scispace - formally typeset
Search or ask a question
Institution

Australian National University

EducationCanberra, Australian Capital Territory, Australia
About: Australian National University is a education organization based out in Canberra, Australian Capital Territory, Australia. It is known for research contribution in the topics: Population & Galaxy. The organization has 34419 authors who have published 109261 publications receiving 4315448 citations. The organization is also known as: The Australian National University & ANU.


Papers
More filters
Journal ArticleDOI
TL;DR: Results in this paper show that if a large neural network is used for a pattern classification problem and the learning algorithm finds a network with small weights that has small squared error on the training patterns, then the generalization performance depends on the size of the weights rather than the number of weights.
Abstract: Sample complexity results from computational learning theory, when applied to neural network learning for pattern classification problems, suggest that for good generalization performance the number of training examples should grow at least linearly with the number of adjustable parameters in the network. Results in this paper show that if a large neural network is used for a pattern classification problem and the learning algorithm finds a network with small weights that has small squared error on the training patterns, then the generalization performance depends on the size of the weights rather than the number of weights. For example, consider a two-layer feedforward network of sigmoid units, in which the sum of the magnitudes of the weights associated with each unit is bounded by A and the input dimension is n. We show that the misclassification probability is no more than a certain error estimate (that is related to squared error on the training set) plus A/sup 3/ /spl radic/((log n)/m) (ignoring log A and log m factors), where m is the number of training patterns. This may explain the generalization performance of neural networks, particularly when the number of training examples is considerably smaller than the number of weights. It also supports heuristics (such as weight decay and early stopping) that attempt to keep the weights small during training. The proof techniques appear to be useful for the analysis of other pattern classifiers: when the input domain is a totally bounded metric space, we use the same approach to give upper bounds on misclassification probability for classifiers with decision boundaries that are far from the training examples.

1,234 citations

Journal ArticleDOI
TL;DR: In this article, a broad range of resonant electromagnetic effects by using two effective coupled oscillators, including the Fano resonance, electromagnetically induced transparency, Kerker and Borrmann effects, and parity-time symmetry breaking, are reviewed.
Abstract: The importance of the Fano resonance concept is recognized across multiple fields of physics. In this Review, Fano resonance is explored in the context of optics, with particular emphasis on dielectric nanostructures and metasurfaces. Rapid progress in photonics and nanotechnology brings many examples of resonant optical phenomena associated with the physics of Fano resonances, with applications in optical switching and sensing. For successful design of photonic devices, it is important to gain deep insight into different resonant phenomena and understand their connection. Here, we review a broad range of resonant electromagnetic effects by using two effective coupled oscillators, including the Fano resonance, electromagnetically induced transparency, Kerker and Borrmann effects, and parity–time symmetry breaking. We discuss how to introduce the Fano parameter for describing a transition between two seemingly different spectroscopic signatures associated with asymmetric Fano and symmetric Lorentzian shapes. We also review the recent results on Fano resonances in dielectric nanostructures and metasurfaces.

1,234 citations

Journal ArticleDOI
01 Apr 1992-Geology
TL;DR: In this article, the detrital zircons in an Archean metaquartzite are 3550 Ma or older; about one-fourth have ages between 3800 and 3850 Ma.
Abstract: Ion microprobe U-Pb analyses of zircons from the Sino-Korean craton have identified remnants of ≥3800 Ma crust at two localities—near Caozhuang east of Beijing, and near Anshan, northeast China. Near Caozhuang, the detrital zircons in an Archean metaquartzite are 3550 Ma or older; about one-fourth have ages between 3800 and 3850 Ma. The ages for detrital zircons with concordant U-Pb ages form a polymodal distribution, interpreted to show that the detritus that formed the quartzite was derived from a terrane containing early Archean rocks of several ages. Near Anshan, sheared gneiss is present in a complex containing ca. 3300 and 3000 Ma granites. Some of the zircons in this gneiss are concordant with a weighted mean 207 Pb/ 206 Pb age of 3804 ±5 Ma (2 σ), interpreted to be the age of the protolith of the gneiss.

1,234 citations

Journal ArticleDOI
TL;DR: In this paper, the Global Mapping Function (GMF) was proposed based on data from the global ECMWF numerical weather model and the coefficients of the GMF were obtained from an expansion of the Vienna mapping function (VMF1) parameters into spherical harmonics on a global grid.
Abstract: [1] Troposphere mapping functions are used in the analyses of Global Positioning System and Very Long Baseline Interferometry observations to map a priori zenith hydrostatic and wet delays to any elevation angle. Most analysts use the Niell Mapping Function (NMF) whose coefficients are determined from site coordinates and the day of year. Here we present the Global Mapping Function (GMF), based on data from the global ECMWF numerical weather model. The coefficients of the GMF were obtained from an expansion of the Vienna Mapping Function (VMF1) parameters into spherical harmonics on a global grid. Similar to NMF, the values of the coefficients require only the station coordinates and the day of year as input parameters. Compared to the 6-hourly values of the VMF1 a slight degradation in short-term precision occurs using the empirical GMF. However, the regional height biases and annual errors of NMF are significantly reduced with GMF.

1,232 citations

Journal ArticleDOI
20 Apr 2007-Science
TL;DR: Olivine phenocrysts' compositions record differences in the contributions of pyroxenite-derived melts in mid-ocean ridge basalts, which imply involvement of 2 to 20% (up to 28%) of recycled crust in mantle melting.
Abstract: Plate tectonic processes introduce basaltic crust (as eclogite) into the peridotitic mantle. The proportions of these two sources in mantle melts are poorly understood. Silica-rich melts formed from eclogite react with peridotite, converting it to olivine-free pyroxenite. Partial melts of this hybrid pyroxenite are higher in nickel and silicon but poorer in manganese, calcium, and magnesium than melts of peridotite. Olivine phenocrysts' compositions record these differences and were used to quantify the contributions of pyroxenite-derived melts in mid-ocean ridge basalts (10 to 30%), ocean island and continental basalts (many >60%), and komatiites (20 to 30%). These results imply involvement of 2 to 20% (up to 28%) of recycled crust in mantle melting.

1,226 citations


Authors

Showing all 34925 results

NameH-indexPapersCitations
Cyrus Cooper2041869206782
Nicholas G. Martin1921770161952
David R. Williams1782034138789
Krzysztof Matyjaszewski1691431128585
Anton M. Koekemoer1681127106796
Robert G. Webster15884390776
Ashok Kumar1515654164086
Andrew White1491494113874
Bernhard Schölkopf1481092149492
Paul Mitchell146137895659
Liming Dai14178182937
Thomas J. Smith1401775113919
Michael J. Keating140116976353
Joss Bland-Hawthorn136111477593
Harold A. Mooney135450100404
Network Information
Related Institutions (5)
University of Oxford
258.1K papers, 12.9M citations

92% related

University College London
210.6K papers, 9.8M citations

91% related

Pennsylvania State University
196.8K papers, 8.3M citations

91% related

University of Edinburgh
151.6K papers, 6.6M citations

91% related

University of Cambridge
282.2K papers, 14.4M citations

91% related

Performance
Metrics
No. of papers from the Institution in previous years
YearPapers
2023280
2022773
20215,261
20205,464
20195,109
20184,825