Distinct Tasks Research Articles

FedLPPA: Learning Personalized Prompt and Aggregation for Federated Weakly-supervised Medical Image Segmentation.

Federated learning (FL) effectively mitigates the data silo challenge brought about by policies and privacy concerns, implicitly harnessing more data for deep model training. However, traditional centralized FL models grapple with diverse multi-center data, especially in the face of significant data heterogeneity, notably in medical contexts. In the realm of medical image segmentation, the growing imperative to curtail annotation costs has amplified the importance of weakly-supervised techniques which utilize sparse annotations such as points, scribbles, etc. A pragmatic FL paradigm shall accommodate diverse annotation formats across different sites, which research topic remains under-investigated. In such context, we propose a novel personalized FL framework with learnable prompt and aggregation (FedLPPA) to uniformly leverage heterogeneous weak supervision for medical image segmentation. In FedLPPA, a learnable universal knowledge prompt is maintained, complemented by multiple learnable personalized data distribution prompts and prompts representing the supervision sparsity. Integrated with sample features through a dual-attention mechanism, those prompts empower each local task decoder to adeptly adjust to both the local distribution and the supervision form. Concurrently, a dual-decoder strategy, predicated on prompt similarity, is introduced for enhancing the generation of pseudo-labels in weakly-supervised learning, alleviating overfitting and noise accumulation inherent to local data, while an adaptable aggregation method is employed to customize the task decoder on a parameter-wise basis. Extensive experiments on four distinct medical image segmentation tasks involving different modalities underscore the superiority of FedLPPA, with its efficacy closely parallels that of fully supervised centralized training. Our code and data will be available at https://github.com/llmir/FedLPPA.

BIM multitask project manager: responsibility, organization and interoperability

Building Information Modelling (BIM) methodology has been supporting the construction activity in all sectors involved: multidisciplinary designs development; construction planning and monitoring; building management and maintenance. A BIM environment aggregates several specialties and related professionals and in order to control and improve the quality of a BIM project, a BIM manager is required. The BIM manager has the responsibility to coordinate all tasks involved in a building design and in the associated activities that are usually workout over the project documents. A BIM platform can be considered as a useful resource for multidisciplinary coordination, extending the BIM 3D modelling scope into the construction planning and after, in the operation and maintenance stages. A BIM manager is then required: to incentive the collaboration at work and to control responsibility given to each professional involved; to guaranty the correction of the amount and quality of the information, included in each BIM model step, and transferred between tasks; to coordinate all designs and the distinct mandatory activities (conflict detection, cost estimation or energetic simulation). The topic of the present study is illustrated with three building cases were distinct specific projects, disciplines and tasks were developed: structural analyses and reinforcement details; quantity take-off of materials and cost estimation; construction scheduling and simulation. The study presents the relevance of a correct coordination within the development of building projects. Although there are limitations in the implementation of BIM methodology in all sectors or activities within the construction industry, BIM has been bringing an important improvement in the quality of a building design, reflected in the quality of the final product. BIM methodology is a current demand in the construction industry supported on advanced technology and in an adequate management of projects.

Learning Content-Weighted Pseudocylindrical Representation for 360° Image Compression.

Learned 360° image compression methods using equirectangular projection (ERP) often confront a non-uniform sampling issue, inherent to sphere-to-rectangle projection. While uniformly or nearly uniformly sampling representations, along with their corresponding convolution operations, have been proposed to mitigate this issue, these methods often concentrate solely on uniform sampling rates, thus neglecting the content of the image. In this paper, we urge that different contents within 360° images have varying significance and advocate for the adoption of a content-adaptive parametric representation in 360° image compression, which takes into account both the content and sampling rate. We first introduce the parametric pseudocylindrical representation and corresponding convolution operation, upon which we build a learned 360° image codec. Then, we model the hyperparameter of the representation as the output of a network, derived from the image's content and its spherical coordinates. We treat the optimization of hyperparameters for different 360° images as distinct compression tasks and propose a meta-learning algorithm to jointly optimize the codec and the metaknowledge, i.e., the hyperparameter estimation network. A significant challenge is the lack of a direct derivative from the compression loss to the hyperparameter network. To address this, we present a novel method to relax the rate-distortion loss as a function of the hyperparameters, enabling gradient-based optimization of the metaknowledge. Experimental results on omnidirectional images demonstrate that our method achieves state-of-the-art performance and superior visual quality.

