InduceKV is a retrieval-based continual adaptation method that uses bilevel selection to build a compact set of inducing KV memories for fixed-footprint updates to multimodal LLMs.
No images, no problem: Retaining knowledge in continual vqa with questions-only memory
4 Pith papers cite this work. Polarity classification is still indexing.
4
Pith papers citing it
citation-role summary
background 1
citation-polarity summary
roles
background 1polarities
background 1representative citing papers
ASR regularizes skill-wise prototype distributions of Fourier power spectra from cross-attention maps to reduce catastrophic forgetting in continual multimodal LLMs without data replay.
AIM applies modality-specific masks to balance stability and plasticity in asymmetric VLMs, achieving SOTA average performance and reduced forgetting on continual VQA v2 and GQA while preserving generalization to novel compositions.
A taxonomy survey of continual learning for vision-language models, grouping methods into multi-modal replay, cross-modal regularization, and parameter-efficient adaptation, with a review of benchmarks and metrics.
citing papers explorer
-
InduceKV: Fixed-Footprint Continual Adaptation of Multimodal LLMs via Inducing KV Memories
InduceKV is a retrieval-based continual adaptation method that uses bilevel selection to build a compact set of inducing KV memories for fixed-footprint updates to multimodal LLMs.
-
Attention-Spectrum Regularization for Replay-Free Continual Multimodal LLMs
ASR regularizes skill-wise prototype distributions of Fourier power spectra from cross-attention maps to reduce catastrophic forgetting in continual multimodal LLMs without data replay.
-
AIM: Asymmetric Information Masking for Visual Question Answering Continual Learning
AIM applies modality-specific masks to balance stability and plasticity in asymmetric VLMs, achieving SOTA average performance and reduced forgetting on continual VQA v2 and GQA while preserving generalization to novel compositions.
-
Continual Learning for VLMs: A Survey and Taxonomy Beyond Forgetting
A taxonomy survey of continual learning for vision-language models, grouping methods into multi-modal replay, cross-modal regularization, and parameter-efficient adaptation, with a review of benchmarks and metrics.