Successful Model Research Articles

Quantifying the Cumulative Effects of Large-Scale Reclamation on Coastal Wetland Degradation

Considering the importance of coastal wetlands as key land resources and the ecological degradation caused by large-scale and multi-stage reclamation, as well as the significant synergistic and superimposed effects of reclamation on wetland degradation in temporal and spatial dimensions, it is vital to conduct in-depth research on the impact mechanisms and cumulative effects of reclamation on wetland degradation. However, the existing methods for evaluating these cumulative effects still have some shortcomings in characterizing the spatiotemporal scale. Consequently, it is urgent to introduce or develop a cumulative effect evaluation method based on remote sensing. Taking the Jiangsu coastal wetland as a typical case study area, the present study constructed a cumulative effect evaluation method based on calculus theory combined with landscape succession modeling and statistical analysis. This method was then used to quantitatively analyze the impacts and cumulative effects of reclamation on wetland degradation in the Jiangsu coastal region from 1980 to 2024. The results show that degradation of the Jiangsu coastal wetlands over the last 45 years covered 2931.54 km2, accounting for 46.92% of the area in 1980. This degradation primarily reflects a shift from natural wetland to constructed wetland. In addition, the reclaimed area of 2119.61 km2 is mainly used for aquaculture and agricultural cultivation. The reclamation rate of Jiangsu showed insignificant fluctuations and significant spatial differences. The reclamation rate of the north counties and cities presented a downward trend, while that of the south counties and cities presented an upward trend. Reclamation has a significant impact on wetland degradation, with a contribution rate of 50.62%. The cumulative effect in the study area reached its maximum value in 2015, except for Nantong City. This study provides a new perspective for quantitatively analyzing the impacts and cumulative effects of coastal wetland reclamation and provides guidance for the effective management and sustainable utilization of coastal wetland resources.

Newly discovered diamictite and cap carbonate couplet in the southern Alxa Block, northwestern China: Implications for stratigraphic correlation and Marinoan glaciation

Neoproterozoic glaciogenic diamictite and cap carbonate couplets have played a pivotal role in understanding glacial-interglacial cycles and establishing regional stratigraphic correlation. The Alxa Block in northwestern China preserves a sequence of Neoproterozoic diamictites along its southern margin, but the age and origin of the succession remain debatable due to the lack of discovery of cap carbonate. We report newly discovered cap carbonates that overlie the diamictites of the Shaohuotonggou Formation in the Longshoushan region in the southern Alxa Block. Based on detailed geological investigations, we identified massive diamictites, stratified diamictites, and cap carbonates in the lower part of the formation. The presence of ice-rafted dropstones, bullet-shaped and facetted clasts, glacial striations, and relatively low chemical index of alteration values of sedimentary matrix support a glaciogenic origin of the diamictites. The 2- to 2.6-m-thick cap carbonates are mainly composed of thinly laminated microcrystalline dolomites and show sheet cracks, cemented breccias, and tepee-like structures at the basal part of the unit. These features and their consistently negative δ13C values (−5.2‰ to −2.2‰) are characteristic of Marinoan-age cap carbonates (ca. 635 Ma). The quasi-continuous deposition of the massive diamictites, stratified diamictites, and cap carbonates suggests that the formation of this couplet was closely related to the Marinoan glaciation and subsequent deglaciation. We propose a three-stage depositional model for the glaciogenic succession and recommend that the diamictite and cap carbonate couplet in the Alxa Block provides a credible mark of the Cryogenian−Ediacaran boundary for further stratigraphic correlation and investigation.

