Filename
stringlengths 22
64
| Paragraph
stringlengths 8
5.57k
|
|---|---|
Processed_Near-Infrared_and_Star-forming_properties_of_Local.txt | Alonso-Herrero, A., & Knapen, J. H. 2001, AJ, 122, 1350 Alonso-Herrero, A., Rieke, G. H., Rieke, M. J., & Scoville, N. Z. |
Processed_Near-Infrared_and_Star-forming_properties_of_Local.txt | Melbourne, J. & Salzer, J. J. 2002, AJ, 123, 2302 Melbourne, J., Koo, D. C., & Le Floc’h, E. 2005, ApJ, 632, 65 Miles, J. W., Houck, J. R., Hayward, T. L., & Ashby, M. L. N. |
Processed_Near-Infrared_and_Star-forming_properties_of_Local.txt | Rigby, J. R., & Rieke, G. H. 2004, ApJ, 606, 237 Roy, A. L., Goss, W. M., Mohan, N. R., & Anantharamaiah, K. R. |
Processed_A_mixture_model_approach_to_infer_land-use_influen.txt | Abstract The assessment of water quality across space and time is of considerable interest for both agricultural and public health reasons. The standard method to assess the water quality of a catchment, or a group of catchments, usually involves collecting point measurements of water quality and other additional information such as the date and time of measurements, rainfall amounts, the land-use and soil-type of the catchment and the elevation. Some of this auxiliary information will be point data, measured at the exact location, whereas other such as land-use will be areal data often in a compositional format. Two problems arise if analysts try to incorporate this information into a statistical model in order to predict (for example) the influence of land-use on water quality. First is the spatial change of support problem that arises when using areal data to predict outcomes at point lo- cations. Secondly, the physical process driving water quality is not compositional, rather it is the observation process that provides compositional data. In this paper we present an approach that accounts for these two issues by using a latent vari- able to identify the land-use that most likely influences water quality. This latent variable is used in a spatial mixture model to help estimate the influence of land- use on water quality. We demonstrate the potential of this approach with data from a water quality research study in the Mount Lofty range, in South Australia. |
Processed_A_mixture_model_approach_to_infer_land-use_influen.txt | Fig. 1 Measurements in the same catchment are not subject to the same land-use influence. sB and sC are specific examples. |
Processed_A_mixture_model_approach_to_infer_land-use_influen.txt | In this article, we demonstrate how a modified version of the latter approach provides an alternative to the methods described previously. The main advantage of our approach is that it deals with the COSP while also capturing the spatial and seasonal effects of land-use type on water quality. The disadvantages of this approach are that it limits the land-use influence to one type per site, for any given water quality parameter, and does not include an effect for the area of the land use type. |
Processed_A_mixture_model_approach_to_infer_land-use_influen.txt | The article is organized as follows. In section 3, we present the model used to infer the spatio-temporal effects of land use types on water quality. In section 4 we present two estimation approaches. The first uses the Expectation-Maximisation (EM, see Dempster et al., 1977) algorithm. The second adopts a Bayesian hierar- chical approach. In section 5 we compare the results obtained with our approach to the results obtained with an existing model drawn from the literature. We apply our method to a dataset from a water monitoring program from sub-catchment of the Mount Lofty Range in South Australia (see Ford et al., 2015, for details of the program). |
Processed_A_mixture_model_approach_to_infer_land-use_influen.txt | Our analysis aims to identify the effect of land-use type on water quality, and thereby provide a mechanism to predict water quality at sites in catchments that have not been monitored despite the aforementioned limitations. This analysis was designed to form part of a wider assessment of the risks of exceeding water quality parameters across the MLR, particularly under low flow conditions Ford et al. (2015). Here we use the concentration of total nitrogen measured at 16 sites from 2008 to 2013 to demonstrate application of the statistical model. |
Processed_A_mixture_model_approach_to_infer_land-use_influen.txt | distributions in the mixture is then equal to the number of land-use types in the entire area. |
