Prior Adaptation Research Articles

A Small-Scale Object Detection Algorithm in Intelligent Transportation Scenarios

In response to the problem of poor detection ability of object detection models for small-scale targets in intelligent transportation scenarios, a fusion method is proposed to enhance the features of small-scale targets, starting from feature utilization and fusion methods. The algorithm is based on the YOLOv4 tiny framework and enhances the utilization of shallow and mid-level features on the basis of Feature Pyramid Network (FPN), improving the detection accuracy of small and medium-sized targets. In view of the problem that the background of the intelligent traffic scene image is cluttered, and there is more redundant information, the Convolutional Block Attention Module (CBAM) is used to improve the attention of the model to the traffic target. To address the problem of data imbalance and prior bounding box adaptation in custom traffic data sets that expand traffic images in COCO and VOC, we propose a Copy-Paste method with an improved generation method and a K-means algorithm with improved distance measurement to enhance the model’s detection ability for corresponding categories. Comparative experiments were conducted on a customized 260-thousand traffic data set containing public traffic images, and the results showed that compared to YOLOv4 tiny, the proposed algorithm improved mAP by 4.9% while still ensuring the real-time performance of the model.

A human isolate of bovine H5N1 is transmissible and lethal in animal models.

The outbreak of clade 2.3.4.4b highly pathogenic avian influenza viruses of the H5N1 subtype (HPAI H5N1) in dairy cows in the US has so far resulted in spillover infections of at least thirteen farm workers1-3, who presented with mild respiratory symptoms or conjunctivitis, and one individual with no known animal exposure who was hospitalized but recovered3,4. Here, we characterized A/Texas/37/2024 (huTX37-H5N1), a virus isolated from the eyes of an infected farm worker who developed conjunctivitis5. huTX37-H5N1 replicated efficiently in primary human alveolar epithelial cells, but less efficiently in corneal epithelial cells. Despite causing mild disease in the infected worker, huTX37-H5N1 was lethal in mice and ferrets and spread systemically with high titres in respiratory and non-respiratory organs. Importantly, in four independent experiments in ferrets, huTX37-H5N1 transmitted via respiratory droplets in 17%-33% of transmission pairs and five of six exposed ferrets that became infected died. PB2-631L (encoded by bovine isolates), promoted influenza polymerase activity in human cells, suggesting a role in mammalian adaptation like that of PB2-627K (encoded by huTX37-H5N1). Additionally, bovine HPAI H5N1 viruses were found to be susceptible to polymerase inhibitors both in vitro and in mice. Thus, HPAI H5N1 virus derived from dairy cattle transmits by respiratory droplets in mammals without prior adaptation and causes lethal disease in animal models.

Pre-and-post impact fall detection based on support vector machines using inertial and barometric pressure data

Objective: This paper presents a novel real-time algorithm for fall detection, which contextualizes falls by identifying activities occurring both pre- and post-impact utilizing machine learning techniques and wearable sensors. Methodology: The activities selected to contextualize fall events included standing, lying, walking, running, climbing stairs, and using the elevator. Data were collected using an inertial measurement unit and a barometric altimeter positioned on the participants’ lower backs. Thirteen healthy subjects were observed performing the activities and fall events were recorded from five healthy subjects. The proposed algorithm combines thresholding and cascade support vector machines (SVMs), whose robustness is enhanced by a verification process of the subject’s posture aimed at determining the occurrence of the fall more accurately. Results: The performance of the algorithm was evaluated in terms of the hit rate (HT) both offline and in real-time. From the activities studied, stairs climbing proved to be the most challenging to detect, with an offline HT of 85% and an online HT of 76 %. The overall offline performance was superior, with an HT of 96 %, compared to the performance achieved online, an HT of 91 %; in both cases the fall detection HT was 100 %. Conclusions: The algorithm can be used to recognize fall events occurring to any user, as it has the advantage of not needing prior adaptation due to the nonlinear nature of the SVMs. The cascade SVMs allow for using small sets of variables, leading to low computational cost and a suitable real-time implementation. These features, in addition to the posture verification process, make our algorithm suitable for activity recognition in non-laboratory environments.

