Predicting Harmful Algal Blooms Using Explainable Deep Learning Models: A Comparative Study DOI Open Access
Bekir Zahit Demiray, Omer Mermer, Özlem Baydaroğlu

и другие.

Water, Год журнала: 2025, Номер 17(5), С. 676 - 676

Опубликована: Фев. 26, 2025

Harmful algal blooms (HABs) have emerged as a significant environmental challenge, impacting aquatic ecosystems, drinking water supply systems, and human health due to the combined effects of activities climate change. This study investigates performance deep learning models, particularly Transformer model, there are limited studies exploring its effectiveness in HAB prediction. The chlorophyll-a (Chl-a) concentration, commonly used indicator phytoplankton biomass proxy for occurrences, is target variable. We consider multiple influencing parameters—including physical, chemical, biological quality monitoring data from stations located west Lake Erie—and employ SHapley Additive exPlanations (SHAP) values an explainable artificial intelligence (XAI) tool identify key input features affecting HABs. Our findings highlight superiority especially Transformer, capturing complex dynamics parameters providing actionable insights ecological management. SHAP analysis identifies Particulate Organic Carbon, Nitrogen, total phosphorus critical factors predictions. contributes development advanced predictive models HABs, aiding early detection proactive management strategies.

Язык: Английский

Artificial Intelligence-Enabled Intelligent Assistant for Personalized and Adaptive Learning in Higher Education DOI Creative Commons
Ramteja Sajja, Yusuf Sermet,

Muhammed Cikmaz

и другие.

Information, Год журнала: 2024, Номер 15(10), С. 596 - 596

Опубликована: Сен. 30, 2024

This paper presents a novel framework, artificial intelligence-enabled intelligent assistant (AIIA), for personalized and adaptive learning in higher education. The AIIA system leverages advanced AI natural language processing (NLP) techniques to create an interactive engaging platform. platform is engineered reduce cognitive load on learners by providing easy access information, facilitating knowledge assessment, delivering support tailored individual needs styles. AIIA’s capabilities include understanding responding student inquiries, generating quizzes flashcards, offering pathways. research findings have the potential significantly impact design, implementation, evaluation of AI-enabled virtual teaching assistants (VTAs) education, informing development innovative educational tools that can enhance outcomes, engagement, satisfaction. methodology, architecture, services, integration with management systems (LMSs) while discussing challenges, limitations, future directions

Язык: Английский

Процитировано

86
U-net-based semantic classification for flood extent extraction using SAR imagery and GEE platform: A case study for 2019 central US flooding DOI Creative Commons
Zhouyayan Li, İbrahim Demir

The Science of The Total Environment, Год журнала: 2023, Номер 869, С. 161757 - 161757

Опубликована: Янв. 21, 2023

Язык: Английский

Процитировано

69
A Review of Generative Adversarial Networks (GANs) and Its Applications in a Wide Variety of Disciplines: From Medical to Remote Sensing DOI Creative Commons
Ankan Dash, Junyi Ye, Guiling Wang

и другие.

IEEE Access, Год журнала: 2023, Номер 12, С. 18330 - 18357

Опубликована: Дек. 22, 2023

We look into Generative Adversarial Network (GAN), its prevalent variants and applications in a number of sectors. GANs combine two neural networks that compete against one another using zero-sum game theory, allowing them to create much crisper discrete outputs. can be used perform image processing, video generation prediction, among other computer vision applications. also utilised for variety science-related activities, including protein engineering, astronomical data remote sensing dehazing, crystal structure synthesis. Other notable fields where have made gains include finance, marketing, fashion design, sports, music. Therefore this article we provide comprehensive overview the wide disciplines. first cover theory supporting GAN, GAN variants, metrics evaluate GANs. Then present how applied twelve domains, ranging from STEM fields, such as astronomy biology, business marketing arts, As result, researchers may grasp work apply their own study. To best our knowledge, provides most survey GAN's different field.

Язык: Английский

Процитировано

66
Platform-independent and curriculum-oriented intelligent assistant for higher education DOI Creative Commons
Ramteja Sajja, Yusuf Sermet, David M. Cwiertny

и другие.

International Journal of Educational Technology in Higher Education, Год журнала: 2023, Номер 20(1)

Опубликована: Июль 23, 2023

Abstract Miscommunication between instructors and students is a significant obstacle to post-secondary learning. Students may skip office hours due insecurities or scheduling conflicts, which can lead missed opportunities for questions. To support self-paced learning encourage creative thinking skills, academic institutions must redefine their approach education by offering flexible educational pathways that recognize continuous this end, we developed an AI-augmented intelligent assistance framework based on powerful language model (i.e., GPT-3) automatically generates course-specific assistants regardless of discipline level. The virtual teaching assistant (TA) system, at the core our framework, serves as voice-enabled helper capable answering wide range questions, from curriculum logistics course policies. By providing with easy access information, TA help improve engagement reduce barriers At same time, it also logistical workload TAs, freeing up time focus other aspects supporting students. Its GPT-3-based knowledge discovery component generalized system architecture are presented accompanied methodical evaluation system’s accuracy performance.

