Category: 7. Maths
-
Detecting shortcut learning for fair medical AI using shortcut testing
Shortcut learning poses significant challenges for machine learning in healthcare, where predictions based on spurious correlations pose significant concerns regarding safety21,22 and fairness3,4,12. However, identifying whether shortcut learning is responsible for model unfairness is challenging, especially when sensitive attributes such as age may be causally linked to the clinical task. In this paper, we propose…
-
Peptide conformational sampling using the Quantum Approximate Optimization Algorithm
Encoding a conformation in qubits In the framework considered here, a self-avoiding walk or peptide conformation is sampled by preparing a quantum state and measuring it in a computational basis; the measured bitstring encodes a specific protein conformation. Several approaches exist to encode a protein conformation in a bitstring, see ref. 10 for a review.…
-
Simple algorithm for judging equivalence of differential-algebraic equation systems
Equivalence judgment methods Figure 1 schematically presents our proposed method for judging the equivalence of two equation groups. The algorithm 1) eliminates variables, 2) checks whether two equation groups have the same set of variables, and 3) judges the equivalence between each equation in one equation group and each equation in the other equation group. If…
-
Conservation laws, solitary wave solutions, and lie analysis for the nonlinear chains of atoms
Riaz, M. B., Atangana, A., Jhangeer, A. & Junaid-U-Rehman, M. Some exact explicit solutions and conservation laws of Chaffee-Infante equation by Lie symmetry analysis. Phys. Scr. 96(8), 084008 (2021). Article ADS Google Scholar Kumar, S., Jadaun, V. & Ma, W. X. Application of the Lie symmetry approach to an extended Jimbo-Miwa equation in (3+ 1)…
-
Atlas-scale single-cell multi-sample multi-condition data integration using scMerge2
scMerge2 Single-cell grouping within one batch Following the same principals as scMerge, the scMerge2 approach begins by grouping the cells that share similar biological signals within each dataset or batch. We can approach this in two ways: one way is to perform unsupervised clustering; the other way is using results from supervised cell type classification.…
-
Shoot gravitropism and organ straightening cooperate to arrive at a mechanically favorable shape in Arabidopsis
Quantitative analysis reveals spatio-temporal characteristics of hyperbending stems To capture the spatio-temporal characteristics of inflorescence stems during bending, we took time-lapse images of wild-type stems and myosin xif xik stems (Fig. 1a,f, Figs. S1, S2). We used ImageJ to extract the centerline of each stem from the images and constructed a continuous curve using spline interpolation.…
-
Unsupervised abnormality detection in neonatal MRI brain scans using deep learning
In this study, we compared AE and VAE-type architectures to identify an effective model for detecting anomalies within the neonatal brain. We utilized image preprocessing methods to ensure homogeneity in the dataset, and postprocessing to ensure the outputs of our models can be interpreted by a radiologist. Our methods were as follows: Anomaly detection using…
-
Life-long phishing attack detection using continual learning
This section covers a detailed discussion of our methodology to identify and mitigate the performance drop. We analyzed our dataset features using low-dimensional principal component analysis (PCA) embedding of samples to visualize the difference in distribution. We transformed features into 1-D and 2-D PCA as shown in Figs. 1 and 2, respectively. Figure 1 shows…
-
Unveiling the drives behind tetracycline adsorption capacity with biochar through machine learning
Statistical results of biochar characteristics This study utilized a combination of box plots and normal distribution curves to illustrate the distribution patterns of continuous data (see Fig. 1). The composite plot comprises two sections—the left segment illustrates the box plot, whereas the right segment manifests the normal distribution curve of the data. The box plot depicts…
-
Numerical simulation for impact of implement of reflector and turbulator within the solar system in existence of nanomaterial
To gain more heat flux from the solar system, parabolic reflector has been applied in present work. Concentrated solar systems can absorb high levels of solar irradiation and can be utilized in industrial applications. To increase the thermal output of the solar unit, the absorber can be equipped with a turbulator. For present work, a…