Domain Constraints Research Articles

Concept-aware embedding for logical query reasoning over knowledge graphs

Logical query reasoning over knowledge graphs (KGs) is an important task for querying some information upon specified conditions. Despite recent advancements, existing methods typically focus on the inherent structure of logical queries and fail to capture the commonality among entities and relations, resulting in cascading errors during multi-hop inference. To mitigate this issue, we resort to inferring relations’ domain constraints based on the commonality of their connected entities implicitly. Specifically, to capture the domain constraints of relations, we treat the set of relations emitted by an entity as its implicit concept information and derive a relation’s domain constraint by aggregating the implicit concept information of its head entities. Employing a geometric-based embedding strategy, we enrich the representations of entities in the query with their implicit concept information. Additionally, we design a straightforward yet effective curriculum learning strategy to refine its reasoning skills. Notably, our model can be integrated into any existing query embedding-based logical query reasoning methods in a plug-and-play manner, enhancing their understanding of the entities as well as relations in queries. Experiments on three widely used datasets show that our model can achieve comparable outcomes and improve the performance of existing logical query reasoning models. Particularly, as a plug-in, it achieves an absolute improvement of the maximum 8.4% Hits@3 compared to the original model on the FB15k dataset, and it surpasses the former state-of-the-art plug-and-play logical query reasoning model in most scenes, exceeding it by up to 2.1% average Hits@3 results.

Initial state diversification for efficient AlphaZero-style training

AlphaZero has achieved superhuman performance in Go, chess, and shogi with a general reinforcement learning (RL) algorithm. This achievement is remarkable because AlphaZero does not rely on any training dataset of strong players. However, AlphaZero-style training requires substantial computational resources. Gumbel AlphaZero, a recently introduced more efficient version of AlphaZero, reduces the computational cost of AlphaZero training. The goal of this study is to further improve the playing strength of Gumbel AlphaZero under a limited amount of computational resources. We focus on the diversity in training games, inspired by procedural generation and domain randomization in RL studies, and propose a novel method, initial state diversification. This method diversifies the initial states of a self-play game to encourage the RL agent to understand the game in a more general manner through diverse experiences. For example, in shogi, the initial state of each self-play game is diversified by rearranging the pieces under realistic domain constraints. Experiments demonstrated that training with initial state diversification improves the playing strength of Gumbel AlphaZero in shogi, within the same computational budget for training.

Transient response improvement of predictive control under time domain constraints

In this paper, a novel predictive controller design approach within augmented structure base on convex optimization of an extra parameter is proposed in order to improve the transient response of discrete-time linear systems subject to input constraints. This methodology starts with the conception of an initial stabilizing predictive controller which ensures a fast closed loop response to the detriment of the other response specifications indexes (settling time, settling error, overshoot and control signal). Then, this controller is redesigned off-line using linear programming solver to meet a certain transient response specifications of the closed loop system determined before the design as time domain constraints. The resulting controlled output will track the reference as fast and as smooth as possible with a low energy consumption. Two examples are given to show the effectiveness of the developed controller.

Adaptive support constraint based on pixel difference distribution for twin-image removal in in-line holography

A tight support domain can enhance the reconstruction effect of phase retrieval algorithms using a support domain constraint. Hence, an adaptive support method based on pixel difference distribution to calculate a more accurate support domain is proposed in this paper. The support domain is continuously updated at each iteration of the phase retrieval algorithm, resulting in a support domain that closely corresponds to the shape and size of the object. Simulations are conducted to investigate the reconstruction effect of complex-valued objects as well as the influence of varying the binarization threshold and the difference window size on the convergence performance in the absence of noise. The changing dynamics of the support domain during the iterative process are also showcased. The experiments demonstrate that the adaptive support method can improve the reconstruction image quality and image contrast under actual conditions. Both simulations and experiments affirm the feasibility of this approach.

The intention of Egyptian healthcare workers to take the monkeypox vaccine: is urgent action required?

