Search

Web applications, widely used by enterprises for business services, require extensive testing to ensure functionality. Performing form testing with random input data often takes a long time to complete. Previously, we introduced a model for automated testing of web applications using reinforcement learning. The model was trained to fill form fields with fixed input values and click buttons. However, the performance of this model was limited by a fixed set of input data and the imprecise detection of successful form submission. This paper proposes a model to address these limitations. First, we use a large language model with data fakers to generate a wide variety of input data. Additionally, whether form submission is successful is partially determined by GPT-4o. Experiments show that our method increases average statement coverage by 2.3% over the previous model and 7.7% to 11.9% compared to QExplore, highlighting its effectiveness.

Objectives: This paper studies the segmentation and detection of small metastatic brain tumors. This study aims to evaluate the feasibility of training a deep neural network for the segmentation and detection of metastatic brain tumors in MRI using a very small dataset of 33 cases, by leveraging large public datasets of primary tumors; Methods: This study explores various methods, including supervised learning, two transfer learning approaches, and self-supervised learning, utilizing U-net and Swin UNETR models; Results: The self-supervised learning approach utilizing the Swin UNETR model yielded the best performance. The Dice score for small brain tumors was approximately 0.19. Sensitivity reached 100%, while specificity was 54.5%. When excluding subjects with hyperintensities, the specificity improved to 80.0%; Conclusions: It is feasible to train a model using self-supervised learning and a small dataset for the segmentation and detection of small brain tumors.

The dynamic flexible job-shop problem (DFJSP) is a realistic and challenging problem that many production plants face. As the product line becomes more complex, the machines may suddenly break down or resume service, so we need a dynamic scheduling framework to cope with the changing number of machines over time. This issue has been rarely addressed in the literature. In this paper, we propose an improved learning-to-dispatch (L2D) model to generate a reasonable and good schedule to minimize the makespan. We formulate a DFJSP as a disjunctive graph and use graph neural networks (GINs) to embed the disjunctive graph into states for the agent to learn. The use of GINs enables the model to handle the dynamic number of machines and to effectively generalize to large-scale instances. The learning agent is a multi-layer feedforward network trained with a reinforcement learning algorithm, called proximal policy optimization. We trained the model on small-sized problems and tested it on various-sized problems. The experimental results show that our model outperforms the existing best priority dispatching rule algorithms, such as shortest processing time, most work remaining, flow due date per most work remaining, and most operations remaining. The results verify that the model has a good generalization capability and, thus, demonstrate its effectiveness.

The advent of intelligent technology has spurred most large companies to introduce business intelligence systems (BIS), but those with low information maturity still have a wait-and-see attitude towards BIS. In order to accelerate the introduction of BIS, this study found and analyzed the critical factors (CFs) considered by companies when introducing BIS. First, the literature on factors considered by companies to introduce BIS was reviewed. The three stages before, during, and after introduction in marketing that organizations undergo during the procurement process were developed into a three-layer hierarchy factor table. An expert questionnaire with pairwise factors was then designed and sent to senior executives in companies that had introduced BIS, and the weights of all factors were calculated by the fuzzy analytic hierarchy process (FAHP) based on the collected questionnaire data. After this, four critical factors—system function integrity, approaching corporate strategy, licensing fee, and information technology maturity—were determined objectively by using the conditions for the acceptable advantage of Vlse Kriterijumska Optimizacija Kompromisno Resenje (VIKOR) and further explored in order to help companies input fewer resources, introduce BIS efficiently, and thus increase their decision-making power.

We are pleased to announce a new article type, “Patent Summary”, for Inventions.[...]

The Laser Direct Synthesis and Patterning (LDSP) technology has advantages in terms of processing time and cost compared to nanomaterials-based laser additive microfabrication processes. In LDSP, a scanning laser on the substrate surface induces chemical reactions in the reactive liquid solution and selectively deposits target material in a preselected pattern on the substrate. In this study, we experimentally investigated the effect of the processing parameters and type and concentration of the additive solvent on the properties and growth rate of the resulting metal film fabricated by this LDSP technology. It was shown that reactive metal ion solutions with substantial viscosity yield metal films with superior physical properties. A numerical analysis was also carried out the first time to investigate the coupled opto-thermo-fluidic transport phenomena and the effects on the metal film growth rate. To complete the simulation, the optical properties of the LDSP deposited metal film with a variety of thicknesses were measured. The characteristics of the temperature field and the thermally induced flow associated with the moving heat source are discussed. It was shown that the processing temperature range of the LDSP is from 330 to 390 K. A semi-empirical model for estimating the metal film growth rate using this process was developed based on these results. From the experimental and numerical results, it is seen that, owing to the increased reflectivity of the silver film as its thickness increases, the growth rate decreases gradually from about 40 nm at initial to 10 nm per laser scan after ten scans. This self-controlling effect of LDSP process controls the thickness and improves the uniformity of the fabricated metal film. The growth rate and resulting thickness of the metal film can also be regulated by adjustment of the processing parameters, and thus can be utilized for controllable additive nano/microfabrication.