Duncan Cameron-Steinke

arXiv2026

Overview of the Canadian Hydrogen Observatory and Radio Transient Detector (CHORD) Project

The Canadian Hydrogen Observatory and Radio-transient Detector (CHORD) is a next-generation wideband radio interferometer currently being constructed and commissioned at the Dominion Radio Astrophysical Observatory in British Columbia, Canada. CHORD is designed for precision 21\,cm cosmology, fast radio transient discovery, spectral line galaxy surveys, and pulsar science using a highly redundant large-N, small-diameter drift-scan array architecture. The telescope consists of a 512-element core array of 6\,m dishes operating from 300--1500\,MHz in drift-scan mode, together with two 64-dish outrigger stations located at the Hat Creek Radio Observatory and the Green Bank Observatory for long-baseline transient localization. The instrument supports multiple simultaneous digital backends for interferometric correlation, FRB detection, pulsar beamforming, and high spectral resolution surveys. CHORD is designed with an emphasis on precision beam control and stable instrumental response, incorporating lessons learned from the Canadian Hydrogen Intensity Mapping Experiment (CHIME) while providing a substantial increase in sensitivity. Initial performance has been evaluated using a three-dish engineering array, and a 64-dish pathfinder array is currently being commissioned. The full array will be commissioned in 2028.

astro-ph.IM↗

arXiv2019

Short-Text Classification Using Unsupervised Keyword Expansion

Short-text classification, like all data science, struggles to achieve high performance using limited data. As a solution, a short sentence may be expanded with new and relevant feature words to form an artificially enlarged dataset, and add new features to testing data. This paper applies a novel approach to text expansion by generating new words directly for each input sentence, thus requiring no additional datasets or previous training. In this unsupervised approach, new keywords are formed within the hidden states of a pre-trained language model and then used to create extended pseudo documents. The word generation process was assessed by examining how well the predicted words matched to topics of the input sentence. It was found that this method could produce 3-10 relevant new words for each target topic, while generating just 1 word related to each non-target topic. Generated words were then added to short news headlines to create extended pseudo headlines. Experimental results have shown that models trained using the pseudo headlines can improve classification accuracy when limiting the number of training examples.

cs.CL↗