Task Analysis Research Articles

A novel model based on a transformer for intent detection and slot filling

Building task-oriented dialogue systems has become a topic of interest in the research community and industry. The task-oriented dialogue system is a closed-domain dialogue system that can perform specific tasks for users. The natural language understanding module of a task-oriented dialogue system is crucial because it is related to a task-oriented dialogue system that provides correctional services for users. The natural language understanding module of a task-oriented dialogue system performs two tasks: intent detection and slot filling. The intent detection task can be regarded as a text classification task; a classification model is trained to predict the intention of the user from the user’s input information. The slot filling task can be regarded as a sequence analysis task; a sequence analysis model is trained to predict the details of the user’s intention. In this paper, we proposed a novel model based on a transformer encoder for intent detection and slot filling. It follows the encoder-decoder structure, including a vanilla Transformer encoder, a bidirectional LSTM encoder, a linear classification decoder for intent detection, and a conditional random field decoder for slot filling. The experimental results on two public datasets show that our proposed model outperforms the existing methods based on the Transformer and can be combined with BERT to achieve better intent detection and slot filling results.

Enhancing the vision–language foundation model with key semantic knowledge-emphasized report refinement

Recently, vision–language representation learning has made remarkable advancements in building up medical foundation models, holding immense potential for transforming the landscape of clinical research and medical care. The underlying hypothesis is that the rich knowledge embedded in radiology reports can effectively assist and guide the learning process, reducing the need for additional labels. However, these reports tend to be complex and sometimes even consist of redundant descriptions that make the representation learning too challenging to capture the key semantic information. This paper develops a novel iterative vision–language representation learning framework by proposing a key semantic knowledge-emphasized report refinement method. Particularly, raw radiology reports are refined to highlight the key information according to a constructed clinical dictionary and two model-optimized knowledge-enhancement metrics. The iterative framework is designed to progressively learn, starting from gaining a general understanding of the patient’s condition based on raw reports and gradually refines and extracts critical information essential to the fine-grained analysis tasks. The effectiveness of the proposed framework is validated on various downstream medical image analysis tasks, including disease classification, region-of-interest segmentation, and phrase grounding. Our framework surpasses seven state-of-the-art methods in both fine-tuning and zero-shot settings, demonstrating its encouraging potential for different clinical applications.

Optimization Algorithms for Real-Time Image Processing in IoT Networks

This study explores the application of optimization algorithms, with a focus on Particle Swarm Optimization (PSO), to enhance real-time image processing in Internet of Things (IoT) networks. Given the constraints of computational power and energy resources in IoT devices, efficient processing of image data is a critical challenge. The proposed PSO-based framework aims to minimize processing time and energy consumption while maintaining high accuracy in image analysis tasks. Experimental results demonstrate that PSO can reduce processing time by 10% and energy consumption by 15% on average, with a slight improvement in accuracy. Comparisons with traditional and other evolutionary optimization techniques reveal that PSO provides a superior balance between efficiency and performance, making it particularly well-suited for real-time IoT applications. The findings suggest that the integration of advanced optimization algorithms like PSO into IoT systems can significantly enhance their operational efficiency, scalability, and effectiveness, particularly in resource-constrained environments.

Executive functions in children with developmental language disorder: a systematic review and meta-analysis.

The purpose of the current study was to assess the differences between children with developmental language disorder (DLD) and typically developing (TD) children in their performance of executive functions from working memory, inhibitory control, and cognitive flexibility. We performed a systematical search of PubMed, Embase, Cochrane, and Web of Science for case control studies (published in English between January 1, 1950, and October 11, 2023) comparing the differences in the performance of executive functions between DLD and TD children. Forty eligible studies were included in the present study (N = 3,168 participants). In comparison with TD children, DLD children exhibited significantly poorer performances in all six verbal working memory tasks (backward digit recall task, SMD -1.4321, 95% CI -2.2692 to -0.5950; listening recall task, SMD -1.4469, 95% CI -1.7737 to -1.1202; counting recall task, SMD -0.9192, 95% CI -1.4089 to -0.4295; digit recall task, SMD -1.2321, 95% CI -1.4397 to -1.0244; word list recall task, SMD -1.1375, 95% CI -1.5579 to -0.7171; non-word recall task, SMD -1.5355, 95% CI -1.8122 to -1.2589). However, regarding inhibitory control and cognitive flexibility, the differences between DLD and TD children depended on specific circumstances. In subgroup analyses of all verbal working memory tasks, DLD children presented notably lower performance than TD children in both the monolingual English and monolingual non-English groups, and in both the preschooler and school-aged groups. This study proves that verbal working memory deficits can be seen as a marker for children with DLD and are not affected by age or language type. https://www.crd.york.ac.uk/PROSPERO/display_record.php?RecordID=391070, CRD42023391070.