Predicting student success with and without library instruction using supervised machine learning methods

PurposeThe main purpose of this analysis was to demonstrate the value of predictive modeling of student success and identify the key groups of students for which library instruction could provide the most impact.Design/methodology/approachData regarding the attendance of library instruction associated with a first-year writing course were combined with student demographic and academic data over a four year period representing over 10,000 students. We applied supervised machine learning methods to determine the most accurate model for predicting student outcomes, including course outcome, persistence and graduation. We also assessed the impact of library instruction on these outcomes.FindingsThe gradient-boosted decision tree model provided the most accurate predictions. The impact of library instruction was modest but still was second only to the previous grade point average (GPA). The value of this metric, however, was greatest for students who were struggling, especially those who were first-generation students, regardless of ethnicity. More notably, the impact of library instruction was substantially greater for specific student demographics, including students with lower cumulative GPAs.Research limitations/implicationsFeatures of the models were limited to high-level academic metrics, some of which may not be very useful in predicting outcomes. Measures more closely related to learning styles, the course or course of study could provide for greater accuracy.Practical implicationsPrediction modeling could allow for a more selective approach to outreach and offers information that the librarian can use to customize instruction sessions and reference interactions.Social implicationsTargeting students who may be at risk of not succeeding in a course has ethical implications either way. If used to bias the subjective assessments, these predictions could produce self-fulfilling prophecies. Conversely, to ignore indicators of possible difficulties the student may have with the material is a disservice to the education of that student.Originality/valueThere are few studies that have incorporated library instruction into models of predicting student outcomes. Library resources and services can play a major role in the success of students, particularly those who have had less exposure to the resources and skills needed to use these resources.

Implementation of Telemedicine in Medical Practice: Opportunities and Barriers

The integration of telemedicine into medical practice represents a transformative shift in healthcare delivery, offering both unprecedented opportunities and encountering significant barriers. This journal article explores the implementation of telemedicine, examining the multifaceted landscape of benefits and challenges associated with its adoption in medical settings. The opportunities presented by telemedicine include enhanced accessibility to healthcare services, especially in remote or underserved areas, improved patient outcomes through timely interventions, and increased efficiency in healthcare delivery. Telemedicine also facilitates remote patient monitoring, teleconsultations, and the exchange of medical information, contributing to a more patient-centric and adaptable healthcare system. However, the implementation of telemedicine faces several barriers, including regulatory complexities, concerns about data security and privacy, technological infrastructure limitations, and the need for healthcare professionals to adapt to new modes of communication. Cultural and organizational resistance further complicates the seamless integration of telemedicine into established medical practices. The article critically analyzes case studies and global implementations of telemedicine, highlighting successful models and lessons learned. It emphasizes the importance of policy frameworks, standardization, and ongoing training for healthcare providers to optimize the benefits of telemedicine while mitigating potential risks.

The Effects and Mechanism of Scutellaria baicalensis Georgi Stems and Leaves Flavonoids on Myelin Sheath Degeneration Induced by Composite Aβ in Rats.

Alzheimer's disease is a degenerative disease of the central nervous system, and its characteristic pathological changes are closely associated with Aβ deposition and neurofibrillary tangles. Many studies have found that malignant changes in the myelin sheath and oligodendrocyte (OL) are accompanied by the occurrence and development of AD. Therefore, any method that can resist myelin sheath and OL disorders may be a potential strategy for AD. To investigate the effects and mechanism of Scutellaria baicalensis Georgi stem and leaf flavonoids (SSFs) on the myelin sheath degeneration induced by Aβ25-35 combined with AlC13 and RHTGF-β1 (composite Aβ) in rats. A rat AD model was established by intracerebroventricular injection of composite Aβ. The Morris water maze was used to screen the memory impairment rat model. The successful model rats were divided into the model group and the 35, 70, and 140 mg/kg SSFS groups. The myelin sheath changes in the cerebral cortex were observed with an electron microscope. The expression of the oligodendrocyte- specific protein claudin 11 was detected with immunohistochemistry. The protein expression levels of myelin oligodendrocyte glycoprotein (MOG), myelin-associated glycoprotein (MAG) and myelin basic protein (MBP), sphingomyelin synthase-1 (SMS1), and sphingomyelinase-2 (SMPD2) were assayed by Western blotting. The intracerebroventricular injection of composite Aβ caused degeneration of the myelin sheath structure and was accompanied by the decreased claudin 11, MOG, MAG, MBP, and SMS1, and increased SMPD2 protein expression in the cerebral cortex. However, 35, 70, and 140 mg/kg SSFs can differentially ameliorate the above abnormal changes induced by composite Aβ. SSFs can alleviate myelin sheath degeneration and increase the protein expression of claudin 11, MOG, MAG, and MBP, and the effective mechanism may be related to the positive regulation of SMS1 and SMPD2 activities.

