Cover Image

GNSP: Gradient Null Space Projection for Preserving Cross-Modal Alignment in VLMs Continual Learning

Fuente: arXiv

Saved in:

Bibliographic Details
Main Authors:	Peng, Tiantian, Liu, Yuyang, Yang, Shuo, Hong, Qiuhe, Tian, YongHong
Format:	Preprint
Published:	2025
Subjects:	Machine Learning Computer Vision and Pattern Recognition
Online Access:	Acceder al recurso
Tags:	Add Tag No Tags, Be the first to tag this record!

Similar Items

Reasoning Portability: Guiding Continual Learning for MLLMs in the RLVR Era
by: Hong, Qiuhe, et al.
Published: (2026)

Continual Learning for VLMs: A Survey and Taxonomy Beyond Forgetting
by: Liu, Yuyang, et al.
Published: (2025)

Visual Prompt Tuning in Null Space for Continual Learning
by: Lu, Yue, et al.
Published: (2024)

AttAnchor: Guiding Cross-Modal Token Alignment in VLMs with Attention Anchors
by: Zhang, Junyang, et al.
Published: (2025)

Mamba-CL: Optimizing Selective State Space Model in Null Space for Continual Learning
by: Cheng, De, et al.
Published: (2024)

DecAlign: Hierarchical Cross-Modal Alignment for Decoupled Multimodal Representation Learning
by: Qian, Chengxuan, et al.
Published: (2025)

NP-LoRA: Null Space Projection for Subject-Style LoRA Fusion
by: Chen, Chuheng, et al.
Published: (2025)

Preserving Cross-Modal Stability for Visual Unlearning in Multimodal Scenarios
by: Li, Jinghan Xu Yuyang Zhang Qixuan Cai Jiancheng Chen Keqiu
Published: (2025)

Learning Modality Knowledge Alignment for Cross-Modality Transfer
by: Ma, Wenxuan, et al.
Published: (2024)

Similarity-as-Evidence: Calibrating Overconfident VLMs for Interpretable and Label-Efficient Medical Active Learning
by: Xie, Zhuofan, et al.
Published: (2026)

Learning in the Null Space: Small Singular Values for Continual Learning
by: Pham, Cuong Anh, et al.
Published: (2026)

Enhancing CLIP Robustness via Cross-Modality Alignment
by: Zhu, Xingyu, et al.
Published: (2025)

BioVLM: Routing Prompts, Not Parameters, for Cross-Modality Generalization in Biomedical VLMs
by: Singha, Mainak, et al.
Published: (2026)

Deep Pre-Alignment for VLMs
by: Yu, Tianyu, et al.
Published: (2026)

StructAlign: Structured Cross-Modal Alignment for Continual Text-to-Video Retrieval
by: Wang, Shaokun, et al.
Published: (2026)

Preserving Localized Patch Semantics in VLMs
by: Esmaeilkhani, Parsa, et al.
Published: (2026)

Taming Modality Entanglement in Continual Audio-Visual Segmentation
by: Hong, Yuyang, et al.
Published: (2025)

Null-LoRA: Low-Rank Adaptation on Null Space
by: Zhang, Yi, et al.
Published: (2025)

Principled Steering via Null-space Projection for Jailbreak Defense in Vision-Language Models
by: Zhu, Xingyu, et al.
Published: (2026)

Personalized Cross-Modal Emotional Correlation Learning for Speech-Preserving Facial Expression Manipulation
by: Chen, Tianshui, et al.
Published: (2026)

Preserving Cross-Modal Consistency for CLIP-based Class-Incremental Learning
by: Chen, Haoran, et al.
Published: (2025)

Guiding Cross-Modal Representations with MLLM Priors via Preference Alignment
by: Zhao, Pengfei, et al.
Published: (2025)

Continual Cross-Modal Generalization
by: Xia, Yan, et al.
Published: (2025)

Causal Disentanglement and Cross-Modal Alignment for Enhanced Few-Shot Learning
by: Jiang, Tianjiao, et al.
Published: (2025)

CrossOver: 3D Scene Cross-Modal Alignment
by: Sarkar, Sayan Deb, et al.
Published: (2025)

Mind the Gap: Preserving and Compensating for the Modality Gap in CLIP-Based Continual Learning
by: Huang, Linlan, et al.
Published: (2025)

Towards Generalizable Deepfake Detection with Spatial-Frequency Collaborative Learning and Hierarchical Cross-Modal Fusion
by: Qiao, Mengyu, et al.
Published: (2025)

Enhancing Audio-Visual Spiking Neural Networks through Semantic-Alignment and Cross-Modal Residual Learning
by: He, Xiang, et al.
Published: (2025)

When Images Speak Louder: Mitigating Language Bias-induced Hallucinations in VLMs through Cross-Modal Guidance
by: Cao, Jinjin, et al.
Published: (2025)

Gradient Alignment for Cross-Domain Face Anti-Spoofing
by: Le, Binh M., et al.
Published: (2024)

Bridge Feature Matching and Cross-Modal Alignment with Mutual-filtering for Zero-shot Anomaly Detection
by: Bai, Yuhu, et al.
Published: (2025)

AMVICC: A Novel Benchmark for Cross-Modal Failure Mode Profiling for VLMs and IGMs
by: Basappa, Aahana, et al.
Published: (2026)

Better Reasoning with Less Data: Enhancing VLMs Through Unified Modality Scoring
by: Xu, Mingjie, et al.
Published: (2025)

iGSP:Implicit Gradient Subspace Projection for Efficient Continual Learning of Vision-Language Models
by: Cui, Xuezhi, et al.
Published: (2026)

To See or To Please: Uncovering Visual Sycophancy and Split Beliefs in VLMs
by: Hong, Rui, et al.
Published: (2026)

Can Text-to-image Model Assist Multi-modal Learning for Visual Recognition with Visual Modality Missing?
by: Feng, Tiantian, et al.
Published: (2024)

Unifying Visual and Semantic Feature Spaces with Diffusion Models for Enhanced Cross-Modal Alignment
by: Zheng, Yuze, et al.
Published: (2024)

CKAA: Cross-subspace Knowledge Alignment and Aggregation for Robust Continual Learning
by: He, Lingfeng, et al.
Published: (2025)

Preserving Multi-Modal Capabilities of Pre-trained VLMs for Improving Vision-Linguistic Compositionality
by: Oh, Youngtaek, et al.
Published: (2024)

Cross-Modality Attack Boosted by Gradient-Evolutionary Multiform Optimization
by: Gong, Yunpeng, et al.
Published: (2024)