# ModalFidelity: Routing Modalities for Deepfake Detection on a Budget

arXiv: https://arxiv.org/abs/2609.38246 · HTML: https://mlsanity.com/paper/2609.38246/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Builds on 21 of its 29 references that are in the map.
- First posted on arXiv 29 Sep 2026.

## Builds on (most cited first, 21 of 21)

- [Learning Transferable Visual Models From Natural Language Supervision](https://mlsanity.com/paper/2103.00020/) (2021)
- [MobileNetV2: Inverted Residuals and Linear Bottlenecks](https://mlsanity.com/paper/1801.04381/) (2018)
- [wav2vec 2.0: A Framework for Self-Supervised Learning of Speech Representations](https://mlsanity.com/paper/2006.11477/) (2020)
- [Outrageously Large Neural Networks: The Sparsely-Gated Mixture-of-Experts Layer](https://mlsanity.com/paper/1701.06538/) (2017)
- [FaceForensics++: Learning to Detect Manipulated Facial Images](https://mlsanity.com/paper/1901.08971/) (2019)
- [Multimodal Machine Learning: A Survey and Taxonomy](https://mlsanity.com/paper/1705.09406/) (2017)
- [DynamicViT: Efficient Vision Transformers with Dynamic Token Sparsification](https://mlsanity.com/paper/2106.02034/) (2021)
- [First Order Motion Model for Image Animation](https://mlsanity.com/paper/2003.00196/) (2020)
- [A Lip Sync Expert Is All You Need for Speech to Lip Generation In The Wild](https://mlsanity.com/paper/2008.10010/) (2020)
- [BranchyNet: Fast Inference via Early Exiting from Deep Neural Networks](https://mlsanity.com/paper/1709.01686/) (2017)
- [Self-critical Sequence Training for Image Captioning](https://mlsanity.com/paper/1612.00563/) (2016)
- [Transfer Learning from Speaker Verification to Multispeaker Text-To-Speech Synthesis](https://mlsanity.com/paper/1806.04558/) (2018)
- [FakeAVCeleb: A Novel Audio-Video Multimodal Deepfake Dataset](https://mlsanity.com/paper/2108.05080/) (2021)
- [AASIST: Audio Anti-Spoofing using Integrated Spectro-Temporal Graph Attention Networks](https://mlsanity.com/paper/2110.01200/) (2021)
- [Automatic speaker verification spoofing and deepfake detection using wav2vec 2.0 and data augmentation](https://mlsanity.com/paper/2202.12233/) (2022)
- [Emotions Don't Lie: An Audio-Visual Deepfake Detection Method Using Affective Cues](https://mlsanity.com/paper/2003.06711/) (2020)
- [AV-Deepfake1M: A Large-Scale LLM-Driven Audio-Visual Deepfake Dataset](https://mlsanity.com/paper/2311.15308/) (2023)
- [Circumventing shortcuts in audio-visual deepfake detection datasets with unsupervised learning](https://mlsanity.com/paper/2412.00175/) (2024)
- [Deepfake Detection that Generalizes Across Benchmarks](https://mlsanity.com/paper/2508.06248/) (2025)
- [SecureSpectra: Safeguarding Digital Identity from Deep Fake Threats via Intelligent Signatures](https://mlsanity.com/paper/2407.00913/) (2024)
- [PhonemeFake: Redefining Deepfake Realism with Language-Driven Segmental Manipulation and Adaptive Bilevel Detection](https://mlsanity.com/paper/2506.22783/) (2025)

## Abstract (arXiv)

Deepfakes no longer need to fake a whole video. Generators that read the transcript now alter only the few seconds in which a video's meaning turns, so a forgery hides in a small, unknown fraction of the video. Yet detectors still read every one-second window of both the audio and image streams, spending nearly all of their compute where nothing was altered. We observe that deciding where to look is far cheaper than looking. We present ModalFidelity, a lightweight router that previews each window and decides, before any forensic detector runs, which stream is worth reading, under a hard compute budget it can never exceed. On AV-Deepfake1M, reading at most a fifth of the windows, it is more accurate than gating after the detectors at 15.9x less compute, and retains over 96% of the accuracy of an oracle that knows where every forgery lies.