Fully automated landmarking and facial segmentation on 3D photographs.

Three-dimensional facial stereophotogrammetry provides a detailed representation of craniofacial soft tissue without the use of ionizing radiation. While manual annotation of landmarks serves as the current gold standard for cephalometric analysis, it is a time-consuming process and is prone to human error. The aim in this study was to develop and evaluate an automated cephalometric annotation method using a deep learning-based approach. Ten landmarks were manually annotated on 2897 3D facial photographs. The automated landmarking workflow involved two successive DiffusionNet models. The dataset was randomly divided into a training and test dataset. The precision of the workflow was evaluated by calculating the Euclidean distances between the automated and manual landmarks and compared to the intra-observer and inter-observer variability of manual annotation and a semi-automated landmarking method. The workflow was successful in 98.6% of all test cases. The deep learning-based landmarking method achieved precise and consistent landmark annotation. The mean precision of 1.69 ± 1.15 mm was comparable to the inter-observer variability (1.31 ± 0.91 mm) of manual annotation. Automated landmark annotation on 3D photographs was achieved with the DiffusionNet-based approach. The proposed method allows quantitative analysis of large datasets and may be used in diagnosis, follow-up, and virtual surgical planning.

Postoperative Global Period Cost Reduction Using 3 Successive Risk-Stratified Pancreatectomy Clinical Pathways.

We hypothesized that iterative revisions of our original 2016 risk-stratified pancreatectomy clinical pathways would be associated with decreased 90-day perioperative costs. From a single-institution retrospective cohort study of consecutive patients with 3 iterations: "version 1" (V1) (October 2016 to January 2019), V2 (February 2019 to October 2020), and V3 (November 2020 to February 2022), institutional data were aggregated using revenue codes and adjusted to constant 2022-dollar value. Grand total perioperative costs (primary endpoint) were the sum of pancreatectomy, inpatient care, readmission, and 90-day global outpatient care. Proprietary hospital-based costs were converted to ratios using the mean cost of all hospital operations as the denominator. Of 814 patients, pathway V1 included 363, V2 229, and V3 222 patients. Accordion Grade 3+ complications decreased with each iteration (V1: 28.4%, V2: 22.7%, and V3: 15.3%). Median length of stay decreased (V1: 6 days, interquartile range [IQR] 5 to 8; V2: 5 [IQR 4 to 6]; and V3: 5 [IQR 4 to 6]) without an increase in readmissions. Ninety-day global perioperative costs decreased by 32% (V1 cost ratio 12.6, V2 10.9, and V3 8.6). Reduction of the index hospitalization cost was associated with the greatest savings (-31%: 9.4, 8.3, and 6.5). Outpatient care costs decreased consistently (1.58, 1.41, and 1.04). When combining readmission and all outpatient costs, total "postdischarge" costs decreased (3.17, 2.59, and 2.13). Component costs of the index hospitalization that were associated with the greatest savings were room or board costs (-55%: 1.74, 1.14, and 0.79) and pharmacy costs (-61%: 2.20, 1.61, and 0.87; all p < 0.001). Three iterative risk-stratified pancreatectomy clinical pathway refinements were associated with a 32% global period cost savings, driven by reduced index hospitalization costs. This successful learning health system model could be externally validated at other institutions performing abdominal cancer surgery.

