Abstract
By a News Reporter-Staff News Editor at Robotics & Machine Learning Daily News Daily News-Current study results on Machine Learn ing have been published. According to news originating from Hawthorn, Australia, by NewsRx correspondents, research stated, "Machine learning photo-z methods, t rained directly on spectroscopic redshifts, provide a viable alternative to trad itional template-fitting methods but may not generalize well on new data that de viates from that in the training set. In this work, we present a Hybrid Algorith m for WI(Y)de-range photo-z estimation with Artificial neural networks and TEmpl ate fitting (hayate), a novel photo-z method that combines template fitting and data-driven approaches and whose training loss is optimized in terms of both red shift point estimates and probability distributions." Financial supporters for this research include Australian Research Council, Aust ralian Research Council. Our news journalists obtained a quote from the research from the Swinburne Unive rsity of Technology, "We produce artificial training data from low-redshift gala xy spectral energy distributions (SEDs) at z<1.3, artifici ally redshifted up to z = 5. We test the model on data from the ZFOURGE surveys, demonstrating that hayate can function as a reliable emulator of eazy for the b road redshift range beyond the region of sufficient spectroscopic completeness. The network achieves precise photo-z estimations with smaller errors (sigma(NMAD )) than eazy in the initial low-z region (z <1.3), while b eing comparable even in the high-z extrapolated regime (1.3 <z<5). Meanwhile, it provides more robust photo-z estimati ons than eazy with the lower outlier rate (eta(0.2 )less than or similar to 1 pe r cent) but runs similar to 100 times faster than the original template-fitting method. We also demonstrate hayate offers more reliable redshift probability den sity functions, showing a flatter distribution of Probability Integral Transform scores than eazy. The performance is further improved using transfer learning w ith spec-z samples."
NETL
NSTL