Epidemiology and biological characteristics of influenza A (H4N6) viruses from wild birds

ABSTRACT During the active surveillance, we isolated nine H4N6 subtype influenza A viruses from wild birds in China. To reveal the epidemiology and biology characteristics of H4 subtype influenza A virus from wild birds, we investigated H4 subtype viruses available in the public source, and found that the H4 viruses have been detected in at least 37 countries to date, and more than 73.6% of the viruses were from wild Anseriformes. Bayesian phylogeographic analysis showed that Mongolia worked as the important transmission centre for Eurasian lineage H4 viruses spreading. Phylogenetic analysis of HA genes indicated that global H4 influenza A viruses were divided into Eurasian and North American lineage, our nine H4N6 isolates fell into the Eurasian lineage. Recombination analysis suggested that nine H4N6 isolates underwent complex gene recombination with various subtypes of influenza A viruses and formed two genotypes. Notably, nine H4N6 isolates acquired mammalian virulence-increasing residues. Two representative H4N6 viruses possessed dual receptor binding specificity, they could efficiently replicate in MDCK and 293 T cells in vitro infection, also could cross the species barrier to infect mice directly without prior adaption in vivo experiments. These findings emphasize the public health issues represented by H4 viruses, and highlight the need to strengthen the active surveillance of H4 viruses from wild birds.

Sirtuin 3 drives sex-specific responses to age-related changes in mouse embryonic fibroblasts

The aging process is a complex phenomenon characterised by a gradual decline in physiological functions and an increased susceptibility to age-related diseases. An important factor in aging is mitochondrial dysfunction, which leads to an accumulation of cellular damage over time. Mitochondrial Sirtuin 3 (Sirt3), an important regulator of energy metabolism, plays a central role in maintaining mitochondrial function. Loss of Sirt3 can lead to reduced energy levels and an impaired ability to repair cellular damage, a hallmark of the aging process. In this study we investigated the impact of Sirt3 loss on mitochondrial function, metabolic responses and cellular aging processes in male and female mouse embryonic fibroblasts (MEF) exposed to etoposide-induced DNA damage, which is commonly associated with cellular dysfunction and senescence. We found that Sirt3 contributes to the sex-specific metabolic response to etoposide treatment. While male MEF exhibited minimal damage suggesting potential prior adaptation to stress due to Sirt3 loss, female MEF lacking Sirt3 experienced higher vulnerability to genotoxic stress, implying a pivotal role of Sirt3 in their resistance to such challenges. These findings offer potential insights into therapeutic strategies targeting Sirt3- and sex-specific signalling pathways in diseases associated with DNA damage that play a critical role in the aging process.

Kernels of Motor Memory Formation: Temporal Generalization in Bimanual Adaptation.

In daily life, we coordinate both simultaneous and sequential bimanual movements to manipulate objects. Our ability to rapidly account for different object dynamics suggests there are neural mechanisms to quickly deal with them. Here we investigate how actions of one arm can serve as a contextual cue for the other arm, and facilitate adaptation. Specifically, we examine the temporal characteristics that underlie motor memory formation and recall, by testing the contextual effects of prior, simultaneous, and post contralateral arm movements in both male and female human participants. To do so, we measure their temporal generalization in three bimanual interference tasks. Importantly, the timing context of the learned action plays a pivotal role in the temporal generalization. While motor memories trained with post adaptation contextual movements generalize broadly, motor memories trained with prior contextual movements exhibit limited generalization, and motor memories trained with simultaneous contextual movements do not generalize to prior or post contextual timings. This highlights temporal tuning in sensorimotor plasticity: different training conditions yield substantially different temporal generalization characteristics. Since these generalizations extend far beyond any variability in training times, we suggest that the observed differences may stem from inherent differences in the use of prior, current and post-adaptation contextual information in the generation of natural behavior. This would imply differences in the underlying neural circuitry involved in learning and executing the corresponding coordinated bimanual movements.Significance Statement This study addresses a fundamental question in the field of sensorimotor neuroscience of how multiple movements are temporally linked within a single motor memory. We examine the temporal generalization of motor memory formation by varying the timing of contextual movements associated with a learned motor memory across a range of prior, current, and post adaptation movement times using a bimanual motor learning task. We observed distinct patterns of temporal generalization based on whether the contextual movements occurred prior to, simultaneously with, or after the adaptation movement. For the first time, our findings reveal that the timing of contextual movements is crucial in the formation and generalization of motor memories.