Язык: Английский

Процитировано

44
GeoAI for Large-Scale Image Analysis and Machine Vision: Recent Progress of Artificial Intelligence in Geography DOI Creative Commons
Wenwen Li, Chia-Yu Hsu

ISPRS International Journal of Geo-Information, Год журнала: 2022, Номер 11(7), С. 385 - 385

Опубликована: Июль 11, 2022

GeoAI, or geospatial artificial intelligence, has become a trending topic and the frontier for spatial analytics in Geography. Although much progress been made exploring integration of AI Geography, there is yet no clear definition its scope research, broad discussion how it enables new ways problem solving across social environmental sciences. This paper provides comprehensive overview GeoAI research used large-scale image analysis, methodological foundation, most recent applications, comparative advantages over traditional methods. We organize this review according to different kinds structured data, including satellite drone images, street views, geo-scientific as well their applications variety analysis machine vision tasks. While tend use diverse types data models, we summarized six major strengths (1) enablement analytics; (2) automation; (3) high accuracy; (4) sensitivity detecting subtle changes; (5) tolerance noise data; (6) rapid technological advancement. As remains rapidly evolving field, also describe current knowledge gaps discuss future directions.

Язык: Английский

Процитировано

64
An ethical decision-making framework with serious gaming: a smart water case study on flooding DOI Creative Commons
Gregory J. Ewing, İbrahim Demir

Journal of Hydroinformatics, Год журнала: 2021, Номер 23(3), С. 466 - 482

Опубликована: Март 18, 2021

Abstract Sensors and control technologies are being deployed at unprecedented levels in both urban rural water environments. Because sensor networks allow for higher-resolution monitoring decision making time space, greater discretization of will an precision impacts, positive negative. Likewise, humans continue to cede direct decision-making powers decision-support technologies, e.g. data algorithms. Systems have ever-greater potential effect human lives, yet, be distanced from decisions. Combined these trends challenge resources management tools incorporate the concepts ethical normative expectations. Toward this aim, we propose Water Ethics Web Engine (WE)2, integrated generalized web framework voting-based preferences into support. We demonstrate with a ‘proof-of-concept’ use case where models learned respond flooding scenarios. Findings indicate that can capture group ‘wisdom’ within making. The methodology system presented here step toward building engage people algorithmic cases considered. share our its cyber components openly research community.

Язык: Английский

Процитировано

62
Terrain feature-aware deep learning network for digital elevation model superresolution DOI
Yifan Zhang, Wenhao Yu, Di Zhu

и другие.

ISPRS Journal of Photogrammetry and Remote Sensing, Год журнала: 2022, Номер 189, С. 143 - 162

Опубликована: Май 17, 2022

Язык: Английский

Процитировано

49
A sheep dynamic counting scheme based on the fusion between an improved-sparrow-search YOLOv5x-ECA model and few-shot deepsort algorithm DOI

Yuanyang Cao,

Jian Chen, Zichao Zhang

и другие.

Computers and Electronics in Agriculture, Год журнала: 2023, Номер 206, С. 107696 - 107696

Опубликована: Фев. 9, 2023

Язык: Английский

Процитировано

32
Temporal and spatial satellite data augmentation for deep learning-based rainfall nowcasting DOI Creative Commons
Özlem Baydaroğlu, İbrahim Demir

Journal of Hydroinformatics, Год журнала: 2024, Номер 26(3), С. 589 - 607

Опубликована: Март 1, 2024

Abstract The significance of improving rainfall prediction methods has escalated due to climate change-induced flash floods and severe flooding. In this study, nowcasting been studied utilizing NASA Giovanni satellite-derived precipitation products the convolutional long short-term memory (ConvLSTM) approach. goal study is assess impact data augmentation on flood nowcasting. Due requirements deep learning-based methods, performed using eight different interpolation techniques. Spatial, temporal, spatio-temporal interpolated are used conduct a comparative analysis results obtained through rainfall. This research examines two catastrophic that transpired in Türkiye Marmara Region 2009 Central Black Sea 2021, which selected as focal case studies. regions prone frequent flooding, which, dense population, devastating consequences. Furthermore, these exhibit distinct topographical characteristics patterns, frontal systems them also dissimilar. nowcast for significant difference. Although significantly reduced error values by 59% one region, it did not yield same effectiveness other region.

Язык: Английский

Процитировано

9
Realistic River Image Synthesis Using Deep Generative Adversarial Networks DOI Creative Commons

Akshat Gautam,

Muhammed Sit, İbrahim Demir

и другие.

Frontiers in Water, Год журнала: 2022, Номер 4

Опубликована: Фев. 23, 2022

In this paper, we demonstrated a practical application of realistic river image generation using deep learning. Specifically, explored generative adversarial network (GAN) model capable generating high-resolution and images that can be used to support modeling analysis in surface water estimation, meandering, wetland loss, other hydrological research studies. First, have created an extensive repository overhead training. Second, incorporated the Progressive Growing GAN (PGGAN), architecture iteratively trains smaller-resolution GANs gradually build up very high resolution generate quality (i.e., 1,024 × 1,024) synthetic imagery. With simpler architectures, difficulties arose terms exponential increase training time vanishing/exploding gradient issues, which PGGAN implementation seemed significantly reduce. The results presented study show great promise high-quality capturing details structure flow research.

Язык: Английский

Процитировано

38