In this work, we propose several enhancements to a geometric transformation based model for anomaly detection in images (GeoTranform). The model assumes that the anomaly class is unknown and that only inlier samples are available for training. We introduce new filter based transformations useful for detecting anomalies in astronomical images, that highlight artifact properties to make them more easily distinguishable from real objects. In addition, we propose a transformation selection strategy that allows us to find indistinguishable pairs of transformations. This results in an improvement of the area under the Receiver Operating Characteristic curve (AUROC) and accuracy performance, as well as in a dimensionality reduction. The models were tested on astronomical images from the High Cadence Transient Survey (HiTS) and Zwicky Transient Facility (ZTF) datasets. The best models obtained an average AUROC of 99.20% for HiTS and 91.39% for ZTF. The improvement over the original GeoTransform algorithm and baseline methods such as One-Class Support Vector Machine, and deep learning based methods is significant both statistically and in practice.

In this paper, we propose an enhanced CNN model for detecting supernovae (SNe). This is done by applying a new method for obtaining rotational invariance that exploits cyclic symmetry. In addition, we use a visualization approach, the layer-wise relevance propagation (LRP) method, which allows finding the relevant pixels in each image that contribute to discriminate between SN candidates and artifacts. We introduce a measure to assess quantitatively the effect of the rotational invariant methods on the LRP relevance heatmaps. This allows comparing the proposed method, CAP, with the original Deep-HiTS model. The results show that the enhanced method presents an augmented capacity for achieving rotational invariance with respect to the original model. An ensemble of CAP models obtained the best results so far on the HiTS dataset, reaching an average accuracy of 99.53%. The improvement over Deep-HiTS is significant both statistically and in practice.

There is strong observational evidence for the presence of large-scale magnetic fields MF in galaxies and clusters, with strength $\sim \mu$G and coherence lenght on the order of Kpc. However its origin remains as an outstanding problem. One of the possible explanations is that they have been generated in the early universe. Recently, it has been proposed that helical primordial magnetic fields PMFs, could be generated during the EW or QCD phase transitions, parity-violating processes and predicted by GUT or string theory. Here we concentrate on the study of two mechanisms to generate PMFs, the first one is the $\nu$MSM which triggers instability in the Maxwell's equations and leads to the generation of helical PMFs. The second one is the usual electroweak baryogenesis scenario. Finally, we calculate the exact power spectra of these helical PMFs and we show its role in the production of gravitational waves finding a scale-invariant on large scales and an oscillatory motion (damping) for $k\eta \gg 1$