ترغب بنشر مسار تعليمي؟ اضغط هنا

What do writing features tell us about AI papers?

119   0   0.0 ( 0 )
 نشر من قبل Zining Zhu
 تاريخ النشر 2021
  مجال البحث الهندسة المعلوماتية
والبحث باللغة English




اسأل ChatGPT حول البحث

As the numbers of submissions to conferences grow quickly, the task of assessing the quality of academic papers automatically, convincingly, and with high accuracy attracts increasing attention. We argue that studying interpretable dimensions of these submissions could lead to scalable solutions. We extract a collection of writing features, and construct a suite of prediction tasks to assess the usefulness of these features in predicting citation counts and the publication of AI-related papers. Depending on the venues, the writing features can predict the conference vs. workshop appearance with F1 scores up to 60-90, sometimes even outperforming the content-based tf-idf features and RoBERTa. We show that the features describe writing style more than content. To further understand the results, we estimate the causal impact of the most indicative features. Our analysis on writing features provides a perspective to assessing and refining the writing of academic articles at scale.



قيم البحث

اقرأ أيضاً

Deeply virtual Compton scattering (DVCS) and timelike Compton scattering (TCS) leading twist amplitudes are intimately related thanks to their analytic properties as a function of $Q^2$. We exploit this feature to use Compton form factors previously extracted from available DVCS data and derive data-driven predictions for TCS observables to be measured in near future experiments. Our results quantitatively illustrate the complementarity of DVCS and TCS experiments.
143 - Richard de Grijs 2008
The diagnostic age versus mass-to-light ratio diagram is often used in attempts to constrain the shape of the stellar initial mass function, and the stability and the potential longevity of extragalactic young to intermediate-age massive star cluster s. Here, we explore the pitfalls associated with this approach and its potential for use with Galactic open clusters. We conclude that for an open cluster to survive for any significant fraction of a Hubble time (in the absence of substantial external perturbations), it is a necessary but not a sufficient condition to be located close to the predicted photometric evolutionary sequences for normal simple stellar populations.
434 - Sylvie Vauclair 2008
Studying the internal structure of exoplanet-host stars compared to that of similar stars without detected planets is particularly important for the understanding of planetary formation. The observed average overmetallicity of stars with planets is a n interesting point in that respect. In this framework, asteroseismic studies represent an excellent tool to determine the structural differences between stars with and without detected planets. It also leads to more precise values of the stellar parameters like mass, gravity, effective temperature, than those obtained from spectroscopy alone. Interestingly enough, the detection of stellar oscillations is obtained with the same instruments as used for the discovery of exoplanets, both from the ground and from space. The time scales however are very different, as the oscillations of solar type stars have periods around five to ten minutes, while the exoplanets orbits may go from a few days up to many years. Here I discuss the asteroseismology of exoplanet-host stars, with a few examples.
Planck data has not found the smoking gun of non-Gaussianity that would have necessitated consideration of inflationary models beyond the simplest canonical single field scenarios. This raises the important question of what these results do imply for more general models, and in particular, multi-field inflation. In this paper we revisit four ways in which two-field scenarios can behave differently from single field models; two-field slow-roll dynamics, curvaton-type behaviour, inflation ending on an inhomogeneous hypersurface and modulated reheating. We study the constraints that Planck data puts on these classes of behaviour, focusing on the latter two which have been least studied in the recent literature. We show that these latter classes are almost equivalent, and extend their previous analyses by accounting for arbitrary evolution of the isocurvature mode which, in particular, places important limits on the Gaussian curvature of the reheating hypersurface. In general, however, we find that Planck bispectrum results only constrain certain regions of parameter space, leading us to conclude that inflation sourced by more than one scalar field remains an important possibility.
297 - C.A.L. Bailer-Jones 2013
Gaia will provide parallaxes and proper motions with accuracy ranging from 10 to 1000 microarcsecond on up to one billion stars. Most of these will be disk stars: for an unreddened K giant at 6 kpc, it will measure the distance accurate to 15% and th e transverse velocity to an accuracy of about 1 km/s. Gaia will observe tracers of Galactic structure across the whole HR diagram, including Cepheids, RR Lyrae, white dwarfs, F dwarfs and HB stars. Onboard low resolution spectrophotometry will permit -- in addition to a Teff estimate -- dwarf/giant discrimination, metallicity measurement and extinction determination. For the first time, then, Gaia will provide us with a 3D spatial/properties map and at least a 2D velocity map of these tracers (RVs will be obtained too for brighter stars.) This will be a goldmine of information from which to learn about the origin and evolution of the Galactic disk. I briefly review the Gaia mission, and then show how the expected astrometric accuracies translate into distance and velocity accuracies and statistics. I examine the impact Gaia should have on a few scientific areas relevant to the Galactic disk. I discuss how a better determination of the spiral arm locations and pattern speed, plus a better reconstruction of the Suns orbit over the past billion years (from integration through the Gaia-measured gravitational potential) will allow us to assess the possible role of spiral arm crossings in ice ages and mass extinctions on the Earth.
التعليقات
جاري جلب التعليقات جاري جلب التعليقات
mircosoft-partner

هل ترغب بارسال اشعارات عن اخر التحديثات في شمرا-اكاديميا