Biocultural Conservation-based Narrative Text: An Effort to Enhance Conservation Literacy of the Local Community in the Sanctuary Area

This study is part of the Faculty of Languages and Arts’ second-year research project in partnership with International Animal Rescue (IAR). The major goal was to raise conservation awareness among the local people dwelling in the sanctuary region of Pematang Gadung Village, Ketapang Regency, West Borneo, to tackle environmental challenges such as illegal logging, poaching, and illegal mining. The research and development process were divided into three stages based on the Successive Approximation Model (SAM): preparation, iterative design, and iterative development. This work focuses primarily on the iterative design phase, expanding on the findings of the research completed in the first year when the local people perceived forest as commodities due to their lack of conservation literacy. Their biocultural conservation, on the other hand, was profoundly based on their indigenous, Muslim Malay ethnic beliefs and traditions. This study aims to construct a biocultural conservation-based narrative text as a first step toward generating conservation material that resonates with the community’s local wisdom using a biocultural approach.&#x0D; Keywords: conservation literacy, biocultural conservation approach, SAM, REB, local community of Pematang Gadung Village

Implementing intensive comprehensive aphasia programs in Brazil: Challenges and solutions

ABSTRACT Aphasia, a communication disability prevalent among stroke survivors, significantly impacts psychosocial well-being and quality of life. Intensive Comprehensive Aphasia Programs (ICAPs) provide intensive and comprehensive treatment, incorporating individual and group therapy, patient and family education, and technology utilization. Despite positive outcomes, global implementation of ICAPs remains limited. In Brazil, preliminary evidence supports the development of intensive and comprehensive aphasia programs, such as the Brazilian Aphasia House—a program inspired by the successful Aphasia House model in the United States. The Brazilian Aphasia House offers personalized and intensive therapy in a simulated domestic environment, encompassing individual and group therapy sessions, and patient and family education, with the goal of enhancing functional communication skills and overall quality of life. Operating as a teaching clinic within a public university, it also serves as a training ground for future therapists to expand ICAP availability in Brazil. Challenges in implementing ICAPs in Brazil include socioeconomic barriers, limited public aphasia awareness, a shortage of specialized professionals, transportation, and session costs. Sustainable implementation requires committed leadership and financial resourcing, with involvement from public institutions and universities. By increasing awareness and nationwide implementation of ICAPs, access to aphasia rehabilitation services can be improved, potentially leading to better outcomes for individuals with aphasia and their caregivers. The Brazilian Aphasia House serves as an exemplar of an ICAP, offering intensive and personalized therapy to enhance communication abilities and promote social participation for people with aphasia.

Ligand-based pharmacophore modeling and machine learning for the discovery of potent aurora A kinase inhibitory leads of novel chemotypes.

Aurora-A (AURKA) is serine/threonine protein kinase involved in the regulation of numerous processes of cell division. Numerous studies have demonstrated strong association between AURKA and cancer. AURKA is overexpressed in many cancers, such as colon, breast and prostate cancers. Consequently, AURKA has emerged as promising target for therapeutic intervention in cancer management. Herein, we describe a computational workflow for the discovery of novel anti-AURKA inhibitory leads starting with ligand-based assessment of the pharmacophoric space of six diverse sets of inhibitors. Subsequently, machine learning/QSAR modeling was coupled with genetic function algorithm to search for the best possible combination of machine learner, ligand-based pharmacophore(s) and molecular descriptors capable of explaining variation in anti-AURKA bioactivities within a collected list of inhibitors. Two learners succeeded in achieving acceptable structure/activity correlations, namely, random forests and extreme gradient boosting (XGBoost). Three pharmacophores emerged in the successful ML models. These were then used as 3D search queries to mine the National Cancer Institute database for novel anti-AURKA leads. Top-ranking 38 hits were assessed in vitro for their anti-AURKA bioactivities. Among them, three compounds exhibited promising dose-response curves, demonstrating experimental IC50 values ranging from sub-micromolar to low micromolar values. Remarkably, two of these compounds are of novel chemotypes.

Liraglutide Improves PCOS Symptoms in Rats by Targeting FDX1.