Augmentations of Forman’s Ricci curvature and their applications in community detection

Abstract The notion of curvature on graphs has recently gained traction in the networks community, with the Ollivier–Ricci curvature (ORC) in particular being used for several tasks in network analysis, such as community detection. In this work, we choose a different approach and study augmentations of the discretization of the Ricci curvature proposed by Forman (AFRC). We empirically and theoretically investigate its relation to the ORC and the un-augmented Forman–Ricci curvature. In particular, we provide evidence that the AFRC frequently gives sufficient insight into the structure of a network to be used for community detection, and therefore provides a computationally cheaper alternative to previous ORC-based methods. Our novel AFRC-based community detection algorithm is competitive with an ORC-based approach.

BERT Model with Fuzzy Logic Optimization on Multivariate Sentiment Analysis Tasks

BERT Model with Fuzzy Logic Optimization on Multivariate Sentiment Analysis Tasks

Lightweight Progressive Fusion Calibration Network for Rotated Object Detection in Remote Sensing Images

Rotated object detection is a crucial task in aerial image analysis. To address challenges such as multi-directional object rotation, complex backgrounds with occlusions, and the trade-off between speed and accuracy in remote sensing images, this paper introduces a lightweight progressive fusion calibration network for rotated object detection (LPFC-RDet). The network comprises three main modules: the Retentive Meet Transformers (RMT) feature extraction block, the Progressive Fusion Calibration module (PFC), and the Shared Group Convolution Lightweight detection head (SGCL). The RMT feature extraction block integrates a retentive mechanism with global context modeling to learn rotation-insensitive features. The PFC module employs pixel-level, local-level, and global-level weights to calibrate features, enhancing feature extraction from occluded objects while suppressing background interference. The SGCL detection head uses decoupled detection tasks and shared group convolution layers to achieve parameter sharing and feature interaction, improving accuracy while maintaining a lightweight structure. Experimental results demonstrate that our method surpasses state-of-the-art detectors on three widely used remote sensing object datasets: HRSC2016, UCAS_AOD, and DOTA.

Riemannian geometry for efficient analysis of protein dynamics data

An increasingly common viewpoint is that protein dynamics datasets reside in a nonlinear subspace of low conformational energy. Ideal data analysis tools should therefore account for such nonlinear geometry. The Riemannian geometry setting can be suitable for a variety of reasons. First, it comes with a rich mathematical structure to account for a wide range of geometries that can be modeled after an energy landscape. Second, many standard data analysis tools developed for data in Euclidean space can be generalized to Riemannian manifolds. In the context of protein dynamics, a conceptual challenge comes from the lack of guidelines for constructing a smooth Riemannian structure based on an energy landscape. In addition, computational feasibility in computing geodesics and related mappings poses a major challenge. This work considers these challenges. The first part of the paper develops a local approximation technique for computing geodesics and related mappings on Riemannian manifolds in a computationally feasible manner. The second part constructs a smooth manifold and a Riemannian structure that is based on an energy landscape for protein conformations. The resulting Riemannian geometry is tested on several data analysis tasks relevant for protein dynamics data. In particular, the geodesics with given start- and end-points approximately recover corresponding molecular dynamics trajectories for proteins that undergo relatively ordered transitions with medium-sized deformations. The Riemannian protein geometry also gives physically realistic summary statistics and retrieves the underlying dimension even for large-sized deformations within seconds on a laptop.

Augmented contour scoring snake for instance segmentation of placental separable villi

Segmenting terminal villous structures as separable instances is a prerequisite task for quantitative and explainable analysis of placental histopathology. Inspired by the fact that villous structure is typically surrounded by syncytiotrophoblast cells which yield critical hints for segmentation, we focus on designing a contour-based instance segmentation method. Previous contour-based methods usually utilize the confidence of object localization (or classification) as the instance score, without contour scores which explicitly measure the contour refinement quality (i.e., the distance discrepancy between the predicted contour and its ground truth). In this paper, we propose an augmented contour Scoring Snake framework, termed as SSnake, which learns both contour deformation and refinement quality estimation. To form contour quality measurement, we devise an axis-aware size-adaptive smooth function mapping from predicted contours to normalized scores in point resolution. Besides, to estimate scores of long-distance cases and strengthen the score learning process, a contour-augmented training scheme is designed for initial box contours. Our method recalibrates instance scores using contour quality by prioritizing instances with finer contours. Experimental results on the in-house separable villi segmentation dataset and a public cell nucleus segmentation dataset demonstrate that our proposed method significantly outperforms all competitors including state-of-the-art approaches. In villi segmentation dataset, our proposed SSnake achieves 3.7% improvements in COCO APm over the baseline DeepSnake. The source code is available at https://github.com/Psilym/SSnake.