Processed_A_mixture_model_approach_to_infer_land-use_influen.txt | As stated in the previous sections, this problem is fully spatio-temporal and pos- sible correlations have to be integrated in the model. We list below the solutions chosen to account for these correlations. |
Processed_A_mixture_model_approach_to_infer_land-use_influen.txt | The model presented by Szpiro et al. (2010) supports the idea that the temporal variation remains globally identical over the spatial domain, making it possible to fit a smooth temporal function prior to fitting the full model. The spatio-temporal residuals ν are then assumed to be independent in both space and time. This as- sumption, however, is very strong. We therefore develop an alternative modelling approach that allows for seasonality over the spatial domain, that can be affected by the land-use type at each measurement location, both in terms of amplitude, phase and shape. Note that the original model assumes that this influence affects amplitude alone. Our new model introduces temporal basis functions for each land-use type. These basis functions are calculated using a seasonal-trend (STL) decomposition approach introduced by Cleveland and Cleveland (1990). In this approach the time series of observations are decomposed into trend and season- ality components using a Loess smoother. In our model, the time series for each land-use is extracted from the time series of each site using the latent variable. Then, for a given land-use type, the STL decomposition is applied to each site time-series associated to that land-use through the latent variable. |
Processed_A_mixture_model_approach_to_infer_land-use_influen.txt | It is possible to make multiple assumptions regarding the spatial structure. The most common approach is to assume that neighbouring sites are correlated with a correlation value that is a function of the Euclidean distance between sites. Re- cent works (Peterson et al., 2007) suggests that this model is not optimal when we dealing with data on a stream network , as the euclidean distance between sites is less relevant than the distance measured along the network, hence Peterson et al. (2007) suggest the use of stream distance rather than euclidean distance. These two approaches limit the correlation structure to the measurements. We believe that some of the spatial dependency observed in the data come from the spatial structure of the land-uses. The solution presented in this paper is to model the spatial dependency through the latent variable of the mixture, making sure that neighbouring sites (in the sense of common border) are more likely to have iden- tical land-uses. In the EM technique, this can be achieved by adding a penalization term. In the Gibbs sampling technique we model the latent variable of the mixture using a Potts model (also called 2-d Ising model). |
Processed_A_mixture_model_approach_to_infer_land-use_influen.txt | Fig. 2 Schematic figure of the model structure. The spatial correlation can be included at different levels of the model, leading to different likelihood and estimation tools. Model A include the spatial correlation at the baseline level ( y0), implying natural spatial correlation without the effect of land- use. This is the model most often used in the literature. Model B, which we use here, introduces the spatial correlation at the covariate level (land-use). Merging the two models can be considered, but one has to carefully monitor for over-fitting by doing so. |
Processed_A_mixture_model_approach_to_infer_land-use_influen.txt | where the temporal basis function f is now land-use (location) dependent, and β can be understood as a random effect parameter for each location-temporal basis pair. z(s) is the latent indicator for which land-use is associated to site s. |
Processed_A_mixture_model_approach_to_infer_land-use_influen.txt | The model described in Eq. 1 is very simple to express in the likelihood format, as all the components are Gaussian. This leads to a linear log-likelihood, for which maximization is performed quite straightforwardly (see Lindstrom et al., 2013, for details about the maximization procedure, including some simplification tricks). The introduction of the mixture model on top of the spatial dependency (see Eq. 5) requires us to use an EM algorithm or a dedicated Bayesian hierarchical model. The following sections provide a description of the two possible estimation solu- tions. |