Polycystic ovary syndrome (PCOS) is a gynecological endocrine disorder characterized by ovulatory disorders, hyperandrogenemia, and polycystic changes in the ovaries. FDX1 is a ferredoxin-reducing protein on human mitochondria that plays an important role in steroid anabolism. Liraglutide, a glucagon-like peptide-1 receptor agonist (GLP-1RA), has recently emerged as a potential therapeutic agent for PCOS. Recent studies have suggested that FDX1 may be associated with the development of PCOS. This study aims to explore the pivotal role of FDX1 in the amelioration of PCOS through liraglutide intervention. A PCOS rat model was induced via subcutaneous DHEA injections. Following successful model establishment, the rats were treated with liraglutide combined with metformin, or with each drug individually, over a six-week period. After 6 weeks of treatment, we assessed changes in body weight, fasting blood glucose, sex hormone levels, estrous cycle regularity, ovarian morphology, FDX1 expression in ovarian tissue, and ovarian ROS levels. PCOS rats exhibited significant increases in body weight and fasting blood glucose levels, disrupted estrous cycles, and polycystic ovarian morphology. FDX1 expression was notably reduced in the ovarian tissues of PCOS rats. Treatment with liraglutide, both alone and in combination with metformin, led to improvements in body weight, fasting blood glucose, sex hormone balance, estrous cycle regularity, ovarian morphology, and ovarian ROS levels. Notably, FDX1 expression was significantly restored in all treatment groups, with the most substantial increase observed in the liraglutide-treated group. This study suggests that FDX1 could serve as a potential biomarker for elucidating the underlying mechanisms of liraglutide's therapeutic effects in PCOS management.

On the Inadequacy of the Current Transgenic Animal Models of Alzheimer's Disease: The Path Forward.

For at least two reasons, the current transgenic animal models of Alzheimer's disease (AD) appear to be patently inadequate. They may be useful in many respects, the AD models; however, they are not. First, they are incapable of developing the full spectrum of the AD pathology. Second, they respond spectacularly well to drugs that are completely ineffective in the treatment of symptomatic AD. These observations indicate that both the transgenic animal models and the drugs faithfully reflect the theory that guided the design and development of both, the amyloid cascade hypothesis (ACH), and that both are inadequate because their underlying theory is. This conclusion necessitated the formulation of a new, all-encompassing theory of conventional AD-the ACH2.0. The two principal attributes of the ACH2.0 are the following. One, in conventional AD, the agent that causes the disease and drives its pathology is the intraneuronal amyloid-β (iAβ) produced in two distinctly different pathways. Two, following the commencement of AD, the bulk of Aβ is generated independently of Aβ protein precursor (AβPP) and is retained inside the neuron as iAβ. Within the framework of the ACH2.0, AβPP-derived iAβ accumulates physiologically in a lifelong process. It cannot reach levels required to support the progression of AD; it does, however, cause the disease. Indeed, conventional AD occurs if and when the levels of AβPP-derived iAβ cross the critical threshold, elicit the neuronal integrated stress response (ISR), and trigger the activation of the AβPP-independent iAβ generation pathway; the disease commences only when this pathway is operational. The iAβ produced in this pathway reaches levels sufficient to drive the AD pathology; it also propagates its own production and thus sustains the activity of the pathway and perpetuates its operation. The present study analyzes the reason underlying the evident inadequacy of the current transgenic animal models of AD. It concludes that they model, in fact, not Alzheimer's disease but rather the effects of the neuronal ISR sustained by AβPP-derived iAβ, that this is due to the lack of the operational AβPP-independent iAβ production pathway, and that this mechanism must be incorporated into any successful AD model faithfully emulating the disease. The study dissects the plausible molecular mechanisms of the AβPP-independent iAβ production and the pathways leading to their activation, and introduces the concept of conventional versus unconventional Alzheimer's disease. It also proposes the path forward, posits the principles of design of productive transgenic animal models of the disease, and describes the molecular details of their construction.

