stable audio-to-image rankings (Safe@1 = 1.0), establishing LUMA-RAG as a practical framework for production multimodal RAG systems.