Processed_A_mixture_model_approach_to_infer_land-use_influen.txt | There are two elements that are not estimated by the model, but need to be in- corporated in order to perform the estimation: the temporal basis; and the neigh- bouring structure. |
Processed_A_mixture_model_approach_to_infer_land-use_influen.txt | for function j, and land-use k. That equation also implies that in the estimation procedure, each time the latent variables are updated, the temporal basis functions are also updated. |
Processed_A_mixture_model_approach_to_infer_land-use_influen.txt | Neighbouring structure In order to compute the spatial dependency of the data, we need to identify the spatial structure. In this paper, we assume that the latent variables are carrying all the spatial dependency, by virtue of their neighbourhood relationship. This structure is achieved through the construction of a Voronoi lat- tice, where the cell centers are the monitoring sites. |
Processed_A_mixture_model_approach_to_infer_land-use_influen.txt | Because δ is unobserved, it is difficult to evaluate (cid:96) (θ ). Instead, the EM maximize the expected (over δ) complete-data likelihood given y and a previously calculated θ . |
Processed_A_mixture_model_approach_to_infer_land-use_influen.txt | The overall principle remains the same, with two steps, one for expectation, one for maximisation. Ambroise and Govaert (1998) proposed an optimisation method based on the fixed point approach to achieve the E-step, which is slightly changed because of the penalization term. |
Processed_A_mixture_model_approach_to_infer_land-use_influen.txt | The main difficulty lies in updating δ. However, quick mixing can be achieved using the scheme of Murray et al. (2006) and the Swendsen-Wang algorithm to simulate from the Potts model. Additional details can be found in Murray et al. (2006); Barbu and Zhu (2007); Everitt (2012); Cucala and Marin (2013). |
Processed_A_mixture_model_approach_to_infer_land-use_influen.txt | The results of three models are presented in Table 1: one mixture model, one mix- ture model with spatial dependency and the CLR model. In the table, we present the maximum a posteriori (MAP) estimates for the land-use types, for the differ- ent temporal basis function: in our model, three basis function are used, constant, trend and seasonal; in the CLR model two temporal basis functions are used only, to match the degrees of freedom of the models. On the bottom side of the ta- ble we display the sum of square errors (SSE, using the MAP) for the prediction power of each model. The three presented models adjust differently to the data. The mixtures approaches have a better fit than the CLR model. Results show that the mixture and the spatial mixture provide almost identical results. This suggests that the estimation of the latent variables demonstrates a natural spatial correla- tion, which didn’t need to be enforced through the spatial modelling. This spatial consistency is also demonstrated in Figure 4. |
Processed_A_mixture_model_approach_to_infer_land-use_influen.txt | Table 1 MAP estimates for the three fitted models. It has been ordered from the lowest baseline value, to the highest, following the mixture models estimates. We observe that the order and the amplitude of the estimates is not the same between the mixture models and the compositional-log-ratio (CLR) transform model. |
Processed_A_mixture_model_approach_to_infer_land-use_influen.txt | Fig. 3 Example of plot of the (normalized) temporal basis functions for the land-uses "nature con- servation" and "irrigated perennial horticulture". The orange line represents the seasonality, the green line represents the trend. |
Processed_A_mixture_model_approach_to_infer_land-use_influen.txt | will happen) to water quality in this area. This also can allow for the identification of change points (observable through the trend plot) and for the prediction of water quality in changing conditions. |
