Ideas Spread Inc. (E-Journals)
Not a member yet
1368 research outputs found
Sort by
Research on Food Volume Estimation Algorithm Based on Deep Learning
Get PDFWith the improvement of people's health awareness, the importance of accurate monitoring of dietary nutrition in the prevention and control of chronic diseases has become increasingly prominent. As one of the core indicators of dietary evaluation, the accurate estimation of food volume is directly related to the calculation accuracy of energy intake. However, in practical applications, due to factors such as different food forms, complex lighting conditions, and occlusion, traditional volume estimation methods have great limitations in terms of portability, real-time and user acceptance.
In recent years, with the development of deep learning technology, especially in the field of computer vision, image-based food volume estimation has become a research hotspot [1]. Although existing studies have solved the problems of food image recognition and volume estimation to a certain extent, there are still challenges in terms of model lightweight, occlusion robustness, complex background adaptability, and single-image estimation accuracy. Especially in the application scenarios for ordinary users and mobile terminals, how to achieve efficient and low-cost volume estimation is a problem that has not been fully solved at present [2]. Food volume estimation based on deep learning can help improve the intelligent level of dietary assessment, promote the development of mobile nutrition applications, and provide basic support for universal health data collection. In this paper, deep learning technology is used to accurately identify and segment the food area in the image, and the food volume is accurately estimated based on the reference object size
Site Suitability Analysis and Treatment Technology for Solid Waste Landfills in Hidden Karst Areas of the Dongting Lake Region
Get PDFThe complex geological conditions of hidden karst in the Dongting Lake region pose significant challenges for the siting and foundation treatment of solid waste landfills. This paper takes a typical landfill in this region as a case study, integrating three-dimensional (3D) geological modeling, karst vulnerability assessment (LEPT), the Analytic Hierarchy Process (AHP), and finite element numerical simulation to systematically conduct site suitability analysis and stability evaluation. First, based on geological survey data, a 3D geological model was constructed to identify the distribution characteristics of weak substrata and hidden karst features. Second, the LEPT method was employed to assess regional karst vulnerability, and the AHP was used to comprehensively grade site suitability based on multiple criteria encompassing natural, environmental, and infrastructural factors. Furthermore, the finite element method was applied to analyze foundation settlement and stability under landfill loading, proposing comprehensive treatment technologies based on deformation coordination and reinforcement. The research results indicate that landfill siting in hidden karst areas should prioritize avoiding high-vulnerability zones; settlement of weak substrata can be effectively controlled through composite foundation treatment and synergistic design with the lining system; the proposed integrated "assessment-design-treatment" framework can provide a theoretical basis and technical reference for the construction of similar landfills in karst regions
The Legal and Regulatory Challenges Posed by the Integration of Predictive Analytics (PA) into International Arbitration
Get PDFThis research examines the rapid adoption of predictive analytics in international ADR. It argues that although the technology produces efficiency gains - reducing costs and timelines while enhancing consistency of decision making - the very efficiency gains carry with them collective and individual legal and ethical risks. This analysis supports the argument that the PA's dependence on complicated, proprietary ML models results in an inescapable legitimacy crisis. The emergence of this crisis can be traced back to fundamental process disputes, particularly the "Black Box" defect, which introduces an insurmountable opacity to algorithms and acute flaws in data governance surrounding confidentiality, reverse engineering risk, and non-compliance with regimes like the GDPR. All of these violations contravene the principles of due process, ultimately jeopardizing the global enforceability of arbitral awards under the New York Convention.
The key contention raised in the proceedings is the tension between the confidentiality of technical information and the non-derogability of a party's right to contest crucial inputs affecting the decision. Due to the lack of regulations, proprietary systems can operate without oversight. These systems will import systemic bias from historical data into future legal outcomes. This structural deficit creates a compelling ethical reason for requiring immediate regulation. The current failure of institutional control over a third-party vendor's technology necessitates a shift from simple ethical advice to enforceable and transnational standards, ensuring the long-term legitimacy of the international ADR system.
