scholarly journals Background Splitting: Finding Rare Classes in a Sea of Background

2021 ◽  
Author(s):  
Ravi Teja Mullapudi ◽  
Fait Poms ◽  
William R. Mark ◽  
Deva Ramanan ◽  
Kayvon Fatahalian
Keyword(s):  
Rare Classes

scholarly journals The Role of Wide Field X-ray Surveys in Astronomy

2012 ◽  
Vol 10 (H16) ◽  
pp. 669-670
Author(s):  
Richard D. Saxton
Keyword(s):  
High Variability ◽  
Wide Field ◽  
Clusters Of Galaxies ◽  
Tidal Disruption ◽  
X Ray ◽  
Sky Survey ◽  
History Of ◽  
X Ray Background ◽  
Rare Classes

AbstractWe review the history of X-ray sky surveys from the early experiments to the catalogues of 105 sources produced by ROSAT, Chandra and XMM-Newton. At bright fluxes the X-ray sky is shared between stars, accreting binaries and extragalactic sources while deeper surveys are dominated by AGN and clusters of galaxies. The X-ray background, found by the earliest missions, has been largely resolved into discrete sources at soft (0.3-2 keV) energies but at higher energies an important fraction still escapes detection. The possible identification of the missing flux with Compton-thick AGN has been probed in recent years by Swift and Integral.Variability seen in objects observed at different epochs has proved to be an excellent discriminator for rare classes of objects. The comparison of ROSAT All Sky Survey (RASS) and ROSAT pointed observations identified several Novae and high variability AGN as well as initiating the observational study of Tidal Disruption events. More recently the XMM-Newton slew survey, in conjunction with archival RASS data, has detected further examples of flaring objects which have been followed-up in near-real time at other wavelengths.

scholarly journals Relieving the Incompatibility of Network Representation and Classification for Long-Tailed Data Distribution

2021 ◽  
Vol 2021 ◽  
pp. 1-10
Author(s):  
Hao Hu ◽  
Mengya Gao ◽  
Mingsheng Wu
Keyword(s):  
Large Scale ◽  
Deep Neural Networks ◽  
State Of The Art ◽  
Data Distribution ◽  
Distribution Problem ◽  
Imbalanced Dataset ◽  
Network Representation ◽  
Knowledge Distillation ◽  
Rare Classes ◽  
And Training

In the real-world scenario, data often have a long-tailed distribution and training deep neural networks on such an imbalanced dataset has become a great challenge. The main problem caused by a long-tailed data distribution is that common classes will dominate the training results and achieve a very low accuracy on the rare classes. Recent work focuses on improving the network representation ability to overcome the long-tailed problem, while it always ignores adapting the network classifier to a long-tailed case, which will cause the “incompatibility” problem of network representation and network classifier. In this paper, we use knowledge distillation to solve the long-tailed data distribution problem and fully optimize the network representation and classifier simultaneously. We propose multiexperts knowledge distillation with class-balanced sampling to jointly learn high-quality network representation and classifier. Also, a channel activation-based knowledge distillation method is also proposed to improve the performance further. State-of-the-art performance on several large-scale long-tailed classification datasets shows the superior generalization of our method.

scholarly journals Identifying Rare Classes with Sparse Training Data

2007 ◽  
pp. 751-760
Author(s):  
Mingwu Zhang ◽  
Wei Jiang ◽  
Chris Clifton ◽  
Sunil Prabhakar
Keyword(s):  
Training Data ◽  
Rare Classes

scholarly journals Classification Algorithm based on MS Apriori for Rare Classes

2012 ◽  
Vol 48 (22) ◽  
pp. 52-56 ◽  
Author(s):  
Devashree Rai ◽  
Kesari Verma ◽  
A. S. Thoke
Keyword(s):  
Classification Algorithm ◽  
Rare Classes

Hmong-Mien Languages

2017 ◽  
Author(s):  
David R. Mortensen
Keyword(s):  
Southeast Asia ◽  
Southern China ◽  
Compact Groups ◽  
Minority Languages ◽  
Special Relationship ◽  
Language Groups ◽  
Chinese Scholars ◽  
Rare Classes ◽  
Internal Relationships ◽  
Mainland Southeast Asia