BackgroundIn light of the ongoing monkeypox (MPOX) epidemic, healthcare workers (HCWs) have been in contact with various diseases. Therefore, they should take appropriate preventive and control measures to maintain their health. This study assessed Egyptian HCWs’ intentions to take MPOX vaccines.MethodsA cross-sectional survey was conducted using social media platforms between September 27 and November 4, 2022. An anonymous online survey using the 5C scale was conducted using convenience and snowball sampling methods to assess the five psychological antecedents of vaccination (i.e., confidence, constraints, complacency, calculation, and collective responsibility).ResultsA total of 399 HCWs with a mean age of 32.6 ± 5.7 participated in this study. Of them, 89.7% were female. The five C psychological antecedents of vaccination were as follows: 55.9% were confident about vaccination, 50.6% were complacent, 56.6% experienced constraints, 60.7% calculated the risk and benefit, and 58.4% had collective responsibility. Multivariate analysis showed that high income level and having information about MPOX were significant predictors of confidence in the MPOX vaccines (adjusted odds ratio ((AOR) = 4.19, 95% CI (1.12– 15.59), P = 0.032). Participants aged 31–45 years and 19–30 years showed significant association (AOR = 2.46, 95% CI (0.85–7.15), P = 0.096) and (AOR = 4.19, 95% CI (1.39–12.64), P = 0.011), respectively. Having an idea about the MPOX vaccines significantly predicted the complacency domain (AOR = 3.77, 95%CI (1.47–9.65, P = 0.006). Moreover, precollege/undergraduate education and having an idea about MPOX vaccination were significant predictors of the constraint domain (AOR = 1.81.95% CI (1.09–2.99, P = 0.020), (AOR = 2.70, 95% CI (1.05–6.95, P = 0.038), respectively). Female sex, having a diploma, postgraduate studies, and having an idea about MPOX vaccine significantly predicted calculation domain (AOR = 2.06, 95% CI (1.05–4.04, P = 0.035), (AOR = 3.98,95% CI (1.33–11.87, P = 0.013), (AOR = 2.02, 95% CI (1.25–3.26, P = 0.004) & (AOR = 2.75. 95% CI (1.05–7.18, P = 0.039), respectively. The only significant predictor of collective responsibility was having a diploma and postgraduate studies (AOR = 3.44, 95% CI (1.21–9.78, P = 0.020), (AOR = 1.90,95% CI (1.17–3.09, P = 0.009).ConclusionsEfforts to control MPOX should focus on promoting protective measures such as the vaccination of HCWs as well as raising their awareness about the updated information regarding the virus and the approved vaccines.

An electronic nose drift compensation algorithm based on semi-supervised adversarial domain adaptive convolutional neural network

Sensor drift is a significant challenge leading to performance degradation in electronic nose(E-nose) systems. Effectively addressing sensor drift represents the most daunting problem in E-nose technology. This work proposes a domain adaptive approach called Semi-Supervised Adversarial Domain Adaptive Convolutional Neural Network (SAD-CNN) to tackle the long-term drift in E-nose and device displacement. SAD-CNN leverages adversarial learning to minimize distribution disparities between the source and target domains. Unlike traditional methods employing projection matrices, SAD-CNN utilizes one-dimensional convolutional neural networks as feature extractors, circumventing the complexities associated with parameter adjustments and matrix calculations in the projection process. During the training process, utilizing pseudo-labels generated by the semi-supervised self-training method to train the model, thereby reducing the labeling costs. Additionally, confidence threshold screening is introduced during the self-training phase to minimize erroneous pseudo-labels. Furthermore, the regenerative Hilbert space’s Maximum Mean Difference combined with Minimum Variance is introduced as a domain constraint function to mitigate distribution discrepancies between domains and enhance feature discriminability across domains. The experimental results demonstrate that the SAD-CNN method outperforms others. Within the long-term drift dataset, the classification accuracies for different scenarios are 78.01 % and 82.53 %, respectively. Meanwhile, the instrument change dataset yields a classification accuracy of 96.45 %.