MOODv2: Masked Image Modeling for Out-of-Distribution Detection.

The crux of effective out-of-distribution (OOD) detection lies in acquiring a robust in-distribution (ID) representation, distinct from OOD samples. While previous methods predominantly leaned on recognition-based techniques for this purpose, they often resulted in shortcut learning, lacking comprehensive representations. In our study, we conducted a comprehensive analysis, exploring distinct pretraining tasks and employing various OOD score functions. The results highlight that the feature representations pre-trained through reconstruction yield a notable enhancement and narrow the performance gap among various score functions. This suggests that even simple score functions can rival complex ones when leveraging reconstruction-based pretext tasks. Reconstruction-based pretext tasks adapt well to various score functions. As such, it holds promising potential for further expansion. Our OOD detection framework, MOODv2, employs the masked image modeling pretext task. Without bells and whistles, MOODv2 impressively enhances 14.30% AUROC to 95.68% on ImageNet and achieves 99.98% on CIFAR-10.

Improving SSVEP-BCI Performance Through Repetitive Anodal tDCS-Based Neuromodulation: Insights From Fractal EEG and Brain Functional Connectivity.

This study embarks on a comprehensive investigation of the effectiveness of repetitive transcranial direct current stimulation (tDCS)-based neuromodulation in augmenting steady-state visual evoked potential (SSVEP) brain-computer interfaces (BCIs), alongside exploring pertinent electroencephalography (EEG) biomarkers for assessing brain states and evaluating tDCS efficacy. EEG data were garnered across three distinct task modes (eyes open, eyes closed, and SSVEP stimulation) and two neuromodulation patterns (sham-tDCS and anodal-tDCS). Brain arousal and brain functional connectivity were measured by extracting features of fractal EEG and information flow gain, respectively. Anodal-tDCS led to diminished offsets and enhanced information flow gains, indicating improvements in both brain arousal and brain information transmission capacity. Additionally, anodal-tDCS markedly enhanced SSVEP-BCIs performance as evidenced by increased amplitudes and accuracies, whereas sham-tDCS exhibited lesser efficacy. This study proffers invaluable insights into the application of neuromodulation methods for bolstering BCI performance, and concurrently authenticates two potent electrophysiological markers for multifaceted characterization of brain states.

Retrieval augmentation of large language models for lay language generation

The complex linguistic structures and specialized terminology of expert-authored content limit the accessibility of biomedical literature to the general public. Automated methods have the potential to render this literature more interpretable to readers with different educational backgrounds. Prior work has framed such lay language generation as a summarization or simplification task. However, adapting biomedical text for the lay public includes the additional and distinct task of background explanation: adding external content in the form of definitions, motivation, or examples to enhance comprehensibility. This task is especially challenging because the source document may not include the required background knowledge. Furthermore, background explanation capabilities have yet to be formally evaluated, and little is known about how best to enhance them. To address this problem, we introduce Retrieval-Augmented Lay Language (RALL) generation, which intuitively fits the need for external knowledge beyond that in expert-authored source documents. In addition, we introduce CELLS, the largest (63k pairs) and broadest-ranging (12 journals) parallel corpus for lay language generation. To evaluate RALL, we augmented state-of-the-art text generation models with information retrieval of either term definitions from the UMLS and Wikipedia, or embeddings of explanations from Wikipedia documents. Of these, embedding-based RALL models improved summary quality and simplicity while maintaining factual correctness, suggesting that Wikipedia is a helpful source for background explanation in this context. We also evaluated the ability of both an open-source Large Language Model (Llama 2) and a closed-source Large Language Model (GPT-4) in background explanation, with and without retrieval augmentation. Results indicate that these LLMs can generate simplified content, but that the summary quality is not ideal. Taken together, this work presents the first comprehensive study of background explanation for lay language generation, paving the path for disseminating scientific knowledge to a broader audience. Our code and data are publicly available at: https://github.com/LinguisticAnomalies/pls_retrieval.

