Action Classification Research Articles

Prioritizing management actions for invasive non-native plants through expert-based knowledge and species distribution models

Given the high number of non-native plants that are being introduced worldwide and the time required to process formal pest risk analyses, a framework for the prioritization of management actions is urgently required. We therefore propose a framework for a replicable and standardized prioritization for management actions (eradication, control and monitoring) of invasive non-native plants, combining expert knowledge, current and future climatic suitability estimated by species distribution models (SDMs), clustering and ordination techniques. Based on expert consultation and using Italy as case study, invasive non-native plant species were selected and three categories of management actions were identified: eradication, control and containment, and monitoring. Finally, two further classes of priorities were proposed for each of the management actions: “high” and “low” priority. Overall, SDMs highlighted a high and very high suitability for Continental and Mediterranean bioregions for most invasive plants. Cluster analysis revealed three distinct clusters with varying levels of suitability for the Italian bioregions. Cluster 1 exhibited a higher suitability across all Italian bioregions, whereas non-native plants grouped in Cluster 2 predominantly featured high suitability in Mediterranean areas. Finally, Cluster 3 showed the lowest suitability values. Two ordination analysis highlighted the variability in bioclimatic suitability for each non-native plant within each cluster, as well as their current distribution pattern. Lastly, a third ordination, integrating bioclimatic suitability and spatial patterns, has allowed the differentiation of management actions for each non-native plant at both national and bioregional scales. Specifically, seven non-native plants were earmarked for eradication action, six for monitoring action, while the remaining species were deemed suitable for control and containment. Our results and the methodology proposed meet the demand for replicable new early warning tools; that is to predict the location of new outbreaks, to establish priorities for eradication, control and containment, and to monitor invasive non-native species.

Evaluating predictors of kinase activity of STK11 variants identified in primary human non-small cell lung cancers.

Critical evaluation of computational tools for predicting variant effects is important considering their increased use in disease diagnosis and driving molecular discoveries. In the sixth edition of the Critical Assessment of Genome Interpretation (CAGI) challenge, a dataset of 28 STK11 rare variants (27 missense, 1 single amino acid deletion), identified in primary non-small cell lung cancer biopsies, was experimentally assayed to characterize computational methods from four participating teams and five publicly available tools. Predictors demonstrated a high level of performance on key evaluation metrics, measuring correlation with the assay outputs and separating loss-of-function (LoF) variants from wildtype-like (WT-like) variants. The best participant model, 3Cnet, performed competitively with well-known tools. Unique to this challenge was that the functional data was generated with both biological and technical replicates, thus allowing the assessors to realistically establish maximum predictive performance based on experimental variability. Three out of the five publicly available tools and 3Cnet approached the performance of the assay replicates in separating LoF variants from WT-like variants. Surprisingly, REVEL, an often-used model, achieved a comparable correlation with the real-valued assay output as that seen for the experimental replicates. Performing variant interpretation by combining the new functional evidence with computational and population data evidence led to 16 new variants receiving a clinically actionable classification of likely pathogenic (LP) or likely benign (LB). Overall, the STK11 challenge highlights the utility of variant effect predictors in biomedical sciences and provides encouraging results for driving research in the field of computational genome interpretation.

Is bank misconduct related to social capital? Evidence from U.S. banks

This paper investigates whether social capital plays a role in bank misconduct. I find that U.S. banks headquartered in high social capital areas, as indicated by the strength of civic norms and the density of social networks, are less likely to face enforcement actions. This relationship is mainly significant for banks with a lower geographical dispersion, and it holds in a range of robustness and endogeneity tests. I run additional tests based on classes of enforcement actions, components of social capital, risk-taking, opacity, and bank actions associated with negative externalities. These tests deliver results supporting the idea that the main findings of the paper are largely attributed to social capital's role in exerting external discipline, which prevents misconduct-related behaviors in banks.