Prior Knowledge, Acceptance, Adaptation, and Challenges Following Stoma Formation among Colorectal Cancer Patients in Northern Peninsular of Malaysia: A Qualitative Study

INTRODUCTION: Stoma formation affects an individual in various ways, including physical, emotional, social, and cognitive functions. Diverse studies report ways of an individual lives with new stoma formation. However, the comprehensive understanding of the entire process by the patient, which includes knowledge before the surgical procedure, as well as the subsequent acceptance, adaptation, and challenges to living with a stoma is lacking. MATERIALS AND METHODS: In-depth interview session were conducted with 12 colorectal cancer patients who have undergone surgical procedures for intestinal stoma formation. The patterns and themes within the data were identified by thematic analysis, involving data familiarisation and coding followed by themes’ generation and refinement of the themes. RESULTS: Four themes and 9 subthemes were identified, which revealed the sufficiency of stoma-related information and understanding prior to surgery as well as positive acceptance of self and family members reflected through their reactions and support. Nonetheless, the challenges were anticipated which highlights the complications of the stoma itself, obstacles surrounding social life, and financial burdens. CONCLUSION: This study provided valuable insights into the experiences of individuals living with a stoma following colorectal cancer surgery. The themes and subthemes highlight the need to address social stigma as well as financial issues to alleviate the burden of stoma-related expenses. Increasing public awareness and improving financial assistance could be a way to enhance the overall quality of life for patients living with stoma

Assessing the Impacts of Adaptation to Native-Range Habitats and Contemporary Founder Effects on Genetic Diversity in an Invasive Fish.

Species invading non-native habitats can cause irreversible environmental damage and economic harm. Yet, how introduced species become widespread invaders remains poorly understood. Adaptation within native-range habitats and rapid adaptation to new environments may both influence invasion success. Here, we examine these hypotheses using 7058 SNPs from 36 native, 40 introduced and 19 farmed populations of tench, a fish native to Eurasia. We examined genetic structure among these populations and accounted for long-term evolutionary history within the native range to assess whether introduced populations exhibited lower genetic diversity than native populations. Subsequent to infer genotype-environment correlations within native-range habitats, we assessed whether adaptation to native environments may have shaped the success of some introduced populations. At the broad scale, two glacial refugia contributed to the ancestry and genomic diversity of tench. However, native, introduced and farmed populations of admixed origin exhibited up to 10-fold more genetic diversity (i.e., observed heterozygosity, expected heterozygosity and allelic richness) compared to populations with predominantly single-source ancestry. The effects of introduction to a new location were also apparent as introduced populations exhibited fewer private alleles (mean = 9.9 and 18.9 private alleles in introduced and native populations, respectively) and higher population-specific Fst compared to native populations, highlighting their distinctiveness relative to the pool of allelic frequencies across tench populations. Finally, introduced populations with varying levels of genetic variation and similar genetic compositions have become established and persisted under strikingly different climatic and ecological conditions. Our results suggest that lack of prior adaptation and low genetic variation may not consistently hinder the success of introduced populations for species with a demonstrated ability to expand their native range.

Adaptation of Saccharomyces to High Glucose Concentrations and Its Impact on Growth Kinetics of Alcoholic Fermentations.

Prior adaptation of Saccharomyces cerevisiae to the fermentation medium ensures its implantation and success in alcoholic fermentations. Fermentation kinetics can be characterized with mathematical models to objectively measure the success of adaptation and growth. The study aims at assessing and comparing two pre-culture procedures using, respectively, one or two adaptation steps, analyzing the impact of different initial glucose concentrations on the fermentation profiles of S. cerevisiae cultures, and assessing the performance of three predictive growth models (Buchanan's, modified Gompertz, and Baranyi and Roberts models) under varied initial glucose concentrations. We concluded that both protocols produced S. cerevisiae pre-cultures with similar viability and biomass increase, which suggests that short protocols may be more cost-effective. Furthermore, the study highlights the need of inoculating a high S. cerevisiae population to minimize the depletion of dissolved oxygen in the medium and to ensure that glucose is predominantly directed toward the ethanol formation at early fermentative steps. This study shows that the relationship between kinetic parameters is model-dependent, which hinders inter-study comparisons and stresses the need for standardized growth models. We advocate for the generalized use of confidence intervals of the kinetic parameters to facilitate objective inter-study comparisons.