Hmong-Mien (also known as Miao-Yao) is a bipartite family of minority languages spoken primarily in China and mainland Southeast Asia. The two branches, called Hmongic and Mienic by most Western linguists and Miao and Yao by Chinese linguists, are both compact groups (phylogenetically if not geographically). Although they are uncontroversially distinct from one another, they bear a strong mutual affinity. But while their internal relationships are reasonably well established, there is no unanimity regarding their wider genetic affiliations, with many Chinese scholars insisting on Hmong-Mien membership in the Sino-Tibetan superfamily, some Western scholars suggesting a relationship to Austronesian and/or Tai-Kradai, and still others suggesting a relationship to Mon-Khmer. A plurality view appears to be that Hmong-Mien bears no special relationship to any surviving language family. Hmong-Mien languages are typical—in many respects—of the non-Sino-Tibetan languages of Southern China and mainland Southeast Asia. However, they possess a number of properties that make them stand out. Many neighboring languages are tonal, but Hmong-Mien languages are, on average, more so (in terms of the number of tones). While some other languages in the area have small-to-medium consonant inventories, Hmong-Mien languages (and especially Hmongic languages) often have very large consonant inventories with rare classes of sounds like uvulars and voiceless sonorants. Furthermore, while many of their neighbors are morphologically isolating, few language groups display as little affixation as Hmong-Mien languages. They are largely head-initial, but they deviate from this generalization in their genitive-noun constructions and their relative clauses (which vary in position and structure, sometimes even within the same language).

Synthesis of substituted benzo[b][1,4]oxazepine derivatives by the reaction of 2-aminophenols with alkynones

2020 ◽  
Vol 18 (3) ◽  
pp. 415-419
Author(s):  
Kohei Oshimoto ◽  
Biao Zhou ◽  
Hiroaki Tsuji ◽  
Motoi Kawatsura
Keyword(s):  
Synthetic Method ◽  
Rare Classes

We have developed a novel synthetic method accessing benzo[b][1,4]oxazepines that are one of the rare classes of benzoxazepine derivatives by reaction of 2-aminophenols with alkynones in 1,4-dioxane at 100 °C.

scholarly journals Two new magnetic cataclysmic variables discovered in the 3XMM catalogue

2018 ◽  
Vol 615 ◽  
pp. A133 ◽  
Author(s):  
N. A. Webb ◽  
A. Schwope ◽  
I. Zolotukhin ◽  
D. Lin ◽  
S. R. Rosen
Keyword(s):  
Orbital Period ◽  
Model Fitting ◽  
Cataclysmic Variables ◽  
Compact Objects ◽  
Clusters Of Galaxies ◽  
X Ray ◽  
Large Numbers ◽  
Magnetic Cataclysmic Variables ◽  
Polar Type ◽  
Rare Classes

Context. X-ray catalogues provide a wealth of information on many source types, ranging from compact objects to galaxies, clusters of galaxies, stars, and even planets. Thanks to the huge volume of X-ray sources provided in the 3XMM catalogue, along with many source specific products, many new examples from rare classes of sources can be identified. Aims. Through visualising spectra and lightcurves from about 80 observations included in the incremental part of the 3XMM catalogue, 3XMM-DR5, as part of the quality control of the catalogue, we identified two new X-ray sources, 3XMM J183333.1+225136 and 3XMM J184916.1+652943, that were highly variable. This work aims to investigate their nature. Methods. Through simple model fitting of the X-ray spectra and analysis of the X-ray lightcurves of 3XMM J183333.1+225136 and 3XMM J184916.1+652943, along with complementary photometry from the XMM-Newton Optical Monitor, Pan-STARRS and the Stella/WiFSIP and Large Binocular Telescope (LBT) spectra, we suggest that the two sources might be magnetic cataclysmic variables (CVs) of the polar type and we determine some of their properties. Results. Both CVs have very hard spectra, showing no soft excess. They are both situated in the local neighbourhood, located within ~1 kpc. 3XMM J183333.1+225136 has an orbital period of 2.15 h. It shows features in the lightcurve that may be a total eclipse of the white dwarf. 3XMM J184916.1+652943 has an orbital period of 1.6 h. Given that only a small sky area was searched to identify these CVs, future sensitive all sky surveys such as the eROSITA project should be very successful at uncovering large numbers of such sources.

Sign in / Sign up

Export Citation Format

Share Document