Transfer_Learning
-
From Predictions to Explanations: Explainable AI for Autism Diagnosis and Identification of Critical Brain Regions
2025-09-04 03:48:10Kush Gupta, Amir Aly, Emmanuel Ifeachor, Rohit ShankarAbstract
Autism spectrum disorder (ASD) is a neurodevelopmental condition characterized by atypical brain maturation. However, the adaptation of transfer learning paradigms in machine learning for ASD research remains notably limited. In this study, we propose a computer-aided diagnostic framework with two modules. This chapter presents a two-module framework combining deep learning and explainable AI for ASD diagnosis. The first module leverages a deep learning model fine-tuned through cross-domain transfer learning for ASD classification. The second module focuses on interpreting the model decisions and identifying critical brain regions. To achieve this, we employed three explainable AI (XAI) techniques: saliency mapping, Gradient-weighted Class Activation Mapping, and SHapley Additive exPlanations (SHAP) analysis. This framework demonstrates that cross-domain transfer learning can effectively address data scarcity in ASD research. In addition, by applying three established explainability techniques, the approach reveals how the model makes diagnostic decisions and identifies brain regions most associated with ASD. These findings were compared against established neurobiological evidence, highlighting strong alignment and reinforcing the clinical relevance of the proposed approach.
Abstract (translated)
自闭症谱系障碍（ASD）是一种神经发育状况，其特征是大脑成熟过程中的非典型发展。然而，在机器学习领域中适应迁移学习范式以应用于自闭症研究仍然非常有限。本研究提出了一种计算机辅助诊断框架，包括两个模块。本章介绍了一个结合深度学习和可解释人工智能的双模块框架用于ASD诊断。第一个模块利用通过跨域迁移学习微调过的深度学习模型来进行自闭症分类。第二个模块则侧重于对模型决策进行解读并识别关键脑区。为此，我们采用了三种可解释性AI（XAI）技术：敏感度映射、梯度加权类激活映射和SHapley Additive exPlanations (SHAP) 分析。该框架表明跨域迁移学习可以有效解决自闭症研究中的数据稀缺问题。此外，通过应用三种已建立的可解释性技术，此方法揭示了模型如何做出诊断决策，并识别与ASD最相关的脑区。这些发现与公认的神经生物学证据进行了比较，显示出高度的一致性，并强调了所提出方法的临床相关性。
URL
https://arxiv.org/abs/2509.10523
PDF
-
Transfer Learning-Based CNN Models for Plant Species Identification Using Leaf Venation Patterns
2025-09-03 21:23:09Bandita Bharadwaj, Ankur Mishra, Saurav BharadwajAbstract
This study evaluates the efficacy of three deep learning architectures: ResNet50, MobileNetV2, and EfficientNetB0 for automated plant species classification based on leaf venation patterns, a critical morphological feature with high taxonomic relevance. Using the Swedish Leaf Dataset comprising images from 15 distinct species (75 images per species, totalling 1,125 images), the models were demonstrated using standard performance metrics during training and testing phases. ResNet50 achieved a training accuracy of 94.11% but exhibited overfitting, reflected by a reduced testing accuracy of 88.45% and an F1 score of 87.82%. MobileNetV2 demonstrated better generalization capabilities, attaining a testing accuracy of 93.34% and an F1 score of 93.23%, indicating its suitability for lightweight, real-time applications. EfficientNetB0 outperformed both models, achieving a testing accuracy of 94.67% with precision, recall, and F1 scores exceeding 94.6%, highlighting its robustness in venation-based classification. The findings underscore the potential of deep learning, particularly EfficientNetB0, in developing scalable and accurate tools for automated plant taxonomy using venation traits.
Abstract (translated)
这项研究评估了三种深度学习架构（ResNet50、MobileNetV2和EfficientNetB0）在基于叶片脉络模式的植物种类自动分类中的有效性。叶脉是一种具有重要分类学意义的关键形态特征。使用包含来自15种不同物种图像的数据集（每个物种75张图片，总计1,125张图片），瑞典叶数据集，在训练和测试阶段分别通过标准性能指标展示了这些模型的性能。 ResNet50在训练过程中达到了94.11%的准确率，但表现出了过拟合的现象，这反映在其较低的测试准确率为88.45%，F1分数为87.82%。MobileNetV2则显示了更好的泛化能力，其测试准确率为93.34%，F1得分为93.23%，表明该模型适合轻量级和实时应用。EfficientNetB0在所有模型中表现最佳，在测试阶段达到了94.67%的准确率，并且精度、召回率以及F1分数都超过了94.6%，这说明其在基于脉络分类的应用中具有较强的鲁棒性。 研究结果强调了深度学习，尤其是EfficientNetB0，在使用叶脉特征开发可扩展和精确的自动化植物分类工具方面的潜力。
URL
https://arxiv.org/abs/2509.03729
PDF
-
Isolated Bangla Handwritten Character Classification using Transfer Learning
2025-09-03 06:46:14Abdul Karim, S M Rafiuddin, Jahidul Islam Razin, Tahira AlamAbstract
Bangla language consists of fifty distinct characters and many compound characters. Several notable studies have been performed to recognize Bangla characters, both handwritten and optical. Our approach uses transfer learning to classify the basic, distinct, as well as compound Bangla handwritten characters while avoiding the vanishing gradient problem. Deep Neural Network techniques such as 3D Convolutional Neural Network (3DCNN), Residual Neural Network (ResNet), and MobileNet are applied to generate an end-to-end classification of all possible standard formations of handwritten characters in the Bangla language. The Bangla Lekha Isolated dataset, which contains 166,105 Bangla character image samples categorized into 84 distinct classes, is used for this classification model. The model achieved 99.82% accuracy on training data and 99.46% accuracy on test data. Comparisons with various state-of-the-art benchmarks of Bangla handwritten character classification show that the proposed model achieves better accuracy in classifying the data.
Abstract (translated)
孟加拉语由五十个不同的字符和许多复合字符组成。已经进行了多项研究来识别手写和光学形式的孟加拉语字符。我们的方法利用迁移学习对基本、独立以及复合孟加拉语手写字符进行分类，同时避免梯度消失问题。通过应用深度神经网络技术（如三维卷积神经网络(3DCNN)、残差神经网络(ResNet)和MobileNet）来生成孟加拉语言所有可能标准的手写字符形式的端到端分类。我们使用了Bangla Lekha Isolated数据集，该数据集中包含了166,105个归类为84种不同类别的孟加拉语字符图像样本。我们的模型在训练数据上达到了99.82%的准确率，在测试数据上则达到了99.46%的准确率。与各种最先进的孟加拉语手写字符分类基准进行比较，所提出的模型在分类数据时取得了更高的准确性。
URL
https://arxiv.org/abs/2509.03061
PDF
-
RiverScope: High-Resolution River Masking Dataset
2025-09-02 16:00:27Rangel Daroya, Taylor Rowley, Jonathan Flores, Elisa Friedmann, Fiona Bennitt, Heejin An, Travis Simmons, Marissa Jean Hughes, Camryn L Kluetmeier, Solomon Kica, J. Daniel V\'elez, Sarah E. Esenther, Thomas E. Howard, Yanqi Ye, Audrey Turcotte, Colin Gleason, Subhransu MajiAbstract
Surface water dynamics play a critical role in Earth's climate system, influencing ecosystems, agriculture, disaster resilience, and sustainable development. Yet monitoring rivers and surface water at fine spatial and temporal scales remains challenging -- especially for narrow or sediment-rich rivers that are poorly captured by low-resolution satellite data. To address this, we introduce RiverScope, a high-resolution dataset developed through collaboration between computer science and hydrology experts. RiverScope comprises 1,145 high-resolution images (covering 2,577 square kilometers) with expert-labeled river and surface water masks, requiring over 100 hours of manual annotation. Each image is co-registered with Sentinel-2, SWOT, and the SWOT River Database (SWORD), enabling the evaluation of cost-accuracy trade-offs across sensors -- a key consideration for operational water monitoring. We also establish the first global, high-resolution benchmark for river width estimation, achieving a median error of 7.2 meters -- significantly outperforming existing satellite-derived methods. We extensively evaluate deep networks across multiple architectures (e.g., CNNs and transformers), pretraining strategies (e.g., supervised and self-supervised), and training datasets (e.g., ImageNet and satellite imagery). Our best-performing models combine the benefits of transfer learning with the use of all the multispectral PlanetScope channels via learned adaptors. RiverScope provides a valuable resource for fine-scale and multi-sensor hydrological modeling, supporting climate adaptation and sustainable water management.
Abstract (translated)
地表水动态在地球气候系统中扮演着关键角色，影响生态系统、农业、灾害抵御能力和可持续发展。然而，在细粒度的空间和时间尺度上监测河流和其他地表水仍然具有挑战性——尤其是在分辨率较低的卫星数据难以捕捉到的狭窄或含泥量高的河流区域。为了解决这个问题，我们介绍了RiverScope，这是一个通过计算机科学和水文学专家合作开发的高分辨率数据集。RiverScope包括1,145张高分辨率图像（覆盖2,577平方公里），并附有专家标注的河流与地表水面掩模，这些标记需要超过100小时的手动注释工作。每一张图像都与其他几种传感器的数据进行了同步注册，如Sentinel-2、SWOT以及SWOT河流数据库（SWORD）。这使得能够评估不同传感器在成本和准确性之间的权衡——这对运营水资源监测来说是一个关键考量因素。 此外，我们建立了第一个全球性的高分辨率基准来估计河流宽度，并达到了7.2米的中位误差值，显著超越了现有的卫星衍生方法。我们在多个网络架构（例如CNNs和transformers）、预训练策略（如监督式学习与自我监督式学习）以及各种培训数据集（如ImageNet及卫星图像）上广泛测试了深度神经网络模型。 我们的最佳表现模型通过转移学习的益处，结合使用所有多光谱PlanetScope通道并通过学习适配器来实现。RiverScope为精细尺度和多传感器水文建模提供了宝贵的资源，并支持气候适应与可持续水资源管理。
URL
https://arxiv.org/abs/2509.02451
PDF
-
Speech transformer models for extracting information from baby cries
2025-09-02 12:34:33Guillem Bonafos, J\'eremy Rouch, L\'eny Lego, David Reby, Hugues Patural, Nicolas Mathevon, R\'emy EmonetAbstract
Transfer learning using latent representations from pre-trained speech models achieves outstanding performance in tasks where labeled data is scarce. However, their applicability to non-speech data and the specific acoustic properties encoded in these representations remain largely unexplored. In this study, we investigate both aspects. We evaluate five pre-trained speech models on eight baby cries datasets, encompassing 115 hours of audio from 960 babies. For each dataset, we assess the latent representations of each model across all available classification tasks. Our results demonstrate that the latent representations of these models can effectively classify human baby cries and encode key information related to vocal source instability and identity of the crying baby. In addition, a comparison of the architectures and training strategies of these models offers valuable insights for the design of future models tailored to similar tasks, such as emotion detection.
Abstract (translated)
使用预训练语音模型的潜在表示进行迁移学习，在标记数据稀缺的任务中表现出卓越性能。然而，这些方法在非语音数据中的适用性以及它们所编码的具体声学特性仍然很大程度上未被探索。在这项研究中，我们探讨了这两个方面。我们在八个婴儿哭声数据集上评估了五种预训练的语音模型，这些数据集中包含来自960个婴儿的总计115小时音频。对于每个数据集，我们对每种模型的所有可用分类任务进行了潜在表示的评估。 我们的研究结果表明，这些模型的潜在表示能够有效地对人类婴儿哭声进行分类，并且编码了与声音来源不稳定性和哭泣婴儿身份相关的关键信息。此外，对比分析这些模型的架构和训练策略为设计未来适用于类似任务（如情感检测）的模型提供了有价值的见解。
URL
https://arxiv.org/abs/2509.02259
PDF
-
HydroVision: Predicting Optically Active Parameters in Surface Water Using Computer Vision
2025-09-02 02:12:52Shubham Laxmikant Deshmukh, Matthew Wilchek, Feras A. BatarsehAbstract
Ongoing advancements in computer vision, particularly in pattern recognition and scene classification, have enabled new applications in environmental monitoring. Deep learning now offers non-contact methods for assessing water quality and detecting contamination, both critical for disaster response and public health protection. This work introduces HydroVision, a deep learning-based scene classification framework that estimates optically active water quality parameters including Chlorophyll-Alpha, Chlorophylls, Colored Dissolved Organic Matter (CDOM), Phycocyanins, Suspended Sediments, and Turbidity from standard Red-Green-Blue (RGB) images of surface water. HydroVision supports early detection of contamination trends and strengthens monitoring by regulatory agencies during external environmental stressors, industrial activities, and force majeure events. The model is trained on more than 500,000 seasonally varied images collected from the United States Geological Survey Hydrologic Imagery Visualization and Information System between 2022 and 2024. This approach leverages widely available RGB imagery as a scalable, cost-effective alternative to traditional multispectral and hyperspectral remote sensing. Four state-of-the-art convolutional neural networks (VGG-16, ResNet50, MobileNetV2, DenseNet121) and a Vision Transformer are evaluated through transfer learning to identify the best-performing architecture. DenseNet121 achieves the highest validation performance, with an R2 score of 0.89 in predicting CDOM, demonstrating the framework's promise for real-world water quality monitoring across diverse conditions. While the current model is optimized for well-lit imagery, future work will focus on improving robustness under low-light and obstructed scenarios to expand its operational utility.
