High-level Structure Research Articles

Phylogenetic, morphological and niche differentiation unveil new species limits for the big brown bat (Eptesicus fuscus).

Phylogeographic accounts of mammals across fragmented landscapes show high levels of genetic, morphological and ecological variation. The big brown bat (Eptesicus fuscus) widely spans mainland landmasses from Canada to Ecuador and Colombia, and the insular Caribbean through The Bahamas and Greater and Lesser Antilles. Given the distribution of E. fuscus, we hypothesized that insular lineages could represent a different species aided by isolation. We assessed species limits by capitalizing on available mitochondrial and genomic data. Novel morphological and spatial datasets were produced to examine limits phenotypically and whether ecological niches could be associated with differences between groups. Phylogenetics strongly supported the Caribbean as unique compared to the mainland. Genomic data indicated high levels of genetic structure within the Caribbean and no detectable admixture of the Caribbean with continental lineages. Similarly, the Caribbean group shows high phenotypic disparity, and niche models revealed differences in habitat suitability between groups, concordant with the phylogenetic results. This study uncovered signals of divergence supporting the Caribbean clade of E. fuscus as unique through an integrative framework. We endorse re-evaluating the taxonomic status of Caribbean big brown bats as Eptesicus dutertreus. This recognition can help promote local conservation plans for insular lineages of big brown bats.

Assessing the role of mentors in mitigating burnout and enhancing professional development in medical education.

To assess the correlation between mentor behaviours and medical student burnout and their professional development within medical education. A cross-sectional study using convenience sampling was conducted among third-, fifth-, and sixth-year medical students (N=307). Participants voluntarily completed anonymous online questionnaires measuring the Mentor Behavior Scale, the Maslach Burnout Inventory-Student Survey, and the Professional Self-Identity Questionnaire. Multivariate regression analysis was performed to analyse relationships between student burnout, mentor behaviours and their impact on professional development. Among participants, 26% (N=80) experienced burnout, which was significantly associated with lower competency support (OR = 2.0, 95% CI: 1.1-3.5, p = 0.016), medication use (OR = 2.1, 95% CI: 1.1-4.0, p = 0.029), and a lower Grade Point Average (OR = 3.3, 95% CI: 1.6-6.9, p = 0.001) compared to non-burnout students. In the development of professional identity, a high level of mentor relationship structure had statistically significant associations with higher scores in key domains of the Professional Self-Identity Questionnaire, including teamwork (OR = 3.9, 95% CI: 1.5-9.9, p < 0.01), communication (OR = 3.4, 95% CI: 1.5-7.7, p < 0.01), ethical awareness (OR = 3.3, 95% CI: 1.4-8.0, p < 0.01), and record use (OR = 2.8, 95% CI: 1.2-6.5, p < 0.05). The impact of mentor behaviours on medical students is evident. Enhancing mentorship by addressing specific mentor behaviours can improve programme quality. Future research should explore the long-term effects and strategies for effectively implementing targeted enhancements in mentor behaviours.

Dysregulated Long Non-coding RNAs in Myasthenia Gravis- A Mini-Review.

Myasthenia gravis (MG) is an acquired autoimmune disease that is mediated by humoral immunity, supplemented by cellular immunity, along with participation of the complement system. The pathogenesis of MG is complex; although autoimmune dysfunction is clearly implicated, the specific mechanism remains unclear. Long non-coding RNAs (lncRNAs) are a class of non-coding RNA molecules with lengths greater than 200 nucleotides, with increasing evidence of their rich biological functions and high-level structure conservation. LncRNAs can directly interact with proteins and microRNAs to regulate the expression of target genes at the transcription and post-transcription levels. In recent years, emerging studies have suggested that lncRNAs play roles in the differentiation of immune cells, secretion of immune factors, and complement production in the human body. This suggests the involvement of lncRNAs in the occurrence and progression of MG through various mechanisms. In addition, the differentially expressed lncRNAs in peripheral biofluid may be used as a biomarker to diagnose MG and evaluate its prognosis. Moreover, with the development of lncRNA expression regulation technology, it is possible to regulate the differentiation of immune cells and influence the immune response by regulating the expression of lncRNAs, which will provide a potential therapeutic option for MG. Here, we review the research progress on the role of lncRNAs in different pathophysiological events contributing to MG, focusing on specific lncRNAs that may largely contribute to the pathophysiology of MG, which could be used as potential diagnostic biomarkers and therapeutic targets.

