Interpretable All-Type Audio Deepfake Detection with Audio LLMs via Frequency-Time Reinforcement Learning

• URL: http://arxiv.org/abs/2601.02983v1
• Date: Tue, 06 Jan 2026 12:50:02 GMT
• Title: Interpretable All-Type Audio Deepfake Detection with Audio LLMs via Frequency-Time Reinforcement Learning
• Abstract summary: Real-world audio deepfake detection requires all-type detectors that generalize across heterogeneous audio.<n>We propose a two-stage training paradigm that cold-starts ALLMs with SFT and then applies GRPO under rule-based frequency-time constraints.<n> Experiments demonstrate FT-GRPO achieves state-of-the-art performance on all-type ADD while producing interpretable, FT-grounded rationales.
• Score: 31.51615040769059
• License: http://creativecommons.org/licenses/by/4.0/
• Abstract: Recent advances in audio large language models (ALLMs) have made high-quality synthetic audio widely accessible, increasing the risk of malicious audio deepfakes across speech, environmental sounds, singing voice, and music. Real-world audio deepfake detection (ADD) therefore requires all-type detectors that generalize across heterogeneous audio and provide interpretable decisions. Given the strong multi-task generalization ability of ALLMs, we first investigate their performance on all-type ADD under both supervised fine-tuning (SFT) and reinforcement fine-tuning (RFT). However, SFT using only binary real/fake labels tends to reduce the model to a black-box classifier, sacrificing interpretability. Meanwhile, vanilla RFT under sparse supervision is prone to reward hacking and can produce hallucinated, ungrounded rationales. To address this, we propose an automatic annotation and polishing pipeline that constructs Frequency-Time structured chain-of-thought (CoT) rationales, producing ~340K cold-start demonstrations. Building on CoT data, we propose Frequency Time-Group Relative Policy Optimization (FT-GRPO), a two-stage training paradigm that cold-starts ALLMs with SFT and then applies GRPO under rule-based frequency-time constraints. Experiments demonstrate that FT-GRPO achieves state-of-the-art performance on all-type ADD while producing interpretable, FT-grounded rationales. The data and code are available online.