Processed_A_mixture_model_approach_to_infer_land-use_influen.txt | In Eq. 5, the random variable ν is representing the uncertainty of the observations. This is usually linked to unknown predictors and measurement errors. However, when the model is over-fitting the data, the uncertainty can be under estimated. This situation is encountered by the CLR model when an additional temporal basis function is added. The resulting SSE and σ2 ν are both estimated to be equal to 0. With the proposed model the over-fitting risk can be monitored by looking at the smoothness of the seasonality and trend curves. Figure 3 shows regular curves, a result consistently overruling the over-fitting risk. An additional feature of our model is interesting for measuring uncertainty. The mixture approach allows us to decided whether a single measurement error is modelled, or if each component of the mixture has a different measurement error. This additional flexibility allows us to identify sites / land-uses where additional sampling effort are required in order to improve the monitoring. |
Processed_A_mixture_model_approach_to_infer_land-use_influen.txt | of the nature of the model, the latent state is estimated for each Voronoi cell in each sub-catchment, allowing us to display a map of most likely land-use types. We display in Figure 4 a map of the sub-catchments, coloured by their estimated latent state and ordered by their baseline values. We notice the strong spatial consistency of the map, with the higher concentration of nitrate located in the same area. |
Processed_A_mixture_model_approach_to_infer_land-use_influen.txt | Fig. 4 Map of some of the sub-catchment in the Mount Lofty region. In the legend, items have been ordered according to Nitrate concentration: lower Nitrate concentration values are shaded green (eg MRP); higher Nitrate concentration are shaded red (eg IPH). Higher nitrate measurements are ob- served near the Cox Creek sub-catchment, with land-use being related to human activities (residential, services, transport and communication, irrigated perennial horticulture). |
Processed_A_mixture_model_approach_to_infer_land-use_influen.txt | if the latter is easily implementable. Spatial misalignment is an important issue that should be considered carefully. Additionally, the presented model provides a helpful insight on the land-use influence when the information about it is limited, particularly in precision. The outputs of the model provide an idea about the land- use and sites association, and associates the variation of the measurements to the most likely land-use influence. Being able to identify these is of prime importance in particular for understanding the real impact of land-use on water quality. Some improvements are possible. First, we limit the number of influence per site to one, as our aim is to identify the main land-use influence per site. This may prove a bit restrictive as some sites may be exposed to more than one significant land- use influence. A simple example for that is when the water source is a river, and the two opposite banks have two different land-use types. Our approach can be generalized to more than one influence, however this would increase the com- putational load significantly, even more so if we consider that different sites may have a different number of influences (and hence adding to the burden the need to estimate the number of influences for each site). The latent variable approach forces the Markov chain to run longer, primarily due to the Metropolis step within the Gibbs sampler. It is important to monitor the behaviour of the samples of the posterior distribution in order to ensure convergence of the chain. Another potential improvement would be the use of an automatic selection tool for the latent space dimension. In this article, we choose which and how many land-use type constitute the latent space. This approach, based on expert knowledge, can be completed by a variable selection approach (of the type LASSO - Raftery and Dean (2004); Städler et al. (2010), or slope heuristic - Baudry et al. (2012)). |
Processed_A_mixture_model_approach_to_infer_land-use_influen.txt | This study was supported by funding from the Goyder Institute for Water Research for the project "Mt Lofty Ranges Water Allocation and Planning" led by Associate Professor Jim Cox. The authors would like to thank SA Water, SA EPA, DEWNR, AMLR NRMB and Dr Leon van der Linden for providing water quality data from the Mt Lofty Ranges for the risk assessment. Thank you also to staff from numer- ous South Australian Government Agencies who provided valuable feedback at presentations given about project outputs. |
Processed_A_mixture_model_approach_to_infer_land-use_influen.txt | Ambroise, C., Dang, V.M., Govaert, G., 1997. Clustering of spatial data by the EM algorithm, in: geoENV I - Geostatistics for Environmental Applications, Quanti- tative Geology and Geostatistics (Vol. 9), pp. 493–504. |