Research on motion planning system for wall-climbing mobile manipulator for large steel structures welding operation

Purpose The purpose of this study is to address the welding demands within large steel structures by presenting a global spatial motion planning algorithm for a mobile manipulator. This algorithm is based on an independently developed wall-climbing robot, which comprises a four-wheeled climbing mobile platform and a six-degree-of-freedom robotic manipulator, ensuring high mobility and operational flexibility. Design/methodology/approach A convex hull feasible domain constraint is developed for motion planning in the mobile manipulator. For extensive spatial movements, connected sequences of convex polyhedra are established between the composite robot’s initial and target states. The composite robot’s path and obstacle avoidance optimization problem are solved by constraining the control points on B-spline curves. A dynamic spatial constraint rapidlye-xploring random trees-connect (RRTC) motion planning algorithm is proposed for the manipulator, which quickly generates reference paths using spherical spatial constraints at the manipulator’s end, eliminating the need for complex nonconvex constraint modeling. Findings Experimental results show that the proposed motion planning algorithm achieves optimal paths that meet task constraints, significantly reducing computation times in task conditions and shortening operation times in non-task conditions. Originality/value The algorithm proposed in this paper holds certain application value for the realization of automated welding operations within large steel structures using mobile manipulator.

Innovative air supply management in fuel cells: An economic predictive control approach without pre-measured OER

Existing fuel cell air supply control methods improve the economic performance of Proton Exchange Membrane Fuel Cells (PEMFC) under steady-state conditions through oxygen excess ratio curve, while ignoring the transient process of the air supply system under varying operating conditions. Furthermore, the oxygen excess ratio curve may exhibit deviations due to fuel cell aging and measurement inaccuracies, which would further diminish the economic performance derived from the control process. Not relying on pre-measuring the oxygen excess ratio curve, an EMPC (Economic Model Predictive Control) method is proposed for fuel cell air supply systems. It incorporates an economic cost function reflecting the economic objectives of the air supply system. By dynamically optimizing the parasitic power of the air compressor, this method enhances the overall economic performance by achieving performance optimization for both transient and steady-state processes of the fuel cell system. The asymptotic stability of PEMFC is achieved through the utilization of terminal penalty functions and terminal domain constraints. Stability and convergence analysis is carried out by employing auxiliary optimization problems and the Lyapunov method. In transient processes, EMPC method can achieve a maximum 4.97 % improvement in the economic performance of PEMFC. A hardware in loop (HIL) experiment was finally conducted to show the real-time capacity of the proposed EMPC method.

Fast Generation of Low PAPR and Low Sidelobe Noise Waveform

Noise radar waveform, as an important waveform in the field of low-probability-of-intercept radars, has been widely studied. However, the general noise waveform has a high peak average power ratio (PAPR), which leads to a reduction of signal-to-noise ratio. Thus, it is necessary to control the noise waveform PAPR to a suitable range. In this paper, the PAPR value reduction problem of low sidelobe noise waveform is studied. First, the proposed new algorithm is given a general overview. Then, the Combined Spectral Shaping and Peak-to-Average Power Ratio Reduction (COSPAR) algorithm is adjusted according to the alternate projection method and phase retrieval algorithm, and an improved COSPAR (ICOSPAR) algorithm with a faster execution speed and higher waveform generation efficiency is derived. The specific method performs waveform projection between the spectral constraint domain and the PAPR constraint domain and then evaluates the computational load of the ICOSPAR algorithm. Simulation results show that the ICOSPAR algorithm is reliable and efficient in generating low PAPRs and low sidelobe noise waveforms.

HEAL: High-Frequency Enhanced and Attention-Guided Learning Network for Sparse-View CT Reconstruction.

X-ray computed tomography (CT) imaging technology has become an indispensable diagnostic tool in clinical examination. However, it poses a risk of ionizing radiation, making the reduction of radiation dose one of the current research hotspots in CT imaging. Sparse-view imaging, as one of the main methods for reducing radiation dose, has made significant progress in recent years. In particular, sparse-view reconstruction methods based on deep learning have shown promising results. Nevertheless, efficiently recovering image details under ultra-sparse conditions remains a challenge. To address this challenge, this paper proposes a high-frequency enhanced and attention-guided learning Network (HEAL). HEAL includes three optimization strategies to achieve detail enhancement: Firstly, we introduce a dual-domain progressive enhancement module, which leverages fidelity constraints within each domain and consistency constraints across domains to effectively narrow the solution space. Secondly, we incorporate both channel and spatial attention mechanisms to improve the network's feature-scaling process. Finally, we propose a high-frequency component enhancement regularization term that integrates residual learning with direction-weighted total variation, utilizing directional cues to effectively distinguish between noise and textures. The HEAL network is trained, validated and tested under different ultra-sparse configurations of 60 views and 30 views, demonstrating its advantages in reconstruction accuracy and detail enhancement.