Cognitive interviewing to assess and adapt three measures of mental health symptoms among people living with HIV in Rakai, Uganda: the Thinking a Lot Questionnaire, the Patient Health Questionnaire 9 (PHQ-9), and the Hopkins Symptoms Checklist (HSCL).

Mental health is conceptualized differently across cultures, making cross-cultural validation of screening tools critical. In Uganda, we used cognitive interviewing to assess and adapt three scales for measuring psychological distress: the Thinking a Lot Questionnaire, the Patient Health Questionnaire 9 (PHQ-9), and the Hopkins Symptoms Checklist (HSCL). We recruited 12 people living with HIV from the Rakai Community Cohort Study (RCCS) and interviewed seven potential users of the scales (four RCCS survey interviewers and three local health workers). Data were analyzed systematically using a team-based matrix approach. The HSCL was generally well understood, with minor clarifications needed. The Thinking a Lot Questionnaire was also well understood, though differences between "how much" and "how often" required specificity. Both included local idioms of distress from prior adaptations. The PHQ-9 performed less well, with many questions interpreted variably or showing unclear local applicability, especially among people living with HIV. For example, questions about trouble concentrating were misunderstood, focusing on examples like newspapers rather than the broader issue of concentration. Future research should explore the validity and utility of commonly used instruments as mental health research expands in Africa.

Swine Influenza Viruses Isolated from 2019 to 2022 in Shandong Province, China, Exemplify the Dominant Genotype.

Swine influenza viruses (SIVs) have been circulating in swine globally and are potential threats to human health. During the surveillance of SIVs in Shandong Province, China, from 2019 to 2022, 21 reassortant G4 genotype Eurasian avian-like (EA) H1N1 subtypes containing genes from the EA H1N1 (HA and NA), 2009 pandemic (pdm/09) H1N1 virus (PB2, PB1, PA, NP, and M), and classical swine (CS) H1N1 (NS) lineages were isolated. The analysis of the key functional amino acid sites in the isolated viruses showed that two mutation sites (190D and 225E) that preferentially bind to the human α2-6 sialic acid receptor were found in HA. In PB2, three mutation sites (271A, 590S, and 591R) that may increase mammalian fitness and a mutation site (431M) that increases pathogenicity in mice were found. A typical human signature marker that may promote infection in humans, 357K, was found in NP. The viruses could replicate efficiently in mouse lungs and turbinates, and one of the H1N1 isolates could replicate in mouse kidneys and brains without prior adaption, which indicates that the viruses potentially pose a threat to human health. Histopathological results showed that the isolated viruses caused typical bronchopneumonia and encephalitis in mice. The results indicate that G4 genotype H1N1 has potential transmissibility to humans, and surveillance should be enhanced, which could provide important information for assessing the pandemic potential of the viruses.

An aggregated dataset of serial morbidity and titer measurements from influenza A virus-infected ferrets

Data from influenza A virus (IAV) infected ferrets provides invaluable information towards the study of novel and emerging viruses that pose a threat to human health. This gold standard model can recapitulate many clinical signs of infection present in IAV-infected humans, support virus replication of human, avian, swine, and other zoonotic strains without prior adaptation, and permit evaluation of virus transmissibility by multiple modes. While ferrets have been employed in risk assessment settings for >20 years, results from this work are typically reported in discrete stand-alone publications, making aggregation of raw data from this work over time nearly impossible. Here, we describe a dataset of 728 ferrets inoculated with 126 unique IAV, conducted by a single research group under a uniform experimental protocol. This collection of morbidity, mortality, and viral titer data represents the largest publicly available dataset to date of in vivo-generated IAV infection outcomes on a per-ferret level.

