# HOLA: Enhancing Audio-visual Deepfake Detection via Hierarchical Contextual Aggregations and Efficient Pre-training

arXiv: https://arxiv.org/abs/2507.22781 · HTML: https://mlsanity.com/paper/2507.22781/

## Key facts

- Cited by 3 papers in the map (data through 14 Oct 2026).
- Builds on 5 of its 13 references that are in the map.
- First posted on arXiv 30 Jul 2025.

## Builds on (most cited first, 5 of 5)

- [Is Space-Time Attention All You Need for Video Understanding?](https://mlsanity.com/paper/2102.05095/) (2021)
- [DF40: Toward Next-Generation Deepfake Detection](https://mlsanity.com/paper/2406.13495/) (2024)
- [AV-Deepfake1M: A Large-Scale LLM-Driven Audio-Visual Deepfake Dataset](https://mlsanity.com/paper/2311.15308/) (2023)
- [HiCMAE: Hierarchical Contrastive Masked Autoencoder for Self-Supervised Audio-Visual Emotion Recognition](https://mlsanity.com/paper/2401.05698/) (2024)
- [FFR_FD: Effective and Fast Detection of DeepFakes Based on Feature Point Defects](https://mlsanity.com/paper/2107.02016/) (2021)

## Built on by (most cited first, 3 of 3)

- [Investigating self-supervised representations for audio-visual deepfake detection](https://mlsanity.com/paper/2511.17181/) (2025)
- [Detecting AI-Generated Video: A Vision-Language Dual-View Survey](https://mlsanity.com/paper/2607.10787/) (2026)
- [GEM-TFL: Bridging Weak and Full Supervision for Forgery Localization through EM-Guided Decomposition and Temporal Refinement](https://mlsanity.com/paper/2603.05095/) (2026)

## Abstract (arXiv)

Advances in Generative AI have made video-level deepfake detection increasingly challenging, exposing the limitations of current detection techniques. In this paper, we present HOLA, our solution to the Video-Level Deepfake Detection track of 2025 1M-Deepfakes Detection Challenge. Inspired by the success of large-scale pre-training in the general domain, we first scale audio-visual self-supervised pre-training in the multimodal video-level deepfake detection, which leverages our self-built dataset of 1.81M samples, thereby leading to a unified two-stage framework. To be specific, HOLA features an iterative-aware cross-modal learning module for selective audio-visual interactions, hierarchical contextual modeling with gated aggregations under the local-global perspective, and a pyramid-like refiner for scale-aware cross-grained semantic enhancements. Moreover, we propose the pseudo supervised singal injection strategy to further boost model performance. Extensive experiments across expert models and MLLMs impressivly demonstrate the effectiveness of our proposed HOLA. We also conduct a series of ablation studies to explore the crucial design factors of our introduced components. Remarkably, our HOLA ranks 1st, outperforming the second by 0.0476 AUC on the TestA set.