Active suspension hierarchical control with parameter uncertainty and external disturbance of electro-hydraulic actuators

In order to improve the ride comfort and handling stability of the electro-hydraulic active suspension system, a hierarchical control strategy is proposed. For the active suspension system with body mass uncertainty and safety constraints, an enhanced constraint adaptive backstepping controller is designed to generate the target force of body vertical motion. When dealing with constraints, nonlinear filters are introduced, backstepping technology and quadratic Lyapunov function are used to combine the control target and domain constraints, and the vertical displacement of the body and the suspension deflection control index are synthesized into a single controlled variable, so that the control variable converging to zero meets the suspension dynamic displacement limit. The ride comfort and safety of the active suspension system are improved. Secondly, stability analysis is conducted on the zero dynamic error system to ensure that all safety performance indicators are bounded. The effects of external disturbances, parameter uncertainties and modelling errors on the force tracking accuracy of asymmetric cylinder electrohydraulic systems are addressed. A backstepping dynamic surface control method based on a nonlinear perturbation observer is proposed, which estimates the composite perturbation terms such as modelling error and external perturbation, and uses the estimated value of the nonlinear observer to design a feedforward compensating control term to offset the effect of the composite perturbation on the system performance. In addition, the dynamic surface control method is used to calculate the derivative of the target force input, which avoids the derivative explosion phenomenon and reduces the computational complexity of the system. Simulation test results show that this method reduces the computational complexity of the system and improves the control performance. And under random and bumpy road conditions, the root-mean-square value of sprung mass acceleration performance index is reduced by 48.354 % and 72.677 % compared with passive suspension, which improves the ride comfort of the vehicle.

A Variational Surface-Evolution Approach to Optimal Transport over Transitioning Compact Supports with Domain Constraints

We examine the optimal mass transport problem in Rn between densities with transitioning compact support by considering the geometry of a continuous interpolating support boundary Γ in space-time within which the mass density evolves according to the fluid dynamical framework of Benamou and Brenier. We treat the geometry of this space-time embedding in terms of points, vectors, and sets in Rn+1=R×Rn and blend the mass density and velocity as well into a space-time solenoidal vector field W|Ω→Rn+1 over a compact set Ω⊂Rn+1. We then formulate a joint optimization for W and its support using the shaped gradient of the space-time surface Γ outlining the support boundary ∂Ω. This easily accommodates spatiotemporal constraints, including obstacles or mandatory regions to visit.

Spatial-Logic-Aware Weakly Supervised Learning for Flood Mapping on Earth Imagery

Flood mapping on Earth imagery is crucial for disaster management, but its efficacy is hampered by the lack of high-quality training labels. Given high-resolution Earth imagery with coarse and noisy training labels, a base deep neural network model, and a spatial knowledge base with label constraints, our problem is to infer the true high-resolution labels while training neural network parameters. Traditional methods are largely based on specific physical properties and thus fall short of capturing the rich domain constraints expressed by symbolic logic. Neural-symbolic models can capture rich domain knowledge, but existing methods do not address the unique spatial challenges inherent in flood mapping on high-resolution imagery. To fill this gap, we propose a spatial-logic-aware weakly supervised learning framework. Our framework integrates symbolic spatial logic inference into probabilistic learning in a weakly supervised setting. To reduce the time costs of logic inference on vast high-resolution pixels, we propose a multi-resolution spatial reasoning algorithm to infer true labels while training neural network parameters. Evaluations of real-world flood datasets show that our model outperforms several baselines in prediction accuracy. The code is available at https://github.com/spatialdatasciencegroup/SLWSL.

DeepSaDe: Learning Neural Networks That Guarantee Domain Constraint Satisfaction

As machine learning models, specifically neural networks, are becoming increasingly popular, there are concerns regarding their trustworthiness, specially in safety-critical applications, e.g. actions of an autonomous vehicle must be safe. There are approaches that can train neural networks where such domain requirements are enforced as constraints, but they either cannot guarantee that the constraint will be satisfied by all possible predictions (even on unseen data) or they are limited in the type of constraints that can be enforced. In this paper, we present an approach to train neural networks which can enforce a wide variety of constraints and guarantee that the constraint is satisfied by all possible predictions. The approach builds on earlier work where learning linear models is formulated as a constraint satisfaction problem (CSP). To make this idea applicable to neural networks, two crucial new elements are added: constraint propagation over the network layers, and weight updates based on a mix of gradient descent and CSP solving. Evaluation on various machine learning tasks demonstrates that our approach is flexible enough to enforce a wide variety of domain constraints and is able to guarantee them in neural networks.

