AI Paper Atlas

Research map / Label-efficient and robust learning / Recognition and long-tail

FANoise: Singular Value-Adaptive Noise Modulation for Robust Multimodal Representation Learning

Jiaoyang Li, Jun Fang, Tianhao Gao, Xiaohui Zhang, Zhiyuan Liu, Chao Liu, Pengzhang Liu, Qixia Jiang

AAAI 2026 · Label-efficient and robust learning › Recognition and long-tail · approach: Multilingual language representation pretraining · 1 citations (OpenAlex)

Read the paperCheck your idea against 137,217 papers

AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
add FANoise sublinear noiseVLM2Vec Phi3.5-V · MMEB (36 datasets)overall Precision@1: 60.1 -> 60.8Improved
add FANoise sublinear noiseVLM2Vec Qwen2-VL-2B · MMEB (36 datasets)overall Precision@1: 60.1 -> 61.1Improved
add FANoise sublinear noiseVLM2Vec LLaVA-1.6-LR · MMEB (36 datasets)overall Precision@1: 55.0 -> 59.2Improved
add FANoise sublinear noiseVLM2Vec LLaVA-1.6-HR · MMEB (36 datasets)overall Precision@1: 62.9 -> 66.4Improved
add FANoise sublinear noiseVLM2Vec Qwen2-VL-7B · MMEB (36 datasets)overall Precision@1: 65.8 -> 66.6Improved
add FANoise sublinear noiseVLM2Vec LLaVA-1.6-LR · MMEB VQAVQA score: 50.3 -> 43.1Hurt
add FANoise sublinear noiseVLM2Vec Qwen2-VL-2B · MMEB groundingGrounding score: 73.4 -> 72.3Hurt
add uniform scaling noise (ablation)VLM2Vec Qwen2-VL-2B · MMEB (small-budget study)overall Precision@1: 60.06 -> 60.93Improved
scale noise linearly with singular values (ablation)VLM2Vec Qwen2-VL-2B · MMEB (small-budget study)overall Precision@1: 60.06 -> 60.69Improved
scale noise sublinearly (sqrt) rather than uniformly (ablation)VLM2Vec Qwen2-VL-2B · MMEB (small-budget study)overall Precision@1: 60.93 -> 61.08Improved
inject noise at output layer instead of input layer (ablation)VLM2Vec Qwen2-VL-2B · MMEB (noise position study)overall Precision@1: 48.46 -> 48.73Improved
increase noise strength alpha too high (ablation)FANoise_ss · MMEB (noise strength study)overall Precision@1Hurt

Closest papers by meaning

Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy