Article
Computer Science, Artificial Intelligence
Yalan Ye, Ziwei Huang, Tongjie Pan, Jingjing Li, Heng Tao Shen
Summary: The study introduces a new unsupervised domain adaptation method that addresses the current issues by matching the distribution of two domains and reducing the classifier's bias towards source samples. Experimental results demonstrate the effectiveness of the method in unsupervised domain adaptation scenarios.
Article
Biochemistry & Molecular Biology
Xihao Chen, Jingya Yu, Shenghua Cheng, Xiebo Geng, Sibo Liu, Wei Han, Junbo Hu, Li Chen, Xiuli Liu, Shaoqun Zeng
Summary: This article proposes an unsupervised method to normalize cytopathology image styles through a two-stage style normalization framework, achieving superior results on six cervical cell datasets from different hospitals and scanners. The method greatly improves the recognition accuracy of lesion cells on unseen cytopathology images, meaningful for model generalization.
COMPUTATIONAL AND STRUCTURAL BIOTECHNOLOGY JOURNAL
(2021)
Article
Computer Science, Artificial Intelligence
Huanjing Yue, Yijia Cheng, Fanglong Liu, Jingyu Yang
Summary: This paper introduces an unsupervised Generative Adversarial Network for removing moire patterns, which is the first attempt at unsupervised learning based moire removal. The method outperforms state-of-the-art demoireing methods on a large set of test images.
Article
Chemistry, Multidisciplinary
Gaoming Yang, Yuanjin Qu, Xianjin Fang
Summary: This work presents a model applied in image generation that can alleviate model collapse and facilitate disentanglement. Compared to previous methods, this approach only makes minor modifications to the original model without changing the training method, and demonstrates superior performance on certain image datasets.
APPLIED SCIENCES-BASEL
(2022)
Article
Engineering, Civil
Can Song, Jin Wu, Lei Zhu, Mei Zhang, Haibin Ling
Summary: In this paper, an unsupervised convolutional neural network method is proposed for parsing nighttime road scenes. The method transfers daytime and nighttime images into a shared feature space using an appearance transferring module, and maps the feature to semantic labels using a segmentation module. Experimental results demonstrate significant improvement of the proposed model on the benchmark dataset.
IEEE TRANSACTIONS ON INTELLIGENT TRANSPORTATION SYSTEMS
(2022)
Article
Computer Science, Artificial Intelligence
Chaoqiang Zhao, Gary G. Yen, Qiyu Sun, Chongzhen Zhang, Yang Tang
Summary: This article proposes a masked generative adversarial network (GAN) for unsupervised monocular depth and ego-motion estimations. The MaskNet and Boolean mask scheme are designed to eliminate the effects of occlusions and impacts of visual field changes. Experiments show that each component proposed in this article contributes to the performance, and both depth and trajectory predictions achieve competitive performance on the KITTI and Make3D data sets.
IEEE TRANSACTIONS ON NEURAL NETWORKS AND LEARNING SYSTEMS
(2021)
Article
Computer Science, Artificial Intelligence
Wenming Cao, Zhongfan Zhang, Cheng Liu, Rui Li, Qianfen Jiao, Zhiwen Yu, Hau-San Wong
Summary: In this paper, an enhanced deep clustering network (EDCN) is proposed, which consists of a Feature Extractor, a Conditional Generator, a Discriminator, and a Siamese Network. The EDCN utilizes adversarial training to generate two types of data and original data, which are used to train the Feature Extractor for effective latent representation learning. A Siamese network is adopted to generate realistic data using pseudo-labels for enhancing the Feature Extractor. Extensive experiments show the effectiveness and superiority of the proposed EDCN.
PATTERN RECOGNITION
(2022)
Article
Computer Science, Artificial Intelligence
Shuai Fu, Jing Chen, Liang Lei
Summary: The domain shift between the training dataset and test dataset often affects the performance of deep learning models. In order to address this issue, we propose a Cooperative Attention Generative Adversarial Network (CAGAN) that generates target samples with given class labels and implements class-level transfer. The model integrates Coupled Generative Adversarial Networks (CoGAN) into a classification network and employs a semantic-consistent loss and attention layers for improved domain adaptation performance.
KNOWLEDGE-BASED SYSTEMS
(2023)
Article
Computer Science, Artificial Intelligence
Mingxu Zhang, Hongxia Wang, Peisong He, Asad Malik, Hanqing Liu
Summary: This study introduces an unsupervised domain adaptation strategy for detecting GAN-generated images in an unknown domain. By constructing a Self-Attention block and a novel loss function, the domain adaptation process is optimized, and experimental results demonstrate high detection accuracy on the generalization problem.
KNOWLEDGE-BASED SYSTEMS
(2022)
Article
Computer Science, Artificial Intelligence
Chunyan She, Tao Chen, Shukai Duan, Lidan Wang
Summary: In this work, a semantic-aware generative adversarial network is proposed to improve the performance limitations of low-light image enhancement (LLIE). By using a pre-trained VGG model to extract semantic information and designing innovative discriminator and image fusion strategies, better results are achieved. The experimental results demonstrate the competitiveness of the proposed model and the effectiveness of each component.
KNOWLEDGE-BASED SYSTEMS
(2023)
Article
Geochemistry & Geophysics
Caijun Ren, Xiangyu Wang, Jian Gao, Xiren Zhou, Huanhuan Chen
Summary: This study introduces a novel change detection framework utilizing Generative Adversarial Network (GAN) to generate better coregistered images, improving the performance of change detection algorithms. Experimental results demonstrate that this method is less sensitive to the issue of unregistered images and effectively utilizes deep learning structures.
IEEE TRANSACTIONS ON GEOSCIENCE AND REMOTE SENSING
(2021)
Article
Chemistry, Multidisciplinary
Yuanbo Yang, Qunbo Lv, Baoyu Zhu, Xuefu Sui, Yu Zhang, Zheng Tan
Summary: This paper proposes an end-to-end one-sided unsupervised image-dehazing network based on a generative adversarial network that directly learns the mapping between haze and haze-free images. The proposed method addresses the information inequality issue between haze and haze-free images and preserves the features of the original image through multi-scale skip connection and feature-fusion module. Experimental results demonstrate that our method outperforms existing dehazing algorithms in terms of quantitative metrics and visual effects.
APPLIED SCIENCES-BASEL
(2022)
Article
Geochemistry & Geophysics
Huanyu Zhou, Qingjie Liu, Dawei Weng, Yunhong Wang
Summary: This research proposes an unsupervised generative adversarial framework for pan sharpening, which learns from full-scale images without ground truths, addressing the difficulties faced by existing methods in dealing with full-scale images. The proposed method greatly improves the pan-sharpening performance on full-scale images, demonstrating its practical value.
IEEE TRANSACTIONS ON GEOSCIENCE AND REMOTE SENSING
(2022)
Article
Agriculture, Multidisciplinary
Federico Magistri, Jan Weyler, Dario Gogoll, Philipp Lottes, Jens Behley, Nik Petrinic, Cyrill Stachniss
Summary: In traditional arable crop fields, tractors uniformly apply large quantities of herbicides and pesticides for weed control and plant protection. Autonomous robots, however, can provide targeted treatments on a per-plant basis, making weed control and plant protection more ecologically friendly. Existing perception systems rely on machine learning techniques, but face performance decay in new field conditions. In this paper, a simple and effective unsupervised domain adaptation approach is proposed, allowing a segmentation pipeline to be adapted to different fields, robots, and crops, achieving high accuracy without extra manual annotations.
COMPUTERS AND ELECTRONICS IN AGRICULTURE
(2023)
Article
Geochemistry & Geophysics
Xiaobo Liu, Xiang Liu, Haoran Dai, Xudong Kang, Antonio Plaza, Wenjie Zu
Summary: This article introduces a multiscale unsupervised architecture based on generative adversarial networks (GANs) for remote sensing image pansharpening called Mun-GAN. Mun-GAN achieves high-resolution fusion of remote sensing images through a multiscale feature extractor, a self-adaptation weighted fusion module, and a nest feature aggregation module. Experimental results compared with other methods demonstrate that Mun-GAN yields better fusion results.
IEEE TRANSACTIONS ON GEOSCIENCE AND REMOTE SENSING
(2023)
Editorial Material
History & Philosophy Of Science
Nicola Liberati
Summary: The article discusses the impact of digital imagination on phenomenology and raises questions generated by the introduction of technologically embedded imagination.
FOUNDATIONS OF SCIENCE
(2022)
Article
Ethics
Nicola Liberati
Summary: This paper examines the impact of digital technologies in Japan and China on intimacy, and demonstrates how digital companions shape our identities and give meaning and value to love.
Article
Philosophy
Dmytro Mykhailov
Summary: In this paper, an 'environmental framework' is developed based on the postphenomenological variation of instrumental realism to address the 'problem of representation.' The framework focuses on three elements of the representational environment and emphasizes the interplay between scientific images and their instrumental environment.
PROMETEICA-REVISTA DE FILOSOFIA Y CIENCIAS
(2022)
Editorial Material
Philosophy
Nicola Liberati, Maurizio Balistreri
PROMETEICA-REVISTA DE FILOSOFIA Y CIENCIAS
(2022)
Editorial Material
Philosophy
Nicola Liberati, Wu Yan
PROMETEICA-REVISTA DE FILOSOFIA Y CIENCIAS
(2022)
Editorial Material
Philosophy
Nicola Liberati, Francesco Verso
PROMETEICA-REVISTA DE FILOSOFIA Y CIENCIAS
(2022)
Article
Social Sciences, Interdisciplinary
Dmytro Mykhailov
Summary: Intelligent algorithms and machine learning techniques pose significant challenges in contemporary value sensitive design, as they blur the causal link between programmers and computer behavior. This paper examines the conceptual tools within the value sensitive design school of thought for evaluating machine learning algorithms in the absence of a causal relation. It investigates computer intentionality conceptually and emphasizes machine learning algorithms technically.
HUMAN AFFAIRS-POSTDISCIPLINARY HUMANITIES & SOCIAL SCIENCES QUARTERLY
(2023)
Article
Philosophy
Dmytro Mykhailov, Nicola Liberati
Summary: This paper introduces phenomenological elements to provide a better framework for addressing the impact of technologies on society. It explores the concept of technological intentionality in relation to technological mediation and highlights the need to clarify their differences. The concept of passive synthesis is applied to technologies to illustrate their inner passive activity and how they can connect to a broader technological environment independently.
PHENOMENOLOGY AND THE COGNITIVE SCIENCES
(2023)
Article
Ethics
Dmytro Mykhailov
Summary: The purpose of this article is to contribute to the theory of technological mediation by introducing a new type of relationship between humans and technology called 'transcending mediation'. The previous postphenomenology theory didn't focus much on how technology mediates human's relation to Transcendence, due to its empirical and anti-metaphysical nature. However, in this paper, the author shows that the empirical element of technology can be balanced by metaphysical findings, using Karl Jaspers's metaphysics of ciphers. The author demonstrates how technology not only mediates our relation to the world but also shapes our relation to Transcendence.
Article
Ethics
Galit Wellner, Dmytro Mykhailov
Summary: This article suggests design principles for ethical algorithms in fighting fake news, highlighting the need for engineering and ethical considerations. Insights from ethics of care are proposed to guide the development of algorithms that generate care. Four algorithmic design principles are offered, including strategy development, stakeholder involvement, user reporting, and user updates. Implementing these principles can make the development process more ethically oriented and improve the ability to fight fake news.
SCIENCE AND ENGINEERING ETHICS
(2023)
Article
Philosophy
Dmytro Y. Mykhailov
Summary: The empirical turn in the mid-1980s had a significant impact on the field of philosophy of technology, leading to the emergence of postphenomenology as a contemporary school of thought. Postphenomenology emphasizes the role of technology in shaping human perception and the world. However, it overlooks the connection between human beings and transcendence. This article explores the possibility of revealing this connection through Karl Jaspers' metaphysics, focusing on the concepts of antinomy and cipher. By extending the theory of technological mediation with Jaspers' ideas, the study identifies the connection between transcendence and modern technology.
Article
Social Sciences, Interdisciplinary
Dmytro Mykhailov
Summary: The paper discusses the moral impact of Intelligent Decision-Support Systems (IDSS) in contemporary medical diagnosis, positioning it as a moral agent with specific agent features and behavior.
HUMAN AFFAIRS-POSTDISCIPLINARY HUMANITIES & SOCIAL SCIENCES QUARTERLY
(2021)