Significance and Optimization of Operating Parameters in Hydrothermal Carbonization Using RSM–CCD

To ascertain the significance of temperature and residence time of hydrothermal carbonization (HTC) in controlling hydrochar production, multiple regression was employed based on central composite design (CCD) to model the responses of mass yield (MY) and higher heating value (HHV). The hydrothermal reaction was explored at temperatures and times ranging from 150 to 250 °C and 0.5 to 3.5 h. Sorghum bagasse (SB) and microalgae (MA) were used to complex the reaction due to their differences in organic constituents. Simultaneously, the operating parameters were optimized by maximizing the response values under domain constraints in the HHV models. The results show that at least temperature and time in the linear system played a significant role in determining the solids recovery and the energy generation of hydrochars (p-values = 0.00), regardless of the biomass type. Moreover, the optimum conditions of SB and MA hydrochars can be achieved by increasing the temperature to the limit of 250 °C and prolonging the time to 3.5 and 3.25 h, respectively. Both respective conditions resulted in maximum HHVs of 27.54 and 35.83 MJ kg−1.

Physics-informed ConvNet: Learning physical field from a shallow neural network

We introduce a novel methodology for solving nonlinear partial differential equation (PDE) on regular or irregular domains using physics-informed ConvNet, which we call the PICN. The network structure consists of three parts: 1) a convolutional neural network for physical field generation, 2) a pre-trained convolutional layer corresponding to the finite-difference filters to estimate differential fields of the generated physical field, and 3) an interpolation network for loss analysis in irregular geometry domains. From a CNN perspective, the physical field is generated by a deconvolution layer and a convolution layer. Unlike the standard Physics-informed Neural Network (PINN) approach, the convolutions corresponding to the finite-difference filters estimate the spatial gradients forming the physical operator and then construct the PDE residual in a PINN-like loss function. The total loss function involving boundary conditions and the physical constraints in irregular geometry domains can be calculated from an efficient linear interpolation network. The theoretical analysis of PICN convergence is performed on a simplified case for solving a one-dimensional physical field, and several examples of nonlinear PDE of solutions with multifrequency characteristics are executed. The theory and examples confirm the effective learning capability of PICN for the physical field solution with high-frequency components, compared to the standard PINN. A series of numerical cases are performed to validate the current PICN, including the solving (and estimation) of nonlinear physical operator equations and recovering physical information from noisy observations. First, the ability of PICN to solve nonlinear PDE has been verified by executing three nonlinear problems including ODE with sine nonlinearity, PDE involving nonlinear sine-square operators, and Schrödinger equation. The proposed PICN has been assessed by solving some nonlinear PDE on irregular domains such as star-shaped domain, bird-like domain, and starfish domain. Moreover, PICN is applied to identify the thermal diffusivity parameters in an anisotropic heat transfer problem from noisy data, and a denoising display of the temperature field from strong noisy data with standard deviations ranging from 0.1 to 0.4. The numerical results demonstrate the high accuracy approximation and fast convergence performance of PICN. The potential advantage in approximating complex physical field with multi-frequency components indicates that PICN may become an alternative efficient neural network solver in physics-informed machine learning. This paper is adapted from the work originally posted on arXiv.com by the same authors (arXiv:2201.10967, Jan 26, 2022). The data and code accompanying this paper are publicly available at https://github.com/zengzhi2015/PICN.

