Search

Hate speech relies heavily on cultural influences, leading to varying individual interpretations. For that reason, we propose a Semantic Componential Analysis (SCA) framework for a cross-cultural and cross-domain analysis of hate speech definitions. We create the first dataset of definitions derived from five domains: online dictionaries, research papers, Wikipedia articles, legislation, and online platforms, which are later analyzed into semantic components. Our analysis reveals that the components differ from definition to definition, yet many domains borrow definitions from one another without taking into account the target culture. We conduct zero-shot model experiments using our proposed dataset, employing three popular open-sourced LLMs to understand the impact of different definitions on hate speech detection. Our findings indicate that LLMs are sensitive to definitions: responses for hate speech detection change according to the complexity of definitions used in the prompt.

The notion of projection families generalizes the classical notions of vector and operator-valued measures. We show that projection families are general enough to extend the Spectral Theorem to Banach algebras and operators between Banach spaces. To this end, we first develop a Smooth Functional Calculus in Banach algebras using the Cauchy-Pompeiu Formula, which is further extended to a Continuous Functional Calculus. We also show that these theorems are proper generalizations of the usual result for operators between Hilbert spaces.

We construct a new kind of measures, called projection families, which generalize the classical notion of vector and operator-valued measures. The maximal class of reasonable functions admits an integral with respect to a projection family, where the integral is defined as an element of the second dual instead of the original space. We show that projection families possess strong enough properties to satisfy the theorems of Monotone Convergence and Dominated convergence, but are much easier to come by than the more restrictive operator-valued measures.

We propose misogyny detection as an Argumentative Reasoning task and we investigate the capacity of large language models (LLMs) to understand the implicit reasoning used to convey misogyny in both Italian and English. The central aim is to generate the missing reasoning link between a message and the implied meanings encoding the misogyny. Our study uses argumentation theory as a foundation to form a collection of prompts in both zero-shot and few-shot settings. These prompts integrate different techniques, including chain-of-thought reasoning and augmented knowledge. Our findings show that LLMs fall short on reasoning capabilities about misogynistic comments and that they mostly rely on their implicit knowledge derived from internalized common stereotypes about women to generate implied assumptions, rather than on inductive reasoning.

We developed a rapid polymorphic screening approach based on contracting sessile microdroplets which offers several advantages: (1) achieves very high supersaturation to facilitate formation of metastable forms (2) allows systematic labeling of samples (3) gives access to statistical distribution of polymorphic selectivity as a function of experimental conditions (4) ensures the formation of crystal for each droplet, addressing the problem of uncrystallized droplets in traditional microfluidics. We studied the competitive nucleation of D-mannitol polymorphs and investigated the effect of droplet volume on polymorphic selectivity. We showed that our observed polymorph distributions at different volumes are qualitatively consistent with the predictions of classical nucleation theory except for very small volumes where thermodynamic confinement or surface effects could play a substantial role. Overall, our microfluidic approach can be a promising tool not only for routine screening of pharmaceutical polymorphs in the industrial context but also in the fundamental understanding of the mechanisms underlying the competitive nucleation of polymorphs.

We show that every operator in $L^{2}$ has an associated measure on a space of functions and prove that it can be used to find solutions to abstract Cauchy problems, including partial differential equations. We find explicit formulas to compute the integral of functions with respect to this measure and develop approximate formulas in terms of a perturbative expansion. We show that this method can be used to represent solutions of classical equations, such as the diffusion and Fokker-Plank equations, as Wiener and Martin-Siggia-Rose-Jansen-de Dominics integrals, and propose an extension to paths in infinite dimensional spaces., Comment: Typos corrected, comments added

We introduce the Guideline-Centered annotation process, a novel data annotation methodology focused on reporting the annotation guidelines associated with each data sample. We identify three main limitations of the standard prescriptive annotation process and describe how the Guideline-Centered methodology overcomes them by reducing the loss of information in the annotation process and ensuring adherence to guidelines. Additionally, we discuss how the Guideline-Centered enables the reuse of annotated data across multiple tasks at the cost of a single human-annotation process.

With the increasing role of Natural Language Processing (NLP) in various applications, challenges concerning bias and stereotype perpetuation are accentuated, which often leads to hate speech and harm. Despite existing studies on sexism and misogyny, issues like homophobia and transphobia remain underexplored and often adopt binary perspectives, putting the safety of LGBTQIA+ individuals at high risk in online spaces. In this paper, we assess the potential harm caused by sentence completions generated by English large language models (LLMs) concerning LGBTQIA+ individuals. This is achieved using QueerBench, our new assessment framework, which employs a template-based approach and a Masked Language Modeling (MLM) task. The analysis indicates that large language models tend to exhibit discriminatory behaviour more frequently towards individuals within the LGBTQIA+ community, reaching a difference gap of 7.2% in the QueerBench score of harmfulness.