Processed_A_mixture_model_approach_to_infer_land-use_influen.txt | Arnold, J.G., Williams, J.R., Maidment, D.R., 1995. Continuous-Time Water and Sediment-Routing Model for Large Basins. Journal of Hydraulic Engineering 121, 171–183. |
Processed_A_mixture_model_approach_to_infer_land-use_influen.txt | Bilmes, J.A., 1998. A gentle tutorial of the EM algorithm and its application to parameter estimation for Gaussian mixture and hidden Markov models. Inter- national Computer Science Institute 4, 1–15. |
Processed_A_mixture_model_approach_to_infer_land-use_influen.txt | Bonyah, E., Munyakazi, L., Asong, D., Bashiru, I.I., 2013. Application of area to point Kriging to breast cancer incidence in Ashanti Region of Ghana. Interna- tional Journal of Medicine and Medical Sciences 5, 67–74. |
Processed_A_mixture_model_approach_to_infer_land-use_influen.txt | Buck, O., Niyogi, D.K., Townsend, C.R., 2004. Scale-dependence of land use effects on water quality of streams in agricultural catchments. Environmental pollution (Barking, Essex : 1987) 130, 287–99. |
Processed_A_mixture_model_approach_to_infer_land-use_influen.txt | Cooley, D., Sain, S.R., 2010. Spatial hierarchical modeling of precipitation ex- tremes from a regional climate model. Journal of Agricultural, Biological, and Environmental Statistics 15, 381–402. |
Processed_A_mixture_model_approach_to_infer_land-use_influen.txt | Cox, J.W., Oliver, D.P., Fleming, N.K., Anderson, J.S., 2012. Off-site transport of nutrients and sediment from three main land-uses in the Mt Lofty Ranges, South Australia. Agricultural Water Management 106, 50–59. |
Processed_A_mixture_model_approach_to_infer_land-use_influen.txt | Everitt, R.G., 2012. Bayesian Parameter Estimation for Latent Markov Random Fields and Social Networks. Journal of Computational and Graphical Statistics 8600, 26. |
Processed_A_mixture_model_approach_to_infer_land-use_influen.txt | Ford, J., Ickowicz, A., Oliver, D., Hayes, K., Kookana, R., 2015. Integrated catch- ment water planning support for Adelaide Mount Lofty Ranges Water Allocation Planning ( GWAP Project ) Task 5 : Tiered Water Quality Risk Assessment. Tech- nical Report 15/4. Goyder Institute for Water Research. Adelaide. |
Processed_A_mixture_model_approach_to_infer_land-use_influen.txt | Hunsaker, C.T., Levine, D.A., 1995. Hierarchical Approaches of Water Quality in Rivers Study processes are important in developing. Sciences-New York 45, 193–203. |
Processed_A_mixture_model_approach_to_infer_land-use_influen.txt | King, R.S., Baker, M.E., Whigham, D.F., Weller, D.E., Jordan, T.E., Kazyak, P.F., Hurd, M.K., 2005. Spatial Considerations for Linking Watershed Land Cover To Ecological Indicators in Streams. Ecological Applications 15, 137–153. |
Processed_A_mixture_model_approach_to_infer_land-use_influen.txt | Lehmann, E.A., Phatak, A., Soltyk, S., Chia, J., Lau, R., Palmer, M., 2013. Bayesian hierarchical modelling of rainfall extremes, in: 20th International Congress on Modelling and Simulation, Adelaide, Australia, 1-6 December 2013, pp. 1–6. Lindstrom, J., Szpiro, A., Sampson, P., Sheppard, L., Oron, a., Richards, M., Larson, T., 2011. A flexible spatio-temporal model for air pollution: Allowing for spatio- temporal covariates. UW Biostatistics Working Paper Series 370, 1–38. |
Processed_A_mixture_model_approach_to_infer_land-use_influen.txt | Lindstrom, J., Szpiro, A., Sampson, P.D., Bergen, S., Sheppard, L., 2013. Spa- tioTemporal : An R Package for Spatio-Temporal Modelling of Air-Pollution. CRAN Vignettes . |
Processed_A_mixture_model_approach_to_infer_land-use_influen.txt | Murray, I., Ghahramani, Z., MacKay, D.J.C., 2006. MCMC for doubly-intractable distributions. Proceedings of the 22nd Annual Conference on Uncertainty in Artificial Intelligence (UAI-06) , 359–366. |
Processed_A_mixture_model_approach_to_infer_land-use_influen.txt | Naim, I., Gildea, D., 2012. Convergence of the EM Algorithm for Gaussian Mix- tures with Unbalanced Mixing Coefficients. Proceedings of the 29th Interna- tional Conference on Machine Learning (ICML-12) , 1655–1662. |
Processed_A_mixture_model_approach_to_infer_land-use_influen.txt | Peterson, E.E., Sheldon, F., Darnell, R., Bunn, S.E., Harch, B.D., 2011. A compari- son of spatially explicit landscape representation methods and their relationship to stream condition. Freshwater Biology 56, 590–610. |