The Multiple Roles of Structured Multipurpose Databases in Fusion Research

Most databases in fusion research are devoted to a single topic, such as energy confinement, H-modes, profiles, or disruptions. To allow for a wider range of analysis, modeling, and validation tasks, the JETPEAK broad-based multipurpose database has been developed for JET. This database currently includes 27 065 stationary state (∂/∂t ≈ 0) samples and nearly 1000 scalar, one-dimensional (profiles), and two-dimensional (R and Z dependent) variables grouped into topical structures. A similar database has been created for the Tokamak à Configuration Variable (TCV), comprising 65 000 samples reaching back to early TCV operation in the 1990s. The breadth and flexibility of these databases allows them to be used for a wide variety of investigations such as modeling tasks, confinement scaling, testing, validation and benchmarking of algorithms and modeling codes, and long-term monitoring of device conditions, as well as for documentation.

Measuring the Company’s Capabilities and Assessing their Impact on Its Economic Results: Analytic Tools

With the increasing uncertainty of the external environment, the capabilities of companies become the driving force behind their operations and development, making the management of these capabilities an important task for management. The aim of the research is to develop a toolkit for measuring a company’s capabilities, assessing their condition, and evaluating their impact on productivity. The author’s concept is presented, according to which the final results of the company’s activity are formed through the mechanism of interaction of the company’s organizational capability with acquired resources, the formation of resource capabilities (production, financial, labor, and market) and their transformation into tangible and intangible products. Based on this model, methodological guidelines have been developed for measuring the capabilities of organizations and assessing their impact on the final results, a value-functional approach to selecting appropriate indicators, and analytical tools. The following methods have been substantiated: modeling, the indicator method, data convolution, scaling, and statistical methods. To measure capabilities, it is proposed to use indicators aggregated into composite indices through a developed algorithm. The use of sales growth rate, market capitalization, profit before taxes, and net cash flow as the final results of the company (productivity indicators) has been justified. To solve analytical tasks at different levels of management, combinations of methods, tools and indicators called “research formats” (detailed, diagnostic, express diagnostic analysis) were developed. The results of the empirical testing of the author’s research are presented: through detailed analysis, the validity of the proposed approach and tools has been established, and through diagnostic analysis, the stability of the selected indicators and constructed mathematical models beyond the observation period was identified. The result is a toolkit for researching crucial company’s capabilities, enabling management to analyze and monitor these capabilities, assess their productivity, make forecasts of productivity indicators, and develop a system for managing the organization’s capabilities.

On the applications of neural ordinary differential equations in medical image analysis

Medical image analysis tasks are characterized by high-noise, volumetric, and multi-modality, posing challenges for the model that attempts to learn robust features from the input images. Over the last decade, deep neural networks (DNNs) have achieved enormous success in medical image analysis tasks, which can be attributed to their powerful feature representation capability. Despite the promising results reported in numerous literature, DNNs are also criticized for several pivotal limits, with one of the limitations is lack of safety. Safety plays an important role in the applications of DNNs during clinical practice, helping the model defend against potential attacks and preventing the model from silent failure prediction. The recently proposed neural ordinary differential equation (NODE), a continuous model bridging the gap between DNNs and ODE, provides a significant advantage in ensuring the model’s safety. Among the variants of NODE, the neural memory ordinary differential equation (nmODE) owns the global attractor theoretically, exhibiting superiority in prompting the model’s performance and robustness during applications. While NODE and its variants have been widely used in medical image analysis tasks, there is a lack of a comprehensive review of their applications, hindering the in-depth understanding of NODE’s working principle and its potential applications. To mitigate this limitation, this paper thoroughly reviews the literature on the applications of NODE in medical image analysis from the following five aspects: segmentation, reconstruction, registration, disease prediction, and data generation. We also summarize both the strengths and downsides of the applications of NODE, followed by the possible research directions. To the best of our knowledge, this is the first review regards the applications of NODE in the field of medical image analysis. We hope this review can draw the researchers’ attention to the great potential of NODE and its variants in medical image analysis.

