跳到正文
HuggingFace Daily Papers(社區熱門論文)·· 19 天前AI 評分50

VeriSpoof 以單一音訊 LLM 統合五項語音鑑證任務

VeriSpoof: Beyond Real or Fake with One Audio LLM

AI 導讀

VeriSpoof 是一款音訊語言模型,以共用權重整合深偽偵測、生成器歸因、錄音條件分析、合成片段定位和來源驗證五項任務。

正文

Published on Nov 3

Authors:

Abstract

VeriSpoof is an audio language model that unifies deepfake detection, generator attribution, source verification, and evidence-based rationale generation with minimal parameter updates.

A speech deepfake detector estimates whether a recording is synthetic, but a forensic analyst also asks which system produced it, whether recordings share a source, and what evidence supports each conclusion. We present VeriSpoof, an audio language model that unifies five tasks with shared weights: deepfake detection, generator attribution (generator, acoustic model and vocoder), recording-condition analysis, synthesized-span localisation, and source verification. Each decision is supported by grounded rationales whose statements can be independently verified. Answers are scored by sentence likelihood, allowing detection and source verification to be thresholded without separate heads. Rationale statements can be checked against waveform measurements, span labels, or generator metadata. Counterfactual examples train decisions to follow the stated evidence. Across the 14 Speech-DF-Arena test sets, VeriSpoof-Plus achieves 2.14% pooled EER, ranking third of 28 systems and first among open-weight systems. VeriSpoof-Base, trained only on official training splits of eight corpora, achieves 4.43% EER, less than half that of the next open-weight detector. With the same weights, VeriSpoof-Plus verifies whether recordings from generators unseen in training share a source at 6.2% EER. Controlled studies show that the audio representation governs detection, the LLM governs recording comparison, and source supervision is necessary for neural-codec fakes. Statement-checkable rationales catch partial fakes missed by one-step answers (99.6% vs. 78.4% on PartialEdit) and determine decisions under edits unseen in training (99.5%). Training updates only 29.7M of 2.85B parameters, fitting a single 24 GB GPU with gradient checkpointing. We release both models, inference code, a demo, and VeriSpoof-Bench, the evaluation suite covering these questions.

View arXiv page View PDF Project page GitHub 3 Add to collection

Get this paper in your agent:

hf papers read 2611.04634

Don't have the latest CLI?

curl -LsSf https://hf.co/cli/install.sh | bash

Models citing this paper 2

AustinXiao/VeriSpoof-Base

Audio-Text-to-Text •

3B •

Updated about 18 hours ago

•

95

•

3

AustinXiao/VeriSpoof-Plus

Audio-Text-to-Text •

3B •

Updated about 17 hours ago

•

1

Datasets citing this paper 1

AustinXiao/VeriSpoof-Bench

Viewer •

Updated about 21 hours ago

•

356k

Spaces citing this paper 1

Collections including this paper 0

No Collection including this paper

Add this paper to a collection to link it from this page.

來源:HuggingFace Daily Papers(社區熱門論文) · huggingface.co