This work introduces Progressive Multimodal Alignment (PMA), a framework that enables the projector to adapt continually while preserving previously learned alignment, and scales across diverse MLLM backbones, demonstrating robust and broadly applicable MCIT performance.
Abstract
Multimodal Large Language Models (MLLMs) rely on a projector to align visual representations with the language embedding space, making it central to cross-modal understanding. In Multimodal Continual Instruction Tuning (MCIT), however, shifting visual distributions and evolving instruction semantics cause this shared projector to drift, leading to projector-level forgetting, an issue largely overlooked by methods that focus primarily on the LLM backbone. We introduce Progressive Multimodal Alignment (PMA), a framework that enables the projector to adapt continually while preserving previously learned alignment. PMA detects multimodal distribution shifts via a lightweight representation descriptor and progressively expands projector experts only when needed. An expandable router integrates expert outputs based on multimodal features, while the original pretrained projector is retained as a stable alignment anchor. This progressive mechanism balances stability and plasticity with sub-linear parameter growth and serves as a method-agnostic add-on to existing MCIT approaches. Extensive experiments on two recent MCIT benchmarks demonstrate that mitigating projector-level forgetting yields consistent gains over prior state-of-the-art methods when combined with PMA. Moreover, PMA scales across diverse MLLM backbones, demonstrating robust and broadly applicable MCIT performance.
This work argues that cross-modal alignment is implicitly captured in the information-compression trajectory, and proposes LLaVAFlow, an information-theoretic distillation framework that preserves alignment flow and enhances both downstream performance and generalization.
Muyao Yuan, Muyan Jiao, Jiangyong Ying et al.· 0 citations
This empirical research establishes the essential groundwork for predictably scaling multimodal foundation models by modeling the influence of data composition on compute laws and allocation exponents and derive an efficiency frontier specifying precise configurations of model size, token count, and data mixture.
The results suggest that competitive MLLM can emerge from alignment alone, reducing multimodal extension to a lightweight projector-training problem that generalizes across modalities and adapts rapidly to each new LLM release.
Xuanru Zhou, Yiwen Shao, Jiahong Li et al.· 1 citation
CoRe-MoE is proposed, a Compact Reusable MoE framework for parameter-efficient continual multimodal instruction tuning that improves final average performance over the strongest competing baseline by up to 5.90 points, while using less than 1% of the trainable parameters required by sequential LoRA for later tasks.
Run-Ze Liu, Naibin Gu, Ming-Xu Ai et al.· 0 citations
Multimodal models often build on architectures designed for generative vision-language modeling, typically combining separately pretrained vision encoders with causal language models. Visual document retrievers such as ColPali repurpose these models as encoders, carrying over the parameter and compute overhead of a VLM for a non-generative task. We introduce NeoMME, a family of 260M and 800M-parameter Multimodal and Multilingual bidirectional Encoders that process multilingual text and raw image patches in a single bidirectional Transformer encoder. Both models are pretrained from scratch with a masked discrete-diffusion text objective, conditioned on visible image patches for multimodal examples. Both support a 16,384-token context, enough to encode up to two standard 4K UHD images. To demonstrate its downstream capabilities, we fine-tune NeoMME with jointly trained dense and late-interaction heads. On the ViDoRe v3 benchmark, the resulting NeoMME-Retriever 260M outperforms all evaluated models strictly below 800M parameters with 0.523 nDCG@10, while NeoMME-Retriever 800M reaches 0.556. At a matched 2048x2048 image input size on an NVIDIA L40S, NeoMME-260M encodes pages with about 2x the throughput of ColModernVBERT. Hierarchical token pooling and asymmetric quantization compress late-interaction multimodal document embeddings by 255x while preserving over 95% of baseline nDCG@10. We contribute NeoMME to Hugging Face Transformers and release the pretrained backbone and retrieval-compatible checkpoints under Apache 2.0 at https://hf.co/collections/Hcompany/neomme.
This work proposes Causal Localization and Retain-Aware Projection (CLRP), a lightweight training-free framework that uses activation patching to identify the layer that causally mediates target-attribute disclosure, then applies a retain-aware projection that removes the target-attribute subspace while preserving same-identity evidence.
Junkai Lin, Junkai Chen, Siqi Hou et al.· 0 citations
We use cookies to run the site and, with your consent, for analytics and to show ads.
See our Cookie Policy.