Misogyny is often expressed through figurative language. Some neutral words can assume a negative connotation when functioning as pejorative epithets. Disambiguating the meaning of such terms might help the detection of misogyny. In order to address such task, we present PejorativITy, a novel corpus of 1,200 manually annotated Italian tweets for pejorative language at the word level and misogyny at the sentence level. We evaluate the impact of injecting information about disambiguated words into a model targeting misogyny detection. In particular, we explore two different approaches for injection: concatenation of pejorative information and substitution of ambiguous words with univocal terms. Our experimental results, both on our corpus and on two popular benchmarks on Italian tweets, show that both approaches lead to a major classification improvement, indicating that word sense disambiguation is a promising preliminary step for misogyny detection. Furthermore, we investigate LLMs' understanding of pejorative epithets by means of contextual word embeddings analysis and prompting.

We study a generalization of the the Starobinsky model adding a term of the form $R^{2p}$ to the Einstien-Hilbert action. We take the power $p$ as a parameter of the model and explore the constraints from CMB plus BAO data through a Bayesian analysis, thus exploring a range of values for the exponent parameter. We incorporate a reheating phase to the model through the background matter content (equation of state) and the duration of this period (number of $e$-foldings of reheating). We find that incorporating information from reheating imposes constraints on cosmological quantities, more stringent than the case of no reheating when tested with the Planck+BAO data. The inferred value of the exponent parameter is statistically consistent with $p=1$, favoring the original Starobinsky potential. Moreover, we report tighter constraints on $p$ and the number of $e$-folds in comparison with previous works. The obtained values for other inflationary observational parameters, such as the scalar spectral index $n_s$ and the scalar amplitude of perturbations $A_s$, are consistent with prior measurements. Finally we present the alternative use of consistency relations in order to simplify the parameter space and test the generalized Starobinsky potential even more efficiently., Comment: 12 pages, 3 figures

We introduce the Wide-field Retrieval of Astrodata Program (WRAP), a tool created to aid astronomers in gathering photometric and astrometric data for point sources that may confuse simple cross-matching algorithms because of their faintness or motion. WRAP allows astronomers to correctly cross-identify objects with proper motion across multiple surveys by wedding the catalog data with its underlying images, thus providing visual confirmation of cross-associations in real time. Developed within the Backyard Worlds: Planet 9 citizen science project, WRAP aims to aid in the characterization of faint, high motion sources by this collaboration (and others)., Comment: 4 pages, 1 figure, Github and DOI link to program

The research aims to assess the impact of foreign direct investment (FDI) and domestic investment on Chile's economic growth. By elucidating the relationship between FDI and domestic investment, the study contributes valuable insights for economic policy formulation and future investments. The findings hold significance in shaping Chile's international perception as an investment destination, potentially influencing its standing in the global economic landscape. Demonstrating that FDI is a significant driver of economic growth could enhance confidence among foreign investors. The project's importance lies in contributing to economic knowledge and guiding strategic decisions for sustainable economic growth in Chile. Understanding the interplay of FDI and domestic investment allows for a balanced approach, promoting stable economic development and mitigating issues like excessive reliance on foreign investment. The study highlights the theory of internationalization as a conceptual framework for understanding the motives and strategies of multinational companies investing abroad. Leveraging data from sources like the Central Bank of Chile, the research analyzes variables such as Chile's economic growth (GDP), FDI, and domestic investment. The hypothesis posits a significant long-term causal relationship between FDI, National Investment (NI), and Chile's Economic Growth (GDP). Statistical analysis using the Eviews 6 software tool confirms that attracting foreign investments and promoting internal investment are imperative for sustainable economic growth in Chile., Comment: in Spanish language

Solubility and interfacial energy are two fundamental parameters underlying the competitive nucleation of polymorphs. However, solubility measurement of metastable phases comes with a risk of solventmediated transformations which can render the results unreliable. In this work, we present a rapid microfluidic technique for measuring aqueous solubility of the metastable form using KDP Phase IV as a model system. This bracketing approach involves analyzing the dissolution behavior of crystals in contact with supersaturated microdroplets generated via evaporation. Then, with the help of our recently developed nucleation time measurement technique, together with Mersmann calculation of interfacial energies from solubilities, we were able to access the interfacial energies of both metastable and stable phases. To gain further insights into the observed nucleation behavior, we employed the Classical Nucleation Theory (CNT) to model the competition of polymorphs using our measured solubility and calculated interfacial energies. The results show that the stable form is favored at lower supersaturation while the metastable form is favored at higher supersaturation, in good agreement with our observations and experimental reports in the literature. Overall, our microfluidic approach allows access to unprecedentedly deep levels of supersaturation and reveals an interesting interplay between thermodynamics and kinetics in polymorphic nucleation. The experimental methods and insights presented herein can be of great interest, notably in the mineral processing and pharmaceutical industry.

The rise of artificial intelligence (AI) has led to various means of integration of AI aimed to provide efficiency in tasks, one of which is career counseling. A key part of getting a job is having a solid resume that passes through the first round of programs and recruiters. It is difficult to find good resources or schedule an appointment with a career counselor to help with editing a resume for a specific role. With the rise of ChatGPT, Bard, and several other AI chat programs it is possible to provide specific, automated feedback on various concerns to suggest places for improvement within the context of career counseling. This paper begins with a quick literature review on the ethical considerations and limitations of AI in career counseling. The authors also have created their own website service, called ResumAI, to test and review the functionality of an AI career counselor. The findings of this study will contribute to the understanding of chat AI ResumAI reviewer programs and sites. The implications of the findings for the field of career counseling, AI development, and ethical practice will be discussed.