Processed_A_mixture_model_approach_to_infer_land-use_influen.txt | Peterson, E.E., Theobald, D.M., Ver Hoef, J.M., 2007. Geostatistical modelling on stream networks: Developing valid covariance matrices based on hydrologic distance and stream flow. Freshwater Biology 52, 267–279. |
Processed_A_mixture_model_approach_to_infer_land-use_influen.txt | Sampson, P.D., Szpiro, A.a., Sheppard, L., Lindström, J., Kaufman, J.D., 2011. Pragmatic estimation of a spatio-temporal air quality model with irregular mon- itoring data. Atmospheric Environment 45, 6593–6606. |
Processed_A_mixture_model_approach_to_infer_land-use_influen.txt | Shen, Z., Hou, X., Li, W., Aini, G., 2014. Relating landscape characteristics to non- point source pollution in a typical urbanized watershed in the municipality of Beijing. Landscape and Urban Planning 123, 96–107. |
Processed_A_mixture_model_approach_to_infer_land-use_influen.txt | Strayer, D.L., Beighley, R.E., Thompson, L.C., Brooks, S., Nilsson, C., Pinay, G., Naiman, R.J., 2003. Effects of Land Cover on Stream Ecosystems: Roles of Empirical Models and Scaling Issues. Ecosystems 6, 407–423. |
Processed_A_mixture_model_approach_to_infer_land-use_influen.txt | Szpiro, A.A., Sampson, P.D., Sheppard, L., Lumley, T., Adar, S.D., Kaufman, J.D., 2010. Predicting intra-urban variation in air pollution concentrations with com- plex spatio-temporal dependencies. Environmetrics 21, 606–631. |
Processed_A_mixture_model_approach_to_infer_land-use_influen.txt | Varcoe, J., van Leeuwen, J.A., Chittleborough, D.J., Cox, J.W., Smernik, R.J., Heitz, A., 2010. Changes in water quality following gypsum application to catchment soils of the Mount Lofty Ranges, South Australia. Organic Geochemistry 41, 116–123. |
Processed_A_mixture_model_approach_to_infer_land-use_influen.txt | Woolrich, M., Behrens, T., Beckmann, C., Smith, S., 2005. Mixture models with adaptive spatial regularization for segmentation with an application to fMRI data. Medical Imaging, IEEE Transactions on 24, 1–11. |
Processed_A_mixture_model_approach_to_infer_land-use_influen.txt | Zhu, L., Carlin, B., Gelfand, A., 2003. Hierarchical regression with misaligned spatial data: relating ambient ozone and pediatric asthma ER visits in Atlanta. Environmetrics , 1–33. |
Processed_Demonstration_of_a_simple_entangling_optical_gate_.txt | We demonstrate a new architecture for an optical entangling gate that is significantly simpler than previous realisations, using partially-polarising beamsplitters so that only a single optical mode-matching condition is required. We demonstrate operation of a controlled-z gate in both continuous-wave and pulsed regimes of operation, fully characterising it in each case using quantum process tomography. We also demonstrate a fully-resolving, nondeterministic optical Bell-state analyser based on this controlled-z gate. This new architecture is ideally suited to guided optics implementations of optical gates. |
Processed_Demonstration_of_a_simple_entangling_optical_gate_.txt | A key resource for using entanglement in quantum in- formation protocols are gates that are capable of entan- gling or disentangling qubits [1]. Entangling gates lie at the heart of quantum computation protocols, for ex- ample, and disentangling gates used in Bell state analy- sers are required for quantum teleportation. Conceptu- ally, the simplest such two-qubit gate is the controlled-z (cz) gate, which in the logical basis produces a π phase shift on the ; ; i i ). This is a maximally entan- ; 10 | i i gling gate which, when coupled with single qubit rota- tions, is universal for quantum computing [2]. |
Processed_Demonstration_of_a_simple_entangling_optical_gate_.txt | In 2001, Knill, Laflamme and Milburn proposed a scheme for linear optical quantum computing which used measurement to non-deterministically realize the opti- cal non-linearity required for two qubit entangling gates [3]. They also showed that deterministic versions of these gates could be achieved using teleportation [4], which requires Bell-state measurement. Since then, there have been a number of demonstrations of quantum logic gates derived from this concept [5, 6, 7, 8, 9], and fur- ther theoretical development of linear optics schemes In particular, there is a recent [10, 11, 12, 13, 14]. suggestion to use non-deterministic cz gates to con- struct cluster-states for demonstrating optical cluster- state quantum computation [15]. |