GENERATING LIGHTWEIGHT BUILDING MODELS WITH PRESERVED STRUCTURAL FEATURES FROM NOISY 3D MESHES

Abstract. Obtaining building instance models directly through photogrammetry and other means results in high polygon counts and structural detail levels. Therefore, it is an important challenge to preserve the features of instant building models and generate lightweight building models from them. In this paper, we propose an improved lightweight reconstruction method for 3D building models based on planar primitives extraction, topology correction, and optimal planes selection. Improvements due to our method arise from three aspects: (1) After plane segmentation based on a simple region growing method, a second plane segmentation is performed on the building model based on the similarity of normal vectors. (2) According to the building structural characteristics, the initial plane segmentation is checked to generate new plane regions within the necessary connection regions. (3) The intersection of candidate planes is improved by enlarging the region of candidate faces, and the topological connection is improved. Furthermore, the topological relations of all planar primitives that have been optimized are recorded in an undirected graph. Finally, a watertight and manifold lightweight model is extracted from the faces of a candidate set by energy minimization. Experiments on different data sets show that the improved method is more reasonable in plane segmentation, and has superior performance in algorithm robustness and necessary structure recovery of buildings. Even when dealing with imperfect data, a watertight model is still obtained.

Amplicon deep sequencing of five highlypolymorphic markers of Plasmodium falciparum reveals high parasite genetic diversity and moderate populationstructure in Ethiopia

BackgroundPlasmodium falciparum genetic diversity can add information on transmission intensity and can be used to track control and elimination interventions.MethodsDried blood spots (DBS) were collected from patients who were recruited for a P. falciparum malaria therapeutic efficacy trial in three malaria endemic sites in Ethiopia from October to December 2015, and November to December 2019. qPCR-confirmed infections were subject to amplicon sequencing of polymorphic markers ama1-D3, csp, cpp, cpmp, msp7. Genetic diversity, the proportion of multiclonal infections, multiplicity of infection, and population structure were analysed.ResultsAmong 198 samples selected for sequencing, data was obtained for 181 samples. Mean MOI was 1.38 (95% CI 1.24–1.53) and 17% (31/181) of infections were polyclonal. Mean He across all markers was 0.730. Population structure was moderate; populations from Metema and Metehara 2015 were very similar to each other, but distinct from Wondogent 2015 and Metehara 2019.ConclusionThe high level of parasite genetic diversity and moderate population structure in this study suggests frequent gene flow of parasites among sites. The results obtained can be used as a baseline for additional parasite genetic diversity and structure studies, aiding in the formulation of appropriate control strategies in Ethiopia.

A Methodological Framework for Assessing the Influence of Process Parameters on Strand Stability and Functional Performance in Fused Filament Fabrication.

With an ever-increasing material and design space available for Fused Filament Fabrication (FFF) technology, fabrication of complex three-dimensional structures with functional performance offers unique opportunities for product customization and performance-driven design. However, ensuring the quality and functionality of FFF-printed parts remains a significant challenge, as material-, process-, and system-level factors introduce variability and potentially hinder the translation of bulk material properties in the respective FFF counterparts. To this end, the present study presents a methodological framework for assessing the influence of process parameters on FFF strand stability and functional performance through a systematic analysis of FFF structural elements (1D stacks of FFF strands and 3D blocks), in terms of dimensional deviation from nominal geometry and resistivity, corresponding to the printability and functionality attributes, respectively. The influence of printing parameters on strand stability was investigated in terms of dimensional accuracy and surface morphology, employing optical microscopy and micro-computed tomography (mCT) for dimensional deviation analysis. In parallel, electrical resistance measurements were carried out to assess the effect of different process parameter combinations and toolpath patterns on functional performance. In low-level structural elements, strand height (H) was found to induce the greatest influence on FFF strand dimensional accuracy and resistivity, with higher H values leading to a reduction in resistivity of up to 38% in comparison with filament feedstock; however, this occurred at the cost of increased dimensional deviation. At higher structural levels, the overall effect of process parameters was found to be less pronounced, indicating that the translation of 1D strand properties to 3D blocks is subject to a trade-off due to competing mechanisms that facilitate/hinder current flow. Overall, the proposed framework enables the quantification of the influence of process parameters on the selected response variables, contributing to the development of standard operating procedures and recommendations for selecting optimal process parameters to achieve the desired process stability and functional performance in FFF.