Potential to use of viral reverse transcriptase inhibitors in oncology

In preparing the review, articles on the functioning of the reverse transcriptase enzyme of endogenous repeat sequences LINE1, the mechanisms of action and antitumor activity of viral reverse transcriptase inhibitors. Articles available in the biomedical literature information databases SciVerse Scopus, PubMed, Web of Science, Russian Science CitationIndex (RSCI) were analyzed. The review used information from 140 publications, of which 95 and 39 were published, respectively, over the last ten and three years, 2 articles present the results of clinical studies, and 45 articles refer to results demonstrating the anticancer properties of the studied compounds in various models in vitro and in vivo. Aim. Based on data on the functional properties of the reverse transcriptase enzyme of endogenous repeat sequences LINE1 (long interspersed nuclear elements 1), analyze the potential use of viral reverse transcriptase inhibitors in oncology, presenting their classification and main mechanisms of action. About 98 % of the human genome consists of repetitive sequences, most of which are represented by mobil genetic elements, the activation of which leads to increased genome instability. These include long (LINE) and short (SINE) interspersed nuclear element repeated DNA sequences interspersed nuclear elements, respectively, which occupy about 45 % of the human genome. Increased expression levels of these sequences in the genome have been identified in many forms of malignant neoplasms. Their transposition occurs due to the expression of LINE1-encoded reverse transcriptase, whichis homologous to viral reverse transcriptase. To date, reverse transcriptase inhibitors of viruses of nucleoside and non-nucleoside structure have been developed and are successfully used in the clinic. These drugs demonstrate an inhibitory effect on both LINE1 reverse transcriptase and telomerase, which provides the tumor cell with the ability to overcome replicative senescence. Due to these properties, these compounds are expected to exhibit both their own antitumor activity and increase the sensitivity of tumor cells to the therapy of malignant neoplasms, which is experimentally confirmed in models of malignant tumors in vitro and in vivo. Use of reverse transcriptase inhibitors in combination therapy seems advisable both to prevent further genome rearrangements caused by LINE1 and to suppress the survival of tumor cells by inhibiting telomerase activity.

Искусственный интеллект как посредник между животным и человеком

The development of technology has changed the position of animals in the modern world in various aspects. However, only the achievements of artificial intelligence in the field of natural languages indicated the possibility of reaching a new level of understanding and relationship with animals. Modern technologies have made it possible to isolate and fi x animal sounds and collect a huge array of audio and video data, and the experience of translation, even in the absence of parallel texts, has indicated the potential for using artificial intelligence to analyze animal sounds. Despite numerous difficulties, including those associated with the difference in the worldview of animals and humans, there are already precedents for translation from the language of animals. The article analyzes the possibilities of using artificial intelligence in conditions of limited data and its current use in the field of animal communication. If for domestic and farm animals, researchers rely on the interpretation of meanings or emotions, then for wild animals, scientists compare sounds and behavior, and rely on the potential of artificial intelligence in solving unstructured problems. Although a number of recent studies report high reliability of “translation” from the language of animals, the very possibility of testing the effectiveness is difficult. Nevertheless, the accelerating emergence of new solutions that facilitate the recognition of the voices of specific animals, the classification of sounds and actions of different animals, etc., indicate the possibility of a qualitative leap in the understanding of animals in the near future. Success in the field of interpretation of animal sounds can lead not only to progress in a large number of areas related to the animal world, but also to a change in the status and position of the animal. At the same time, the achievements raise ethical questions related to the possibility of using new technologies to the detriment of animals and people.

Mitigating human fall injuries: A novel system utilizing 3D 4-stream convolutional neural networks and image fusion

Unintentional human falls, especially in seniors, lead to serious injuries, fatalities, and reduced standard of life. Vision-based fall detection methods have demonstrated their usefulness in timely fall response, helping to lessen such injuries. This paper presents an automated vision-based fall detection system that triggers immediate fall reporting. By incorporating human segmentation and image fusion in the pre-processing stage, the system enhances the accuracy of human action classification, thereby ensuring precise fall alerts. It further employs the innovative 4-stream 3D convolutional neural network (4S-3DCNN) model to learn different but consecutive spatial and temporal features. The system processes video input or live surveillance, segmenting human presence every 32 frames using a fine-tuned deep-learning model and applying a three-level image fusion to accentuate movement differences. This technique produces four pre-processed images, input to the 4S-3DCNN model for classification. Consecutive detection of “Falling” and “Fallen” actions triggers an alert for immediate intervention. The original 4S-3DCNN model is an end-to-end trained deep learning model with a fully connected layer serving as a classifier. The research also evaluates the performance of combining the 4S-3DCNN model with Autoencoders and Support Vector Machines (SVM) networks as classifiers. The SVM classifier demonstrated ideal fall detection performance with 100% accuracy using the MCFD, URFD, and Le2i FDD datasets. The proposed system is vital for detecting and preventing falls and reducing healthcare expenses and productivity losses.

Dual-branch Cross-scale Feature Interaction for Temporal Action Detection

Temporal Action Detection (TAD) is aim to predict action boundary and category simultaneously. Most existing RGB-based methods model temporal dependency using pyramid-style features without interaction among different scales, which usually result in inaccurate prediction for long-term actions. The reason is that features at different scales involve information with different granularity, which is suitable for either prediction of action boundary or category. In this paper, we present a novel Dual-branch Cross-scale Feature Interaction (DCFI) method that directly exchanges different scale information from both temporal and spatial perspective for TAD. To be specific, in one branch, a cross-scale temporal transformer module is devised to enable both semantic and temporal communications among different scale features with a merge-to-split mechanism. While the other branch designs a cross-scale spatial mixer module to mine the most salient spatial difference between consecutive and long-term frames via a scale-mixer. Benefiting from these two modules, DCFI achieves comprehensive temporal as well as spatial interaction across all feature scales, and thus accurately predicts the boundaries of different time-span action instances. Extensive experiments on two challenging benchmarks, i.e.., THUMOS-14 and ActivityNet-1.3, demonstrate that our DCFI achieves new state-of-the-art performance with only RGB.

Does Phage Therapy Need a Pan-Phage?

The emergence of multidrug-resistant bacteria is undoubtedly one of the most serious global health threats. One response to this threat that has been gaining momentum over the past decade is 'phage therapy'. According to this, lytic bacteriophages are used for the treatment of bacterial infections, either alone or in combination with antimicrobial agents. However, to ensure the efficacy and broad applicability of phage therapy, several challenges must be overcome. These challenges encompass the development of methods and strategies for the host range manipulation and bypass of the resistance mechanisms developed by pathogenic bacteria, as has been the case since the advent of antibiotics. As our knowledge and understanding of the interactions between phages and their hosts evolves, the key issue is to define the host range for each application. In this article, we discuss the factors that affect host range and how this determines the classification of phages into different categories of action. For each host range group, recent representative examples are provided, together with suggestions on how the different groups can be used to combat certain types of bacterial infections. The available methodologies for host range expansion, either through sequential adaptation to a new pathogen or through genetic engineering techniques, are also reviewed.

Simulation of artificial intelligence robots in dance action recognition and interaction process based on machine vision

Artificial intelligence robots based on machine vision have shown great potential in dance action recognition and interaction. In order to perform action recognition, a large amount of dance action video data was collected as the basis for training and testing the model. In motion recognition, various computer vision techniques are used to extract motion features from images or videos to capture key information of dance movements. Machine learning technology was adopted to construct an action classification model. By using the data collected in the early stage and the extracted features, an efficient and accurate classifier has been studied and trained. It can classify the input dance actions into different categories, segment continuous video frames into different action segments, and ensure accurate differentiation of different dance actions during the recognition process, avoiding misjudgment or confusion. By accurately capturing and analyzing the posture information of the human body, we can better understand and simulate dance movements, and enhance the performance and realism of artificial intelligence robots in dance interactions. The experimental results show that the system can accurately recognize various dance movements and achieve good interaction with users.

Beyond target chemicals: updating the NORMAN prioritisation scheme to support the EU chemicals strategy with semi-quantitative suspect/non-target screening data

BackgroundPrioritisation of chemical pollutants is a major challenge for environmental managers and decision-makers alike, which is essential to help focus the limited resources available for monitoring and mitigation actions on the most relevant chemicals. This study extends the original NORMAN prioritisation scheme beyond target chemicals, presenting the integration of semi-quantitative data from retrospective suspect screening and expansion of existing exposure and risk indicators. The scheme utilises data retrieved automatically from the NORMAN Database System (NDS), including candidate substances for prioritisation, target and suspect screening data, ecotoxicological effect data, physico-chemical data and other properties. Two complementary workflows using target and suspect screening monitoring data are applied to first group the substances into six action categories and then rank the substances using exposure, hazard and risk indicators. The results from the ‘target’ and ‘suspect screening’ workflows can then be combined as multiple lines of evidence to support decision-making on regulatory and research actions.ResultsAs a proof-of-concept, the new scheme was applied to a combined dataset of target and suspect screening data. To this end, > 65,000 substances on the NDS, of which 2579 substances supported by target wastewater monitoring data, were retrospectively screened in 84 effluent wastewater samples, totalling > 11 million data points. The final prioritisation results identified 677 substances as high priority for further actions, 7455 as medium priority and 326 with potentially lower priority for actions. Among the remaining substances, ca. 37,000 substances should be considered of medium priority with uncertainty, while it was not possible to conclude for 19,000 substances due to insufficient information from target monitoring and uncertainty in the identification from suspect screening. A high degree of agreement was observed between the categories assigned via target analysis and suspect screening-based prioritisation. Suspect screening was a valuable complementary approach to target analysis, helping to prioritise thousands of substances that are insufficiently investigated in current monitoring programmes.ConclusionsThis updated prioritisation workflow responds to the increasing use of suspect screening techniques. It can be adapted to different environmental compartments and can support regulatory obligations, including the identification of specific pollutants in river basins and the marine environments, as well as the confirmation of environmental occurrence levels predicted by modelling tools.Graphical

Cell Painting unravels insecticidal modes of action on Spodoptera frugiperda insect cells

The “Cell Painting” technology utilizes multiplexed fluorescent staining of various cell organelles, to produce high-content microscopy images of cells for multidimensional phenotype assessment. The phenotypic profiles extracted from those images can be analyzed upon perturbations with biologically active molecules to annotate the mode of action or biological activity by comparison with reference profiles of already known mechanisms of action, ultimately enabling the determination of on-target and off-target effects. This approach is already described in various human cell cultures, the most commonly used being the U2OS cell line, yet allows broad applications in additional areas of chemical-biological research. Here we describe for the first time the application and adaptation of Cell Painting to an insect cell line, the Sf9 cells from Spodoptera frugiperda. By adjusting image acquisition and analysis models, specific phenotypic profiles were obtained in a dose-dependent manner for 20 reference compounds, including representatives for the most relevant insecticidal modes of action categories (nerve & muscle, respiration and growth & development). Through a dimensionality-reduction method, both calculations of phenotypic half maximal inhibition concentration (IC50) values as well as similarity analysis of the obtained profiles by hierarchical clustering were performed. By Cell Painting effects on the phenotype could be obtained at higher sensitivity than in other assay formats, such as cytotoxicity assessments. More importantly, these analyses provide insight into mechanistic determinants of biological activity. Compounds with similar modes of action showed a high degree of proximity in a hierarchical clustering analysis while being distinct from actives with an unrelated mode of action. In essence, we provide strong evidence on the impact of Cell Painting mechanistic understanding of insecticides with regards to determinants of efficacy and safety utilizing an insect cell model system.

Anomaly detection for high-content image-based phenotypic cell profiling.

High-content image-based phenotypic profiling combines automated microscopy and analysis to identify phenotypic alterations in cell morphology and provide insight into the cell's physiological state. Classical representations of the phenotypic profile can not capture the full underlying complexity in cell organization, while recent weakly machine-learning based representation-learning methods are hard to biologically interpret. We used the abundance of control wells to learn the in-distribution of control experiments and use it to formulate a self-supervised reconstruction anomaly-based representation that encodes the intricate morphological inter-feature dependencies while preserving the representation interpretability. The performance of our anomaly-based representations was evaluated for downstream tasks with respect to two classical representations across four public Cell Painting datasets. Anomaly-based representations improved reproducibility, Mechanism of Action classification, and complemented classical representations. Unsupervised explainability of autoencoder-based anomalies identified specific inter-feature dependencies causing anomalies. The general concept of anomaly-based representations can be adapted to other applications in cell biology.

Enhancing Single-Frame Supervision for Better Temporal Action Localization.

Temporal action localization aims to identify the boundaries and categories of actions in videos, such as scoring a goal in a football match. Single-frame supervision has emerged as a labor-efficient way to train action localizers as it requires only one annotated frame per action. However, it often suffers from poor performance due to the lack of precise boundary annotations. To address this issue, we propose a visual analysis method that aligns similar actions and then propagates a few user-provided annotations (e.g., boundaries, category labels) to similar actions via the generated alignments. Our method models the alignment between actions as a heaviest path problem and the annotation propagation as a quadratic optimization problem. As the automatically generated alignments may not accurately match the associated actions and could produce inaccurate localization results, we develop a storyline visualization to explain the localization results of actions and their alignments. This visualization facilitates users in correcting wrong localization results and misalignments. The corrections are then used to improve the localization results of other actions. The effectiveness of our method in improving localization performance is demonstrated through quantitative evaluation and a case study.

Non-Hausdorff germinal groupoids for actions of countable groups

We study conditions under which the germinal groupoid associated to a minimal equicontinuous action of a countable group on a Cantor set has non-Hausdorff topology. We develop a new criterion, which serves as an obstruction for the étale topology on the groupoid to be non-Hausdorff. We use this and other criteria to study the topology of germinal groupoids for a few classes of actions. In particular, we give examples of families of contracting and non-contracting self-similar groups, which are amenable and whose actions have associated germinal groupoids with non-Hausdorff topology.

A Video Action Recognition Method via Dual-Stream Feature Fusion Neural Network with Attention

Video action recognition is a technique for automatically determining the category of a video action. It is necessary to design an efficient video action recognition algorithm to predict video labels. This work proposes a video action recognition model based on dual-stream information fusion with attention mechanisms (DSIFAM), which consists of three different sub-modules. First, this proposes an improved keyframe extraction method (IKFE). Based on K-means clustering, this uses convolutional features to calculate the similarity between video frames instead of pixel points. After obtaining preliminary clustering results, the method performs secondary optimization to obtain more representative keyframes. Second, this proposes a video action recognition model based on dual-stream information fusion (DSIF). The method introduces ConvLSTM in the spatial stream and uses P3D instead of the original convolutional network in the temporal stream, which can better extract spatial-temporal information and improve the classification performance. Third, this designs a multi-scale attention mechanism (MSAM) to enhance the feature extraction stage and obtain higher quality classification features. The resulting features are more prominent and have stronger representation capabilities. Finally, this work conducts systematic experiments on different datasets, the results verify the superiority of DSIFAM for video action recognition.

Subject review: Human activity recognition based on depth images

Human activity recognition (HAR) has becomes a well-known area of study for researchers due to its multiple possible applications. HAR systems are applied in many areas including supervisory nursing care, medical supervision, surveillance, human- machine interaction, gaming and entertainment, etc. However, developing a robust HAR system that incorporates a comprehensive and scalable set of feature spaces is a challenging task. Much research has been done in this area to develop efficient HAR systems. In this article, a review of the developments in Human Activity Recognition (HAR) will be presented. Human activity recognition works on two basic principles; feature extraction from sensor data and classification of features into action classes. The review presented in this article is hence divided into two parts. The first part deals with feature representation in HAR systems such as Space-time Features, Frequency Features, Local Descriptors, Optical Flow Based, and Skeleton Joints. The second part deals with the classification approaches used in the literature such as Bayesian, and HMM.

The Body as Medium - Embodied Communication and Body Issues in Cloud Fitness

The "body turn" of "cloud fitness" has brought a new theoretical logic to communication. In this paper, we make a preliminary attempt at embodied communication in the context of the new Crown pneumonia epidemic, exploring the issues of body movement and moving bodies on "Shake" and "b-station". Here, the author argues that understanding embodied communication is inseparable from reflecting on technology, and that it is a form of communication in which people, media and technology interact. The article explores the changes of the philosophical category of bodily action in the virtual space of the Internet by tracing the origins of embodied communication; then understands the communication of the body as a medium through the "cloud space" and "cloud body"; and in the whole process of communication, the physical attributes of the body give the online and offline behaviors of the body a new meaning and a new meaning and a new meaning. In the whole process of communication, the physical attributes of the body bring tension to the online and offline actors; in the future, with the advent of the scene-oriented era, the body will have multiple symbolic attributes.

Evolving malware detection through instant dynamic graph inverse reinforcement learning

The rapid development and increasing evolution of malware necessitate novel defensive techniques with high accuracy and minimal false positives to safeguard information systems against potential threats. Unfortunately, current malware detection methods, primarily relying on deep learning to identify malicious fingerprints from extensive training sets, prove ineffective against few-shot and evolving malware variants. To address these challenges, this paper introduces MalIRL, which designs a model-free inverse reinforcement learning (IRL) mechanism to automatically capture the evolving attack intent of malware. Specifically, MalIRL explores six representative categories of malware actions and employs sliding windows to organically divide the massive malware execution event stream into multiple attack stages, achieving a reduced action space and state space. To model dynamic malicious environments, MalIRL proposes an instant dynamic heterogeneous graph representation learning technique. This technique learns state representations of different malware attack stages, enhancing detection accuracy and efficiency by incrementally capturing the newly added contextual semantics of diverse malware entities and relations. In experiments with three real-world malware datasets, MalIRL surpasses existing state-of-the-art methods, particularly in few-shot malware detection scenarios, MalIRL exhibits performance benefits of up to 18.9%.

Implementation of a health information technology safety classification system in the Veterans Health Administration's Informatics Patient Safety Office.

Implement the 5-type health information technology (HIT) patient safety concern classification system for HIT patient safety issues reported to the Veterans Health Administration's Informatics Patient Safety Office. A team of informatics safety analysts retrospectively classified 1year of HIT patient safety issues by type of HIT patient safety concern using consensus discussions. The processes established during retrospective classification were then applied to incoming HIT safety issues moving forward. Of 140 issues retrospectively reviewed, 124 met the classification criteria. The majority were HIT failures (eg, software defects) (33.1%) or configuration and implementation problems (29.8%). Unmet user needs and external system interactions accounted for 20.2% and 10.5%, respectively. Absence of HIT safety features accounted for 2.4% of issues, and 4% did not have enough information to classify. The 5-type HIT safety concern classification framework generated actionable categories helping organizations effectively respond to HIT patient safety risks.

Perceiving Actions via Temporal Video Frame Pairs

Video action recognition aims at classifying the action category in given videos. In general, semantic-relevant video frame pairs reflect significant action patterns such as object appearance variation and abstract temporal concepts like speed, rhythm, and so on. However, existing action recognition approaches tend to holistically extract spatiotemporal features. Though effective, there is still a risk of neglecting the crucial action features occurring across frames with a long-term temporal span. Motivated by this, in this article, we propose to perceive actions via frame pairs directly and devise a novel Nest Structure with frame pairs as basic units. Specifically, we decompose a video sequence into all possible frame pairs and hierarchically organize them according to temporal frequency and order, thus transforming the original video sequence into a Nest Structure. Through naturally decomposing actions, the proposed structure can flexibly adapt to diverse action variations such as speed or rhythm changes. Next, we devise a Temporal Pair Analysis module (TPA) to extract discriminative action patterns based on the proposed Nest Structure. The designed TPA module consists of a pair calculation part to calculate the pair features and a pair fusion part to hierarchically fuse the pair features for recognizing actions. The proposed TPA can be flexibly integrated into existing backbones, serving as a side branch to capture various action patterns from multi-level features. Extensive experiments show that the proposed TPA module can achieve consistent improvements over several typical backbones, reaching or updating CNN-based SOTA results on several challenging action recognition benchmarks.