Evolution of H6N6 viruses in China between 2014 and 2019 involves multiple reassortment events

ABSTRACT H6N6 avian influenza viruses (AIVs) have been widely detected in wild birds, poultry, and even mammals. Recently, H6N6 viruses were reported to be involved in the generation of H5 and H7 subtype viruses. To investigate the emergence, evolutionary pattern, and potential for an epidemic of H6N6 viruses, the complete genomes of 198 H6N6 viruses were analyzed, including 168 H6N6 viruses deposited in the NCBI and GISAID databases from inception to January 2019 and 30 isolates collected from China between November 2014 and January 2019. Using phylogenetic analysis, the 198 strains of H6N6 viruses were identified as 98 genotypes. Molecular clock analysis indicated that the evolution of H6N6 viruses in China was constant and not interrupted by selective pressure. Notably, the laboratory isolates reassorted with six subtype viruses: H6N2, H5N6, H7N9, H5N2, H4N2, and H6N8, resulting in nine novel H6N6 reassortment events. These results suggested that H6N6 viruses can act as an intermediary in the evolution of H5N6, H6N6, and H7N9 viruses. Animal experiments demonstrated that the 10 representative H6N6 viruses showed low pathogenicity in chickens and were capable of infecting mice without prior adaptation. Our findings suggest that H6N6 viruses play an important role in the evolution of AIVs, and it is necessary to continuously monitor and evaluate the potential epidemic of the H6N6 subtype viruses.

Genetic and virological characteristics of a reassortant avian influenza A H6N1 virus isolated from wild birds at a live-bird market in Egypt.

The first detection of a human infection with avian influenza A/H6N1 virus in Taiwan in 2013 has raised concerns about this virus. During our routine surveillance of avian influenza viruses (AIVs) in live-bird markets in Egypt, an H6N1 virus was isolated from a garganey duck and was characterized. Phylogenetic analysis indicated that the Egyptian H6N1 strain A/Garganey/Egypt/20869C/2022(H6N1) has a unique genomic constellation, with gene segments inherited from different subtypes (H5N1, H3N8, H7N3, H6N1, and H10N1) that have been detected previously in AIVs from Egypt and some Eurasian countries. We examined the replication of kinetics of this virus in different mammalian cell lines (A549, MDCK, and Vero cells) and compared its pathogenicity to that of the ancestral H6N1 virus A/Quail/HK/421/2002(H6N1). The Egyptian H6N1 virus replicated efficiently in C57BL/6 mice without prior adaptation and grew faster and reached higher titers than in A549 cells than the ancestral strain. These results show that reassortant H6 AIVs might pose a potential threat to human health and highlight the need to continue surveillance of H6 AIVs circulating in nature.

Knowledge-driven domain adaptation strategy for rotating machinery fault diagnosis under varying working condition

Due to the frequent switch of the working condition, fault diagnosis model for rotating machinery established on the training set (the source domain) cannot be effectively applied to the test set (the target domain). Traditional domain adaptation methods address this issue by performing feature alignment between the source and target domain, which ignores the positive guidance provided by prior knowledge. In this paper, we propose a prior knowledge-driven domain adaptation (PKDA) method for varying working condition fault diagnosis of rotating machinery, where a self-supervised learning framework is designed to integrate expert prior knowledge and structural prior knowledge. First, an expert prior knowledge guidance module is designed to extract features with physical significance. Then we align the features between the source and target domain by the max mean discrepancy metric. Besides, a progressive Shannon entropy minimization strategy is proposed to realize the feature distribution structure of intra-class compression and inter-class separation, which can effectively integrate the structural prior knowledge. In this way, PKDA can effectively utilize the prior knowledge to achieve better performance in varying working condition fault diagnosis tasks. The effectiveness of the proposed method is illustrated by an open-source rolling bearing fault dataset from Case Western Reserve University and an open-source gearbox fault dataset from Southeast University.

Evolution and biological characteristics of the circulated H8N4 avian influenza viruses1

The circulating avian influenza viruses in wild birds have a high possibility of spillover into domestic birds or mammals at the wild bird-domestic bird or bird-mammal interface. H8N4 viruses primarily circulate in migratory wild waterfowl and have rarely been identified in domestic birds. In this study, we summarized the spatial and temporal distribution of global H8 viruses, specified their natural reservoirs, and performed detailed evolutionary analysis on the dominant H8N4 viruses. Here, we also reported a novel H8N4 virus isolated from a Eurasian coot sample from a wetland in eastern China in 2022. Animal infection studies indicated that the wild bird-originated H8N4 virus can replicate and transmit efficiently in ducks but has not adapted to chickens. Additionally, this naturally isolated H8N4 virus can replicate in mice without prior adaptation. These results indicate that H8 viruses exist mainly in the wild duck reservoir and pose a high infection risk to domestic ducks. Therefore, the active surveillance of influenza viruses at the wild and domestic waterfowl interface will contribute to monitoring the circulation of these viruses.

