University of British Columbia

About: University of British Columbia is a education organization based out in Vancouver, British Columbia, Canada. It is known for research contribution in the topics: Population & Health care. The organization has 89939 authors who have published 209679 publications receiving 9226862 citations. The organization is also known as: UBC & The University of British Columbia.

Auto-WEKA: combined selection and hyperparameter optimization of classification algorithms

Abstract: Many different machine learning algorithms exist; taking into account each algorithm's hyperparameters, there is a staggeringly large number of possible alternatives overall. We consider the problem of simultaneously selecting a learning algorithm and setting its hyperparameters, going beyond previous work that attacks these issues separately. We show that this problem can be addressed by a fully automated approach, leveraging recent innovations in Bayesian optimization. Specifically, we consider a wide range of feature selection techniques (combining 3 search and 8 evaluator methods) and all classification approaches implemented in WEKA's standard distribution, spanning 2 ensemble methods, 10 meta-methods, 27 base classifiers, and hyperparameter settings for each classifier. On each of 21 popular datasets from the UCI repository, the KDD Cup 09, variants of the MNIST dataset and CIFAR-10, we show classification performance often much better than using standard selection and hyperparameter optimization methods. We hope that our approach will help non-expert users to more effectively identify machine learning algorithms and hyperparameter settings appropriate to their applications, and hence to achieve improved performance.

Olaparib in patients with recurrent high-grade serous or poorly differentiated ovarian carcinoma or triple-negative breast cancer: a phase 2, multicentre, open-label, non-randomised study

Abstract: Summary Background Olaparib (AZD2281) is a small-molecule, potent oral poly(ADP-ribose) polymerase (PARP) inhibitor. We aimed to assess the safety and tolerability of this drug in patients without BRCA1 or BRCA2 mutations with advanced triple-negative breast cancer or high-grade serous and/or undifferentiated ovarian cancer. Methods In this phase 2, multicentre, open-label, non-randomised study, women with advanced high-grade serous and/or undifferentiated ovarian carcinoma or triple-negative breast cancer were enrolled and received olaparib 400 mg twice a day. Patients were stratified according to whether they had a BRCA1 or BRCA2 mutation or not. The primary endpoint was objective response rate by Response Evaluation Criteria In Solid Tumors (RECIST). All patients who received treatment were included in the analysis of toxic effects, and patients who had measurable lesions at baseline were included in the primary efficacy analysis. This trial is registered at ClinicalTrials.gov, number NCT00679783. Findings 91 patients were enrolled (65 with ovarian cancer and 26 breast cancer) and 90 were treated between July 8, 2008, and Sept 24, 2009. In the ovarian cancer cohorts, 64 patients received treatment. 63 patients had target lesions and therefore were evaluable for objective response as per RECIST. In these patients, confirmed objective responses were seen in seven (41%; 95% CI 22–64) of 17 patients with BRCA1 or BRCA2 mutations and 11 (24%; 14–38) of 46 without mutations. No confirmed objective responses were reported in patients with breast cancer. The most common adverse events were fatigue (45 [70%] of patients with ovarian cancer, 13 [50%] of patients with breast cancer), nausea (42 [66%] and 16 [62%]), vomiting (25 [39%] and nine [35%]), and decreased appetite (23 [36%] and seven [27%]). Interpretation Our study suggests that olaparib is a promising treatment for women with ovarian cancer and further assessment of the drug in clinical trials is needed. Funding AstraZeneca.

Three-year Wilkinson Microwave Anisotropy Probe (WMAP) observations: temperature analysis

