Modeling detailed chemical kinetics is a primary challenge in combustion simulations. We present a novel framework to enforce physical constraints, specifically total mass and elemental conservation, during the reaction of ML models’ training for the reduced composition space chemical kinetics of large chemical mechanisms in combustion. In these models, the transport equations for a subset of representative species are solved with the ML approaches, while the remaining nonrepresentative species are “recovered” with a separate artificial neural network trained on data. Given the strong correlation between full and reduced solution vectors, our method utilizes a small neural network to establish an accurate and physically consistent mapping. By leveraging this mapping, we enforce physical constraints in the training process of the ML model for reduced composition space chemical kinetics. The framework is demonstrated here for methane, CH4, and oxidation. The resulting solution vectors from our deep operator networks (DeepONet)-based approach are accurate and align more consistently with physical laws.

This study investigates the integration of literal completions of idiomatic multiword expressions (MWEs) into two linguistic contexts: one promoting a literal interpretation and the other a figurative one, requiring reinterpretation to align with figurative bias. Sixteen Italian idioms were distributed in two groups by their Potential Idiomatic Ambiguity (PIA) score, an index of literal plausibility, decomposability and transparency. Using experimental dialogues, the study tested whether high-PIA idioms receive higher acceptability ratings across both contexts than low-PIA idioms. Eighty-four Italian-speaking participants rated idiom literal completions within literal and figurative contexts. Results show that literal completions of high-PIA idioms integrate better across contexts, while those of low-PIA idioms receive lower ratings and have longer combined reading and rating times. This supports hybrid models of idiom processing, emphasizing the role of idiomatic features and context in balancing figurative and compositional interpretations. This study also marks an initial effort to experimentally trace systematicity within idiomatic wordplay, challenging the idea that it lacks relevance for linguistic research while outlining limitations and directions for future work.

Eddy currents play a significant role in the evolution of tokamak plasmas and must therefore be correctly taken into account in time-dependent simulations. In this paper, a computational method for solving the evolution of tokamak plasma considering eddy currents utilising VMEC (Hirshman & Whitson, Phys. Fluids, vol. 26, 1983, pp. 3553–3568), a commonly used static magnetohydrodynamic equilibrium solver, is proposed. This method is convenient since it does not modify the equilibrium solver internally and achieves convergence calculation through external processing. By allowing the components of the magnetic field to be treated separately, this method provides convergence for cases with displacements in arbitrary directions, which has been difficult to achieve with the previous methods.

Macro- and microbotanical remains recovered from post-Teotihuacan occupations in quarry tunnels east of the Sun Pyramid, Teotihuacan, contribute to understanding lifeways in the surrounding valley after the partial abandonment of the city. Plant remains associated with domestic and ritual contexts from the excavations directed by Linda Manzanilla (1993–1996) are relevant to subsistence questions, aspects of surrounding vegetation, landscape exploitation, and the possibility of less-intensive agricultural production during the Epiclassic and Early Postclassic occupations.

Standard approaches to neoclassical theory do not extend into regions of strong gradients in tokamaks such as the pedestal and internal transport barriers. Here, we calculate the modifications to neoclassical electron physics inside strong gradient regions of large aspect ratio tokamaks in the banana regime. We show that these modifications are due to the different ion flow and the strong poloidal variation of the potential. We also provide a physical interpretation of the mechanisms that drive poloidal asymmetries and hence a poloidal electric field. We apply our model to two specific example cases of pedestal profiles, calculating the neoclassical electron flux and the bootstrap current. We find that, depending on the ion flow, weak gradient neoclassical theory overestimates or underestimates the neoclassical electron transport and the bootstrap current in regions with strong gradients. We show that the determination of the mean parallel flow is more complex than in weak gradient neoclassical theory. For vanishing turbulence, we can determine the radial electric field for a given flow profile in the pedestal.

Homophonous morphs have been reported to show differences in acoustic duration in languages such as English and German. How common these differences are across languages, and what factors influence the extent of temporal differences, is still an open question, however. This paper investigates the role of morphological disambiguation in predicting the acoustic duration of homophones using data from a diverse sample of 37 languages. Results indicate a low overall contribution of morphological affiliation compared to other well-studied effects on duration such as speech rate and Final Lengthening. It is proposed that two factors increase the importance of homophony avoidance for the acoustic shape of morphs: crowdedness (i.e. the number of competing homophones) and segmental make-up, in particular the presence of an alveolar fricative. These findings offer an empirically broad perspective on the interplay between morphology and phonetics and align with the view of language as an adaptive and efficient system.

The present study examines whether presenting words in song versus spoken sentences can lead to differences in word learning in 47–50-month-old children. This work extends previous findings on this topic and evaluates whether the location of pitch changes within the song may contribute to how well the words are learned. Using a Preferential Looking Paradigm, 32 children were taught the names of objects, either in spoken sentences or in a song that followed an unfamiliar melody. In both conditions, the novel word was emphasized by a pitch change. Looking patterns indicated that children learned the names of the novel items better when the words were trained in the spoken sentence compared to the song condition. The findings are discussed in relation to theories of word learning, and how differences in the characteristics between speech and song may relate to variability in how well new words are acquired.

In the past decade, researchers have been increasingly interested in understanding the process of language learning, in addition to the effect of instructional interventions on L2 performance gains (i.e., learning products). One goal of such investigations is to reveal the interplay between learning conditions, processes, and outcomes where, for example, certain conditions can promote attention to the learning targets, which in turn facilitates learning. However, the statistical modeling approach taken often does not align with the conceptualization of the complex relationships between these variables. Thus, in this paper, we introduce mediation analysis to SLA research. We offer a step-by-step, contextualized tutorial on the practical application of mediation analysis in three different research scenarios, each addressing a different research design using either simulated or open-source datasets. Our overall goal is to promote the use of statistical techniques that are consistent with the theorization of language learning processes as mediators.

Random-effects meta-analyses with only a few studies often face challenges in accurately estimating between-study heterogeneity, leading to biased effect estimates and confidence intervals with poor coverage. This issue is especially the case when dealing with rare diseases. To address this problem for normally distributed outcomes, two new approaches have been proposed to provide confidence limits of the global mean: one based on fiducial inference, and the other involving two modifications of the signed log-likelihood ratio test statistic in order to have improved performance with small numbers of studies. The performance of the proposed methods was evaluated numerically and compared with the Hartung–Knapp–Sidik–Jonkman approach and its modification to handle small numbers of studies. The simulation results indicated that the proposed methods achieved coverage probabilities closer to the nominal level and produced shorter confidence intervals compared to those based on existing methods. Two real examples are used to illustrate the proposed methods.