# Specialist-Generalist Fusion with Outcome-Supervised Rationales for Deepfake Detection

arXiv: https://arxiv.org/abs/2605.31192 · HTML: https://mlsanity.com/paper/2605.31192/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Builds on 36 of its 52 references that are in the map.
- First posted on arXiv 29 May 2026.

## Builds on (most cited first, 30 of 36)

- [Attention Is All You Need](https://mlsanity.com/paper/1706.03762/) (2017)
- [Learning Transferable Visual Models From Natural Language Supervision](https://mlsanity.com/paper/2103.00020/) (2021)
- [DeepSeekMath: Pushing the Limits of Mathematical Reasoning in Open Language Models](https://mlsanity.com/paper/2402.03300/) (2024)
- [EfficientNet: Rethinking Model Scaling for Convolutional Neural Networks](https://mlsanity.com/paper/1905.11946/) (2019)
- [QLoRA: Efficient Finetuning of Quantized LLMs](https://mlsanity.com/paper/2305.14314/) (2023)
- [Xception: Deep Learning with Depthwise Separable Convolutions](https://mlsanity.com/paper/1610.02357/) (2016)
- [FaceForensics++: Learning to Detect Manipulated Facial Images](https://mlsanity.com/paper/1901.08971/) (2019)
- [CNN-generated images are surprisingly easy to spot... for now](https://mlsanity.com/paper/1912.11035/) (2019)
- [Celeb-DF: A Large-scale Challenging Dataset for DeepFake Forensics](https://mlsanity.com/paper/1909.12962/) (2019)
- [Thinking in Frequency: Face Forgery Detection by Mining Frequency-aware Clues](https://mlsanity.com/paper/2007.09355/) (2020)
- [Face X-ray for More General Face Forgery Detection](https://mlsanity.com/paper/1912.13458/) (2019)
- [Detecting Deepfakes with Self-Blended Images](https://mlsanity.com/paper/2204.08376/) (2022)
- [Generalizing Face Forgery Detection with High-frequency Features](https://mlsanity.com/paper/2103.12376/) (2021)
- [Spatial-Phase Shallow Learning: Rethinking Face Forgery Detection in Frequency Domain](https://mlsanity.com/paper/2103.01856/) (2021)
- [Learning Self-Consistency for Deepfake Detection](https://mlsanity.com/paper/2012.09311/) (2020)
- [DeepfakeBench: A Comprehensive Benchmark of Deepfake Detection](https://mlsanity.com/paper/2307.01426/) (2023)
- [Two-branch Recurrent Network for Isolating Deepfakes in Videos](https://mlsanity.com/paper/2008.03412/) (2020)
- [Exploring Temporal Coherence for More General Video Face Forgery Detection](https://mlsanity.com/paper/2108.06693/) (2021)
- [The DeepFake Detection Challenge (DFDC) Dataset](https://mlsanity.com/paper/2006.07397/) (2020)
- [SIDA: Social Media Image Deepfake Detection, Localization and Explanation with Large Multimodal Model](https://mlsanity.com/paper/2412.04292/) (2024)
- [Synthetic Depth-of-Field with a Single-Camera Mobile Phone](https://mlsanity.com/paper/1806.04171/) (2018)
- [AntifakePrompt: Prompt-Tuned Vision-Language Models are Fake Image Detectors](https://mlsanity.com/paper/2310.17419/) (2023)
- [Generalizing Deepfake Video Detection with Plug-and-Play: Video-Level Blending and Spatiotemporal Adapter Tuning](https://mlsanity.com/paper/2408.17065/) (2024)
- [SeeABLE: Soft Discrepancies and Bounded Contrastive Learning for Exposing Deepfakes](https://mlsanity.com/paper/2211.11296/) (2022)
- [Beyond the Spectrum: Detecting Deepfakes via Re-Synthesis](https://mlsanity.com/paper/2105.14376/) (2021)
- [Rethinking Vision-Language Model in Face Forensics: Multi-Modal Interpretable Forged Face Detector](https://mlsanity.com/paper/2503.20188/) (2025)
- [Ivy-Fake: A Unified Explainable Framework and Benchmark for Image and Video AIGC Detection](https://mlsanity.com/paper/2506.00979/) (2025)
- [Controllable Guide-Space for Generalizable Face Forgery Detection](https://mlsanity.com/paper/2307.14039/) (2023)
- [Unlocking the Capabilities of Large Vision-Language Models for Generalizable and Explainable Deepfake Detection](https://mlsanity.com/paper/2503.14853/) (2025)
- [Detecting Generated Images by Real Images Only](https://mlsanity.com/paper/2311.00962/) (2023)

## Abstract (arXiv)

Generalizable deepfake detection requires complementary forensic and semantic visual evidence. Specialist encoders capture subtle manipulation traces but can overfit to source-specific statistics, whereas MLLMs provide broader visual-semantic representations but may overlook fine forensic artifacts. We propose a two-stage detector in which an MLLM directly fuses patch-level features from a frozen forensic encoder with those from its native vision encoder and produces the authenticity decision itself. This specialist--generalist alignment provides the main cross-domain performance gain. We subsequently introduce outcome-supervised rationale tuning. The model generates a free-form visual rationale before its decision and is optimized using only the binary authenticity label and a format constraint, without task-specific rationale annotations. Rationale generation is optional at inference, so the tuned model can still return a direct binary score. On DF40, specialist--generalist alignment improves average cross-domain AUC from 89.85 to 93.32±0.44. Across six paired runs, rationale tuning obtains 93.50±0.42 and improves five of six paired models; the mean difference is small and not statistically conclusive. Results on SID-Set and legacy benchmarks, together with fusion, output-order, and continued-training controls, demonstrate the value of complementary visual representations and show that label-only rationale tuning can add an optional explanation mode while approximately preserving direct detection performance.