Abstract (translated)
计算机视觉领域的持续进步，尤其是在模式识别和场景分类方面的发展，已经为环境监测带来了新的应用。深度学习现在提供了一种非接触式的方法来评估水质并检测污染，这对于灾难应对和公共卫生保护至关重要。本文介绍了HydroVision，这是一个基于深度学习的场景分类框架，可以从表面水体的标准红绿蓝（RGB）图像中估算出包括叶绿素-A、总叶绿素、溶解有机物（CDOM）、藻青素、悬浮颗粒物和浑浊度在内的光活性水质参数。HydroVision支持早期污染趋势检测，并在外部环境压力、工业活动以及不可抗力事件期间增强监管机构的监测能力。 该模型是在超过50万张由美国地质调查局水文图像可视化信息系统（2022年至2024年间收集）提供的季节性变化图像上训练出来的。这种方法利用广泛可用的RGB图像作为传统多光谱和高光谱遥感的一种可扩展且成本效益高的替代方案。通过迁移学习，评估了四种最先进的卷积神经网络（VGG-16、ResNet50、MobileNetV2、DenseNet121）和一个视觉变换器，以确定最佳的架构。其中，DenseNet121实现了最高的验证性能，在预测CDOM时获得了R²得分为0.89的成绩，这表明该框架在不同条件下进行实际水质监测方面具有前景。 尽管当前模型针对光线充足的图像进行了优化，未来的工作将重点放在改进低光和受阻情况下的鲁棒性上，以扩大其操作实用性。
URL
https://arxiv.org/abs/2509.01882
PDF
-
TransMatch: A Transfer-Learning Framework for Defect Detection in Laser Powder Bed Fusion Additive Manufacturing
2025-09-01 20:15:26Mohsen Asghari Ilani, Yaser Mike BanadAbstract
Surface defects in Laser Powder Bed Fusion (LPBF) pose significant risks to the structural integrity of additively manufactured components. This paper introduces TransMatch, a novel framework that merges transfer learning and semi-supervised few-shot learning to address the scarcity of labeled AM defect data. By effectively leveraging both labeled and unlabeled novel-class images, TransMatch circumvents the limitations of previous meta-learning approaches. Experimental evaluations on a Surface Defects dataset of 8,284 images demonstrate the efficacy of TransMatch, achieving 98.91% accuracy with minimal loss, alongside high precision, recall, and F1-scores for multiple defect classes. These findings underscore its robustness in accurately identifying diverse defects, such as cracks, pinholes, holes, and spatter. TransMatch thus represents a significant leap forward in additive manufacturing defect detection, offering a practical and scalable solution for quality assurance and reliability across a wide range of industrial applications.
Abstract (translated)
激光粉末床熔合（LPBF）中的表面缺陷对增材制造部件的结构完整性构成了重大风险。本文介绍了TransMatch，这是一种将迁移学习和半监督少样本学习相结合的新框架，旨在解决增材制造缺陷数据标签稀缺的问题。通过有效利用标记和未标记的新类图像，TransMatch克服了先前元学习方法的限制。 在包含8,284张图像的表面缺陷数据集上进行的实验评估表明，TransMatch具有显著的效果，在保持最小损失的情况下实现了98.91%的准确性，并且对于多个缺陷类别都达到了高精度、召回率和F1分数。这些发现强调了其能够准确识别各种类型的缺陷（如裂纹、针孔、孔洞和飞溅）的能力。 因此，TransMatch在增材制造缺陷检测方面代表了一个重要的进步，为确保广泛工业应用中的质量和可靠性提供了实际且可扩展的解决方案。
URL
https://arxiv.org/abs/2509.01754
PDF
-
CSRM-LLM: Embracing Multilingual LLMs for Cold-Start Relevance Matching in Emerging E-commerce Markets
2025-09-01 15:51:30Yujing Wang, Yiren Chen, Huoran Li, Chunxu Xu, Yuchong Luo, Xianghui Mao, Cong Li, Lun Du, Chunyang Ma, Qiqi Jiang, Yin Wang, Fan Gao, Wenting Mo, Pei Wen, Shantanu Kumar, Taejin Park, Yiwei Song, Vijay Rajaram, Tao Cheng, Sonu Durgia, Pranam KolariAbstract
As global e-commerce platforms continue to expand, companies are entering new markets where they encounter cold-start challenges due to limited human labels and user behaviors. In this paper, we share our experiences in Coupang to provide a competitive cold-start performance of relevance matching for emerging e-commerce markets. Specifically, we present a Cold-Start Relevance Matching (CSRM) framework, utilizing a multilingual Large Language Model (LLM) to address three challenges: (1) activating cross-lingual transfer learning abilities of LLMs through machine translation tasks; (2) enhancing query understanding and incorporating e-commerce knowledge by retrieval-based query augmentation; (3) mitigating the impact of training label errors through a multi-round self-distillation training strategy. Our experiments demonstrate the effectiveness of CSRM-LLM and the proposed techniques, resulting in successful real-world deployment and significant online gains, with a 45.8% reduction in defect ratio and a 0.866% uplift in session purchase rate.
Abstract (translated)
随着全球电子商务平台的不断扩展，公司进入新市场时面临着冷启动挑战，这些挑战主要是由于缺乏人类标注和用户行为数据。本文分享了Coupang在新兴电子商务市场的相关性匹配中实现竞争力的冷启动性能的经验。具体而言，我们提出了一种冷启动相关性匹配（CSRM）框架，该框架利用多语言大型语言模型（LLM）来解决三个挑战：(1) 通过机器翻译任务激活LLM的跨语言迁移学习能力；(2) 通过基于检索的查询增强提高查询理解，并将电子商务知识融入其中；(3) 通过多次自我蒸馏训练策略减轻训练标签错误的影响。我们的实验表明，CSRM-LLM及其提出的技巧非常有效，实现了成功的真实世界部署和显著的在线收益，包括缺陷率降低了45.8%，会话购买率提高了0.866%。
URL
https://arxiv.org/abs/2509.01566
PDF
-
Preserving Vector Space Properties in Dimensionality Reduction: A Relationship Preserving Loss Framework
2025-09-01 07:31:11Eddi Weinwurm, Alexander KovalenkoAbstract
Dimensionality reduction can distort vector space properties such as orthogonality and linear independence, which are critical for tasks including cross-modal retrieval, clustering, and classification. We propose a Relationship Preserving Loss (RPL), a loss function that preserves these properties by minimizing discrepancies between relationship matrices (e.g., Gram or cosine) of high-dimensional data and their low-dimensional embeddings. RPL trains neural networks for non-linear projections and is supported by error bounds derived from matrix perturbation theory. Initial experiments suggest that RPL reduces embedding dimensions while largely retaining performance on downstream tasks, likely due to its preservation of key vector space properties. While we describe here the use of RPL in dimensionality reduction, this loss can also be applied more broadly, for example to cross-domain alignment and transfer learning, knowledge distillation, fairness and invariance, dehubbing, graph and manifold learning, and federated learning, where distributed embeddings must remain geometrically consistent.
Abstract (translated)
维度减少可能会扭曲向量空间的性质，如正交性和线性独立性，这些性质对于跨模态检索、聚类和分类等任务至关重要。我们提出了一种关系保持损失（RPL），这是一种通过最小化高维数据的关系矩阵（例如Gram矩阵或余弦相似度矩阵）与其低维嵌入之间的差异来保留这些属性的损失函数。RPL 用于训练神经网络进行非线性投影，并且其误差界限是基于矩阵扰动理论得出的。初步实验表明，尽管减少了嵌入维度，但 RPL 能够在很大程度上保持下游任务的表现，这可能是由于它保留下了关键的向量空间属性。虽然这里描述了 RPL 在降维中的应用，但它也可以更广泛地使用，例如用于跨域对齐和迁移学习、知识蒸馏、公平性和不变性、去中心化处理、图与流形学习以及联邦学习等场景，在这些场景中分布式嵌入必须保持几何一致性。
URL
https://arxiv.org/abs/2509.01198
PDF
-
MATL-DC: A Multi-domain Aggregation Transfer Learning Framework for EEG Emotion Recognition with Domain-Class Prototype under Unseen Targets
2025-09-01 05:08:04Guangli Li, Canbiao Wu, Zhehao Zhou, Na Tian, Zhen LiangAbstract
Emotion recognition based on electroencephalography (EEG) signals is increasingly becoming a key research hotspot in affective Brain-Computer Interfaces (aBCIs). However, the current transfer learning model greatly depends on the source domain and target domain data, which hinder the practical application of emotion recognition. Therefore, we propose a Multi-domain Aggregation Transfer Learning framework for EEG emotion recognition with Domain-Class prototype under unseen targets (MATL-DC). We design the feature decoupling module to decouple class-invariant domain features from domain-invariant class features from shallow features. In the model training stage, the multi-domain aggregation mechanism aggregates the domain feature space to form a superdomain, which enhances the characteristics of emotional EEG signals. In each superdomain, we further extract the class prototype representation by class features. In addition, we adopt the pairwise learning strategy to transform the sample classification problem into the similarity problem between sample pairs, which effectively alleviates the influence of label noise. It is worth noting that the target domain is completely unseen during the training process. In the inference stage, we use the trained domain-class prototypes for inference, and then realize emotion recognition. We rigorously validate it on the publicly available databases (SEED, SEED-IV and SEED-V). The results show that the accuracy of MATL-DC model is 84.70\%, 68.11\% and 61.08\%, respectively. MATL-DC achieves comparable or even better performance than methods that rely on both source and target domains. The source code is available at this https URL.
Abstract (translated)
基于脑电图（EEG）信号的情感识别在情感型脑机接口（aBCIs）中正逐渐成为关键的研究热点。然而，目前的迁移学习模型严重依赖于源域和目标域的数据，这阻碍了情感识别的实际应用。因此，我们提出了一种基于未知目标领域下的多域聚合迁移学习框架（MATL-DC），用于EEG情绪识别，并采用领域-类别原型设计。我们设计了一个特征解耦模块，从浅层特征中分离出类不变领域的特征和领域不变的类别特征。在模型训练阶段，多域聚合机制将各个领域的特征空间聚集在一起形成一个超域，从而增强了情感EEG信号的特点。每个超域中，进一步通过类别特征提取类别原型表示。此外，我们采用了成对学习策略，将样本分类问题转化为样本对之间的相似性问题，有效缓解了标签噪声的影响。值得注意的是，在训练过程中目标领域是完全未见的。在推理阶段，我们使用训练好的领域-类别原型进行推理，然后实现情绪识别。 我们在公开可用的数据集（SEED、SEED-IV和SEED-V）上对其进行了严格的验证。结果表明，MATL-DC模型的准确率分别为84.70%、68.11%和61.08%，相较于依赖于源域与目标域的方法，MATL-DC取得了相当或更好的性能表现。 该研究的源代码可在提供的链接中获取。
URL
https://arxiv.org/abs/2509.01135
PDF
-
GraViT: Transfer Learning with Vision Transformers and MLP-Mixer for Strong Gravitational Lens Discovery
2025-08-29 20:26:04Ren\'e Parlange, Juan C. Cuevas-Tello, Octavio Valenzuela, Omar de J. Cabrera-Rosas, Tom\'as Verdugo, Anupreeta More, Anton T. JaelaniAbstract
Gravitational lensing offers a powerful probe into the properties of dark matter and is crucial to infer cosmological parameters. The Legacy Survey of Space and Time (LSST) is predicted to find O(10^5) gravitational lenses over the next decade, demanding automated classifiers. In this work, we introduce GraViT, a PyTorch pipeline for gravitational lens detection that leverages extensive pretraining of state-of-the-art Vision Transformer (ViT) models and MLP-Mixer. We assess the impact of transfer learning on classification performance by examining data quality (source and sample size), model architecture (selection and fine-tuning), training strategies (augmentation, normalization, and optimization), and ensemble predictions. This study reproduces the experiments in a previous systematic comparison of neural networks and provides insights into the detectability of strong gravitational lenses on that common test sample. We fine-tune ten architectures using datasets from HOLISMOKES VI and SuGOHI X, and benchmark them against convolutional baselines, discussing complexity and inference-time analysis.
Abstract (translated)
引力透镜效应为探索暗物质的性质提供了强有力的工具，并且对于推断宇宙学参数至关重要。根据预测，为期十年的时间和空间遗产调查（LSST）将发现大约10^5个引力透镜现象，这需要自动分类器的支持。在这项工作中，我们介绍了一个名为GraViT的PyTorch管道，用于探测引力透镜效应，该系统利用了最先进的视觉变换器(ViT)模型和MLP-Mixer的广泛预训练。 我们在数据质量（源质量和样本大小）、模型架构（选择和微调）、训练策略（增强、归一化和优化）以及集成预测方面评估了迁移学习对分类性能的影响。这项研究重现了先前关于神经网络系统比较实验的结果，并提供了在公共测试样本上强引力透镜检测可行性的见解。 我们使用HOLISMOKES VI和SuGOHI X的数据集，对十种架构进行了微调，并与卷积基线进行了基准测试，讨论了复杂性和推理时间的分析。
URL
https://arxiv.org/abs/2509.00226
PDF
-
Unsupervised Video Continual Learning via Non-Parametric Deep Embedded Clustering
2025-08-29 16:49:03Nattapong Kurpukdee, Adrian G. BorsAbstract
We propose a realistic scenario for the unsupervised video learning where neither task boundaries nor labels are provided when learning a succession of tasks. We also provide a non-parametric learning solution for the under-explored problem of unsupervised video continual learning. Videos represent a complex and rich spatio-temporal media information, widely used in many applications, but which have not been sufficiently explored in unsupervised continual learning. Prior studies have only focused on supervised continual learning, relying on the knowledge of labels and task boundaries, while having labeled data is costly and not practical. To address this gap, we study the unsupervised video continual learning (uVCL). uVCL raises more challenges due to the additional computational and memory requirements of processing videos when compared to images. We introduce a general benchmark experimental protocol for uVCL by considering the learning of unstructured video data categories during each task. We propose to use the Kernel Density Estimation (KDE) of deep embedded video features extracted by unsupervised video transformer networks as a non-parametric probabilistic representation of the data. We introduce a novelty detection criterion for the incoming new task data, dynamically enabling the expansion of memory clusters, aiming to capture new knowledge when learning a succession of tasks. We leverage the use of transfer learning from the previous tasks as an initial state for the knowledge transfer to the current learning task. We found that the proposed methodology substantially enhances the performance of the model when successively learning many tasks. We perform in-depth evaluations on three standard video action recognition datasets, including UCF101, HMDB51, and Something-to-Something V2, without using any labels or class boundaries.
Abstract (translated)
我们提出了一种无监督视频学习的真实场景，在这种场景下，当连续执行多项任务时，既没有提供任务边界也没有提供标签。此外，针对未充分研究的无监督视频持续学习问题，我们也提供了一个非参数化的解决方案。视频代表了复杂且丰富的时空媒体信息，在许多应用中被广泛使用，但在无监督持续学习领域尚未得到充分探索。先前的研究仅关注有监督的持续学习，并依赖于标签和任务边界的知识，而获得标注数据的成本高昂且不切实际。为了解决这一缺口，我们研究了无监督视频连续学习（uVCL）。相较于处理图像，由于视频处理在计算和内存需求方面的要求更高，因此uVCL提出了更大的挑战。 为了便于对uVCL的研究，我们引入了一种通用的基准实验协议，该协议考虑到了每个任务中未结构化的视频数据类别的学习。我们建议使用无监督视频变换网络提取的深度嵌入式视频特征来进行核密度估计（KDE），以作为非参数化概率数据表示方法。我们还提出了一个新颖性检测标准，用于判断新任务数据，并动态启用内存簇扩展，旨在捕捉在连续任务学习过程中的新知识。 我们利用从先前任务中转移的学习成果作为当前学习任务的知识传递的初始状态。研究发现，所提出的这种方法显著提升了模型在连续执行多项任务时的表现。我们在三个标准视频动作识别数据集上进行了深入评估，包括UCF101、HMDB51和Something-to-Something V2，且未使用任何标签或类别边界信息。
URL
https://arxiv.org/abs/2508.21773
PDF
-
Integrating Large Language Models with Network Optimization for Interactive and Explainable Supply Chain Planning: A Real-World Case Study
2025-08-29 13:34:55Saravanan VenkatachalamAbstract
This paper presents an integrated framework that combines traditional network optimization models with large language models (LLMs) to deliver interactive, explainable, and role-aware decision support for supply chain planning. The proposed system bridges the gap between complex operations research outputs and business stakeholder understanding by generating natural language summaries, contextual visualizations, and tailored key performance indicators (KPIs). The core optimization model addresses tactical inventory redistribution across a network of distribution centers for multi-period and multi-item, using a mixed-integer formulation. The technical architecture incorporates AI agents, RESTful APIs, and a dynamic user interface to support real-time interaction, configuration updates, and simulation-based insights. A case study demonstrates how the system improves planning outcomes by preventing stockouts, reducing costs, and maintaining service levels. Future extensions include integrating private LLMs, transfer learning, reinforcement learning, and Bayesian neural networks to enhance explainability, adaptability, and real-time decision-making.
Abstract (translated)
这篇论文提出了一种结合传统网络优化模型与大型语言模型（LLM）的综合框架，旨在为供应链规划提供互动、可解释且角色感知型的决策支持。所提出的系统通过生成自然语言摘要、情境化可视化和量身定制的关键绩效指标（KPI），弥合了复杂的运筹学输出与业务利益相关者理解之间的差距。 核心优化模型采用混合整数公式，解决多个周期内多物品在分销中心网络中的战术性库存重新分配问题。技术架构集成了AI代理、RESTful API和动态用户界面，以支持实时交互、配置更新以及基于模拟的洞察分析。 一个案例研究展示了该系统如何通过防止缺货、降低成本和服务水平来改善规划结果。未来的扩展包括整合私有LLM、迁移学习、强化学习和贝叶斯神经网络，从而增强解释性、适应性和实时决策能力。
URL
https://arxiv.org/abs/2508.21622
PDF
-
E-ConvNeXt: A Lightweight and Efficient ConvNeXt Variant with Cross-Stage Partial Connections
2025-08-28 16:17:19Fang Wang, Huitao Li, Wenhan Chao, Zheng Zhuo, Yiran Ji, Chang Peng, Yupeng SunAbstract
Many high-performance networks were not designed with lightweight application scenarios in mind from the outset, which has greatly restricted their scope of application. This paper takes ConvNeXt as the research object and significantly reduces the parameter scale and network complexity of ConvNeXt by integrating the Cross Stage Partial Connections mechanism and a series of optimized designs. The new network is named E-ConvNeXt, which can maintain high accuracy performance under different complexity configurations. The three core innovations of E-ConvNeXt are : (1) integrating the Cross Stage Partial Network (CSPNet) with ConvNeXt and adjusting the network structure, which reduces the model's network complexity by up to 80%; (2) Optimizing the Stem and Block structures to enhance the model's feature expression capability and operational efficiency; (3) Replacing Layer Scale with channel attention. Experimental validation on ImageNet classification demonstrates E-ConvNeXt's superior accuracy-efficiency balance: E-ConvNeXt-mini reaches 78.3% Top-1 accuracy at 0.9GFLOPs. E-ConvNeXt-small reaches 81.9% Top-1 accuracy at 3.1GFLOPs. Transfer learning tests on object detection tasks further confirm its generalization capability.
Abstract (translated)
许多高性能网络最初并未考虑轻量级应用场景的需求，这极大地限制了它们的应用范围。本文以ConvNeXt为研究对象，在集成Cross Stage Partial Connections（CSP）机制和一系列优化设计的基础上，大幅减少了ConvNeXt的参数规模与网络复杂度，并将新网络命名为E-ConvNeXt。该网络能够在不同的复杂度配置下保持高精度性能。 E-ConvNeXt的三大核心创新包括： 1. 将Cross Stage Partial Network（CSPNet）与ConvNeXt相结合并调整网络结构，最多可减少模型的80%网络复杂度； 2. 优化Stem和Block结构，增强模型的特征表达能力和操作效率； 3. 使用通道注意力机制替代Layer Scale。 在ImageNet分类任务上的实验验证表明了E-ConvNeXt优秀的精度与效率平衡：E-ConvNeXt-mini版本以0.9GFLOPs达到78.3%的Top-1准确率，而E-ConvNeXt-small版本则以3.1GFLOPs实现了81.9%的Top-1准确率。在目标检测任务上的迁移学习测试进一步证实了其泛化能力。
URL
https://arxiv.org/abs/2508.20955
PDF
-
Is the medical image segmentation problem solved? A survey of current developments and future directions
2025-08-27 01:12:45Guoping Xu, Jayaram K. Udupa, Jax Luo, Songlin Zhao, Yajun Yu, Scott B. Raymond, Hao Peng, Lipeng Ning, Yogesh Rathi, Wei Liu, You ZhangAbstract
Medical image segmentation has advanced rapidly over the past two decades, largely driven by deep learning, which has enabled accurate and efficient delineation of cells, tissues, organs, and pathologies across diverse imaging modalities. This progress raises a fundamental question: to what extent have current models overcome persistent challenges, and what gaps remain? In this work, we provide an in-depth review of medical image segmentation, tracing its progress and key developments over the past decade. We examine core principles, including multiscale analysis, attention mechanisms, and the integration of prior knowledge, across the encoder, bottleneck, skip connections, and decoder components of segmentation networks. Our discussion is organized around seven key dimensions: (1) the shift from supervised to semi-/unsupervised learning, (2) the transition from organ segmentation to lesion-focused tasks, (3) advances in multi-modality integration and domain adaptation, (4) the role of foundation models and transfer learning, (5) the move from deterministic to probabilistic segmentation, (6) the progression from 2D to 3D and 4D segmentation, and (7) the trend from model invocation to segmentation agents. Together, these perspectives provide a holistic overview of the trajectory of deep learning-based medical image segmentation and aim to inspire future innovation. To support ongoing research, we maintain a continually updated repository of relevant literature and open-source resources at this https URL
Abstract (translated)
过去二十年来，医学图像分割领域取得了快速进展，这主要得益于深度学习技术的应用，它使得细胞、组织、器官及各种病理在不同成像模式下的准确高效边界划分成为可能。这一进步引发了一个基本问题：目前的模型在多大程度上克服了持久存在的挑战？还有哪些差距需要填补？在这项工作中，我们对医学图像分割进行了深入回顾，追踪了过去十年中的进展和关键发展。我们考察了核心原则，包括多尺度分析、注意力机制以及先验知识集成，在分割网络的编码器、瓶颈层、跳跃连接和解码器组件中均有体现。我们的讨论围绕七个关键维度展开：(1) 从监督学习到半监督/无监督学习的转变；(2) 从器官分割向病灶任务的关注变化；(3) 多模态集成与领域适应的进步；(4) 基础模型和迁移学习的作用；(5) 从确定性划分向概率划分的演进；(6) 从二维到三维及四维分割的发展；以及 (7) 从调用模型向分割代理的趋势。这些视角共同提供了一个基于深度学习医学图像分割的整体概览，并旨在激发未来的创新。为了支持持续的研究，我们维护了一份不断更新的相关文献和开源资源库，可在此处的[URL]访问。 请注意，在原文中提到的具体网址（this https URL）未给出实际链接地址，请根据上下文或使用原始文档来查找该链接的实际位置。
URL
https://arxiv.org/abs/2508.20139
PDF
-
One Joke to Rule them All? On the possibility of Generalizing Humor
2025-08-26 19:55:40Mor Turgeman, Chen Shani, Dafna ShahafAbstract
Humor is a broad and complex form of communication that remains challenging for machines. Despite its broadness, most existing research on computational humor traditionally focused on modeling a specific type of humor. In this work, we wish to understand whether competence on one or more specific humor tasks confers any ability to transfer to novel, unseen types; in other words, is this fragmentation inevitable? This question is especially timely as new humor types continuously emerge in online and social media contexts (e.g., memes, anti-humor, AI fails). If Large Language Models (LLMs) are to keep up with this evolving landscape, they must be able to generalize across humor types by capturing deeper, transferable mechanisms. To investigate this, we conduct a series of transfer learning experiments across four datasets, representing different humor tasks. We train LLMs under varied diversity settings (1-3 datasets in training, testing on a novel task). Experiments reveal that models are capable of some transfer, and can reach up to 75% accuracy on unseen datasets; training on diverse sources improves transferability (1.88-4.05%) with minimal-to-no drop in in-domain performance. Further analysis suggests relations between humor types, with Dad Jokes surprisingly emerging as the best enabler of transfer (but is difficult to transfer to). We release data and code.
Abstract (translated)
幽默是一种广泛而复杂的交流形式，对于机器来说仍具挑战性。尽管其范围很广，但现有关于计算幽默的研究大多集中于建模特定类型的幽默上。在这项工作中，我们希望探讨掌握一种或多种具体幽默任务的能力是否能够转移到未见过的新类型中；换句话说，这种碎片化的研究方式是不可避免的吗？这个问题在当下尤为及时，因为新的幽默形式不断出现在在线和社交媒体环境中（例如，表情包、反幽默、AI失误）。如果大型语言模型（LLMs）要跟上这一不断变化的发展态势，就必须能够跨不同类型的幽默进行泛化，捕捉更深层次且可转移的机制。为调查这一点，我们进行了四组代表不同幽默任务的数据集上的迁移学习实验，在训练时使用不同的多样性和数据设置（在1-3个数据集中训练，对一个全新的任务进行测试）。实验结果表明模型能够在一定程度上实现跨类型迁移，并能达到75%的未见过数据集上的准确率；在多样化来源上进行训练可以提高迁移能力（提升幅度为1.88%-4.05%），同时对特定领域的表现几乎没有或没有下降。进一步分析发现，不同的幽默类型之间存在一定的关联性，“爸爸笑话”出人意料地成为了最佳的迁移促进者（但很难被迁移到其他类型的幽默中）。我们将在后续公开数据和代码以供研究使用。
URL
https://arxiv.org/abs/2508.19402
PDF
-
Amplifying Emotional Signals: Data-Efficient Deep Learning for Robust Speech Emotion Recognition
2025-08-26 19:08:54Tai VuAbstract
Speech Emotion Recognition (SER) presents a significant yet persistent challenge in human-computer interaction. While deep learning has advanced spoken language processing, achieving high performance on limited datasets remains a critical hurdle. This paper confronts this issue by developing and evaluating a suite of machine learning models, including Support Vector Machines (SVMs), Long Short-Term Memory networks (LSTMs), and Convolutional Neural Networks (CNNs), for automated emotion classification in human speech. We demonstrate that by strategically employing transfer learning and innovative data augmentation techniques, our models can achieve impressive performance despite the constraints of a relatively small dataset. Our most effective model, a ResNet34 architecture, establishes a new performance benchmark on the combined RAVDESS and SAVEE datasets, attaining an accuracy of 66.7% and an F1 score of 0.631. These results underscore the substantial benefits of leveraging pre-trained models and data augmentation to overcome data scarcity, thereby paving the way for more robust and generalizable SER systems.
Abstract (translated)
语音情感识别（SER）在人机交互中面临着一个重大且持久的挑战。虽然深度学习已经推动了口语处理技术的进步，但在有限的数据集上实现高性能仍然是一个重要障碍。本文通过开发和评估一系列机器学习模型来应对这一问题，这些模型包括支持向量机（SVM）、长短时记忆网络（LSTM）和卷积神经网络（CNN），以进行人类语音情感的自动分类。我们证明，通过战略性地采用迁移学习和创新的数据增强技术，我们的模型能够在数据集相对较小的情况下实现令人印象深刻的性能表现。 最有效的模型是我们采用的ResNet34架构，在结合RAVDESS和SAVEE数据集后建立了一个新的性能基准，达到了66.7%的准确率和0.631的F1分数。这些结果强调了利用预训练模型和数据增强技术克服数据不足的重要性，并为构建更强大、更具通用性的SER系统铺平了道路。
URL
https://arxiv.org/abs/2509.00077
PDF
-
Efficient Multi-Source Knowledge Transfer by Model Merging
2025-08-26 18:31:38Marcin Osial, Bartosz W\'ojcik, Bartosz Zieli\'nski, Sebastian CygertAbstract
While transfer learning is an advantageous strategy, it overlooks the opportunity to leverage knowledge from numerous available models online. Addressing this multi-source transfer learning problem is a promising path to boost adaptability and cut re-training costs. However, existing approaches are inherently coarse-grained, lacking the necessary precision for granular knowledge extraction and the aggregation efficiency required to fuse knowledge from either a large number of source models or those with high parameter counts. We address these limitations by leveraging Singular Value Decomposition (SVD) to first decompose each source model into its elementary, rank-one components. A subsequent aggregation stage then selects only the most salient components from all sources, thereby overcoming the previous efficiency and precision limitations. To best preserve and leverage the synthesized knowledge base, our method adapts to the target task by fine-tuning only the principal singular values of the merged matrix. In essence, this process only recalibrates the importance of top SVD components. The proposed framework allows for efficient transfer learning, is robust to perturbations both at the input level and in the parameter space (e.g., noisy or pruned sources), and scales well computationally.
Abstract (translated)
虽然迁移学习是一种有利的策略，但它忽略了利用在线众多可用模型知识的机会。解决多源迁移学习问题是一条提高适应性和削减再训练成本的有希望的道路。然而，现有的方法本质上是粗粒度的，缺乏从大量源模型或参数数量高的模型中提取精细知识和融合知识所需的精度与效率。我们通过利用奇异值分解（SVD）来克服这些限制：首先将每个源模型分解为其基本的一级分量。随后的聚合阶段则仅选择所有来源中最显著的部分，从而解决以前的方法在效率和精度上的局限性。为了最大程度地保存并有效利用合成的知识库，我们的方法通过微调合并矩阵的主要奇异值来适应目标任务，这一过程仅重新校准顶级SVD成分的重要性。所提出的框架能够实现高效的迁移学习，并且对输入层面和参数空间中的扰动（例如嘈杂或剪枝后的源模型）具有鲁棒性，同时在计算上也表现良好。
URL
https://arxiv.org/abs/2508.19353
PDF
-
C-Flat++: Towards a More Efficient and Powerful Framework for Continual Learning
2025-08-26 09:39:09Wei Li, Hangjie Yuan, Zixiang Zhao, Yifan Zhu, Aojun Lu, Tao Feng, Yanan SunAbstract
Balancing sensitivity to new tasks and stability for retaining past knowledge is crucial in continual learning (CL). Recently, sharpness-aware minimization has proven effective in transfer learning and has also been adopted in continual learning (CL) to improve memory retention and learning efficiency. However, relying on zeroth-order sharpness alone may favor sharper minima over flatter ones in certain settings, leading to less robust and potentially suboptimal solutions. In this paper, we propose \textbf{C}ontinual \textbf{Flat}ness (\textbf{C-Flat}), a method that promotes flatter loss landscapes tailored for CL. C-Flat offers plug-and-play compatibility, enabling easy integration with minimal modifications to the code pipeline. Besides, we present a general framework that integrates C-Flat into all major CL paradigms and conduct comprehensive comparisons with loss-minima optimizers and flat-minima-based CL methods. Our results show that C-Flat consistently improves performance across a wide range of settings. In addition, we introduce C-Flat++, an efficient yet effective framework that leverages selective flatness-driven promotion, significantly reducing the update cost required by C-Flat. Extensive experiments across multiple CL methods, datasets, and scenarios demonstrate the effectiveness and efficiency of our proposed approaches. Code is available at this https URL.
Abstract (translated)
在连续学习（CL）中，平衡对新任务的敏感性和保持过去知识的稳定性至关重要。最近，敏锐度感知最小化在迁移学习中被证明是有效的，并已被用于连续学习以改善记忆保留和学习效率。然而，在某些情况下，仅依赖零阶敏锐度可能会倾向于选择更尖锐而非平坦的极小值，从而导致解决方案不够稳健且可能次优。为此，本文提出了**Continual Flatness (C-Flat)** 方法，该方法旨在促进适合连续学习的更加平缓的损失景观。C-Flat 具有即插即用兼容性，可轻松集成到代码流程中而无需大量修改。 此外，我们还提出了一种通用框架，将 C-Flat 整合到所有主要的 CL 方法中，并进行了全面比较，对比了基于损失极小值优化器和基于平坦极小值的 CL 方法。我们的结果显示，在各种设置下，C-Flat 一致性地提高了性能表现。 此外，我们还引入了 C-Flat++，这是一种高效且有效的框架，利用选择性平滑驱动提升，显著减少了由 C-Flat 所需的更新成本。在多个 CL 方法、数据集和场景上的广泛实验展示了我们的方法的有效性和效率。代码可在提供的链接中获取。
URL
https://arxiv.org/abs/2508.18860
PDF
-
Feature-Space Planes Searcher: A Universal Domain Adaptation Framework for Interpretability and Computational Efficiency
2025-08-26 05:39:21Zhitong Cheng, Yiran Jiang, Yulong Ge, Yufeng Li, Zhongheng Qin, Rongzhi Lin, Jianwei MaAbstract
Domain shift, characterized by degraded model performance during transition from labeled source domains to unlabeled target domains, poses a persistent challenge for deploying deep learning systems. Current unsupervised domain adaptation (UDA) methods predominantly rely on fine-tuning feature extractors - an approach limited by inefficiency, reduced interpretability, and poor scalability to modern architectures. Our analysis reveals that models pretrained on large-scale data exhibit domain-invariant geometric patterns in their feature space, characterized by intra-class clustering and inter-class separation, thereby preserving transferable discriminative structures. These findings indicate that domain shifts primarily manifest as boundary misalignment rather than feature degradation. Unlike fine-tuning entire pre-trained models - which risks introducing unpredictable feature distortions - we propose the Feature-space Planes Searcher (FPS): a novel domain adaptation framework that optimizes decision boundaries by leveraging these geometric patterns while keeping the feature encoder frozen. This streamlined approach enables interpretative analysis of adaptation while substantially reducing memory and computational costs through offline feature extraction, permitting full-dataset optimization in a single computation cycle. Evaluations on public benchmarks demonstrate that FPS achieves competitive or superior performance to state-of-the-art methods. FPS scales efficiently with multimodal large models and shows versatility across diverse domains including protein structure prediction, remote sensing classification, and earthquake detection. We anticipate FPS will provide a simple, effective, and generalizable paradigm for transfer learning, particularly in domain adaptation tasks. .
Abstract (translated)
领域转移（Domain Shift）是指在从标记的源域过渡到未标记的目标域时，模型性能下降的问题。这对部署深度学习系统构成了持续挑战。目前的无监督领域适应（UDA）方法主要依赖于微调特征提取器，这种方法由于效率低下、可解释性差以及难以扩展至现代架构而受到限制。我们的分析发现，基于大规模数据预训练的模型在它们的特征空间中表现出域不变的几何模式，表现为类内聚类和类间分离，从而保留了可转移的判别结构。这些发现表明领域转换主要表现为边界错位而非特征退化。与微调整个预训练模型相比（这可能会引入不可预测的特征扭曲），我们提出了一种新颖的领域适应框架——Feature-space Planes Searcher (FPS)，它通过利用这些几何模式来优化决策边界，同时保持特征编码器冻结不变。这种简化的途径不仅可以实现对适应过程的理解性分析，并且通过离线特征提取大幅减少内存和计算成本，在单次计算周期中即可进行全数据集优化。在公共基准测试中的评估表明，FPS能够达到与最新方法相当或更好的性能。此外，FPS可以高效地扩展到多模态大型模型，并表现出跨蛋白质结构预测、远程传感分类以及地震检测等多样化领域的灵活性和通用性。我们预期FPS将为迁移学习（尤其是领域适应任务）提供一种简单、有效且可推广的范式。
URL
https://arxiv.org/abs/2508.18693
PDF
