• Digital Humanities and Cultural Heritage
• 2012-2021close
• EE close
• arXiv.org e-Print Archive close
• Digital Humanities and Cultural Her... close

Certain models of cosmic inflation produce strong cosmological perturbations at short length scales, which may later collapse into primordial black holes. To find the statistics of these strong perturbations and the ensuing black holes, it is necessary to go beyond linear perturbation theory. Stochastic inflation provides a way to take the leading non-linear effects into account. In this contribution, I discuss recent progress in numerical computations of stochastic inflation. A numerical approach can include more of the non-linearities than an analytical calculation, and can be applied to single-field inflationary models with any potential. Comment: 5 pages, 2 figures. Based on Phys.Rev.Lett. 127 (2021) 10, 101302, arXiv:2012.06551. Proceedings of the 17th International Conference on Topics in Astroparticle and Underground Physics (TAUP 2021)

We present two scale invariant models of inflation in which the addition of quadratic in curvature terms in the usual Einstein-Hilbert action, in the context of Palatini formulation of gravity, manages to reduce the value of the tensor-to-scalar ratio. In both models the Planck scale is dynamically generated via the vacuum expectation value of the scalar fields. Comment: 10 pages, 5 figures, Contribution to Proceedings of HEP 2021, Thessaloniki, Greece, Based on: 2006.09124 and 2104.04550. arXiv admin note: text overlap with arXiv:2104.04550

Manual subject indexing in libraries is a time-consuming and costly process and the quality of the assigned subjects is affected by the cataloguer's knowledge on the specific topics contained in the book. Trying to solve these issues, we exploited the opportunities arising from artificial intelligence to develop Kratt: a prototype of an automatic subject indexing tool. Kratt is able to subject index a book independent of its extent and genre with a set of keywords present in the Estonian Subject Thesaurus. It takes Kratt approximately 1 minute to subject index a book, outperforming humans 10-15 times. Although the resulting keywords were not considered satisfactory by the cataloguers, the ratings of a small sample of regular library users showed more promise. We also argue that the results can be enhanced by including a bigger corpus for training the model and applying more careful preprocessing techniques. Comment: This is a preprint version. It has 12 pages, 5 figures, 3 tables

Automatic emotion recognition has become a trending research topic in the past decade. While works based on facial expressions or speech abound, recognizing affect from body gestures remains a less explored topic. We present a new comprehensive survey hoping to boost research in the field. We first introduce emotional body gestures as a component of what is commonly known as "body language" and comment general aspects as gender differences and culture dependence. We then define a complete framework for automatic emotional body gesture recognition. We introduce person detection and comment static and dynamic body pose estimation methods both in RGB and 3D. We then comment the recent literature related to representation learning and emotion recognition from images of emotionally expressive gestures. We also discuss multi-modal approaches that combine speech or face with body gestures for improved emotion recognition. While pre-processing methodologies (e.g., human detection and pose estimation) are nowadays mature technologies fully developed for robust large scale analysis, we show that for emotion recognition the quantity of labelled data is scarce. There is no agreement on clearly defined output spaces and the representations are shallow and largely based on naive geometrical representations.

It is commonly assumed that the energy density of the Universe was dominated by radiation between reheating after inflation and the onset of matter domination 54,000 years later. While the abundance of light elements indicates that the Universe was radiation dominated during Big Bang Nucleosynthesis (BBN), there is scant evidence that the Universe was radiation dominated prior to BBN. It is therefore possible that the cosmological history was more complicated, with deviations from the standard radiation domination during the earliest epochs. Indeed, several interesting proposals regarding various topics such as the generation of dark matter, matter-antimatter asymmetry, gravitational waves, primordial black holes, or microhalos during a nonstandard expansion phase have been recently made. In this paper, we review various possible causes and consequences of deviations from radiation domination in the early Universe - taking place either before or after BBN - and the constraints on them, as they have been discussed in the literature during the recent years. Comment: 67 pages, 18 figures. v2: Discussion and references added. Accepted for publication in The Open Journal of Astrophysics

This paper investigates the use of automatically collected web audio data for the task of spoken language recognition. We generate semi-random search phrases from language-specific Wikipedia data that are then used to retrieve videos from YouTube for 107 languages. Speech activity detection and speaker diarization are used to extract segments from the videos that contain speech. Post-filtering is used to remove segments from the database that are likely not in the given language, increasing the proportion of correctly labeled segments to 98%, based on crowd-sourced verification. The size of the resulting training set (VoxLingua107) is 6628 hours (62 hours per language on the average) and it is accompanied by an evaluation set of 1609 verified utterances. We use the data to build language recognition models for several spoken language identification tasks. Experiments show that using the automatically retrieved training data gives competitive results to using hand-labeled proprietary datasets. The dataset is publicly available. Comment: Accepted at IEEE Spoken Language Technology Workshop (SLT) 2021

Translating text into a language unknown to the text's author, dubbed outbound translation, is a modern need for which the user experience has significant room for improvement, beyond the basic machine translation facility. We demonstrate this by showing three ways in which user confidence in the outbound translation, as well as its overall final quality, can be affected: backward translation, quality estimation (with alignment) and source paraphrasing. In this paper, we describe an experiment on outbound translation from English to Czech and Estonian. We examine the effects of each proposed feedback module and further focus on how the quality of machine translation systems influence these findings and the user perception of success. We show that backward translation feedback has a mixed effect on the whole process: it increases user confidence in the produced translation, but not the objective quality. 9 pages (excluding references); to appear at NAACL-HWT 2021