Expert-Annotated Dataset to Study Cyberbullying in Polish Language

We introduce the first dataset of harmful and offensive language collected from the Polish Internet. This dataset was meticulously curated to facilitate the exploration of harmful online phenomena such as cyberbullying and hate speech, which have exhibited a significant surge both within the Polish Internet as well as globally. The dataset was systematically collected and then annotated using two approaches. First, it was annotated by two proficient layperson volunteers, operating under the guidance of a specialist in the language of cyberbullying and hate speech. To enhance the precision of the annotations, a secondary round of annotations was carried out by a team of adept annotators with specialized long-term expertise in cyberbullying and hate speech annotations. This second phase was further overseen by an experienced annotator, acting as a super-annotator. In its initial application, the dataset was leveraged for the categorization of cyberbullying instances in the Polish language. Specifically, the dataset serves as the foundation for two distinct tasks: (1) a binary classification that segregates harmful and non-harmful messages and (2) a multi-class classification that distinguishes between two variations of harmful content (cyberbullying and hate speech), as well as a non-harmful category. Alongside the dataset itself, we also provide the models that showed satisfying classification performance. These models are made accessible for third-party use in constructing cyberbullying prevention systems.

Baptist Boys and Girls

Abstract This paper will analyze the publications for children of Woman’s Missionary Union (WMU), auxiliary to the Southern Baptist Convention, from 1953 to 1957. These include Tell, a monthly magazine for girls, and Ambassador Life, a monthly magazine for boys. The paper will argue that these magazines urged girls to develop social graces while challenging boys to physical activity. In addition, both Ambassador Life and Tell generally encouraged their readers to adopt what editors saw as gender-specific roles, positing distinctive tasks for boys (and men) and for girls (and women). Overall, these emphases reflected conservative Southern Baptist understandings of gender and helped shape the views of a generation (or more) of Southern Baptists.

FDNet: Feature decoupling for single-stage pose estimation in complex scenes

Severe error accumulation is a key reason that makes existing multi-person pose estimation challenging in complex scenes. Single-stage pose estimation methods detect human and keypoints in parallel, avoiding serial error accumulation compared to two-stage methods. However, two distinct tasks share identical features in the existing method, which causes error accumulation. Therefore, we propose a Feature Decoupling Network (FDNet) to further extend the pipeline of single-stage methods and reconstruct task-specific features, including Feature Decoupling Module (FDM), Human Aware Loss (HAL) and Keypoint Aware Loss (KAL). FDM can adaptively perceive spatial and channel features and allocate separate feature domains for various tasks. To learn distinctive feature representation for compact human bodies, HAL and KAL measure and suppress feature similarities and distances for different human and keypoints, thus alienating feature interference. Experiments on crowded datasets show that our method is superior, outperforming the state-of-the-art method CID by 1.5%, 0.8% on OCHuman and CrowdPose.

Edge Learning for Large-Scale Internet of Things With Task-Oriented Efficient Communication

In the Internet of Things (IoT) networks, edge learning for data-driven tasks provides intelligent applications and services. As the network size becomes large, different users may generate distinct datasets. Thus, to suit multiple edge learning tasks for large-scale IoT networks, this paper performs efficient communication under the task-oriented principle by using the collaborative design of wireless resource allocation and edge learning error prediction. In particular, we start with multi-user scheduling to alleviate co-channel interference in dense networks. Then, we perform optimal power allocation in parallel for different learning tasks. Thanks to the high parallelization of the designed algorithm, extensive experimental results corroborate that the multi-user scheduling and task-oriented power allocation improve the performance of distinct edge learning tasks efficiently compared with the state-of-the-art benchmark algorithms.

A Comparative Study and Component Analysis of Query Plan Representation Techniques in ML4DB Studies

Query plan is widely used as input in machine learning for databases (ML4DB) research, with query plan representation as a critical step. However, existing studies typically focus on one task, and propose a novel design to represent query plans along with a ML4DB framework, without comparing with other representation methods designed for a different task. This raises a critical question: How do we select a query plan representation method in a ML4DB system? To address this question, we perform a comparative study on ten representation methods on three distinct ML4DB tasks: cost estimation, index selection and query optimization. Our extensive experiments not only verify the interchangeability of representation methods across different tasks, but also identify consistently high-performing models. Further, we dissect the query plan representation into two core components: feature encoding and tree model, and evaluate the impact of design choices for each in different scenarios. Our results show that the findings for tasks optimizing absolute errors are different from findings for tasks optimizing relative errors. Some findings challenge widely-held assumptions, i.e., one finding shows that tree models do not significantly impact cost estimation results, but only play a significant role to optimize relative performance. Practical guidelines and future directions are provided based on the findings of the study.

