Cosmology

Abstract:

We present emission line ratios from a sample of 26 Lyman break galaxies from $z\sim5.5-9.5$ with $-17.05.5$ spectra. We find that the emission line ratios exhibited by these $z\sim5.5-9.5$ galaxies occupy clearly distinct regions of line-ratio space compared to typical z~0-3 galaxies, instead being more consistent with extreme populations of lower-redshift galaxies. This is best illustrated by the [OIII]/[OII] ratio, tracing interstellar medium (ISM) ionisation, in which we observe more than half of our sample to have [OIII]/[OII]>10. Our high signal-to-noise spectra reveal more than an order of magnitude of scatter in line ratios such as [OII]/H$\beta$ and [OIII]/[OII], indicating significant diversity in the ISM conditions within the sample. We find no convincing detections of [NII] in our sample, either in individual galaxies, or a stack of all G395M/F290LP spectra. The emission line ratios observed in our sample are generally consistent with galaxies with extremely high ionisation parameters (log $U\sim-1.5$), and a range of metallicities spanning from $\sim0.1\times Z_\odot$ to higher than $\sim0.3\times Z_\odot$, suggesting we are probing low-metallicity systems undergoing periods of rapid star-formation, driving strong radiation fields. These results highlight the value of deep observations in constraining the properties of individual galaxies, and hence probing diversity within galaxy population.Comment: 20 pages, 9 figures, submitted to Astronomy & Astrophysics, updated values in table

Abstract:

When choosing between competing symbolic models for a data set, a human will naturally prefer the “simpler” expression or the one which more closely resembles equations previously seen in a similar context. This suggests a non-uniform prior on functions, which is, however, rarely considered within a symbolic regression (SR) framework. In this paper we develop methods to incorporate detailed prior information on both functions and their parameters into SR. Our prior on the structure of a function is based on a ngram language model, which is sensitive to the arrangement of operators relative to one another in addition to the frequency of occurrence of each operator. We also develop a formalism based on the Fractional Bayes Factor to treat numerical parameter priors in such a way that models may be fairly compared though the Bayesian evidence, and explicitly compare Bayesian, Minimum Description Length and heuristic methods for model selection. We demonstrate the performance of our priors relative to literature standards on benchmarks and a real-world dataset from the field of cosmology.