AI Paper Atlas

Research map / Federated learning and privacy / Privacy attacks and secure learning

ShadowFPT: Backdooring Federated Prompt Tuning via Shadow Triggers

Kun Zhai, Teng Li, Yunhao Feng, Xingjun Ma

NeurIPS 2026 · Federated learning and privacy › Privacy attacks and secure learning · approach: Defending federated learning against poisoning · 0 citations (OpenAlex)

Read the paperCheck your idea against 137,217 papers

AI-extracted summary. Written by AI from the paper's full text. Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Proposes ShadowFPT, a backdoor attack on federated prompt tuning of vision-language models in which a malicious client exploits the shared frozen backbone to implant backdoor behavior through the representation space while keeping its updates ordinary-looking.

Problem

Federated prompt tuning freezes a shared pretrained vision-language backbone and trains only lightweight prompts across clients, which creates a backdoor risk that existing defenses and attacks may not address.

Approach

Since the backbone is shared and frozen, a malicious client induces backdoor behavior via the representation space rather than through conventional poisoned prompt updates, while keeping its updates looking benign. Details beyond the abstract are not available.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
add Shadow Trigger pretraining and injection (ShadowFPT-II) versus strongest baselinefederated prompt tuning backdoor · Caltech101 FPT-LASR (%): 19.81 -> 90.36Improved
pretrain trigger with auxiliary public data (ShadowFPT-II) instead of local data (ShadowFPT-I)ShadowFPT · 9 dataset-setting pairsASR (%): 95.63 -> 97.67Improved
use a pretrained Shadow Trigger instead of a random trigger (ablation)ShadowFPT · Caltech101 FPT-LASR (%): 21.32 -> 49.75Improved
add backdoor injection in FPT with adapted trigger (ablation)ShadowFPT · Caltech101 FPT-LASR (%): 49.75 -> 90.36Improved
add layer-wise visual alignment to target-text alignment (ablation)Shadow Trigger pretraining · Caltech101 FPT-LASR (%): 66.38 -> 90.36Improved
add target-text alignment to layer-wise alignment (ablation)Shadow Trigger pretraining · Caltech101 FPT-LASR (%): 59.25 -> 90.36Improved
apply ShadowFPT-II under robust aggregation/defenses (Krum, LaSA, FLAME, CRFL)ShadowFPT · defense rulesASR gain over best baseline: +73.97, +76.55, +84.75, +74.55Improved
ShadowFPT-II across prompt modalities FPT-L/V/VLShadowFPT · Avg over datasetsASR (%): 27.65 -> 96.42 (FPT-L)Improved
ShadowFPT-II on fine-grained Oxford Flowers with textual promptsShadowFPT · Oxford Flowers FPT-Lclean accuracy (%): 66.63 -> 60.85Hurt
ShadowFPT-II clean-accuracy cost across defensesShadowFPT · Avg over datasets and settingsclean accuracy (pp): -1.71 averageMixed

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