Search results for "Machine learning"
showing 10 items of 1464 documents
Which Is Which? Evaluation of Local Descriptors for Image Matching in Real-World Scenarios
2019
Matching with local image descriptors is a fundamental task in many computer vision applications. This paper describes the WISW contest held within the framework of the CAIP 2019 conference, aimed at benchmarking recent descriptors in challenging planar and non-planar real image matching scenarios. According to the contest results, the descriptors submitted to the competition, most of which based on deep learning, perform significantly better than the current state-of-the-art in image matching. Nonetheless, there is still room for improvement, especially in the case of non-planar scenes.
Do trained assessors generalize their knowledge to new stimuli?
2005
Previous work showed that trained assessors are better at discriminating and describing familiar chemico-sensorial stimuli than novices. In this study, we evaluated whether this superiority holds true for new stimuli. We first trained a group of subjects to characterize beer flavors over a two year period. After training was accomplished, we compared the performance of these trained assessors with the performance of novice subjects for discrimination and matching tasks. The tasks were performed using both well-learned and new beers. Trained assessors outperformed novices in the discrimination task for learned beers but not for new beers. But on the matching task, trained assessors outperfor…
Boosting the supercapacitive behavior of CoAl-layered double hydroxides via tuning the metal composition and interlayer space
2020
Layered double hydroxides (LDHs) are promising supercapacitor materials due to their wide chemical versatility, earth abundant metals and high specific capacitances. Many parameters influencing the supercapacitive performance have been studied such as the chemical composition, the synthetic approaches, and the interlayer anion. However, no systematic studies about the effect of the basal space have been carried out. Here, two-dimensional (2D) CoAl-LDHs were synthesized through anion exchange reactions using surfactant molecules in order to increase the interlayer space (ranging from 7.5 to 32.0 Å). These compounds exhibit similar size and dimensions but different basal space to explore excl…
Boosting the Performance of One-Step Solution-Processed Perovskite Solar Cells Using a Natural Monoterpene Alcohol as a Green Solvent Additive
2021
The perovskite film is the core of a perovskite solar cell (PSC), and its quality is crucial for the performance of such devices. The morphology, crystallinity, and surface coverage of the perovskite layer greatly affect the power conversion efficiency (PCE), hysteresis, and long-term stability of PSCs. The incorporation of appropriate solvent additives in the perovskite precursor solution is an effective strategy to control the film morphology and reduce the defects and grain boundaries. However, the commonly used solvent additives are environmentally harmful and highly toxic. In this work, α-terpineol (a nontoxic, eco-friendly, and low-cost monoterpene alcohol) is employed for the first t…
Deep-Learning-Enabled Fast Optical Identification and Characterization of 2D Materials.
2020
© 2020 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim Advanced microscopy and/or spectroscopy tools play indispensable roles in nanoscience and nanotechnology research, as they provide rich information about material processes and properties. However, the interpretation of imaging data heavily relies on the “intuition” of experienced researchers. As a result, many of the deep graphical features obtained through these tools are often unused because of difficulties in processing the data and finding the correlations. Such challenges can be well addressed by deep learning. In this work, the optical characterization of 2D materials is used as a case study, and a neural-network-based algorithm is de…
Search of a topological pattern to evaluate toxicity of heterogeneous compounds.
2001
Abstract Molecular connectivity has been applied to the search of mathematical models able to predict the carcinogenic and teratogenic activity of a wide group of structurally heterogeneous compounds. Through the linear discriminant analysis and the diagrams of distribution of pharmacological activity, the classification criteria that minimizes the percentage of error are established. The easiness and speed of the calculation of the descriptors used in this work make the models developed useful in data bases containing a huge number of compounds.
Statistical criteria for early-stopping of support vector machines
2007
This paper proposes the use of statistical criteria for early-stopping support vector machines, both for regression and classification problems. The method basically stops the minimization of the primal functional when moments of the error signal (up to fourth order) become stationary, rather than according to a tolerance threshold of primal convergence itself. This simple strategy induces lower computational efforts and no significant differences are observed in terms of performance and sparsity.
Implementing some Evolutionary Computing Methods for Determining the Optimal Parameters in the Turning Process
2015
In this paper, we comparatively present two heuristics search methods – Simulated Annealing and Weighted Sum Genetic Algorithm, in order to find optimal cutting parameters in turning operation. We consider five different constraints aiming to achieve minimum total cost of machining. We developed a customizable software application in Microsoft Visual Studio with C# source code, flexible and extensible that implements the optimization methods. The experiments are based on real data gathered from S.C. “Compa” S.A Sibiu, a company that manufactures automotive components and targets improving of product quality and reducing cost and production time. The obtained results show that, although the …
Using box indices in supporting comparison in multiobjective optimization
2009
Because of the conflicting nature of criteria or objectives, solving a multiobjective optimization problem typically requires interaction with a decision maker who can specify preference information related to the objectives in the problem in question. Due to the difficulties of dealing with multiple objectives, the way information is presented plays a very important role. Questions posed to the decision maker must be simple enough and information shown must be easy to understand. For this purpose, visualization and graphical representations can be useful and constitute one of the main tools used in the literature. In this paper, we propose to use box indices to represent information relate…
Determining the Difficulty of Landscapes by PageRank Centrality in Local Optima Networks
2016
The contribution of this study is twofold: First, we show that we can predict the performance of Iterated Local Search (ILS) in different landscapes with the help of Local Optima Networks (LONs) with escape edges. As a predictor, we use the PageRank Centrality of the global optimum. Escape edges can be extracted with lower effort than the edges used in a previous study. Second, we show that the PageRank vector of a LON can be used to predict the solution quality (average fitness) achievable by ILS in different landscapes.