EMG and IMU Data Fusion for Locomotion Mode Classification in Transtibial Amputees

Despite recent advancements in prosthetic technology, lower-limb amputees often remain limited to passive prostheses, which leads to an asymmetric gait and increased energy expenditure. Developing active prostheses with effective control systems is important to improve mobility for these individuals. This study presents a machine-learning-based approach to classify five distinct locomotion tasks: ground-level walking (GWL), ramp ascent (RPA), ramp descent (RPD), stairs ascent (SSA), and stairs descent (SSD). The dataset comprises fused electromyographic (EMG) and inertial measurement unit (IMU) signals from twenty non-amputated and five transtibial amputated participants. EMG sensors were strategically positioned on the thigh muscles, while IMU sensors were placed on various leg segments. The performance of two classification algorithms, support vector machine (SVM) and long short-term memory (LSTM), were evaluated on segmented data. The results indicate that SVM models outperform LSTM models in accuracy, precision, and F1 score in the individual evaluation of amputee and non-amputee datasets for 80–20 and 50–50 data distributions. In the 80–20 distribution, an accuracy of 95.46% and 95.35% was obtained with SVM for non-amputees and amputees, respectively. An accuracy of 93.33% and 93.30% was obtained for non-amputees and amputees by using LSTM, respectively. LSTM models show more robustness and inter-population generalizability than SVM models when applying domain-adaptation techniques. Furthermore, the average classification latency for SVM and LSTM models was 19.84 ms and 37.07 ms, respectively, within acceptable limits for real-time applications. This study contributes to the field by comprehensively comparing SVM and LSTM classifiers for locomotion tasks, laying the foundation for the future development of real-time control systems for active transtibial prostheses.

Multi-task Pre-training Language Model for Semantic Network Completion

Semantic networks, exemplified by the knowledge graph, serve as a means to represent knowledge by leveraging the structure of a graph. While the knowledge graph exhibits promising potential in the field of natural language processing, it suffers from incompleteness. This article focuses on the task of completing knowledge graphs by predicting linkages between entities, which is fundamental yet critical. Traditional methods based on translational distance struggle when dealing with unseen entities. In contrast, semantic matching presents itself as a potential solution due to its ability to handle such cases. However, semantic matching-based approaches necessitate large-scale datasets for effective training, which are typically unavailable in practical scenarios, hindering their competitive performance. To address this challenge, we propose a novel architecture for knowledge graphs known as LP-BERT, which incorporates a language model. LP-BERT consists of two primary stages: multi-task pre-training and knowledge graph fine-tuning. During the pre-training phase, the model acquires relationship information from triples by predicting either entities or relations through three distinct tasks. In the fine-tuning phase, we introduce a batch-based triple-style negative sampling technique inspired by contrastive learning. This method significantly increases the proportion of negative sampling while maintaining a nearly unchanged training time. Furthermore, we propose a novel data augmentation approach that leverages the inverse relationship of triples to enhance both the performance and robustness of the model. To demonstrate the effectiveness of our proposed framework, we conduct extensive experiments on three widely used knowledge graph datasets: WN18RR, FB15k-237, and UMLS. The experimental results showcase the superiority of our methods, with LP-BERT achieving state-of-the-art performance on the WN18RR and FB15k-237 datasets.

Effectiveness of the Krotov method in finding controls for open quantum systems

We apply the Krotov method (KM) for open and closed quantum systems to find optimized controls to manipulate qubit/qutrit systems in the presence of the external environment. In the case of unitary optimization, the KM is first applied to a quantum system neglecting its interaction with the environment. The resulting controls from the unitary optimization are then used to drive the system along with the environmental noise. In the case of non-unitary optimization, the KM already takes into account the noise during the optimization process. We consider two distinct computational tasks: target-state preparation and quantum gate implementation. These tasks are carried out in simple qubit/qutrit systems and also in systems presenting leakage states. For the state preparation cases, the controls from the non-unitary optimization outperform the controls from the unitary optimization. However, as we show here, this is not always true for the implementation of quantum gates. There are some situations where the unitary optimization performs equally well compared to the non-unitary optimization. We verify that these situations correspond to either the absence of leakage states or to the effects of dissipation being spread uniformly over the system, including non-computational levels. For such cases, the quantum gate implementation must cover the entire Hilbert space and there is no way to dodge dissipation. On the other hand, if the subspace containing the computational levels and its complement are differently affected by dissipation, the non-unitary optimization becomes effective.

