Filter and export

Publications

What is a Publication?
1549 Publications visible to you, out of a total of 1549
Non-crossing quantile regression neural network as a calibration tool for ensemble weather forecasts

Abstract

Not specified

Authors: Mengmeng Song, Dazhi Yang, Sebastian Lerch, Xiang’ao Xia, Gokhan Mert Yagli, Jamie M. Bright, Yanbo Shen, Bai Liu, Xingli Liu, Martin János Mayer

Date Published: 1st Mar 2024

Publication Type: Journal

Effect of nuclear reactions rates and core boundary mixing on the seismology of red clump stars

Abstract (Expand)

Context: Modeling of the stars in the red clump (RC), that is, core helium-burning stars that have gone through a He flash, is challenging because of the uncertainties associated with the physical … processes in their core and during the helium flash. By probing the internal stellar structure, asteroseismology allows us to constrain the core properties of RC stars and eventually, to improve our understanding of this evolutionary phase. Aims: We aim to quantify the impact on the seismic properties of the RC stars of the two main core modeling uncertainties: core boundary mixing, and helium-burning nuclear reaction rates. Methods: Using the MESA stellar evolution code, we computed models with different core boundary mixing as well as different 3α and 12C(α, γ)16O nuclear reaction rates. We investigated the impact of these parameters on the period spacing ΔΠ, which is a probe of the region around the core. Results: We find that different core boundary mixing schemes yield significantly different period spacings, with differences of 30 s between the maximum ΔΠ value computed with semiconvection and maximal overshoot. We show that an increased rate of 12C(α, γ)16O lengthens the core helium-burning phase, which extends the range of ΔΠ covered by the models during their evolution. This results in a difference of 10 s between the models computed with a nominal rate and a rate multiplied by 2, which exceeds the observational uncertainties. The effect of changing the 3α reaction rate is comparatively small. Conclusions: The core boundary mixing is the main source of uncertainty in the seismic modeling of RC stars. Moreover, the effect of the 12C(α, γ)16O is non-negligible, even though it is difficult to distinguish from the effect of the mixing. This degeneracy could be seen more frequently in the future in the new seismic data from the PLATO mission and through theoretical constraints from numerical simulations.

Authors: Anthony Noll, Sarbani Basu, Saskia Hekker

Date Published: 1st Mar 2024

Publication Type: Journal

Towards explainable evaluation metrics for machine translation

Abstract (Expand)

Unlike classical lexical overlap metrics such as BLEU, most current evaluation metrics for machine translation (for example, COMET or BERTScore) are based on black-box large language models. They often … achieve strong correlations with human judgments, but recent research indicates that the lower-quality classical metrics remain dominant, one of the potential reasons being that their decision processes are more transparent. To foster more widespread acceptance of novel high-quality metrics, explainability thus becomes crucial. In this concept paper, we identify key properties as well as key goals of explainable machine translation metrics and provide a comprehensive synthesis of recent techniques, relating them to our established goals and properties. In this context, we also discuss the latest state-of-the-art approaches to explainable metrics based on generative models such as ChatGPT and GPT4. Finally, we contribute a vision of next-generation approaches, including natural language explanations. We hope that our work can help catalyze and guide future research on explainable evaluation metrics and, mediately, also contribute to better and more transparent machine translation systems.

Authors: Christoph Leiter, Piyawat Lertvittayakumjorn, Marina Fomicheva, Wei Zhao, Yang Gao, Steffen Eger

Date Published: 1st Mar 2024

Publication Type: Journal

1100 days in the life of the supernova 2018ibb. The best pair-instability supernova candidate, to date

Abstract

Not specified

Authors: Steve Schulze, Claes Fransson, Alexandra Kozyreva, Ting-Wan Chen, Ofer Yaron, Anders Jerkstrand, Avishay Gal-Yam, Jesper Sollerman, Lin Yan, Tuomas Kangas, Giorgos Leloudas, Conor M. B. Omand, Stephen J. Smartt, Yi Yang, Matt Nicholl, Nikhil Sarin, Yuhan Yao, Thomas G. Brink, Amir Sharon, Andrea Rossi, Ping Chen, Zhihao Chen, Aleksandar Cikota, Kishalay De, Andrew J. Drake, Alexei V. Filippenko, Christoffer Fremling, Laurane Fréour, Johan P. U. Fynbo, Anna Y. Q. Ho, Cosimo Inserra, Ido Irani, Hanindyo Kuncarayakti, Ragnhild Lunnan, Paolo Mazzali, Eran O. Ofek, Eliana Palazzi, Daniel A. Perley, Miika Pursiainen, Barry Rothberg, Luke J. Shingles, Ken Smith, Kirsty Taggart, Leonardo Tartaglia, WeiKang Zheng, Joseph P. Anderson, Letizia Cassara, Eric Christensen, S. George Djorgovski, Lluı́s Galbany, Anamaria Gkini, Matthew J. Graham, Mariusz Gromadzki, Steven L. Groom, Daichi Hiramatsu, D. Andrew Howell, Mansi M. Kasliwal, Curtis McCully, Tomás E. Müller-Bravo, Simona Paiano, Emmanouela Paraskeva, Priscila J. Pessi, David Polishook, Arne Rau, Mickael Rigault, Ben Rusholme

Date Published: 1st Mar 2024

Publication Type: Journal

Substituting density functional theory in reaction barrier calculations for hydrogen atom transfer in proteins

Abstract (Expand)

Hydrogen atom transfer (HAT) reactions are important in many biological systems. As these reactions are hard to observe experimentally, it is of high interest to shed light on them using simulations. … Here, we present a machine learning model based on graph neural networks for the prediction of energy barriers of HAT reactions in proteins. As input, the model uses exclusively non-optimized structures as obtained from classical simulations. It was trained on more than 17 000 energy barriers calculated using hybrid density functional theory. We built and evaluated the model in the context of HAT in collagen, but we show that the same workflow can easily be applied to HAT reactions in other biological or synthetic polymers. We obtain for relevant reactions (small reaction distances) a model with good predictive power (R2 ∼ 0.9 and mean absolute error of <3 kcal mol−1). As the inference speed is high, this model enables evaluations of dozens of chemical situations within seconds. When combined with molecular dynamics in a kinetic Monte-Carlo scheme, the model paves the way toward reactive simulations.

Authors: Kai Riedmiller, Patrick Reiser, Elizaveta Bobkova, Kiril Maltsev, Ganna Gryn’ova, Pascal Friederich, Frauke Gräter

Date Published: 14th Feb 2024

Publication Type: Journal

Easy uncertainty quantification (EasyUQ): Generating predictive distributions from single-valued model output

Abstract

Not specified

Authors: Eva-Maria Walz, Alexander Henzi, Johanna Ziegel, Tilmann Gneiting

Date Published: 8th Feb 2024

Publication Type: Journal

Testing quantile forecast optimality

Abstract

Not specified

Authors: Jack Fosten, Daniel Gutknecht, Marc-Oliver Pohle

Date Published: 8th Feb 2024

Publication Type: Journal

Powered by
 (v.1.14.2)
About HITS Publication SEEK | About HITS-Publication-SEEK | Funding and Programmes | Credits
Copyright © 2008 - 2023 The University of Manchester and HITS gGmbH