Editing Motion Graphics Video via Motion Vectorization and Transformation

Motion graphics videos are widely used in Web design, digital advertising, animated logos and film title sequences, to capture a viewer's attention. But editing such video is challenging because the video provides a low-level sequence of pixels and frames rather than higher-level structure such as the objects in the video with their corresponding motions and occlusions. We present a motion vectorization pipeline for converting motion graphics video into an SVG motion program that provides such structure. The resulting SVG program can be rendered using any SVG renderer (e.g. most Web browsers) and edited using any SVG editor. We also introduce a program transformation API that facilitates editing of a SVG motion program to create variations that adjust the timing, motions and/or appearances of objects. We show how the API can be used to create a variety of effects including retiming object motion to match a music beat, adding motion textures to objects, and collision preserving appearance changes.

Speed control with low complexity for multiple autonomous vehicles in roundabouts

The paper introduces a high level speed control method for the coordination of multiple autonomous vehicles (AVs) in roundabout scenarios. The aim of the control method is to guarantee collision-free motion of the AVs, and similarly, to minimize their traveling time. In the method a priority-based ordering process of the AVs is used, which enforces the time-efficient motion of the AVs. The collision-free motion is guaranteed through an optimization-based method including control input constraints. The ordering process and the optimization form a low complexity solution, which requires low computation effort. The proposed control strategy is involved in the high level of a hierarchical control structure. The effectiveness of the proposed control strategy is illustrated by simulation examples and Hardware-in-the-Loop demonstration.

More pieces for the puzzle: novel information on the genetic diversity and population structure of Steno bredanensis (Artiodactyla: Delphinidae) in Central America and the Caribbean Sea

Introduction: The rough-toothed dolphin (Steno bredanensis) inhabits oceanic waters of tropical latitudes and exhibits philopatry in some oceanic islands. However, the species has been observed in shallow coastal waters in a few areas. Particularly in Central America, the rough-toothed dolphin has been reported by occasional records and strandings. For instance, the first confirmed record of this species in the Panama’s Caribbean was on July 17, 2012, in a coastal region of the Chiriquí Lagoon during a bottlenose dolphin monitoring survey. Similarly, the first rough-toothed dolphin mass stranding reported for the Pacific of Panama was on April 20, 2016, at the Ostional Beach, where 60 dolphins stranded and ten died. These sightings and events offered a valuable opportunity to obtain samples to conduct genetic studies, which are scarce in the region.&#x0D; Objective: In this study, we present the first assessment of genetic diversity for rough-toothed dolphins based on mitochondrial DNA Control Region (mtDNA-CR) in the Panamanian Pacific and the Wider Caribbean.&#x0D; Methods: Samples were collected in Colombia (N=5), Panama (N-Caribbean=1, N-Pacific=9), and Puerto Rico (N=3) from free-ranging and stranded individuals. DNA was extracted from each sample, and a mtDNA segment of around 534 to 748 bp was amplified through the PCR reaction. The obtained sequences were compared with rough-toothed dolphin haplotypes previously published in NCBI (N=70), from the Atlantic, Indian, and the Pacific Oceans.&#x0D; Results: Our findings showed significant population structure among ocean basins (strong differentiation with ΦST data), and high genetic diversity within each phylogroup. Only the Atlantic Ocean showed high genetic differentiation within the basin, detecting three phylogroups: the Caribbean, northern, and southern Atlantic.&#x0D; Conclusions: These findings support previous genetic studies that indicate high levels of population structure among ocean basins, although this species seems to be widely dispersed. However, samples from Panama and the Caribbean appear to show connectivity between highly differentiated Atlantic and Pacific Oceans. Therefore, our results highlight the need for more research to assess the rough-toothed dolphin genetic and population status in Central America, as the piece of the puzzle needed to clarify its taxonomy and genetic differentiation worldwide. This information is needed due to the rough-toothed dolphin IUCN categorization as “Least Concern” and its classification into appendix II according to CITES. While individuals are potentially threatened by incidental fishing, no management units are currently used to conserve this species despite its high genetic differentiation.

Phylogeography, genetic diversity, and intraspecific genetic structure of the black-horned capuchin (Sapajus nigritus).

