Artículos de conferencia
Permanent URI for this collection
Browse
Recent Submissions
- ItemWeather on other worlds: BD variability and the VVV(UNIVERSIDAD NACIONAL AUTONOMA MEXICO INSTITUTO ASTRONOMIA, 2014) Kurtev, R.; Metchev, S.; Heinze, A.; Gromadzki, M.; Ivanov, V. D.; Minniti, D.; Beamin Muhlenbrock, Juan Carlos; Borissova, J.; Mateus, A; GregorioHetem, J; Fernandes, RCVarious evidences point to the presence of clouds in ultra-cool atmospheres. An important ambiguity remains as to whether all variability in ultra-cool dwarfs is caused by patchy clouds, or other fenomena like magnetic activity and auroras. Simultaneous multi-wavelength photometric and/or spectroscopic monitoring could help to reveal this enigma.
- ItemVVV IR high proper motion stars(UNIVERSIDAD NACIONAL AUTONOMA MEXICO INSTITUTO ASTRONOMIA, 2015) Kurtev, R.; Gromadzki, M.; Beamin Muhlenbrock, Juan Carlos; Pena, K.; Folkes, S.; Ivanov, V. D.; Borissova, J.; Kuhn, M.; Villanueva, V.; Minniti, D.; Mendez, R.; Lucas, P.; Smith, L.; Pinfield, D.; Antonova, A.; Vieira, K; VanAltena, W; Mendez, RAWe used the VISTA Variables en Via Lactea (VVV) survey to search for large proper motion (PM) objects in the zone of avoidance in the Milky Way bulge and southern Galactic disk. This survey is multi-epoch and already spans a period of more than four years, giving us an excellent opportunity for proper motion and parallax studies. We found around 1700 PM objects with PM>30 mas yr(-1). The majority of them are early and mid M-dwarfs. There are also few later spectral type objects, as well as numerous new K- and G-dwarfs. 75 of the stars have PM>300 mas(-1) and 189 stars have PM>200 mas(-1). There are only 42 previously known stars in the VVV area with proper motion PM>200 mas(-1). We also found three dM+WD binaries and new members of the immediate solar vicinity of 25 pc. We generated a catalog which will be a complementary to the existing catalogs outside this zone.
- ItemVista variables in the via lactea (VVV): first results and perspectives(2011) Saito, R. K.; Minniti, D.; Dekany, I.; Hempel, M.; Alonso-Garcia, J.; Toledo, I.; Beamin Muhlenbrock, Juan Carlos; Angeloni, R.; Lucas, P. W.; Emerson, J. P.VISTA Variables in the Via Lactea (VVV) is a public ESO near-IR variability survey scanning the Milky Way Bulge and an adjacent section of the mid-plane. The survey will take 1929 hours of observations with the 4 m VISTA telescope during five years (2010-2014), covering similar to 10(9) point sources across an area of 520 deg(2). Here we address the first results obtained from the VVV Survey as well as a glimpse into the possibilities for using a deep near-IR atlas in five passbands and a catalogue of more than 10(6) variable point sources. We expect to use the data to find planetary transits of late-type main-sequence stars. We discuss the planet searches and future follow-ups
- ItemPEERING THROUGH THE DUST: PRECISE ASTROMETRY IN THE GALACTIC MID-PLANE WITH THE VVV SURVEY(2015) Lucas, PW; Smart, RL; Jones, HRA; Kurtev, R; Beamin Muhlenbrock, Juan Carlos; Borissova, J.; Gromadzki, Grzegorz; Ivanov, Valentin; Minniti, Dante; Pinfield, DJGaia will see little of the Galactic mid-plane and nuclear bulge due to high extinction at optical wavelengths. To study the structure and kinematics of the inner Galaxy we must look to longer wavelengths. The Vista Variables in the Via Lactea (VVV, Minniti et al. 2010) survey currently provides just over 4 years of observations covering approximately 560 square degrees of the Galactic bulge and plane. Typically each source is observed 50-150 times in the Ks band over this period. Using these data we provide relative proper motions for approximately 200 million unique sources down to Ks similar to 16 with uncertainties approaching 1 mas yr(-1). In addition, we fit a solution of the parallactic motion of all sources with significant proper motion and discover a number of new nearby brown dwarfs. These results will allow us to identify faint common proper motion companions to stars with Gaia parallaxes, increasing the number of brown dwarf benchmark objects. Our absolute astrometric calibration precision is currently similar to 2 mas yr(-1), based on PPMXL. The Gaia absolute astrometric reference grid will allow us to precisely anchor our results and measure the streaming motions of stars in the bulge. Finally, we anticipate that the catalogue could provide kinematic distances to the numerous optically invisible high amplitude variable stars that VVV is discovering.
- ItemPrevalence of major depressive disorder among immigrants of the metropolitan region of santiago, chile(2020) Errazuriz Concha, AntoniaModern machine learning pipelines are limited due to data availability, storage quotas, privacy regulations, and expensive annotation processes. These constraints make it difficult or impossible to train and update large-scale models on such dynamic annotated sets. Continual learning directly approaches this problem, with the ultimate goal of devising methods where a deep neural network effectively learns relevant patterns for new (unseen) classes, without significantly altering its performance on previously learned ones. In this paper, we address the problem of continual learning for video data. We introduce PIVOT, a novel method that leverages extensive knowledge in pre-trained models from the image domain, thereby reducing the number of trainable parameters and the associated forgetting. Unlike previous methods, ours is the first approach that effectively uses prompting mechanisms for continual learning without any in-domain pre-training. Our experiments show that PIVOT improves state-of-the-art methods by a significant 27% on the 20-task ActivityNet setup.