Showing 1–7 of 7 results for author: Hay, G

Function Extrapolation with Neural Networks and Its Application for Manifolds

Authors: Guy Hay, Nir Sharon

Abstract: This paper addresses the problem of accurately estimating a function on one domain when only its discrete samples are available on another domain. To answer this challenge, we utilize a neural network, which we train to incorporate prior knowledge of the function. In addition, by carefully analyzing the problem, we obtain a bound on the error over the extrapolation domain and define a condition nu… ▽ More This paper addresses the problem of accurately estimating a function on one domain when only its discrete samples are available on another domain. To answer this challenge, we utilize a neural network, which we train to incorporate prior knowledge of the function. In addition, by carefully analyzing the problem, we obtain a bound on the error over the extrapolation domain and define a condition number for this problem that quantifies the level of difficulty of the setup. Compared to other machine learning methods that provide time series prediction, such as transformers, our approach is suitable for setups where the interpolation and extrapolation regions are general subdomains and, in particular, manifolds. In addition, our construction leads to an improved loss function that helps us boost the accuracy and robustness of our neural network. We conduct comprehensive numerical tests and comparisons of our extrapolation versus standard methods. The results illustrate the effectiveness of our approach in various scenarios. △ Less

Submitted 17 May, 2024; originally announced May 2024.

Comments: 32 pages, 11 figures

MSC Class: 65K05

Marginal Laplacian Score

Authors: Guy Hay, Ohad Volk

Abstract: High-dimensional imbalanced data poses a machine learning challenge. In the absence of sufficient or high-quality labels, unsupervised feature selection methods are crucial for the success of subsequent algorithms. Therefore, we introduce a Marginal Laplacian Score (MLS), a modification of the well known Laplacian Score (LS) tailored to better address imbalanced data. We introduce an assumption th… ▽ More High-dimensional imbalanced data poses a machine learning challenge. In the absence of sufficient or high-quality labels, unsupervised feature selection methods are crucial for the success of subsequent algorithms. Therefore, we introduce a Marginal Laplacian Score (MLS), a modification of the well known Laplacian Score (LS) tailored to better address imbalanced data. We introduce an assumption that the minority class or anomalous appear more frequently in the margin of the features. Consequently, MLS aims to preserve the local structure of the dataset's margin. We propose its integration into modern feature selection methods that utilize the Laplacian score. We integrate the MLS algorithm into the Differentiable Unsupervised Feature Selection (DUFS), resulting in DUFS-MLS. The proposed methods demonstrate robust and improved performance on synthetic and public datasets. △ Less

Submitted 2 February, 2024; v1 submitted 29 November, 2023; originally announced November 2023.

Comments: 10 pages

ACM Class: I.5.0

SORTAD: Self-Supervised Optimized Random Transformations for Anomaly Detection in Tabular Data

Authors: Guy Hay, Pablo Liberman

Abstract: We consider a self-supervised approach to anomaly detection in tabular data. Random transformations are applied to the data, and then each transformation is identified based on its output. These predicted transformations are used to identify anomalies. In tabular data this approach faces many challenges that are related to the uncorrelated nature of the data. These challenges affect the transforma… ▽ More We consider a self-supervised approach to anomaly detection in tabular data. Random transformations are applied to the data, and then each transformation is identified based on its output. These predicted transformations are used to identify anomalies. In tabular data this approach faces many challenges that are related to the uncorrelated nature of the data. These challenges affect the transformations that should be used, as well as the use of their predictions. To this end, we propose SORTAD, a novel algorithm that is tailor-made to solve these challenges. SORTAD optimally chooses random transformations that help the classification process, and have a scoring function that is more sensitive to the changes in the transformations classification prediction encountered in tabular data. SORTAD achieved state-of-the-art results on multiple commonly used anomaly detection data sets, as well as in the overall results across all data sets tested. △ Less

Submitted 18 November, 2023; originally announced November 2023.

Comments: 16 pages, 7 figures

ACM Class: I.5.1

Recent Developments in Measuring Signal and Noise in Phased Array Feeds at CSIRO

Authors: A. P. Chippendale, D. McConnell, K. Bannister, N. Nikolic, A. W. Hotan, K. W. Smart, R. D. Shaw, D. B. Hayman, S. G. Hay

Abstract: We describe recent developments in measuring both signal and noise in phased array feeds for radio astronomy at CSIRO. We introduce new techniques including aperture array noise measurements with beamforming weights matched to a reflector's focal field. Weights are calculated via antenna-range and in-reflector measurements. We also describe the separation of system temperature and aperture efficie… ▽ More We describe recent developments in measuring both signal and noise in phased array feeds for radio astronomy at CSIRO. We introduce new techniques including aperture array noise measurements with beamforming weights matched to a reflector's focal field. Weights are calculated via antenna-range and in-reflector measurements. We also describe the separation of system temperature and aperture efficiency via drift scans. △ Less

Submitted 12 February, 2016; v1 submitted 1 February, 2016; originally announced February 2016.