Recent molecular studies have clarified the overarching taxonomy of capuchin monkeys, but intraspecific genetic diversity remains unexplored for most capuchin species. One example is Sapajus nigritus, the southernmost capuchin monkey, found in Brazil and Argentina; its phenotypic diversity has been recognized as two geographic subspecies, but the intraspecific genetic structure of this taxon is poorly known. Here, we sampled across most of this species' geographic distribution, producing a newly sequenced data set for genetic analyses that included 78 individuals from 14 populations. We investigated the intraspecific diversity, genetic structure, and evolutionary history using three mitochondrial markers. Our results indicated that S. nigritus populations exhibited high levels of genetic structure. We found strong support for two monophyletic clades within this species with a deep phylogenetic split, and clear separation from other related taxa. Vicariance events seem to have played a prevalent role in shaping S. nigritus genetic differentiation. The Paraíba do Sul River may have driven the deep divergence between southern and northern clades, whereas the Tietê River may have had a weaker, more recent effect on the divergence of populations within the southern clade.

Hierarchical WaveFunction Collapse

Video game developers are increasingly utilising procedural content generation (PCG) techniques in order to generate more content far quicker than if it were designed. Although promising, much of the successful work to date has been achieved in simple 2D environments or has required significant hand-designed effort. This is due to the difficult nature of defining plausible metrics, fitness functions or reward functions which can quantify the quality of generated levels. Our work aims to avoid this difficulty by utilising minimal human design to build up constraints, and generating diverse levels that maintain these constraints. We achieve this by hierarchically applying the recent WaveFunction collapse (WFC) algorithm. Our approach allows designers to specify larger-scale components, and additional constraints that are difficult to enforce using standard WFC. We empirically demonstrate that our approach does indeed incorporate these higher-level structures, and is more controllable than our baselines. Despite these benefits, our levels do not suffer from a lack of diversity. Finally, we illustrate the scalability and flexibility of our approach by applying it to both 2D and 3D domains.

Common but not connected: high genetic structure and cryptic genetic diversity in the ubiquitous biting midge Culicoides peregrinus Kieffer.

The biting midge Culicoides peregrinus Kieffer is a significant pest and vector species, and knowledge of its genetic diversity and genetic structure is critically important for designing an effective control program. However, such information is limited to only small sample-size DNA barcoding studies. Therefore, in this study, we used mitochondrial cytochrome c oxidase I (COI) to examine genetic structure and diversity of C. peregrinus from northeastern Thailand. In addition, we also inferred genetic relationships between C. peregrinus from Thailand and those reported from other countries across the geographic range of the species. Maximum intraspecific genetic divergence (3.83%) within Thai specimens was relatively high compared to other Culicoides species. Genetic structure analysis revealed that 71% (32 from 45) of population comparisons were highly significantly different. A high level of genetic structure among populations, even between those in close geographic proximity (22 km geographic distance) suggested that there has been little or no movement between local populations. This is possibly due to the ability to exploit diverse types of breeding site and a generalist feeding habit which enables C. peregrinus to complete its life cycle within cattle pens. Genetic relationships between Thai C. peregrinus and those reported from other countries revealed three genetically divergent lineages (A, B and C) associated with geographic origins. Specimens from Thailand + China formed lineage A, those from Australia formed lineage B and India + Bangladesh belonged to lineage C. These genetically divergent lineages also agree with morphological variation of the wing pale marking spots. Further investigation using independent genetic loci from nuclear genes will be very useful to resolve taxonomic status of these divergent lineages.

Hand-blink reflex modulation: The role of primary emotions and attachment dimensions.

The hand-blink reflex (HBR) is a subcortical response elicited by the electrical stimulation of the median nerve. HBR magnitude is enhanced when the stimulated hand is close to the face and is modulated by high-level structures according to the perceived threat magnitude. Psychological factors may contribute to threat evaluation and possibly to HBR amplitude modulation. In this study, we assessed distinctively emotional and relational aspects of personality and evaluated their associations with the HBR response, or lack thereof, in healthy subjects. Seventy-one volunteers filled the Experiences in Close Relationships Scale, the Affective Neuroscience Personality Scales, and the State-Trait Anxiety Inventory Form Y questionnaires and underwent HBR recording. We found that the HBR could be evoked only in 50.7% of all subjects (responders). Non-responders subjects showed higher scores in the avoidance dimension (p = .005), and lower scores in the care dimension (p = .008), compared with responders. In responders, regression analysis showed a negative association of HBR amplitude (difference in near vs. far responses) with anger dimension and a positive association with state anxiety (R2 = 0.239). A positive association also emerged with HBR latency and fear dimension (R2 = 0.419). We conclude that primary emotional and relational factors may play an important role in the modulation of brainstem circuits mediating the HBR response. Our results may also contribute to the question about the absence of the HBR in about half of the subjects since high-level cognitive processes seem to play an important role in the differentiation between responder and non-responder.