Processed_Demonstration_of_a_simple_entangling_optical_gate_.txt | a significant factor in the gate’s performance. We fully characterize the operation in both regimes using quan- tum process tomography, and we also demonstrate the use of this kind of gate for fully-resolving Bell measure- ments. This simple entangling optical gate is promising for micro-optics or guided optics implementations where extremely good non-classical interference is realisable. |
Processed_Demonstration_of_a_simple_entangling_optical_gate_.txt | FIG. 1: (a) Interferometric cz gate based on the approach of Refs. [13, 14]. Gate operation is enabled by transforming each qubit from polarisation to spatial encoding, and back again. This requires high interferometric stability and spatio- temporal mode-matching for correct operation. (b) Partially- polarising beam splitter (PPBS) gate. The qubits can remain polarisation-encoded, since the vertically-polarised modes are completely reflected by the first PPBS, and do not interact. Nonclassical interference occurs between the horizontally- polarised modes, with η=1/3. The subsequent PPBSs give the required losses in the cV and tV modes as shown in (a). |
Processed_Demonstration_of_a_simple_entangling_optical_gate_.txt | FIG. 2: Quantum process tomography of the cz gate. Real components of the χ matrix for the: (a) ideal; (b) CW; and (c) pulsed cz gate. The imaginary components of the experimental matrices are not shown: a few elements are on the order of 0.05, the average is ∼0.005. |
Processed_Demonstration_of_a_simple_entangling_optical_gate_.txt | arbitrary input state, the entire circuit performs the + δ V H transformation: α | i → | | | 1 3 [ ]+..., where H and HV i i V refer to horizontal and vertical polarisation, and the terms not shown correspond to the failure mode of the gate (i.e., the control and target output ports do not each contain one photon). With probability 1/9 the circuit performs the cz operation (using the logic-basis defini- tions, 0 H). After the network of 1/3 beam- splitters, the two spatial modes of the control and target must be recombined to return to polarisation encoded qubits. Since the phase relationship between the logical modes must be maintained throughout this operation, interferometric stability is required between the control and target modes. Inherently stable interferometers have previously been used [6, 7] to achieve this—however these may not be suitable for scaling to large numbers in micro- or integrated-optical realisations. Here we take an alter- native approach which does not require interferometric stability. |
Processed_Demonstration_of_a_simple_entangling_optical_gate_.txt | To test multi-qubit circuits, multi-photon sources are required. The current gold standard for generating two or more photons is pulsed parametric downconversion: pump power densities far greater than those possible with CW sources lead to significantly higher probabilities of multi-photon events. However, the short pump pulses lead to a larger bandwidth of downconverted photons, and to more stringent temporal mode-matching require- ments. Thus any new gate architecture should be shown to be compatible with both CW and pulsed sources. We tested the PPBS architecture with both CW and femtosecond-pulsed sources, which produce pairs of en- ergy degenerate single photons via spontaneous paramet- ric downconversion in a β-Barium-Borate (BBO) crystal (Table I). The photon pairs were collected into single mode optical fibers to improve the spatial mode, and in- jected into the cz gate [Fig. 1(b)]. In the pulsed case, mode-matching was also improved by collecting the gate output into single mode fibers. A pair of half- and quarter-wave plates at the output of each fiber was used for input state preparation. A coincidence window of 5 ns was used and no correction for accidental counts ∼ was made. The gates were completely characterised via quantum process tomography [7, 18]. |
Processed_Demonstration_of_a_simple_entangling_optical_gate_.txt | A convenient representation of the measured process is the χ matrix, which is a complete and unique descrip- tion of the process relative to a given basis of operators. The χ matrix for ideal cz gate operation in the Pauli basis is shown in Fig. 2(a) (all the components are real). |
