Yu Li, Xiaoran Shang, Qizhi Pei, Yun Zhu, Xin Gao, Honglin Lin, Zhanping Zhong, Zhuoshi Pan, Zheng Liu, Xiaoyang Wang, Conghui He, Dahua Lin, Feng Zhao, Lijun Wu. Proceedings of the 64th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers). 2026.

Zheng Liu, Honglin Lin, Xiaoyang Wang, Xin Gao, Yu Li, Mengzhang Cai, Yun Zhu, Zhanping Zhong, Qizhi Pei, Zhuoshi Pan, Xiaoran Shang, Conghui He, Bin Cui, Wentao Zhang, Lijun Wu. Proceedings of the 64th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers). 2026.

Existing unsupervised distillation-based methods rely on the differences between encoded and decoded features to locate abnormal regions in test images. However, the decoder trained only on normal samples still reconstructs abnormal patch features well, degrading performance. This issue is particularly pronounced in unsupervised multi-class anomaly detection tasks. We attribute this behavior to ‘over-generalization’ (OG) of decoder: the significantly increasing diversity of patch patterns in multi-class training enhances the model generalization on normal patches, but also inadvertently broadens its generalization to abnormal patches. To mitigate ‘OG’, we propose a novel approach that leverages class-agnostic learnable prompts to capture common textual normality across various visual patterns, and then apply them to guide the decoded features towards a ‘normal’ textual representation, suppressing ‘over-generalization’ of the decoder on abnormal patterns. To further improve performance, we also introduce a gated mixture-of-experts module to specialize in handling diverse patch patterns and reduce mutual interference between them in multi-class training. Our method achieves competitive performance on the MVTec AD and VisA datasets, demonstrating its effectiveness.

This paper aims at generating anomalous images and their segmentation labels to address the lack of real-world anomaly samples and privacy issues. Departing from conventional approaches that use masks solely to guide the generation of anomaly images, we propose a dual-branch training strategy for the generative model. This strategy enables the simultaneous production of anomaly images and masks, with an alignment regularization loss that ensures the coherence between the generated images and their masks. During inference, only the image-generation branch is activated to produce synthetic samples for training the downstream segmentation model. Furthermore, we propose to integrate the well-trained generative model into the training of segmentation models, utilizing a generative feedback loss to refine the segmentation model's performance. Experiments show our method's IoU metrics exceed previous methods by 5.03%, 5.68% and 16.63% on Real-IAD (industrial), polyp (medical), and Floor Dirty (indoor) datasets. The code is publicly accessible at https://github.com/huan-yin/anomaly-alignment.

Image-level weakly supervised semantic segmentation has received increasing attention due to its low annotation cost. Existing methods mainly rely on Class Activation Mapping (CAM) to obtain pseudo-labels for training semantic segmentation models. In this work, we are the first to demonstrate that long-tailed distribution in training data can cause the CAM calculated through classifier weights over-activated for head classes and under-activated for tail classes due to the shared features among head- and tail- classes. This degrades pseudo-label quality and further influences final semantic segmentation performance. To address this issue, we propose a Shared Feature Calibration (SFC) method for CAM generation. Specifically, we leverage the class prototypes which carry positive shared features and propose a Multi-Scaled Distribution-Weighted (MSDW) consistency loss for narrowing the gap between the CAMs generated through classifier weights and class prototypes during training. The MSDW loss counterbalances over-activation and under-activation by calibrating the shared features in head-/tail-class classifier weights. Experimental results show that our SFC significantly improves CAM boundaries and achieves new state-of-the-art performances. The project is available at https://github.com/Barrett-python/SFC.

Most continual segmentation methods tackle the prob-lem as a per-pixel classification task. However, such a paradigm is very challenging, and we find query-based seg-menters with built-in objectness have inherent advantages compared with per-pixel ones, as objectness has strong transfer ability and forgetting resistance. Based on these findings, we propose CoMasTRe by disentangling continual segmentation into two stages: forgetting-resistant continual objectness learning and well-researched continual classi-fication. CoMasTRe uses a two-stage segmenter learning class-agnostic mask proposals at the first stage and leaving recognition to the second stage. During continual learning, a simple but effective distillation is adopted to strengthen objectness. To further mitigate the forgetting of old classes, we design a multi-label class distillation strategy suited for segmentation. We assess the effectiveness of CoMas-TRe on PASCAL VOC and ADE20K. Extensive experiments show that our method outperforms per-pixel and query-based methods on both datasets. Code will be available at https://github.com/jordangong/CoMasTRe.