Sowing Seeds of Sovereignty: A Qualitative Exploration of Chapter Leader Experiences Who Belong to an African, Caribbean, and Black Women's Agriculture Advocacy Collective.

African, Caribbean, and Black (ACB) populations are disproportionately impacted by food insecurity since the COVID-19 pandemic, with reports including limited access to healthy, affordable, and available food resulting in the exacerbation of existing health disparities and poorer living conditions and environments. Many interventions addressing food insecurity among ACB populations incorporate downstream models that emphasize consumer-level solutions (e.g., food donation or assistance programs). While useful for developing short-term solutions, these approaches do not meaningfully engage broader structures and root-causes-and, in many ways, reify existing inequities among these populations. Missing from mainstream dialogue is how food insecurity inequities at the individual level are often generated at higher structural levels and the unique assets of a targeted community or population to grow/distribute their own food and restore frayed transnational cultural foodways. This paper documents a case study of an international, Black women-led agriculture education collective headquartered in the South-Central region of the United States. Drawing on in-depth interviews with affiliate chapter leaders in the United States, US Virgin Islands, and Africa (n = 22), the research investigates how chapter leaders describe unique opportunities and barriers to farming and developing food insecurity solutions. Emergent themes from the data analysis were: (1) diasporically linked barriers, (a) infrastructural/political barriers and (b) patriarchal barriers; (2) mobilizing strategies; and (3) cultural and context-specific solutions. Findings from this study contribute important insights for food security development research centering community-based and grassroots organizing models addressing inequities in food production and health.

Phases of Enhanced Exhumation During the Cretaceous and Cenozoic Orogenies in the Eastern European Alps: New Insights From Thermochronological Data and Thermokinematic Modeling

AbstractAustroalpine nappes in the Eastern European Alps have preserved the record of orogenies in the Cretaceous and Cenozoic but their cooling and exhumation history remains poorly constrained. Here we use low‐temperature thermochronology and thermokinematic modeling to unravel the exhumation history of the Austroalpine nappes in the Gurktal Alps. Our data reveal marked differences between the exhumation of units located at different positions within the nappe stack and relative to the Adriatic indenter. Units located at a high structural level and farther away from the indenter cooled through the zircon fission track closure temperature in the Late Cretaceous and have resided at depths of ≤5–6 km since the Oligocene, as indicated by apatite fission track ages of 35–30 Ma. Thermokinematic modeling constrained that these units experienced enhanced exhumation (∼0.60 km/Ma) between ∼99 and ∼83 Ma due to syn‐ to late‐orogenic Late Cretaceous extension. After a phase of slow exhumation (∼0.02 km/Ma), the exhumation rate increased to ∼0.16 km/Ma at ∼34 Ma due to the onset of the Europe‐Adria collision. In contrast, zircon fission track ages from units at a lower structural level and near the indenter indicate cooling during the Eocene; apatite fission track ages cluster at ∼15 Ma. These units were rapidly exhumed (∼0.76 km/Ma) from ∼44 to ∼39 Ma during an Eocene phase of shortening prior to the Europe‐Adria collision. After slow exhumation (∼0.13 km/Ma) between ∼39 and ∼18 Ma, the exhumation rate increased to ∼0.27 km/Ma in the wake of Miocene escape tectonics in the Eastern Alps.

The influence of classroom context on AAC device use for nonspeaking school-aged autistic children

Purpose: Language learning, including on a speech-generating device (SGD), relies on having communication opportunities to use language. Yet, children who use SGDs do not consistently use their devices throughout the day. As a first step towards increasing device use, it is important to determine how the numerous contexts (e.g. recess, lunch, academic) that exist throughout the school day impact the frequency and type of communication opportunities that children have. Within this study, we used complex adaptive systems theory to investigate whether there were differences in how frequently nonspeaking autistic children classified as emerging communicators (i.e. children who do not consistently produce two-word utterances independently and do not communicate for a variety of communicative functions) used their SGDs to communicate and the type of communication they produced (i.e. spontaneous, prompted, or imitation) across different classroom contexts. Method: Fourteen autistic children who use an SGD as their primary mode of communication were videotaped up to nine times throughout different school days. The videos were coded for the type of device use (i.e. whether the child’s use of the device was spontaneous, prompted, or imitation) and classroom context, which were categorised by the inherent level of support and directiveness. Result: There was a significant difference in SGD use across classroom contexts. Overall, the children communicated more and engaged in more spontaneous, prompted, and imitated communication in classroom contexts with high levels of structure (e.g. tabletop work) compared to contexts with low levels of structure and directiveness (e.g. free play). Conclusion: This study highlights the need to increase communication opportunities across all school contexts. Creating appropriate communication opportunities in all contexts, especially those with less structure, will help ensure that communication does not become context-bound.