Abstract: We present new full-sky temperature maps in five frequency bands from 23 to 94 GHz, based on data from the first 3 years of the WMAP sky survey. The new maps are consistent with the first-year maps and are more sensitive. The 3 year maps incorporate several improvements in data processing made possible by the additional years of data and by a more complete analysis of the polarization signal. These include several new consistency tests as well as refinements in the gain calibration and beam response models. We employ two forms of multifrequency analysis to separate astrophysical foreground signals from the CMB, each of which improves on our first-year analyses. First, we form an improved "Internal Linear Combination" (ILC) map, based solely on WMAP data, by adding a bias-correction step and by quantifying residual uncertainties in the resulting map. Second, we fit and subtract new spatial templates that trace Galactic emission; in particular, we now use low-frequency WMAP data to trace synchrotron emission instead of the 408 MHz sky survey. The WMAP point source catalog is updated to include 115 new sources whose detection is made possible by the improved sky map sensitivity. We derive the angular power spectrum of the temperature anisotropy using a hybrid approach that combines a maximum likelihood estimate at low l (large angular scales) with a quadratic cross-power estimate for l > 30. The resulting multifrequency spectra are analyzed for residual point source contamination. At 94 GHz the unmasked sources contribute 128 ± 27 μK2 to l(l + 1)Cl/2π at l = 1000. After subtracting this contribution, our best estimate of the CMB power spectrum is derived by averaging cross-power spectra from 153 statistically independent channel pairs. The combined spectrum is cosmic variance limited to l = 400, and the signal-to-noise ratio per l-mode exceeds unity up to l = 850. For bins of width Δl/l = 3%, the signal-to-noise ratio exceeds unity up to l = 1000. The first two acoustic peaks are seen at l = 220.8 ± 0.7 and l = 530.9 ± 3.8, respectively, while the first two troughs are seen at l = 412.4 ± 1.9 and l = 675.2 ± 11.1. The rise to the third peak is unambiguous; when the WMAP data are combined with higher resolution CMB measurements, the existence of a third acoustic peak is well established. Spergel et al. use the 3 year temperature and polarization data to constrain cosmological model parameters. A simple six-parameter ΛCDM model continues to fit CMB data and other measures of large-scale structure remarkably well. The new polarization data produce a better measurement of the optical depth to reionization, τ = 0.089 ± 0.03. This new and tighter constraint on τ helps break a degeneracy with the scalar spectral index, which is now found to be ns = 0.960 ± 0.016. If additional cosmological data sets are included in the analysis, the spectral index is found to be ns = 0.947 ± 0.015.

How young and old adults listen to and remember speech in noise

Abstract: Two experiments using the materials of the Revised Speech Perception in Noise (SPIN-R) Test [Bilger et al., J. Speech Hear. Res. 27, 32-48 (1984)] were conducted to investigate age-related differences in the identification and the recall of sentence-final words heard in a babble background. In experiment 1, the level of the babble was varied to determine psychometric functions (percent correct word identification as a function of S/N ratio) for presbycusics, old adults with near-normal hearing, and young normal-hearing adults, when the sentence-final words were either predictable (high context) or unpredictable (low context). Differences between the psychometric functions for high- and low-context conditions were used to show that both groups of old listeners derived more benefit from supportive context than did young listeners. In experiment 2, a working memory task [Daneman and Carpenter, J. Verb. Learn. Verb. Behav. 19, 450-466 (1980)] was added to the SPIN task for young and old adults. Specifically, after listening to and identifying the sentence-final words for a block of n sentences, the subjects were asked to recall the last n words that they had identified. Old subjects recalled fewer of the items they had perceived than did young subjects in all S/N conditions, even though there was no difference in the recall ability of the two age groups when sentences were read. Furthermore, the number of items recalled by both age groups was reduced in adverse S/N conditions. The resutls were interpreted as supporting a processing model in which reallocable processing resources are used to support auditory processing when listening becomes difficult either because of noise, or because of age-related deterioration in the auditory system. Because of this reallocation, these resources are unavailable to more central cognitive processes such as the storage and retrieval functions of working memory, so that "upstream" processing of auditory information is adversely affected.

The Simons Observatory : Science goals and forecasts

Abstract: The Simons Observatory (SO) is a new cosmic microwave background experiment being built on Cerro Toco in Chile, due to begin observations in the early 2020s. We describe the scientific goals of the experiment, motivate the design, and forecast its performance. SO will measure the temperature and polarization anisotropy of the cosmic microwave background in six frequency bands centered at: 27, 39, 93, 145, 225 and 280 GHz. The initial configuration of SO will have three small-aperture 0.5-m telescopes and one large-aperture 6-m telescope, with a total of 60,000 cryogenic bolometers. Our key science goals are to characterize the primordial perturbations, measure the number of relativistic species and the mass of neutrinos, test for deviations from a cosmological constant, improve our understanding of galaxy evolution, and constrain the duration of reionization. The small aperture telescopes will target the largest angular scales observable from Chile, mapping ≈ 10% of the sky to a white noise level of 2 μK-arcmin in combined 93 and 145 GHz bands, to measure the primordial tensor-to-scalar ratio, r, at a target level of σ(r)=0.003. The large aperture telescope will map ≈ 40% of the sky at arcminute angular resolution to an expected white noise level of 6 μK-arcmin in combined 93 and 145 GHz bands, overlapping with the majority of the Large Synoptic Survey Telescope sky region and partially with the Dark Energy Spectroscopic Instrument. With up to an order of magnitude lower polarization noise than maps from the Planck satellite, the high-resolution sky maps will constrain cosmological parameters derived from the damping tail, gravitational lensing of the microwave background, the primordial bispectrum, and the thermal and kinematic Sunyaev-Zel'dovich effects, and will aid in delensing the large-angle polarization signal to measure the tensor-to-scalar ratio. The survey will also provide a legacy catalog of 16,000 galaxy clusters and more than 20,000 extragalactic sources.