# ADOPD: Reference-Privileged On-Policy Distillation for MLLM-Based Industrial Anomaly Detection

arXiv: https://arxiv.org/abs/2608.09789 · HTML: https://mlsanity.com/paper/2608.09789/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Placed in Detection and dense prediction distillation, one of 12 directions in Knowledge distillation.
- Builds on 25 of its 31 references that are in the map.
- First posted on arXiv 10 Aug 2026.

## Placed in

- Knowledge distillation / [Detection and dense prediction distillation](https://mlsanity.com/distill/detection-and-dense-prediction-distillation/)

## Builds on (most cited first, 25 of 25)

- [Qwen2.5-VL Technical Report](https://mlsanity.com/paper/2502.13923/) (2025)
- [Improved Baselines with Visual Instruction Tuning](https://mlsanity.com/paper/2310.03744/) (2023)
- [Qwen3-VL Technical Report](https://mlsanity.com/paper/2511.21631/) (2025)
- [HybridFlow: A Flexible and Efficient RLHF Framework](https://mlsanity.com/paper/2409.19256/) (2024)
- [InternVL3: Exploring Advanced Training and Test-Time Recipes for Open-Source Multimodal Models](https://mlsanity.com/paper/2504.10479/) (2025)
- [Towards Total Recall in Industrial Anomaly Detection](https://mlsanity.com/paper/2106.08265/) (2021)
- [GLM-4.5V and GLM-4.1V-Thinking: Towards Versatile Multimodal Reasoning with Scalable Reinforcement Learning](https://mlsanity.com/paper/2507.01006/) (2025)
- [SPot-the-Difference Self-Supervised Pre-training for Anomaly Detection and Segmentation](https://mlsanity.com/paper/2207.14315/) (2022)
- [AnomalyCLIP: Object-agnostic Prompt Learning for Zero-shot Anomaly Detection](https://mlsanity.com/paper/2310.18961/) (2023)
- [AnomalyGPT: Detecting Industrial Anomalies Using Large Vision-Language Models](https://mlsanity.com/paper/2308.15366/) (2023)
- [Revisiting On-Policy Distillation: Empirical Failure Modes and Simple Fixes](https://mlsanity.com/paper/2603.25562/) (2026)
- [Real-IAD: A Real-World Multi-View Dataset for Benchmarking Versatile Industrial Anomaly Detection](https://mlsanity.com/paper/2403.12580/) (2024)
- [Self-Distilled RLVR](https://mlsanity.com/paper/2604.03128/) (2026)
- [Privileged Information Distillation for Language Models](https://mlsanity.com/paper/2602.04942/) (2026)
- [Scaling Reasoning Efficiently via Relaxed On-Policy Distillation](https://mlsanity.com/paper/2603.11137/) (2026)
- [Vision-OPD: Learning to See Fine Details for Multimodal LLMs via On-Policy Self-Distillation](https://mlsanity.com/paper/2605.18740/) (2026)
- [MMAD: A Comprehensive Benchmark for Multimodal Large Language Models in Industrial Anomaly Detection](https://mlsanity.com/paper/2410.09453/) (2024)
- [AnomalyR1: A GRPO-based End-to-end MLLM for Industrial Anomaly Detection](https://mlsanity.com/paper/2504.11914/) (2025)
- [PKU-GoodsAD: A Supermarket Goods Dataset for Unsupervised Anomaly Detection and Segmentation](https://mlsanity.com/paper/2307.04956/) (2023)
- [IAD-R1: Reinforcing Consistent Reasoning in Industrial Anomaly Detection](https://mlsanity.com/paper/2508.09178/) (2025)
- [V-Zero: Answer-Label-Free On-Policy Distillation with Contrastive Evidence Gating for Fine-Grained Visual Reasoning](https://mlsanity.com/paper/2606.25319/) (2026)
- [Visual Contrastive Self-Distillation](https://mlsanity.com/paper/2607.21556/) (2026)
- [Towards Explainable Industrial Anomaly Detection via Knowledge-Guided Latent Reasoning](https://mlsanity.com/paper/2602.09850/) (2026)
- [AD-Copilot: A Vision-Language Assistant for Industrial Anomaly Detection via Visual In-context Comparison](https://mlsanity.com/paper/2603.13779/) (2026)
- [OPD-IAD: From Language Judgment to Industrial Anomaly Detection via On-Policy Self-Distillation](https://mlsanity.com/paper/2607.18850/) (2026)

## Abstract (arXiv)

Industrial anomaly detection (IAD) requires identifying fine-grained deviations from normal visual patterns. Multimodal large language models (MLLMs) can improve recognition accuracy by comparing query images with references at inference time, but these benefits rely on additional retrieval and processing. We investigate whether the benefits of reference comparison can instead be internalized in the model parameters. Access to references during training allows a reference-aware teacher to supervise a query-only student. However, the teacher may favor plausible responses based on query cues or language priors rather than valid visual information. We propose ADOPD, a reference-privileged on-policy distillation framework. The teacher evaluates student-generated rollouts under matched and mismatched references. The matched-reference teacher-to-student log-ratio defines the token-level learning direction, specifying what the student should learn. The likelihood gap between the two reference views estimates reference-specific support and calibrates the sequence-level weight. ADOPD achieves 77.31% average accuracy on the MMAD benchmark under zero-shot inference, improving the Qwen3-VL-4B backbone by 6.14 points and outperforming its one-shot setting by 2.64 points. Experiments show that ADOPD learns a fine-grained anomaly inspection strategy from reference comparison. The project will be available at https://github.com/withTai/ADOPD.