Search

Machine learning has been widely applied to clearly defined problems of astronomy and astrophysics. However, deep learning and its conceptual differences to classical machine learning have been largely overlooked in these fields. The broad hypothesis behind our work is that letting the abundant real astrophysical data speak for itself, with minimal supervision and no labels, can reveal interesting patterns that may facilitate discovery of novel physical relationships. Here, as the first step, we seek to interpret the representations a deep convolutional neural network chooses to learn, and find correlations in them with current physical understanding. We train an encoder–decoder architecture on the self-supervised auxiliary task of reconstruction to allow it to learn general representations without bias towards any specific task. By exerting weak disentanglement at the information bottleneck of the network, we implicitly enforce interpretability in the learned features. We develop two independent statistical and information-theoretical methods for finding the number of learned informative features, as well as measuring their true correlation with astrophysical validation labels. As a case study, we apply this method to a data set of ∼270 000 stellar spectra, each of which comprising ∼300 000 dimensions. We find that the network clearly assigns specific nodes to estimate (notions of) parameters such as radial velocity and effective temperature without being asked to do so, all in a completely physics-agnostic process. This supports the first part of our hypothesis. Moreover, we find with high confidence that there are ∼4 more independently informative dimensions that do not show a direct correlation with our validation parameters, presenting potential room for future studies.

The new generation of deep photometric surveys requires unprecedentedly precise shape and photometry measurements of billions of galaxies to achieve their main science goals. At such depths, one major limiting factor is the blending of galaxies due to line-of-sight projection, with an expected fraction of blended galaxies of up to 50%. Current deblending approaches are in most cases either too slow or not accurate enough to reach the level of requirements. This work explores the use of deep neural networks to estimate the photometry of blended pairs of galaxies in monochrome space images, similar to the ones that will be delivered by the Euclid space telescope. Using a clean sample of isolated galaxies from the CANDELS survey, we artificially blend them and train two different network models to recover the photometry of the two galaxies. We show that our approach can recover the original photometry of the galaxies before being blended with $\sim$7% accuracy without any human intervention and without any assumption on the galaxy shape. This represents an improvement of at least a factor of 4 compared to the classical SExtractor approach. We also show that forcing the network to simultaneously estimate a binary segmentation map results in a slightly improved photometry. All data products and codes will be made public to ease the comparison with other approaches on a common data set., Comment: 16 pages, 12 figures, submitted to MNRAS, comments welcome

Open clusters (OCs) are popular tracers of the structure and evolutionary history of the Galactic disk. The OC population is often considered to be complete within 1.8 kpc of the Sun. The recent Gaia Data Release 2 (DR2) allows the latter claim to be challenged. We perform a systematic search for new OCs in the direction of Perseus using precise and accurate astrometry from Gaia DR2. We implement a coarse-to-fine search method. First, we exploit spatial proximity using a fast density-aware partitioning of the sky via a k-d tree in the spatial domain of Galactic coordinates, (l, b). Secondly, we employ a Gaussian mixture model in the proper motion space to quickly tag fields around OC candidates. Thirdly, we apply an unsupervised membership assignment method, UPMASK, to scrutinise the candidates. We visually inspect colour-magnitude diagrams to validate the detected objects. Finally, we perform a diagnostic to quantify the significance of each identified overdensity in proper motion and in parallax space We report the discovery of 41 new stellar clusters. This represents an increment of at least 20% of the previously known OC population in this volume of the Milky Way. We also report on the clear identification of NGC 886, an object previously considered an asterism. This letter challenges the previous claim of a near-complete sample of open clusters up to 1.8 kpc. Our results reveal that this claim requires revision, and a complete census of nearby open clusters is yet to be found., Comment: accepted for publication in A&A

Large sky surveys are increasingly relying on image subtraction pipelines for real-time (and archival) transient detection. In this process one has to contend with varying point-spread function (PSF) and small brightness variations in many sources, as well as artefacts resulting from saturated stars and, in general, matching errors. Very often the differencing is done with a reference image that is deeper than individual images and the attendant difference in noise characteristics can also lead to artefacts. We present here a deep-learning approach to transient detection that encapsulates all the steps of a traditional image-subtraction pipeline – image registration, background subtraction, noise removal, PSF matching and subtraction – in a single real-time convolutional network. Once trained, the method works lightening-fast and, given that it performs multiple steps in one go, the time saved and false positives eliminated for multi-CCD surveys like Zwicky Transient Facility and Large Synoptic Survey Telescope will be immense, as millions of subtractions will be needed per night.