Gradient Prior Guidance and Image Adaptation Enhancement for Semi-Supervised SAR Ship Instance Segmentation

Gradient Prior Guidance and Image Adaptation Enhancement for Semi-Supervised SAR Ship Instance Segmentation

A Curriculum-style Self-training Approach for Source-Free Semantic Segmentation.

Source-free domain adaptation has developed rapidly in recent years, where the well-trained source model is adapted to the target domain instead of the source data, offering the potential for privacy concerns and intellectual property protection. However, a number of feature alignment techniques in prior domain adaptation methods are not feasible in this challenging problem setting. Thereby, we resort to probing inherent domain-invariant feature learning and propose a curriculum-style self-training approach for source-free domain adaptive semantic segmentation. In particular, we introduce a curriculum-style entropy minimization method to explore the implicit knowledge from the source model, which fits the trained source model to the target data using certain information from easy-to-hard predictions. We then train the segmentation network by the proposed complementary curriculum-style self-training, which utilizes the negative and positive pseudo labels following the curriculum-learning manner. Although negative pseudo-labels with high uncertainty cannot be identified with the correct labels, they can definitely indicate absent classes. Moreover, we employ an information propagation scheme to further reduce the intra-domain discrepancy within the target domain, which could act as a standard post-processing method for the domain adaptation field. Furthermore, we extend the proposed method to a more challenging black-box source model scenario where only the source model's predictions are available. Extensive experiments validate that our method yields state-of-the-art performance on source-free semantic segmentation tasks for both synthetic-to-real and adverse conditions datasets. The code and corresponding trained models are released at https://github.com/yxiwang/ATP.

Incrementally Adapting Pretrained Model Using Network Prior for Multi-Focus Image Fusion.

Multi-focus image fusion can fuse the clear parts of two or more source images captured at the same scene with different focal lengths into an all-in-focus image. On the one hand, previous supervised learning-based multi-focus image fusion methods relying on synthetic datasets have a clear distribution shift with real scenarios. On the other hand, unsupervised learning-based multi-focus image fusion methods can well adapt to the observed images but lack the general knowledge of defocus blur that can be learned from paired data. To avoid the problems of existing methods, this paper presents a novel multi-focus image fusion model by considering both the general knowledge brought by the supervised pretrained backbone and the extrinsic priors optimized on specific testing sample to improve the performance of image fusion. To be specific, the Incremental Network Prior Adaptation (INPA) framework is proposed to incrementally integrate features extracted from the pretrained strong baselines into a tiny prior network (6.9% parameters of the backbone network) to boost the performance for test samples. We evaluate our method on both synthetic and real-world public datasets (Lytro, MFI-WHU, and Real-MFF) and show that our method outperforms existing supervised learning-based methods and unsupervised learning based methods.

CALDA: Improving Multi-Source Time Series Domain Adaptation With Contrastive Adversarial Learning.

Unsupervised domain adaptation (UDA) provides a strategy for improving machine learning performance in data-rich (target) domains where ground truth labels are inaccessible but can be found in related (source) domains. In cases where meta-domain information such as label distributions is available, weak supervision can further boost performance. We propose a novel framework, CALDA, to tackle these two problems. CALDA synergistically combines the principles of contrastive learning and adversarial learning to robustly support multi-source UDA (MS-UDA) for time series data. Similar to prior methods, CALDA utilizes adversarial learning to align source and target feature representations. Unlike prior approaches, CALDA additionally leverages cross-source label information across domains. CALDA pulls examples with the same label close to each other, while pushing apart examples with different labels, reshaping the space through contrastive learning. Unlike prior contrastive adaptation methods, CALDA requires neither data augmentation nor pseudo labeling, which may be more challenging for time series. We empirically validate our proposed approach. Based on results from human activity recognition, electromyography, and synthetic datasets, we find utilizing cross-source information improves performance over prior time series and contrastive methods. Weak supervision further improves performance, even in the presence of noise, allowing CALDA to offer generalizable strategies for MS-UDA.