With focus on the cosmological evolution of linear perturbations of matter and geometry, we calculate the equivalent expressions to that of the Newtonian and Synchronous gauges within the framework of Unimodular Gravity, being these two gauges commonly used and implemented in Boltzmann codes. An important aspect of our analysis is the inclusion of the energy-momentum current violation, as well as its perturbations. Moreover, for the first time we demonstrate that it is possible to fix both gauges consistently, although as it has been already noticed in previous literature, neither of them is recovered in the sense of the dynamics given in General Relativity for matter and metric fluctuations. Specifically, we show that since the unimodular constraint at the level of linear perturbations lead to only one degree of freedom of scalar modes of metric fluctuations, the dynamics in Unimodular Gravity forces to keep the anisotropic stress in the Newtonian gauge, whereas the cold dark matter comoving frame can not be set in the Synchronous gauge. The physical implications on the density contrast of cold dark matter is reviewed, and the Sachs-Wolfe effect is obtained and compared with previous results in the literature of cosmological perturbations in Unimodular Gravity., Comment: 32 pages, new text and references added. Version accepted for publication in JCAP

We present a numerical analysis of the cosmological evolution of scalar field dark matter (SFDM) in the Boltzmann code $\texttt{CLASS}$, based on a dynamical system analysis of previous works. We show a detailed study of the evolution of the different dynamical variables, and in particular of the energy density and its corresponding linear perturbations. The numerical results are in good agreement with those of the original SFDM equations of motion, and have better accuracy than other approaches. In addition, we calculate the temperature and matter power spectra and discuss the reliability of their numerical results. We also give simple examples in which we can put constraints on the field mass using recent likelihoods incorporated in the Monte Carlo Markov Chain sampler $\texttt{MontePython}$., Comment: 20 pages, 12 figures

We develop novel annotation guidelines for sentence-level subjectivity detection, which are not limited to language-specific cues. We use our guidelines to collect NewsSD-ENG, a corpus of 638 objective and 411 subjective sentences extracted from English news articles on controversial topics. Our corpus paves the way for subjectivity detection in English and across other languages without relying on language-specific tools, such as lexicons or machine translation. We evaluate state-of-the-art multilingual transformer-based models on the task in mono-, multi-, and cross-language settings. For this purpose, we re-annotate an existing Italian corpus. We observe that models trained in the multilingual setting achieve the best performance on the task., Comment: LREC-COLING 2024, pages 273-285

This paper concerns the identification of continuous-time systems in state-space form that are subject to Lebesgue sampling. Contrary to equidistant (Riemann) sampling, Lebesgue sampling consists of taking measurements of a continuous-time signal whenever it crosses fixed and regularly partitioned thresholds. The knowledge of the intersample behavior of the output data is exploited in this work to derive an expectation-maximization (EM) algorithm for parameter estimation of the state-space and noise covariance matrices. For this purpose, we use the incremental discrete-time equivalent of the system, which leads to EM iterations of the continuous-time state-space matrices that can be computed by standard filtering and smoothing procedures. The effectiveness of the identification method is tested via Monte Carlo simulations., Comment: 6 pages, 3 figures

Nucleation, the birth of a stable cluster from disorder, is inherently stochastic. Yet up to date, there are no quantitative studies on NaCl nucleation that accounts for its stochastic nature. Here, we report the first stochastic treatment of NaCl-water nucleation kinetics. Using a recently developed microfluidic system and evaporation model, our measured interfacial energies extracted from a modified Poisson distribution of nucleation time show an excellent agreement with theoretical predictions. Furthermore, analysis of nucleation parameters in 0.5 pL, 1.5pL and 5.5 pL microdroplets reveals an interesting interplay between kinetic confinement and shifting of nucleation mechanisms. Overall, our findings highlight the need to treat nucleation stochastically rather deterministically to bridge the gap between theory and experiment.

We perform a Bayesian study of a generalization of the basic $\alpha$-attractor T model given by the potential $V(\phi)=V_0\left[1-\text{sech}^{p}\left(\phi/\sqrt{6\alpha}M_{pl}\right)\right]$ where $\phi$ is the inflaton field and the parameter $\alpha$ corresponds to the inverse curvature of the scalar manifold in the conformal or superconformal realizations of the attractor models. Such generalization is characterized by the power $p$ which includes the basic or base model for $p=2$. Once the priors for the parameters of the $\alpha$-attractor potential are set by numerical exploration, we perform the corresponding statistical analysis for the cases $p=1\, , 2\, , 3\, ,4$, and derive posteriors. Considering the original $\alpha$-attractor potential as the base model, we calculate the evidence for our generalization, and conclude that the $p=4$ model is preferred by the CMB data. We also present constraints for the parameter $\alpha$. Interestingly, all the cases studied prefer a specific value for the tensor-to-scalar ratio given by $r\simeq 0.0025$., Comment: 17 pages, 8 figures, 4 tables. Minor amendments to match the version published in JCAP