General human action recognition requires understanding of various visual cues. In this paper, we propose a network architecture that computes and integrates the most important visual cues for action recognition: pose, motion, and the raw images. For the integration, we introduce a Markov chain model which adds cues successively. The resulting approach is efficient and applicable to action classification as well as to spatial and temporal action localization. The two contributions clearly improve the performance over respective baselines. The overall approach achieves state-of-the-art action classification performance on HMDB51, J-HMDB and NTU RGB+D datasets. Moreover, it yields state-of-the-art spatio-temporal action localization results on UCF101 and J-HMDB., 10 pages, 7 figures, ICCV 2017 submission

Recent work has shown good recognition results in 3D object recognition using 3D convolutional networks. In this paper, we show that the object orientation plays an important role in 3D recognition. More specifically, we argue that objects induce different features in the network under rotation. Thus, we approach the category-level classification task as a multi-task problem, in which the network is trained to predict the pose of the object in addition to the class label as a parallel task. We show that this yields significant improvements in the classification results. We test our suggested architecture on several datasets representing various 3D data sources: LiDAR data, CAD models, and RGB-D images. We report state-of-the-art results on classification as well as significant improvements in precision and speed over the baseline on 3D detection., BMVC'17 version. Added some experiments + auto-alignment of Modelnet40

Object recognition approaches have recently been extended to yield, aside of the object class output, also viewpoint or pose. Training such approaches typically requires additional viewpoint or keypoint annotation in the training data or, alternatively, synthetic CAD models. In this paper, we present an approach that creates a dataset of images annotated with bounding boxes and viewpoint labels in a fully automated manner from videos. We assume that the scene is static in order to reconstruct 3D surfaces via structure from motion. We automatically detect when the reconstruction fails and normalize for the viewpoint of the 3D models by aligning the reconstructed point clouds. Exemplarily for cars we show that we can expand a large dataset of annotated single images and obtain improved performance when training a viewpoint regressor on this joined dataset.

Hough transform based object detectors divide an object into a number of patches and combine them using a shape model. For efficient combination of patches into the shape model, the individual patches are assumed to be independent of one another. Although this independence assumption is key for fast inference, it requires the individual patches to have a high discriminative power in predicting the class and location of objects. In this paper, we argue that the sparsity of the appearance of a patch in its neighborhood can be a very powerful measure to increase the discriminative power of a local patch and incorporate it as a sparsity potential for object detection. Further, we show that this potential shall depend on the appearance of the patch to adapt to the statistics of the neighborhood specific to the type of appearance (e.g. texture or structure) it represents. We have evaluated our method on challenging datasets including the PASCAL VOC 2007 dataset and show that using the proposed sparsity potential result in a substantial improvement in the detection accuracy.

Image-based License Plate Recognition (LPR) algorithms are the core modules of many Intelligent Transportation Systems (ITS). Different algorithms and approaches have been proposed so far. All of these methods have the following three steps in common: License Plate Localization, Character Segmentation & Character Recognition. There are many real-world issues encountered during the design of each step, including different plate formats, time-variant illumination conditions and etc. To have a reliable operator-free system, all of these need to be overcome. One of such issues which is the main focus of this article and so far has not been addressed in any previous work is the presence of characters corrupted by misplaced rivets/screws. In this paper we present a simple, yet effective technique based on traditional pattern matching methods which when combined with modern character recognition techniques, can bring up the success rates of current systems closer to 100%.

This paper proposes a solution to reduce the energy consumption by bulk power consumers in distributed networks at the times that the network is to experience a sudden peak in consumption graph. This is done using an intelligent system in addition to one variable multi-tariff kWh-meter. Finally the results of applying this method, is estimated using the samples gathered from the energy consumption information of several hospitals. (5 pages)