This research proposes a comprehensive three-pillar regulatory framework to shift the ecosystem from opaque behaviour to auditable transparency. The framework requires (1) the tribunal to disclose the model specification used and its methodology mandatorily; (2) conducting a data provenance, quality and bias audit of the Data Protection model, that is not conducted by the software vendor, of all commercial PA models; and (3) mandating algorithmic contestability, where any parties has the right to contest the outcome of an algorithmic output through a forensic analysis carried out by an expert. Future research needs to look at empirical analyses assessing the relationship between PA use and award challenge rates. Additionally, future research should develop open-source, auditable PA models that non-commercial institutions can utilize. Ultimately, effective governance of algorithmic arbitration relies on international institutions, such as UNCITRAL, to incorporate transparency and accountability into the arbitral procedure in a meaningful manner
A Study of Public Health Theory and Practice Curriculum Reform for Research Integrity Integration Through Research Project Based Pedagogy in Graduate Education
Get PDFThis study addresses critical gaps in public health graduate education where conventional teaching diverges from authentic research practice, and fragmented research integrity training fails to provide coherent guidance throughout the research lifecycle. We propose a pedagogical reform concept of "full-cycle research project embedding," centered on the core course "Public Health Theory and Practice" and synergistically aligned with methodological courses like "Epidemiology." The "Research Lifecycle-Embedded Model" structures research integrity education across four phases—design, implementation, analysis, and translation—cultivating competencies in risk identification, ethical adherence, critical reflection, and professional responsibility. Innovations include authentic sub-projects from national research programs, a "Fieldwork Integrity Log" for real-time ethical challenge documentation, peer-driven "integrity workshops," and a three-tiered portfolio-based assessment system evaluated through multi-stakeholder collaboration. The evaluation framework shifts focus from knowledge testing to behavioral observation, linking integrity performance to academic advancement. While limitations include subjective assessment methods and limited project repositories, this model offers a replicable framework for transforming public health graduate training from fragmented, post-hoc instruction to integrated, pre-emptive professional development—achieved through systematically embedding authentic research projects throughout the "Public Health Theory and Practice" curriculum—preparing graduates to competently navigate ethical complexities in research-to-policy translation
Beyond Equity To Accountability: The Paris Agreement’s Normative Shift In Pursuing Climate Justice
Get PDFDespite the Paris Agreement’s 2015 normative shift by introducing “climate justice” into a core international climate instrument, its capacity to deliver substantive justice remains constrained by its foundational compromises. This analysis argues that while the Agreement advances climate governance through a hybrid responsibility framework—integrating common but differentiated responsibilities, the 1.5°C goal, and an enhanced transparency system—it fails to establish a legally rigorous accountability architecture sufficient to address historical inequities and enforce collective action. Key normative gaps include the ambiguous treatment of historical responsibility, the lack of explicit adoption of the polluter-pays principle, and the absence of a binding compliance mechanism with meaningful consequences. Furthermore, its hybrid approach, which blends top-down goals with bottom-up nationally determined contributions (NDCs), creates critical legal ambiguities and potential loopholes that may undermine compliance compared to the clearer, albeit contentious, top-down model of the Kyoto Protocol. To transform its normative promise into effective justice, the Agreement must institutionalize quantifiable historical responsibility metrics, establish a dual liability rule combining polluter-pays and beneficiary-pays principles, and create a binding accountability system with financial consequences for non-compliance. The operationalization of the loss and damage fund at COP28 represents a nascent but critical step toward acknowledging historical responsibility; however, its success hinges on integrating these principles into the core of the Agreement's implementation and review mechanisms. Only through such substantive reforms can the Paris Agreement’s framework evolve from a platform for voluntary cooperation into an instrument of enforceable climate justice
A Corpus-Based Contrastive Analysis of Metadiscourse in High School and University EFL Writing
Get PDFExisting research on metadiscourse in EFL writing has concentrated on the level of higher education, with relatively fewer studies investigating the comparative differences between high school and university stages. This study conducts a corpus-based contrastive analysis of metadiscourse in 440 argumentative essays written by Chinese high school and university learners. Adopting metadiscourse model, the results finds a shift in writing patterns. Texts by high school students reflect a tendency of enumeration and subjective assumption; university students show a significantly higher use of evidentials and engagement markers, indicating a transformation toward academic objectivity and dialogic interaction. The findings suggest that the transition to university writing involves a shift from list-making to logical structuring and the building of writer-reader relationships. The study concludes with pedagogical implications for bridging the gap in EFL writing instruction
Applications of Metaverse Technology in Future Experimentation
Get PDFThe metaverse is an interactive virtual–reality space built on technologies including artificial intelligence, big data, blockchain, virtual reality, simulation, and digital twins. In recent years, it has become a focal point across various fields. This paper first analyzes the current application status of the metaverse in the military domain. Building on this foundation, it further explores its potential applications in military experimentation and analyzes the challenges and issues the metaverse may face in this field, aiming to provide guidance for related technological research
Asymptotic Stability of Linear Neutral Differential Equations with only Memory-type Neutral Terms
Get PDFIn this paper, we investigate the asymptotic stability of linear neutral differential equations featuring only a memory-type neutral term, without additional internal or boundary damping. By combining the properties of a generalized positive definite kernel (abbreviated to GPDK) with the classical multiplier method, an auxiliary system is constructed to estimate the kinetic energy, potential energy, and convolution terms in the energy, thus proving that the energy function is integrable and decays to zero at least by the rate of . The ideas presented in this paper can be applied to other neutral differential problems and improve the related decay estimates
A Hybrid CNN - TKAN Architecture for Solving Partial Differential Equations
Get PDFThis paper presents a Physics-Informed Convolutional Temporal Kolmogorov-Arnold Network (PhyCTKAN) to address the challenge of solving spatiotemporal partial differential equations (PDEs). The proposed method enhances the Temporal Kolmogorov - Arnold Network (TKAN) through convolution operations and integrates an encoder - decoder architecture for dimensionality reduction and spatial domain mapping. Solutions are iteratively computed at each time step via an autoregressive process based on Euler discretization. Experimental results on nonlinear spatiotemporal systems demonstrate that PhyCTKAN achieves superior long-term prediction accuracy compared to the PhyCRNet model, while maintaining high computational precision even with increased time step sizes
Analysis of Cardiovascular Function Stability in Anesthesia Patients in High-Altitude Areas
Get PDFObjective: To explore the stability of cardiovascular function and influencing factors in anesthesia patients in high-altitude areas. Methods: Select anesthesia patients who meet the inclusion criteria in high-altitude areas of Aba County, record their basic information in detail, and use a multifunctional monitor to continuously monitor cardiovascular function indicators, including heart rate, blood pressure, cardiac output, etc., before surgery, after anesthesia induction, at different stages of surgery, and after surgery. Analyze the impact of different factors on the stability of cardiovascular function. Results: It was found that the cardiovascular function indicators of patients showed dynamic changes during anesthesia surgery, and factors such as age, gender, surgical type, and anesthesia method all affected their stability. Conclusion: This study provides theoretical basis and practical guidance for cardiovascular risk management in anesthesia surgery in high-altitude areas