Optimizing task-analysis instruction: Effects of descriptions and pictures of antecedent stimuli and outcomes.

Little research has examined specific instructional variables that influence the development and effectiveness of task-analysis instruction. We conducted two experiments using text-based task analyses to teach college students to create single-subject reversal design graphs. In Experiment 1, we tested the effects of presenting antecedent and outcome stimuli on graphing performance (accuracy, yield, time to completion). Different groups of participants experienced graphing tutorials with descriptions and pictures of (a) responses; (b) antecedent stimuli and responses; (c) responses and outcomes of correct responses; and (d) antecedent stimuli, responses, and outcomes. In Experiment 2, we compared tutorials with and without pictures. Collectively, the results suggest that graphing accuracy was positively affected by task analyses that included pictures and descriptions of antecedent stimuli and that adding outcome stimuli further benefited graphing accuracy. These results suggest critical instructional elements that should be included in future task analyses of graphing or other complex behavior chains.

Investigating and improving student understanding of the basics of quantum computing

Quantum information science and engineering (QISE) is a rapidly developing field that leverages the skills of experts from many disciplines to utilize the potential of quantum systems in a variety of applications. It requires talent from a wide variety of traditional fields, including physics, engineering, chemistry, and computer science, to name a few. To prepare students for such opportunities, it is important to give them a strong foundation in the basics of QISE, in which quantum computing plays a central role. In this study, we discuss the development, validation, and evaluation of a Quantum Interactive Learning Tutorial, on the basics and applications of quantum computing. These include an overview of key quantum mechanical concepts relevant to quantum computation (including ways a quantum computer is different from a classical computer), properties of single- and multiqubit systems, and the basics of single-qubit quantum gates. The tutorial uses guided inquiry-based teaching-learning sequences. Its development and validation involved conducting cognitive task analysis from both expert and student perspectives and using common student difficulties as a guide. For example, before engaging with the tutorial, after traditional lecture-based instruction, one reasoning primitive that was common in student responses is that a major difference between an N-bit classical and N-qubit quantum computer is that various things associated with a number N for a classical computer should be replaced with the number 2N for a quantum computer (e.g., 2N qubits must be initialized and 2N bits of information are obtained as the output of the computation on the quantum computer). This type of reasoning primitive also led many students to incorrectly think that there are only N distinctly different states available when computation takes place on a classical computer. Research suggests that this type of reasoning primitive has its origins in students learning that quantum computers can provide exponential advantage for certain problems, e.g., Shor’s algorithm for factoring products of large prime numbers, and that the quantum state during the computation can be in a superposition of 2N linearly independent states. The inquiry-based learning sequences in the tutorial provide scaffolding support to help students develop a functional understanding. The final version of the validated tutorial was implemented in two distinct courses offered by the physics department with slightly different student populations and broader course goals. Students’ understanding was evaluated after traditional lecture-based instruction on the requisite concepts and again after engaging with the tutorial. We analyze and discuss their improvement in performance on concepts covered in the tutorial. Published by the American Physical Society 2024

A New Student Performance Prediction Method Based on Belief Rule Base with Automated Construction

Student performance prediction (SPP) is a pivotal task in educational analytics, enabling proactive interventions and optimized resource allocation by educators. Traditional SPP models are often hindered by their complexity and lack of interpretability. This study introduces a novel SPP framework, the Belief Rule Base with automated construction (Auto–BRB), designed to address these issues. Firstly, reference values are derived through data mining techniques. The model employs an IF–THEN rule-based system integrated with evidential reasoning to ensure both transparency and interpretability. Secondly, parameter optimization is achieved using the Projected Covariance Matrix Adaptive Evolution Strategy (P–CMA–ES), significantly enhancing model accuracy. Moreover, the Akaike Information Criterion (AIC) is then applied to fine-tune the balance between model accuracy and complexity. Finally, case studies on SPP have shown that the Auto–BRB model has an advantage over traditional models in terms of accuracy, while maintaining good interpretability. Therefore, Auto–BRB has excellent application effects in educational data analysis.