Multi-task peer-to-peer learning using an encoder-only transformer model

Peer-to-peer (P2P) learning is a decentralized approach to organizing the collaboration between end devices known as agents. Agents contain heterogeneous data, and that heterogeneity is disrupting the convergence and accuracy of the collectively learned models. A common technique to mitigate the negative impact of heterogeneous data is to arrange the learning process in a multi-task setting where each task, although it has the same learning objective, is learned separately. However, the multi-task technique can also be applied to solve distinct learning tasks. This paper presents and evaluates a novel approach that utilizes an encoder-only transformer model to enable collaboration between agents learning two distinct Natural Language Processing (NLP) tasks. The evaluation of the approach studied revealed that collaboration among agents, even when working towards separate objectives, can result in mutual benefits, mainly when the connections between agents are carefully considered. The multi-task collaboration led to a statistically significant increase of 11.6% in the mean relative accuracy compared to the baseline results for individual tasks. To our knowledge, this is the first study demonstrating a successful and beneficial collaboration between two distinct NLP tasks in a peer-to-peer setting.

Behavioral deficits after mild traumatic brain injury by fluid percussion in rats

Mild traumatic brain injury (TBI) can lead to various disorders, encompassing cognitive and psychiatric complications. While pre-clinical studies have long investigated behavioral alterations, the fluid percussion injury (FPI) model still lacks a comprehensive behavioral battery that includes psychiatric-like disorders. To address this gap, we conducted multiple behavioral tasks over two months in adult male Wistar rats, focusing on mild FPI. Statistical analyses revealed that both naive and sham animals exhibited an increase in sweet liquid consumption over time. In contrast, the TBI group did not show any temporal changes, although mild FPI did induce a statistically significant decrease in sucrose consumption compared to control groups during the chronic phase. Additionally, social interaction tasks indicated reduced contact time in TBI animals. The elevated plus maze task demonstrated an increase in open-arm exploration following fluid percussion. Nonetheless, no significant differences were observed in the acute and chronic phases for the forced swim and light–dark box tasks. Evaluation of three distinct memory tasks in the chronic phase revealed that mild FPI led to long-term memory deficits, as assessed by the object recognition task, while the surgical procedure itself resulted in short-term spatial memory deficits, as evaluated by the Y-maze task. Conversely, working memory remained unaffected in the water maze task. Collectively, these findings provide a nuanced characterization of behavioral deficits induced by mild FPI.

MedGCN: An IoT-edge thrombus graph convolutional network for accurate prediction and prescription diagnosis of vascular occlusive diseases from unstructured clinical reports

The timely and accurate prediction and diagnosis of vascular occlusive diseases are pivotal in enhancing patient outcomes. This research addresses this critical healthcare challenge by introducing MedGCN, a novel thrombus graph convolutional network model. MedGCN is specifically designed for the precise prediction of prescriptions and diagnoses based on unstructured clinical diagnostic reports. The model synergistically blends OpenAI’s GPT4 and a uniquely designed Cross Fusion Graph Convolutional Network (CF-GCN) to ensure the meticulous fusion of knowledge. We delve into nine distinct learning tasks, encompassing both prescription and diagnosis and employ a multi-label classification GCN pretraining technique to assess them. Our evaluation underscores MedGCN’s robust predictive prowess across various tasks. By amalgamating cutting-edge AI paradigms with IoT edge computing, this research not only bolsters the efficacy of medical decision-making but also champions patient privacy. The methodologies and findings presented herein hold immense potential for deployment in IoT frameworks, thus proffering swift and precise assistance in medical decision-making and addressing a paramount need in the contemporary healthcare landscape.

