Anomaly detection in the Zwicky Transient Facility DR3. (arXiv:2012.01419v1 [astro-ph.IM])
<a href="http://arxiv.org/find/astro-ph/1/au:+Malanchev_K/0/1/0/all/0/1">K. L. Malanchev</a>, <a href="http://arxiv.org/find/astro-ph/1/au:+Pruzhinskaya_M/0/1/0/all/0/1">M. V. Pruzhinskaya</a>, <a href="http://arxiv.org/find/astro-ph/1/au:+Korolev_V/0/1/0/all/0/1">V. S. Korolev</a>, <a href="http://arxiv.org/find/astro-ph/1/au:+Aleo_P/0/1/0/all/0/1">P. D. Aleo</a>, <a href="http://arxiv.org/find/astro-ph/1/au:+Kornilov_M/0/1/0/all/0/1">M. V. Kornilov</a>, <a href="http://arxiv.org/find/astro-ph/1/au:+Ishida_E/0/1/0/all/0/1">E. E. O. Ishida</a>, <a href="http://arxiv.org/find/astro-ph/1/au:+Krushinsky_V/0/1/0/all/0/1">V. V. Krushinsky</a>, <a href="http://arxiv.org/find/astro-ph/1/au:+Mondon_F/0/1/0/all/0/1">F. Mondon</a>, <a href="http://arxiv.org/find/astro-ph/1/au:+Sreejith_S/0/1/0/all/0/1">S. Sreejith</a>, <a href="http://arxiv.org/find/astro-ph/1/au:+Volnova_A/0/1/0/all/0/1">A. A. Volnova</a>, <a href="http://arxiv.org/find/astro-ph/1/au:+Belinski_A/0/1/0/all/0/1">A. A. Belinski</a>, <a href="http://arxiv.org/find/astro-ph/1/au:+Dodin_A/0/1/0/all/0/1">A. V. Dodin</a>, <a href="http://arxiv.org/find/astro-ph/1/au:+Tatarnikov_A/0/1/0/all/0/1">A. M. Tatarnikov</a>, <a href="http://arxiv.org/find/astro-ph/1/au:+Zheltoukhov_S/0/1/0/all/0/1">S. G. Zheltoukhov</a>

We present results from applying the SNAD anomaly detection pipeline to the
third public data release of the Zwicky Transient Facility (ZTF DR3). The
pipeline is composed of 3 stages: feature extraction, search of outliers with
machine learning algorithms and anomaly identification with followup by human
experts. Our analysis concentrates in three ZTF fields, comprising more than
2.25 million objects. A set of 4 automatic learning algorithms was used to
identify 277 outliers, which were subsequently scrutinised by an expert. From
these, 188 (68%) were found to be bogus light curves — including effects from
the image subtraction pipeline as well as overlapping between a star and a
known asteroid, 66 (24%) were previously reported sources whereas 23 (8%)
correspond to non-catalogued objects, with the two latter cases of potential
scientific interest (e. g. 1 spectroscopically confirmed RS Canum Venaticorum
star, 4 supernovae candidates, 1 red dwarf flare). Moreover, using results from
the expert analysis, we were able to identify a simple bi-dimensional relation
which can be used to aid filtering potentially bogus light curves in future
studies. We provide a complete list of objects with potential scientific
application so they can be further scrutinised by the community. These results
confirm the importance of combining automatic machine learning algorithms with
domain knowledge in the construction of recommendation systems for astronomy.
Our code is publicly available at https://github.com/snad-space/zwad

We present results from applying the SNAD anomaly detection pipeline to the
third public data release of the Zwicky Transient Facility (ZTF DR3). The
pipeline is composed of 3 stages: feature extraction, search of outliers with
machine learning algorithms and anomaly identification with followup by human
experts. Our analysis concentrates in three ZTF fields, comprising more than
2.25 million objects. A set of 4 automatic learning algorithms was used to
identify 277 outliers, which were subsequently scrutinised by an expert. From
these, 188 (68%) were found to be bogus light curves — including effects from
the image subtraction pipeline as well as overlapping between a star and a
known asteroid, 66 (24%) were previously reported sources whereas 23 (8%)
correspond to non-catalogued objects, with the two latter cases of potential
scientific interest (e. g. 1 spectroscopically confirmed RS Canum Venaticorum
star, 4 supernovae candidates, 1 red dwarf flare). Moreover, using results from
the expert analysis, we were able to identify a simple bi-dimensional relation
which can be used to aid filtering potentially bogus light curves in future
studies. We provide a complete list of objects with potential scientific
application so they can be further scrutinised by the community. These results
confirm the importance of combining automatic machine learning algorithms with
domain knowledge in the construction of recommendation systems for astronomy.
Our code is publicly available at https://github.com/snad-space/zwad

http://arxiv.org/icons/sfx.gif