Characteristics of opinions in the societal and non-societal domains

With the increasing availability of user opinions on the web, understanding the distinct nature of opinions in societal and non-societal contexts becomes crucial for opinion mining and sentiment analysis tasks. Societal topics, encompassing social unrest, terrorist acts, and government policies, differ significantly from non-societal topics like product reviews, movie reviews, and restaurant reviews. Given the regional specificity of societal issues and the lack of sentiment-annotated resources for them, this paper highlights the need to comprehend the differences in opinions between these domains for effective sentiment analysis. Through statistical text and network analysis, it investigates word usage, sentiment word association, and homogeneity in societal versus non-societal contexts. The study also explores graph-based analysis as a novel approach to sentiment analysis, considering its advantage in easily expanding context through the addition of nodes, as opposed to the complexity of inserting relevant tokens in text. The findings suggest that while non-societal sentiment resources might not be directly applicable to societal domains, graph-based analysis offers promising avenues for sentiment analysis in diverse societal topics.

Two-layer visual analytics of truckers’ risk-coping social network

Within organizations, managers’ specific responsibilities and domain expertise shape their interests in the output of social network analysis. Our proposed visualization approach is tailored to meet the operation-directed needs and preferences for visual analysis of specific tasks. This method prioritizes an overall geographical map with focal-contextual dynamics within the network. To enable a comprehensive and in-depth understanding of pinpointed focal areas, we customize an analytical framework for analyzing inter-community networks. We extract focal sub-networks from specific nodes to create graph visualization for detailed analysis, represent rich types of domain-specific graphic properties, and provide direct zoom+filtering interactions to allow easy pattern recognition and knowledge discovery. We applied our approach to visualizing the data from interactions among 300 city-based truck communities on the largest occupational platform for truckers in China. We also conduct a case study to demonstrate that our approach is effective in supporting managers’ network analysis and knowledge discovery.

Immersive analytics with augmented reality in meteorology: an exploratory study on ontology and linked data

Although Augmented Reality (AR) has been extensively studied in supporting Immersive Analytics (IA), there are still many challenges in visualising and interacting with big and complex datasets. To deal with these datasets, most AR applications utilise NoSQL databases for storing and querying data, especially for managing large volumes of unstructured or semi-structured data. However, NoSQL databases have limitations in their reasoning and inference capabilities, which can result in insufficient support for certain types of queries. To fill this gap, we aim to explore and evaluate whether an intelligent approach based on ontology and linked data can facilitate visual analytics tasks with big datasets on AR interface. We designed and implemented a prototype of this method for meteorological data analytics. An experiment was conducted to evaluate the use of a semantic database with linked data compared to a conventional approach in an AR-based immersive analytics system. The results significantly highlight the performance of semantic approach in helping the users analysing meteorological datasets and their subjective appreciation in working with the AR interface, which is enhanced with ontology and linked data.

TimeCSL: Unsupervised Contrastive Learning of General Shapelets for Explorable Time Series Analysis

Unsupervised (a.k.a. Self-supervised) representation learning (URL) has emerged as a new paradigm for time series analysis, because it has the ability to learn generalizable time series representation beneficial for many downstream tasks without using labels that are usually difficult to obtain. Considering that existing approaches have limitations in the design of the representation encoder and the learning objective, we have proposed Contrastive Shapelet Learning (CSL), the first URL method that learns the general-purpose shapelet-based representation through unsupervised contrastive learning, and shown its superior performance in several analysis tasks, such as time series classification, clustering, and anomaly detection. In this paper, we develop TimeCSL, an end-to-end system that makes full use of the general and interpretable shapelets learned by CSL to achieve explorable time series analysis in a unified pipeline. We introduce the system components and demonstrate how users interact with TimeCSL to solve different analysis tasks in the unified pipeline, and gain insight into their time series by exploring the learned shapelets and representation.

Tailorable Sampling for Progressive Visual Analytics.

Progressive visual analytics (PVA) allows analysts to maintain their flow during otherwise long-running computations by producing early, incomplete results that refine over time, for example, by running the computation over smaller partitions of the data. These partitions are created using sampling, whose goal it isto draw samples of the dataset such that the progressive visualization becomes as useful as possible as soon as possible. What makes the visualization useful depends on the analysis task and, accordingly, some task-specific sampling methods have been proposed for PVA to address this need. However, as analysts see more and more of their data during the progression, the analysis task at hand often changes, which means that analysts need to restart the computation to switch the sampling method, causing them to lose their analysis flow. This poses a clear limitation to the proposed benefits of PVA. Hence, we propose a pipeline for PVA-sampling that allows tailoring the data partitioning to analysis scenarios by switching out modules in a way that does not require restarting the analysis. To that end, we characterize the problem of PVA-sampling, formalize the pipeline in terms of data structures, discuss on-the-fly tailoring, and present additional examples demonstrating its usefulness.