Industrial processes with multiple operating grades have become increasingly important in satisfying the requirements of agile manufacturing and a diversified market. However, because unknown distribution discrepancy process data collected from different grades, development reliable quality prediction models is still intractable, especially for limited measurements. In this study, novel framework an adversarial transfer learning (ATL)-based soft sensing method was designed inferring...

Soft sensor modeling for dynamic processes has become a trending topic and pending challenge in industrial data analysis, especially limited labeled scenarios. Alternatively, augmentation strategies provide way to address the deficiency of samples. However, current time-series methods do not consider spatiotemporal dependencies among samples during generation procedure. To issue, denoising diffusion probabilistic model (TimeDDPM) is proposed construct soft finite First, long short-term...

As the acquisition of variables that measure quality is typically challenging, labeled samples for building a model soft sensors are often inadequate. Additionally, owing to installation redundant sensors, high-dimensional process data with strong correlations acquired. Therefore, in this study, sample selection strategy active learning (AL), referred as latent-enhanced variational adversarial AL (LVAAL), developed enhance prediction performance limited data. The LVAAL method uses minimax...

In this paper, we study Reinforcement Learning from Demonstrations (RLfD) that improves the exploration efficiency of (RL) by providing expert demonstrations. Most existing RLfD methods require demonstrations to be perfect and sufficient, which yet is unrealistic meet in practice. To work on imperfect demonstrations, first define an setting for a formal way, then point out previous suffer two issues terms optimality convergence, respectively. Upon theoretical findings have derived, tackle...

In the present study, to understand mechanism of Mn on inhibiting Fe-caused Mg corrosion, corrosion behaviour commercial pure and Mg-6Mn alloy in 0.6 M NaCl solution is investigated. It found that alloy, Fe impurity incorporated into form (Fe) phase with as solid solute. The initial galvanic cannot be reduced through converting Fe-rich phase, since also has relatively strong cathodic activity much larger volume fraction than phase. However, activation inhibited. even decreases for Mg-Mn...

Quality prediction is beneficial to intelligent inspection, advanced process control, operation optimization, and product quality improvements of complex industrial processes. Most the existing work obeys assumption that training samples testing follow similar data distributions. The is, however, not true for practical multimode processes with dynamics. In practice, traditional approaches mostly establish a model using from principal operating mode (POM) abundant samples. inapplicable other...

Supervised latent variable regression methods such as partial least squares (PLS) and dynamic PLS have found wide applications in data analytics, quality prediction, fault monitoring various industries. In this article, we tackle the unbalanced problem of sparse measurement abundant process control systems to make use all samples for modeling. A novel semi-supervised (SemiDLVR) method is proposed prediction quality-relevant monitoring. The SemiDLVR integrates Laplacian manifold...

Most software defect prediction models usually assume that enough historical training instances with labels are available. Additionally, the data and predicted should share same features to ensure accuracy. However, in practice, there many datasets different granularities containing information dimensions. Therefore, it is valuable effectively use small scale dimensions of as improve performance model. We propose a heterogeneous orienting multiview transfer learning for prediction, denoted...

As a new classification platform, deep learning has recently received increasing attention from researchers and been successfully applied to many domains. In some domains, like bioinformatics robotics, it is very difficult construct large-scale well-annotated dataset due the expense of data acquisition costly annotation, which limits its development. Transfer relaxes hypothesis that training must be independent identically distributed (i.i.d.) with test data, motivates us use transfer solve...

Targeting the problem of traditional sparrow search algorithms being prone to falling into local optima, a new algorithm called Chaotic Sparrow Search Algorithm with Manta Ray Spiral Foraging (abbreviated as MSSA) is proposed. The Logistic-Sine-Cosine chaotic map and elite Reverse learning strategy are fused initialize population. It experimentally demonstrated that this hybrid outperforms population after random initialization in reducing ineffective individuals. In vigilante update stage,...

Abstract The quantum many-body problems are important for condensed matter physics, however solving the challenging because Hilbert space grows exponentially with size of problem. recently developed deep learning methods provide a promising new route to solve long-standing problems. We report that based simulation can achieve solutions competitive precision spin <?CDATA $J1$?> <mml:math xmlns:mml="http://www.w3.org/1998/Math/MathML" overflow="scroll"> <mml:mi>J</mml:mi> <mml:mn>1</mml:mn>...

Recent studies have yielded some interesting insights into the impacts of a property's walking accessibility on housing affordability and equity from potential property owner's perspective, limited attention was paid renter's perspective. This study investigates eight types other variables second-hand residential (SHP) price rental (RRP) rent. It uses sample 6,603 SHPs 3,566 RRPs that were collected in Shanghai, China, 2021. A modified floating catchment method is used to quantify...

Although recent transfer learning soft sensors show promising applications in multigrade chemical processes, good prediction performance mainly relies on available target domain data, which is difficult to achieve for a start-up grade. Additionally, only employing single global model inadequate characterize the inner relationship of process variables. A just-in-time adversarial (JATL) sensing method developed enhance performance. The distribution discrepancies variables between two different...

This paper studies Learning from Observations (LfO) for imitation learning with access to state-only demonstrations. In contrast Demonstration (LfD) that involves both action and state supervision, LfO is more practical in leveraging previously inapplicable resources (e.g. videos), yet challenging due the incomplete expert guidance. this paper, we investigate its difference LfD theoretical perspectives. We first prove gap between actually lies disagreement of inverse dynamics models imitator...