Comments: 5 pages, 7 figures, accepted for publication in Antennas and Propagation (EuCAP), 2016 10th European Conference on (updated abstract)

Report number: EP16564

Journal ref: 2016 10th European Conference on Antennas and Propagation (EuCAP), Davos, Switzerland, 2016, pp. 1-5

Measured Aperture-Array Noise Temperature of the Mark II Phased Array Feed for ASKAP

Authors: A. P. Chippendale, A. J. Brown, R. J. Beresford, G. A. Hampson, R. D. Shaw, D. B. Hayman, A. Macleod, A. R. Forsyth, S. G. Hay, M. Leach, C. Cantrall, M. L. Brothers, A. W. Hotan

Abstract: We have measured the aperture-array noise temperature of the first Mk. II phased array feed that CSIRO has built for the Australian Square Kilometre Array Pathfinder telescope. As an aperture array, the Mk. II phased array feed achieves a beam equivalent noise temperature less than 40 K from 0.78 GHz to 1.7 GHz and less than 50 K from 0.7 GHz to 1.8 GHz for a boresight beam directed at the zenith.… ▽ More We have measured the aperture-array noise temperature of the first Mk. II phased array feed that CSIRO has built for the Australian Square Kilometre Array Pathfinder telescope. As an aperture array, the Mk. II phased array feed achieves a beam equivalent noise temperature less than 40 K from 0.78 GHz to 1.7 GHz and less than 50 K from 0.7 GHz to 1.8 GHz for a boresight beam directed at the zenith. We believe these are the lowest reported noise temperatures over these frequency ranges for ambient-temperature phased arrays. The measured noise temperature includes receiver electronics noise, ohmic losses in the array, and stray radiation from sidelobes illuminating the sky and ground away from the desired field of view. This phased array feed was designed for the Australian Square Kilometre Array Pathfinder to demonstrate fast astronomical surveys with a wide field of view for the Square Kilometre Array. △ Less

Submitted 6 November, 2015; v1 submitted 17 September, 2015; originally announced September 2015.

Comments: 5 pages, 4 figures, accepted for publication in Antennas and Propagation (ISAP), 2015 International Symposium on

Report number: EP155528

Journal ref: 2015 International Symposium on Antennas and Propagation (ISAP), Hobart, TAS, 2015, pp. 1-4

Measured Sensitivity of the First Mark II Phased Array Feed on an ASKAP Antenna

Authors: A. P. Chippendale, A. J. Brown, R. J. Beresford, G. A. Hampson, A. Macleod, R. D. Shaw, M. L. Brothers, C. Cantrall, A. R. Forsyth, S. G. Hay, M. Leach

Abstract: This paper presents the measured sensitivity of CSIRO's first Mk. II phased array feed (PAF) on an ASKAP antenna. The Mk. II achieves a minimum system-temperature-over-efficiency $T_\mathrm{sys}/η$ of 78 K at 1.23 GHz and is 95 K or better from 835 MHz to 1.8 GHz. This PAF was designed for the Australian SKA Pathfinder telescope to demonstrate fast astronomical surveys with a wide field of view fo… ▽ More This paper presents the measured sensitivity of CSIRO's first Mk. II phased array feed (PAF) on an ASKAP antenna. The Mk. II achieves a minimum system-temperature-over-efficiency $T_\mathrm{sys}/η$ of 78 K at 1.23 GHz and is 95 K or better from 835 MHz to 1.8 GHz. This PAF was designed for the Australian SKA Pathfinder telescope to demonstrate fast astronomical surveys with a wide field of view for the Square Kilometre Array (SKA). △ Less

Submitted 1 September, 2015; originally announced September 2015.

Comments: 5 pages, 2 figures, accepted for publication in Electromagnetics in Advanced applications (ICEAA), 2015 International Conference on

A Quark Model Calculation of gamma gamma to pi pi Including Final State Interactions

Authors: H. G. Blundell, S. Godfrey, G. Hay, E. S. Swanson

Abstract: A quark model calculation of the processes gamma gamma -> pi+ pi- and gamma gamma -> pi0 pi0 is performed. At tree level, only charged pions couple to the initial state photons and neutral pions are not expected in the final state. However, a small but significant gamma gamma -> pi0 pi0 cross section is observed. We demonstrate that this may be accounted for by a rotation in isospin space induce… ▽ More A quark model calculation of the processes gamma gamma -> pi+ pi- and gamma gamma -> pi0 pi0 is performed. At tree level, only charged pions couple to the initial state photons and neutral pions are not expected in the final state. However, a small but significant gamma gamma -> pi0 pi0 cross section is observed. We demonstrate that this may be accounted for by a rotation in isospin space induced by final state interactions. The resulting pi+ pi- cross section is in good agreement with experiment while the pi0 pi0 cross section is in qualitative agreement with the data. △ Less

Submitted 8 December, 1999; v1 submitted 30 August, 1999; originally announced August 1999.

Comments: 22 Revtex pages, 5 postscript figures

Journal ref: Phys.Rev.C61:025203,2000