An approach based on deep learning methods to detect the condition of solar panels in solar power plants

Solar panels are increasingly popular due to global energy shortages and rising costs. However, managing large or elevated panel systems requires regular oversight, leading to potential time and cost challenges. This study was focused on developing an AI-based drone for panel detection to address these issues and facilitate the control process. A low-cost system for AI-based identification of dusty, broken, and healthy solar panels was created using a Raspberry Pi 4B board and camera. The study proposed a Histogram Equalization (HE)-based preprocessing technique to improve the AI model. Firstly, the trainings were performed with YOLOv5 without the proposed method at epoch values of 100, 150, and 200 in order to see the effectiveness of the proposed method more clearly. As a result of these trainings, the highest F1 score was obtained as 80 %. In the second step, three deep learning algorithms - YOLOv5, YOLOv7, and YOLOv8 - with epochs of 100, 150, and 200 respectively, were used for training with the proposed method. A detailed comparative analysis of the developed models was carried out regarding their performance metrics. The YOLOv5l was obtained as the most successful panel detection model with an F1 score of 97 % at 150 epochs. The model with the best performance metrics was used in a real-time test application with an AI-based drone. F1 score results were obtained between 90 % and 97 %, mainly supporting the success rate obtained in real-time application. The results strongly support the effectiveness of this proposed method for panel detection tasks, showcasing its high efficacy and promising potential.

A problem case set for teaching psychophysics and psychophysical modelling

ABSTRACT Psychophysics can be considered a methodological core for quantitative studies of perception, as well as attention, memory, decision-making, etc. In typical introductory “Sensation and Perception” courses, students usually get a very basic idea of psychophysics and its main concepts. Here, I present an educational problem set for advanced practice in psychophysical data analysis and modelling. It is designed for students interested in a deeper understanding of perception and other cognitive processes and can be used as part of advanced undergraduate and master courses. Each problem case from the set includes a pregenerated data set and a research question that the student should answer by creating a psychophysical model (or several alternative models) and testing it against the data. Importantly, a successful model requires not only basic modelling and coding skills but also encourages deeper understanding of the logic of the method and basic concepts used to interpret behaviour in a psychophysical experiment. The article also presents a demonstration case with a detailed tutorial on its solution that can be useful for both students and instructors.

Predicting students' success level in an examination using advanced linear regression and extreme gradient boosting

This research employs a hybrid approach, integrating advanced linear regression and extreme gradient boosting (XGBoost), to forecast student success rates in exams within the dynamic educational landscape. Utilizing Kaggle-sourced data, the study crafts a model amalgamating advanced linear regression and XGBoost, subsequently assessing its performance against the primary dataset. The findings showcase the model's efficacy, yielding an accuracy of 0.680 on the fifth test and underscoring its adeptness in predicting students' exam success. The discussion underscores XGBoost's prowess in managing data intricacies and non-linear features, complemented by advanced linear regression offering valuable coefficient interpretations for linear relationships. This research innovatively contributes by harmonizing two distinct methods to create a predictive model for students' exam success. The conclusion emphasizes the merits of an ensemble approach in refining prediction accuracy, recognizing, however, the study's limitations in terms of dataset constraints and external factors. In essence, this study enhances comprehension of predicting student success, offering educators insights to identify and support potentially struggling students.

A successful and ideal randomized orthogonal additive response model

A successful and ideal randomized orthogonal additive response model

Exploring the Intricate Connections between the Influence of Fishing on Marine Biodiversity and Their Delivery of Ecological Services Driven by Different Management Frameworks

The goal of this project is to assess the state of marine biodiversity in the Natura 2000 marine network in Spain and analyze the impact of fishing policies on it. The study focuses on three marine demarcations in Spain: North Atlantic, South Atlantic, and the Strait of Gibraltar and Alborán Sea. The research uses the DPSIR (Drivers of Change-Pressure-State-Impact-Response) framework, a multi-criteria analysis approach to understand the relationships between biodiversity loss, fishing, marine ecosystem integrity, and the provision of ecosystem services. Our results revealed a significant decline in biodiversity since 1985 in the four marine areas studied; this loss was more pronounced in the Gulf of Cadiz (in Andalusia) and less intense in the northern regions of Spain (Galicia and Asturias). This trend aligns with the global degradation of marine ecosystems and loss of biodiversity caused by human activities. The main drivers behind this decline are economic factors promoting industrial fishing and overexploitation. However, there has been a reduction in the industrial fishing fleet since the 2000s, supported by the European Maritime and Fisheries Fund for transitioning toward sustainable fishing methods. Despite the increase in regulations and the establishment of marine protected areas, these measures have not been effective enough to stop the loss of marine biodiversity. The results highlight the importance of combining administrative measures such as creating marine protected areas and implementing fisheries management regulations with the preservation of cultural services provided by these ecosystems. Successful governance models that involve collaboration between fishermen and decision-makers have been observed in northern Galicia and the Gulf of Cádiz.