Turning the tables on online exam cheating via language mediation tasks

Abstract This article addresses the issue of foreign language testing in the online mode and proposes a strategy to reduce the possibility of academic misconduct, specifically cheating, by designing tasks that assess the language skill of mediation in Legal English exams. Such tasks require test-takers to generate original answers, which can significantly reduce the risk of copying and pasting. Additionally, mastering the skill of mediation is a valuable asset for future lawyers. The authors introduce two original distinct intra-language mediation tasks that they developed for assessing English for Legal Purposes. One is aimed at first-year students with limited Legal English experience, while the other is intended for more advanced and experienced students. The article analyses student responses and evaluates their performance. The authors follow the action research approach, which involves a circle of observing, reflecting and acting, and raise questions concerning the originality of answers, production of a measurable sample of the target language, objectivity of assessment, and whether the practice in class impacts the success rate. The article also discusses challenges encountered during the process and how they were addressed. It can be concluded that mediation tasks are practical for online testing but could be used equally well for traditional classroom assessment.

DiSSCo Prepare Project: Increasing the Implementation Readiness Levels of the European Research Infrastructure

The Distributed System of Scientific Collections (DiSSCo) is a new world-class Research Infrastructure (RI) for Natural Science Collections. The DiSSCo RI aims to create a new business model for one European collection that digitally unifies all European natural science assets under common access, curation, policies and practices that ensure that all the data is easily Findable, Accessible, Interoperable and Reusable (FAIR principles). DiSSCo represents the largest ever formal agreement between natural history museums, botanic gardens and collection-holding institutions in the world. DiSSCo entered the European Roadmap for Research Infrastructures in 2018 and launched its main preparatory phase project (DiSSCo Prepare) in 2020. DiSSCo Prepare is the primary vehicle through which DiSSCo reaches the overall maturity necessary for its construction and eventual operation. DiSSCo Prepare raises DiSSCo’s implementation readiness level (IRL) across the five dimensions: technical, scientific, data, organisational and financial. Each dimension of implementation readiness is separately addressed by specific Work Packages (WP) with distinct targets, actions and tasks that will deliver DiSSCo’s Construction Masterplan. This comprehensive and integrated Masterplan will be the product of the outputs of all of its content related tasks and will be the project’s final output. It will serve as the blueprint for construction of the DiSSCo RI, including establishing it as a legal entity. DiSSCo Prepare builds on the successful completion of DiSSCo’s design study, ICEDIG and the outcomes of other DiSSCo-linked projects such as SYNTHESYS+ and MOBILISE. This paper is an abridged version of the original DiSSCo Prepare grant proposal. It contains the overarching scientific case for DiSSCo Prepare, alongside a description of our major activities.

Examining Workflow for Simulated Pediatric Emergency Medical Services Care

Purpose: Workflow analysis is an important approach for understanding clinical care in complex systems. This is particularly true for low frequency events such as the care of children in the out-of-hospital setting. The purpose of this study was to characterize emergency medical service (EMS) workflow in the care of children during simulated emergency, prehospital encounters.
 Methods: This is a secondary analysis exploring high-fidelity videorecorded simulations, performed by emergency personnel. Two scenarios were used in 19 simulations which consisted of a 15-month and a 1-month old with respiratory decompensation and shock requiring intravenous fluid, respiratory support, and medication administration. One trained investigator performed review of the videos of teams EMS practitioners, quantifying the sequence and number of tasks performed and time to completion of the simulated intervention. The variance in sequence of tasks was quantified using the Levenshtein distance. We quantified the proportion of time with no activity (idle time) and temporal overlap (team multitasking time).
 Results: We identified 17 types of distinct tasks performed during the simulation. There was high variability across simulations in the sequences, types, and number of tasks performed. Team multitasking was noted in all scenarios, with a mean of 99% multitasking ratio (range: 52-202%). Mean proportion of idle time was 4% (range: 0-11%). Weight estimate, intravenous or intraosseous access, radio report, blood glucose level (BGL), medication administration, pulse check and respiration check were observed in all videos. Other tasks were observed in only a proportion of scenarios with varying frequencies. Median number of differences in sequence of tasks between scenarios was 15.
 Conclusions: Our analysis found a high percentage of team multitasking and significant variability in frequency and sequence of task completion. Further research is needed to assess the reasons and effect of this degree of variability on the efficiency and effectiveness of prehospital interventions.