Subscribe to the gold package and get unlimited access to Shamra Academy
Register a new userWhat We Can Learn from the Running of the Spectral Index if no Tensors are Detected in the Cosmic Microwave Background Anisotropy
129
0
0.0
(
0
)
Ask ChatGPT about the research
No Arabic abstract
In this paper we operate under the assumption that no tensors from inflation will be measured in the future by the dedicated experiments and argue that, while for single-field slow-roll models of inflation the running of the spectral index will be hard to be detected, in multi-field models the running can be large due to its strong correlation with non-Gaussianity. A detection of the running might therefore be related to the presence of more than one active scalar degree of freedom during inflation.
rate research
Read More
We compute the spectral distortions of the Cosmic Microwave Background (CMB) polarization induced by non-linear effects in the Compton interactions between CMB photons and cold intergalactic electrons. This signal is of the $y$-type and is dominated by contributions arising from the reionized era. We stress that it is not shadowed by the thermal SZ effect which has no equivalent for polarization. We decompose its angular dependence into $E$- and $B$-modes, and we calculate the corresponding power spectra, both exactly and using a suitable Limber approximation that allows a simpler numerical evaluation. We find that $B$-modes are of the same order of magnitude as $E$-modes. Both spectra are relatively flat, peaking around $ell=280$, and their overall amplitude is directly related to the optical depth to reionization. Moreover, we find this effect to be one order of magnitude larger than the non-linear kinetic Sunyaev-Zeldovich effect in galaxy clusters. Finally, we discuss how to improve the detectability of our signal by cross-correlating it with other quantities sourced by the flow of intergalactic electrons.
The LIGO/Virgo gravitational-wave (GW) interferometers have to-date detected ten merging black hole (BH) binaries, some with masses considerably larger than had been anticipated. Stellar-mass BH binaries at the high end of the observed mass range (with chirp mass ${cal M} gtrsim 25 M_{odot}$) should be detectable by a space-based GW observatory years before those binaries become visible to ground-based GW detectors. This white paper discusses some of the synergies that result when the same binaries are observed by instruments in space and on the ground. We consider intermediate-mass black hole binaries (with total mass $M sim 10^2 -10^4 M_{odot}$) as well as stellar-mass black hole binaries. We illustrate how combining space-based and ground-based data sets can break degeneracies and thereby improve our understanding of the binarys physical parameters. While early work focused on how space-based observatories can forecast precisely when some mergers will be observed on the ground, the reverse is also important: ground-based detections will allow us to dig deeper into archived, space-based data to confidently identify black hole inspirals whose signal-to-noise ratios were originally sub-threshold, increasing the number of binaries observed in both bands by a factor of $sim 4 - 7$.
STPpol, POLARBEAR and BICEP2 have recently measured the cosmic microwave background (CMB) B-mode polarization in various sky regions of several tens of square degrees and obtained BB power spectra in the multipole range 20-3000, detecting the components due to gravitational lensing and to inflationary gravitational waves. We analyze jointly the results of these three experiments and propose modifications of their analysis of the spectra to include in the model, in addition to the gravitational lensing and the inflationary gravitational waves components, also the effects induced by the cosmic polarization rotation (CPR), if it exists within current upper limits. Although in principle our analysis would lead also to new constraints on CPR, in practice these can only be given on its fluctuations <{delta}{alpha}^2>, since constraints on its mean angle are inhibited by the de-rotation which is applied by current CMB polarization experiments, in order to cope with the insufficient calibration of the polarization angle. The combined data fits from all three experiments (with 29% CPR-SPTpol correlation, depending on theoretical model) gives constraint <{delta}{alpha}^2>^1/2 < 27.3 mrad (1.56{deg}) with r = 0.194 pm 0.033. These results show that the present data are consistent with no CPR detection and the constraint on CPR fluctuation is about 1.5{deg}. This method of constraining the cosmic polarization rotation is new, is complementary to previous tests, which use the radio and optical/UV polarization of radio galaxies and the CMB E-mode polarization, and adds a new constraint for the sky areas observed by SPTpol, POLARBEAR and BICEP2.
We investigate the effects of multi-task learning using the recently introduced task of semantic tagging. We employ semantic tagging as an auxiliary task for three different NLP tasks: part-of-speech tagging, Universal Dependency parsing, and Natural Language Inference. We compare full neural network sharing, partial neural network sharing, and what we term the learning what to share setting where negative transfer between tasks is less likely. Our findings show considerable improvements for all tasks, particularly in the learning what to share setting, which shows consistent gains across all tasks.
Learning problems form an important category of computational tasks that generalizes many of the computations researchers apply to large real-life data sets. We ask: what concept classes can be learned privately, namely, by an algorithm whose output does not depend too heavily on any one input or specific training example? More precisely, we investigate learning algorithms that satisfy differential privacy, a notion that provides strong confidentiality guarantees in contexts where aggregate information is released about a database containing sensitive information about individuals. We demonstrate that, ignoring computational constraints, it is possible to privately agnostically learn any concept class using a sample size approximately logarithmic in the cardinality of the concept class. Therefore, almost anything learnable is learnable privately: specifically, if a concept class is learnable by a (non-private) algorithm with polynomial sample complexity and output size, then it can be learned privately using a polynomial number of samples. We also present a computationally efficient private PAC learner for the class of parity functions. Local (or randomized response) algorithms are a practical class of private algorithms that have received extensive investigation. We provide a precise characterization of local private learning algorithms. We show that a concept class is learnable by a local algorithm if and only if it is learnable in the statistical query (SQ) model. Finally, we present a separation between the power of interactive and noninteractive local learning algorithms.
Log in to be able to interact and post comments
comments
Fetching comments
Sign in to be able to follow your search criteria