Generating Role-Playing Game Quests With GPT Language Models

Quests represent an integral part of role-playing games (RPGs). While evocative, narrative-rich quests are still mostly hand-authored, player demands towards more and richer game content, as well as business requirements for continuous player engagement necessitate alternative, procedural quest generation methods. While existing methods produce mostly uninteresting, mechanical quest descriptions, recent advances in AI have brought forth generative language models with promising computational storytelling capabilities. We leverage two of the most successful Transformer models, GPT-2 and GPT-3, to procedurally generate RPG video game quest descriptions. We gathered, processed and openly published a data set of 978 quests and their descriptions from six RPGs. We fine-tuned GPT-2 on this data set with a range of optimizations informed by several mini studies. We validated the resulting Quest-GPT-2 model via an online user study involving 349 RPG players. Our results indicate that one in five quest descriptions would be deemed acceptable by a human critic, yet the variation in quality across individual quests is large. We provide recommendations on current applications of Quest-GPT-2. This is complemented by case-studies on GPT-3 to highlight the future potential of state-of-the-art natural language models for quest generation.

A decision model based on gene expression programming for discretionary lane-changing move

ABSTRACT This study focuses on modeling Discretionary Lane-Changing (DLC), which accounts for the majority of lane-change moves in traffic flows. A binary decision model for lane-changing moves was improved with the method of Gene Expression Programming (GEP). The decision to prefer GEP is due to its high performance in a variety of engineering solutions in the literature. The GEP model was trained with Next Generation SIMulation (NGSIM) trajectory data gathered at the I-80 Freeway in Emeryville, California, and then tested with data gathered at the U.S. Highway 101 in LA, California. The test results indicate that the model made decisions of “change lane” with 92.98% accuracy, and “do not change lane” with 99.65% accuracy. A sensitivity analysis was also conducted to discover potential limits of the performance of the GEP model. The performance of this model was compared with other high-performance decision models developed with the NGSIM's DLC data in the literature and with TransModeler's gap acceptance model. This comparison indicates that the GEP model is the most successful decision model for discretionary lane-changing moves. The GEP model has a high potential to be applied in DLC decision support systems in (semi-) automated vehicles, as well as traffic simulation software.

Learning Disentangled Graph Convolutional Networks Locally and Globally.

Graph convolutional networks (GCNs) emerge as the most successful learning models for graph-structured data. Despite their success, existing GCNs usually ignore the entangled latent factors typically arising in real-world graphs, which results in nonexplainable node representations. Even worse, while the emphasis has been placed on local graph information, the global knowledge of the entire graph is lost to a certain extent. In this work, to address these issues, we propose a novel framework for GCNs, termed LGD-GCN, taking advantage of both local and global information for disentangling node representations in the latent space. Specifically, we propose to represent a disentangled latent continuous space with a statistical mixture model, by leveraging neighborhood routing mechanism locally. From the latent space, various new graphs can then be disentangled and learned, to overall reflect the hidden structures with respect to different factors. On the one hand, a novel regularizer is designed to encourage interfactor diversity for model expressivity in the latent space. On the other hand, the factor-specific information is encoded globally via employing a message passing along these new graphs, in order to strengthen intrafactor consistency. Extensive evaluations on both synthetic and five benchmark datasets show that LGD-GCN brings significant performance gains over the recent competitive models in both disentangling and node classification. Particularly, LGD-GCN is able to outperform averagely the disentangled state-of-the-arts by 7.4% on social network datasets.