Characterization of unsaturated lipids with high structural specificity by coupling epoxidation with tandem mass spectrometry

Comprehensive characterization of lipid structures is of significance, as they play a key role in numerous biological processes. Mass spectrometry (MS) emerges as a preferred tool for lipid analysis, providing good performance in sensitivity and accuracy. However, characterization of unsaturated lipids still has certain difficulties due to structural diversity. In this work, combining epoxidation and tandem MS was developed for unsaturated lipids analysis with high structural specificity. By using the proposed method, it could simultaneously produce high-abundance diagnostic ions specific to lipid C=C locations and sn-positions, facilitating lipid isomers identification. Determining C=C location and sn-position isomers of lipid in a single spectrum can be achieved, which is valuable for discovery of great structural diversity. The capability of this approach in characterizing unsaturated lipids at high structural level has been illustrated with different standards including a lipid extract from complex biological sample. The relative quantitation capability of C=C location and sn-position isomers was also demonstrated based on diagnostic ion intensities.

The impact of natural resources on technology innovation from cross-country panel data: A comparative analysis and policy-level insights

The study aims to examine the direct impact of natural resources on technology innovation by considering the critical role of five dimensions of technology innovation (ICT development, R&D activity, skills, industry activity and access to finance). For this purpose, the common correlated effect mean group (CCEMG) estimator is applied to estimate cross-country panel data encompassing 102 countries from 2008 to 2018. Under the global sample, the empirical evidence strongly supports that natural resources hinder ICT development and R&D activity but have a positive impact on industry activity and access to finance, which shows a traceable path of the resource blessing. In addition, for overall technology innovation, ICT development, R&D and industry activity, the resource curse is effectively transformed into a resource blessing in high-income countries, which benefits from the high level of manufacturing, modern services and industrial structure. When considering the chronological evolution, natural resources show a negative effect on overall technology innovation in the first period and a positive effect in the second period in low-income countries. On this basis, challenges and initiatives in three phases are proposed to better achieve resource blessing, including resource regulation, human resource reserve, international trade adjustment, industrial structure transformation and financial investment optimization. Lastly, policy-level insights on the different dimensions of technology innovation are proposed to provide guidance and valuable reference for policymakers.

How can Chinese cities escape from carbon lock-in? The role of low-carbon city policy

In this study, we critically analyze the effect of low-carbon city policy on carbon lock-in for Chinese cities, based on a balanced panel dataset of 283 Chinese prefectural cities between 2003 and 2017, as well as the difference-in-differences method. Several key findings are highlighted: (1) Low-carbon city policy exerts negative impact on carbon lock-in, which means that the policy successfully inhibits carbon lock-in. (2) The dynamic effect of low-carbon city policy is constant and becomes more prominent over time. (3) The impact of low-carbon city policy on carbon lock-in is heterogeneous and asymmetric. This policy is more effective in cities in the eastern regions, in cities where there is strong government administrative power, and where there is a high level of industry structure upgrading as well as technology innovation. (4) Low-carbon city policy indirectly influences carbon lock-in through environmental regulation, informatization, and human capital effects, which are three effective mediators. Several specific policy implications are put forward according to the above conclusions to overcome the reinforced carbon lock-in situation.

Curriculum learning based overcomplete U-Net for liver tumor segmentation from computed tomography images

In this paper, we have proposed an overcomplete U-Net to perform liver tumor segmentation jointly using a curriculum learning strategy. Liver tumor segmentation is the most prominent and primary step in treating liver cancer and can also help doctors with proper diagnosis and therapy planning. However, it is challenging because of variations in shape, position, and depth of tumors and adjacent boundaries with internal organs around the liver. We have presented a promising solution by designing a U-Net-based segmentation network with two branches: an overcomplete branch to fine grade the small structures and an undercomplete branch to fine grade the high-level structures. This combination allows the network to learn all types of tumor artifacts more accurately. We also changed the conventional learning paradigm to curriculum learning where the input images are fed to the network from easy to hard ones to achieve faster convergence. Finally, our network segments the tumors directly from the whole medical images without the need for segmented liver region of interests (ROIs). The proposed network achieved a DICE score of 75% in tumor segmentation which is a decent value when compared with some existing deep learning methods for liver tumor segmentation.