Related papers

• TEMPO: Temporally-grounded Multi-task Post-training for Large Audio-Language Models [58.80225070159013]
  Large audio-language models (LALMs) describe audio at the clip level but cannot assign timestamps to the events, speakers, or sounds they identify.<n>We present TEMPO, the first unified model to handle audio, speech, and music timestamping tasks.
  arXiv  Detail & Related papers  (2026-08-30T19:49:21Z)
• An Efficient vLLM-Based Inference Pipeline for Unified Audio Understanding and Generation [63.50393485685279]
  We present a vLLM-based inference pipeline for unified speech understanding and generation.<n>We overcome the shared intuition-Free Guidance (CFG) halves throughput.<n>By co-scheduling paired conditional and unconditional requests within a continuous batch, our CFG implementation sustains 80% of non-CFG throughput.
  arXiv  Detail & Related papers  (2026-07-02T12:55:11Z)
• Temporal Contrastive Decoding: A Training-Free Method for Large Audio-Language Models [56.91801348360746]
  Large audio-language models (LALMs) generalize across speech, sound, and music.<n> unified decoders can exhibit a emphtemporal smoothing bias<n>We propose emphTemporal Contrastive Decoding (TCD), a training-free decoding method for unified LALMs.
  arXiv  Detail & Related papers  (2026-04-16T02:30:41Z)
• SONAR: Spectral-Contrastive Audio Residuals for Generalizable Deepfake Detection [6.042897432654865]
  Spectral-cONtrastive Audio Residuals (AR) is a frequency-guided framework for deepfake audio detectors.<n>AR disentangles an audio signal into complementary representations.<n> evaluated on the ASVspoof 2021 and in-the-wild benchmarks.
  arXiv  Detail & Related papers  (2025-11-26T12:16:38Z)
• End-to-end Audio Deepfake Detection from RAW Waveforms: a RawNet-Based Approach with Cross-Dataset Evaluation [8.11594945165255]
  We propose an end-to-end deep learning framework for audio deepfake detection that operates directly on raw waveforms.<n>Our model, RawNetLite, is a lightweight convolutional-recurrent architecture designed to capture both spectral and temporal features without handcrafted preprocessing.
  arXiv  Detail & Related papers  (2025-04-29T16:38:23Z)
• Pseudo-Autoregressive Neural Codec Language Models for Efficient Zero-Shot Text-to-Speech Synthesis [64.12708207721276]
  We introduce a novel pseudo-autoregressive (PAR) language modeling approach that unifies AR and NAR modeling.<n>Building on PAR, we propose PALLE, a two-stage TTS system that leverages PAR for initial generation followed by NAR refinement.<n>Experiments demonstrate that PALLE, trained on LibriTTS, outperforms state-of-the-art systems trained on large-scale data.
  arXiv  Detail & Related papers  (2025-04-14T16:03:21Z)
• Detect All-Type Deepfake Audio: Wavelet Prompt Tuning for Enhanced Auditory Perception [19.10177637063233]
  Existing countermeasures (CMs) perform well in single-type audio deepfake detection (ADD), but their performance declines in cross-type scenarios.<n>We are the first to comprehensively establish an all-type ADD benchmark to evaluate current CMs, incorporating cross-type deepfake detection across speech, sound, singing voice, and music.<n>Considering the auditory perception of different audio types,we propose the wavelet prompt tuning (WPT)-SSL method to capture type-invariant auditory deepfake information.
  arXiv  Detail & Related papers  (2025-04-09T10:18:45Z)
• Fast Streaming Transducer ASR Prototyping via Knowledge Distillation with Whisper [3.717584661565119]
  We demonstrate that streaming Transformer-Transducer (TT) models can be trained from scratch without supervised data. This allows training a robust ASR model just in one stage and does not require large data and computational budget. We validate the proposed framework on 6 languages from CommonVoice and propose multiple filters to filter out hallucinated PLs.
  arXiv  Detail & Related papers  (2024-09-20T13:38:59Z)
• VALL-E R: Robust and Efficient Zero-Shot Text-to-Speech Synthesis via Monotonic Alignment [101.2489492032816]
  VALL-E R is a robust and efficient zero-shot Text-to-Speech system. This research has the potential to be applied to meaningful projects, including the creation of speech for those affected by aphasia.
  arXiv  Detail & Related papers  (2024-06-12T04:09:44Z)
• Frequency-Aware Deepfake Detection: Improving Generalizability through Frequency Space Learning [81.98675881423131]
  This research addresses the challenge of developing a universal deepfake detector that can effectively identify unseen deepfake images. Existing frequency-based paradigms have relied on frequency-level artifacts introduced during the up-sampling in GAN pipelines to detect forgeries. We introduce a novel frequency-aware approach called FreqNet, centered around frequency domain learning, specifically designed to enhance the generalizability of deepfake detectors.
  arXiv  Detail & Related papers  (2024-03-12T01:28:00Z)
• High-Fidelity Speech Synthesis with Minimal Supervision: All Using Diffusion Models [56.00939852727501]
  Minimally-supervised speech synthesis decouples TTS by combining two types of discrete speech representations. Non-autoregressive framework enhances controllability, and duration diffusion model enables diversified prosodic expression.
  arXiv  Detail & Related papers  (2023-09-27T09:27:03Z)
• Fully Automated End-to-End Fake Audio Detection [57.78459588263812]
  This paper proposes a fully automated end-toend fake audio detection method. We first use wav2vec pre-trained model to obtain a high-level representation of the speech. For the network structure, we use a modified version of the differentiable architecture search (DARTS) named light-DARTS.
  arXiv  Detail & Related papers  (2022-08-20T06:46:55Z)
• FastDiff: A Fast Conditional Diffusion Model for High-Quality Speech Synthesis [90.3069686272524]
  This paper proposes FastDiff, a fast conditional diffusion model for high-quality speech synthesis. FastDiff employs a stack of time-aware location-variable convolutions of diverse receptive field patterns to efficiently model long-term time dependencies. Based on FastDiff, we design an end-to-end text-to-speech synthesizer, FastDiff-TTS, which generates high-fidelity speech waveforms.
  arXiv  Detail & Related papers  (2022-04-21T07:49:09Z)
• SpecGrad: Diffusion Probabilistic Model based Neural Vocoder with Adaptive Noise Spectral Shaping [51.698273019061645]
  SpecGrad adapts the diffusion noise so that its time-varying spectral envelope becomes close to the conditioning log-mel spectrogram. It is processed in the time-frequency domain to keep the computational cost almost the same as the conventional DDPM-based neural vocoders.
  arXiv  Detail & Related papers  (2022-03-31T02:08:27Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.