Semi-supervised semantic segmentation allows model to mine effective supervision from unlabeled data to complement label-guided training. Recent research has primarily focused on consistency regularization techniques, exploring perturbation-invariant training at both the image and feature levels. In this work, we proposed a novel feature-level consistency learning framework named Density-Descending Feature Perturbation (DDFP). Inspired by the low-density separation assumption in semi-supervised learning, our key insight is that feature density can shed a light on the most promising direction for the segmentation classifier to explore, which is the regions with lower density. We propose to shift features with confident predictions towards lower-density regions by perturbation injection. The perturbed features are then super-vised by the predictions on the original features, thereby compelling the classifier to explore less dense regions to effectively regularize the decision boundary. Central to our method is the estimation of feature density. To this end, we introduce a lightweight density estimator based on normalizing flow, allowing for efficient capture of the feature density distribution in an online manner. By extracting gradients from the density estimator, we can determine the direction towards less dense regions for each feature. The proposed DDFP outperforms other designs on feature-level perturbations and shows state of the art performances on both Pascal VOC and Cityscapes dataset under various partition protocols. The project is available at https://github.com/Gavinwxy/DDFP.

Recent semi-supervised semantic segmentation methods combine pseudo labeling and consistency regularization to enhance model generalization from perturbation-invariant training. In this work, we argue that adequate supervision can be extracted directly from the geometry of feature space. Inspired by density-based unsupervised clustering, we propose to leverage feature density to locate sparse regions within feature clusters defined by label and pseudo labels. The hypothesis is that lower-density features tend to be under-trained compared with those densely gathered. Therefore, we propose to apply regularization on the structure of the cluster by tackling the sparsity to increase intra-class compactness in feature space. With this goal, we present a Density-Guided Contrastive Learning (DGCL) strategy to push anchor features in sparse regions toward cluster centers approximated by high-density positive keys. The heart of our method is to estimate feature density which is defined as neighbor compactness. We design a multi-scale density estimation module to obtain the density from multiple nearest-neighbor graphs for robust density modeling. Moreover, a unified training framework is proposed to combine label-guided self-training and density-guided geometry regularization to form complementary supervision on unlabeled data. Experimental results on PAS-CAL VOC and Cityscapes under various semi-supervised settings demonstrate that our proposed method achieves state-of-the-art performances. The project is available at https://github.com/Gavinwxy/DGCL.

In industrial anomaly detection, data efficiency and the ability for fast migration across products become the main concerns when developing detection algorithms. Existing methods tend to be data-hungry and work in the one-model-one-category way, which hinders their effectiveness in real-world industrial scenarios. In this paper, we propose a few-shot anomaly detection strategy that works in a low-data regime and can generalize across products at no cost. Given a defective query sample, we propose to utilize a few normal samples as a reference to reconstruct its normal version, where the final anomaly detection can be achieved by sample alignment. Specifically, we introduce a novel regression with distribution regularization to obtain the optimal transformation from support to query features, which guarantees the reconstruction result shares visual similarity with the query sample and meanwhile maintains the property of normal samples. Experimental results show that our method significantly outperforms previous state-of-the-art at both image and pixel-level AUROC performances from 2 to 8-shot scenarios. Besides, with only a limited number of training samples (less than 8 samples), our method reaches competitive performance with vanilla AD methods which are trained with extensive normal samples. The code is available at https://github.com/FzJun26th/FastRecon.

Recent semi-supervised semantic segmentation methods focus on mining extra supervision from unlabeled data by generating pseudo labels. However, noisy labels are inevitable in this process which prevent effective self-supervision. This paper proposes that noisy labels can be corrected based on semantic connections among features. Since a segmentation classifier produces both high and low-quality predictions, we can trace back to feature encoder to investigate how a feature in a noisy group is related to those in the confident groups. Discarding the weak predictions from the classifier, rectified predictions are assigned to the wrongly predicted features through the feature relations. The key to such an idea lies in mining reliable feature connections. With this goal, we propose a class-agnostic relation network to precisely capture semantic connections among features while ignoring their semantic categories. The feature relations enable us to perform effective noisy label corrections to boost self-training performance. Extensive experiments on PASCAL VOC and Cityscapes demonstrate the state-of-the-art performances of the proposed methods under various semi-supervised settings.

Lab-based teaching in which hands-on experiments are to be conducted by students takes an important part for a wide range of engineering and science disciplines. In our current practice, the lab-based teaching involves live demonstration and tutorials after the off-line lab manual review. This has become particularly problematic when the number of students is large and insufficiency on the lab-supporting system becomes a common issue. In the meantime, even with a small number of students, it can be interesting to prepare the lab in a one-to-one tutorial. Well-designed video tutorials eliminate the time and space constraints on learning and provide comprehensive details to students to enable them focus on deepening the understanding of concepts, rather than spending majority of time on trouble shooting during the lab. With the full technical support from the Digital Learning Resources Hub at Xi'an Jiaotong-Liverpool University, we propose to involve student volunteers to generate a series of customized video tutorials for our electronics lab-based teaching practice. From the viewpoint of students themselves, these video tutorials are carefully designed based on students' learning needs to seamlessly integrate a wide range of theoretical and practical information. Under the supervision of staff, volunteers will be able to repeat their learning cycles with a different role and enhance their own understanding and knowledge structures, promoting the student-centered education model. Some scenery-based video tutorials will be used in the online quizzes questions to better prepare the students. The generated video tutorials will be shared across a number of electronic engineering modules to further investigate the effectiveness of these video tutorials. The effectiveness can be further explored using online questionnaires and online quizzes and evaluated by comparative studies.