Notice: The reproducibility variables underlying each score are classified using an automated LLM-based pipeline, validated against a manually labeled dataset. LLM-based classification introduces uncertainty and potential bias; scores should be interpreted as estimates. Full accuracy metrics and methodology are described in Coakley et alK. L. Coakley, T. Snelleman, H. Hoos, and O. E. Gundersen, "The embrace of open science: An analysis of a decade of AI research and 56 800 conference papers," Under Review, 2026..
On the SAC-BL Algorithm for Anomaly Detection
Authors: Xinsong Ma, Jie Wu, Weiwei Liu
NeurIPS 2025 | Venue PDF | LLM Run Details
| Reproducibility Variable | Result | LLM Response |
|---|---|---|
| Research Type | Experimental | Finally, extensive experimental results demonstrate the superiority of our method over SAC-BL algorithm on multiple visual anomaly detection benchmarks. |
| Researcher Affiliation | Academia | Xinsong Ma School of Computer Science Wuhan University EMAIL Wu School of Computer Science Wuhan University EMAIL Liu School of Computer Science Wuhan University EMAIL |
| Pseudocode | Yes | Algorithm 1: Practical SAC-SBL Algorithm |
| Open Source Code | No | Open source code will be provided at a later date. |
| Open Datasets | Yes | Experiments are conducted on three widely used anomaly detection datasets. The MVTec dataset (Bergmann et al., 2019b) consists of 5354 images across 15 object and texture categories. The Vis A dataset (Zou et al., 2022) is an industrial anomaly dataset comprising 10821 images from 12 objects across 3 domains. The BTAD dataset (Mishra et al., 2021) includes 2830 images of 3 industrial products, showcasing body and surface defects. |
| Dataset Splits | Yes | For example, in the widely used AD benchmark MVTec (Bergmann et al., 2019b), the testing set for the class Pill contains only 167 images. |
| Hardware Specification | Yes | All experiments are conducted on a workstation with eight NVIDIA Ge Force GTX 3090 GPUs and two 2.2GHZ Intel CPUs. |
| Software Dependencies | No | The paper does not explicitly mention specific software dependencies with version numbers like Python 3.8, PyTorch 1.9, etc., only general descriptions of frameworks implicitly via citations. |
| Experiment Setup | Yes | In SAC-SBL, we repeat the stochastic perturbation 100 times and take the meaning value. |