An Empirical Study of the Code Generation of Safety-Critical Software Using LLMs

In the digital era of increasing software complexity, improving the development efficiency of safety-critical software is a challenging task faced by academia and industry in domains such as nuclear energy, aviation, the automotive industry, and rail transportation. Recently, people have been excited about using pre-trained large language models (LLMs) such as ChatGPT and GPT-4 to generate code. Professionals in the safety-critical software field are intrigued by the code generation capabilities of LLMs. However, there is currently a lack of systematic case studies in this area. Aiming at the need for automated code generation in safety-critical domains such as nuclear energy and the automotive industry, this paper conducts a case study on generating safety-critical software code using GPT-4 as the tool. Practical engineering cases from the industrial domain are employed. We explore different approaches, including code generation based on overall requirements, specific requirements, and augmented prompts. We propose a novel prompt engineering method called Prompt-FDC that integrates basic functional requirements, domain feature generalization, and domain constraints. This method improves code completeness from achieving 30% functions to 100% functions, increases the code comment rate to 26.3%, and yields better results in terms of code compliance, readability, and maintainability. The code generation approach based on LLMs also introduces a new software development process and V-model lifecycle for safety-critical software. Through systematic case studies, we demonstrate that, with appropriate prompt methods, LLMs can auto-generate safety-critical software code that meets practical engineering application requirements. It is foreseeable that LLMs can be applied to various engineering domains to improve software safety and development efficiency.

Extensive measurement in social choice

Extensive measurement is the standard measurement‐theoretic approach for constructing a ratio scale. It involves the comparison of objects that can be concatenated in an additively representable way. This paper studies the implications of extensively measurable welfare for social choice theory. We do this in two frameworks: an Arrovian framework with a fixed population and no interpersonal comparisons, and a generalized framework with variable populations and full interpersonal comparability. In each framework we use extensive measurement to introduce novel domain restrictions, independence conditions, and constraints on social evaluation. We prove a welfarism theorem for these domains and characterize the social welfare functions that satisfy the axioms of extensive measurement at both individual and social levels. The main results are simple axiomatizations of strong dictatorship in the Arrovian framework and classical utilitarianism in the generalized framework.

Unsupervised Domain Adaptation Fundus Image Segmentation via Multi-Scale Adaptive Adversarial Learning.

Segmentation of the Optic Disc (OD) and Optic Cup (OC) is crucial for the early detection and treatment of glaucoma. Despite the strides made in deep neural networks, incorporating trained segmentation models for clinical application remains challenging due to domain shifts arising from disparities in fundus images across different healthcare institutions. To tackle this challenge, this study introduces an innovative unsupervised domain adaptation technique called Multi-scale Adaptive Adversarial Learning (MAAL), which consists of three key components. The Multi-scale Wasserstein Patch Discriminator (MWPD) module is designed to extract domain-specific features at multiple scales, enhancing domain classification performance and offering valuable guidance for the segmentation network. To further enhance model generalizability and explore domain-invariant features, we introduce the Adaptive Weighted Domain Constraint (AWDC) module. During training, this module dynamically assigns varying weights to different scales, allowing the model to adaptively focus on informative features. Furthermore, the Pixel-level Feature Enhancement (PFE) module enhances low-level features extracted at shallow network layers by incorporating refined high-level features. This integration ensures the preservation of domain-invariant information, effectively addressing domain variation and mitigating the loss of global features. Two publicly accessible fundus image databases are employed to demonstrate the effectiveness of our MAAL method in mitigating model degradation and improving segmentation performance. The achieved results outperform current state-of-the-art (SOTA) methods in both OD and OC segmentation.

The complexity of ethical assessment: Interdisciplinary challenge for character education.

The assessment of learning in the ethical domain is one of the most complex aspects to attend in the educational context. In recent years, character education has contributed greatly to different social disciplines, such as education or nursing. However, the development of this approach has run up against several obstacles and limitations, as there is little evidence regarding its long-term effectiveness or its evaluation. This essay aims to identify some of the main difficulties to assess learning in the ethical domain, as obstacles and possible constraints to Aristotelian-based character education. Methodology is analytical and of a philosophical-educational nature, based on which an argumentative analysis is constructed from the bibliographical review of the contributions of classical and contemporary authors. Results show the existence of four major problems and ten associated subproblems, both of an external nature linked to the contextual factors of assessment, and of an internal nature that affect the essence of the process itself, which highlight the difficulty of carrying out assessments of an ethical kind. Far from proposing a pessimistic position, we argue a realistic vision that allows educators at different educational levels to be aware of the limits and critical points of evaluation in the ethical domain.