Processed_Demonstration_of_a_simple_entangling_optical_gate_.txt | The experimental results for the CW gate are shown in Fig. 2(b), those for the pulsed gate in Fig. 2(c). By us- ing the method of Ref. [7], we are guaranteed physical χ matrices requiring no extra normalisation. In the pulsed case, the form is similar to the ideal. In the cw case, the major deviation from ideal operation is the larger than expected ii term (0.36 instead of 0.25). This arises from imperfect mode matching at the first PPBS, resulting in imperfect nonclassical interference—effectively, the con- trol and target qubits do not interfere and are simply transmitted through the circuit unchanged. |
Processed_Demonstration_of_a_simple_entangling_optical_gate_.txt | Gate performances can be quantified by calculating the process fidelity FP = T r[χmeas χideal] or the aver- age gate fidelity, which is the fidelity between expected and actual output states, averaged over all pure inputs, F = (4FP + 1)/5 [7, 19]. The CW and pulsed gates have process fidelities of 74.6 0.1% re- 0.3% and 84.0 ± spectively; and average gate fidelities of 79.7 0.2% and ± 0.1% [20]. Despite more stringent temporal mode- 87.2 matching requirements in the pulsed regime, the extra spatial filtering led to better gate operation, equivalent to the previous best demonstration [7]. |
Processed_Demonstration_of_a_simple_entangling_optical_gate_.txt | A potential drawback of the PPBS architecture is that the beamsplitting ratios are fixed at manufacture—in contrast to schemes where the setting of a HWP deter- mines the effective beamsplitter reflectivity [6, 7]. While the PPBSs for the CW gate (optimized for 702.2 nm) were measured to be within 0.01 of the required reflec- ± tivities, for the pulsed gate (820 nm), the values for the 0.01; nor- three PPBSs were η = 0.28, 0.28 and 0.29 ( malised to output power). Modelling a gate using 0.28 re- flectivities, we find the optimum process fidelity that can be obtained is F 0.28 P = 96%—near ideal. As originally an- ticipated in Ref. [13], the cz gate is relatively forgiving of the exact splitting ratios—making it an eminently suit- able gate to be realised with a PPBS architecture. Perfor- mance of the PPBS gates are limited almost exclusively by mode matching, primarily spatial, making these gates promising candidates for micro-or integrated-optical im- plementations, where nonclassical mode matching in ex- cess of 99% can be expected [21]. |
Processed_Demonstration_of_a_simple_entangling_optical_gate_.txt | FIG. 3: The cz gate operating as a Bell-state analyser. (a) The two qubit entangled state at the output of the fibres and (b) transformed to the φ+ Bell-state. (c) The measured truth table: the average probability of success is 0.78 ± 0.03. (d-g) Transformation of near-maximally entangled states to near-separable states by a cz gate Bell-state analyser. (d) The input Bell-states determined from (e) the measured input states with the second qubit rotated by a Hadamard. (f) The measured output states, (g) transformed by applying local rotations to each qubit (see text). |
Processed_Demonstration_of_a_simple_entangling_optical_gate_.txt | of the form . We use quantum state V V i tomography [22, 23] to characterize the source state 0.01 and linear en- [Fig. 3(a)]. The tangle T = 0.93 tropy SL = 0.05 0.01 show that this state is highly entangled and highly pure, the fidelity with a maximally entangled state is F = 98.0 0.4%. We determine that ± ϕ=-2.094 radians, and by using the input quarter- and half-wave plates [Fig. 1(b)] to perform appropriate single qubit unitaries on each qubit we can transform the state of Fig. 3(a) to any desired maximally entangled state of linear polarisation. Figure 3(b) shows the example where with fidelity we have produced the state | Fφ+ = 96.1 0.01. 0.2%; T = 0.96 ± To quantify the performance of the cz gate as a Bell- state analyser, we produced the four maximally entan- V A gled states: , i ± | i )/√2. where D V i These are just the usual four Bell states, with the sec- ond qubit rotated by a Hadamard so that they can be discriminated by the cz gate. The four experimentally produced density matrices are shown in Fig. 3(e): the average of their fidelities is ¯F = 95.8 0.7%; the average of the tangles and linear entropies are ¯T = 0.94 0.02 and ¯SL = 0.04 0.01, respectively. For ease of visuali- sation, we have numerically rotated these states into the more familiar form by applying a Hadamard gate to the second qubit [Fig. 3(d)]. |
Processed_Demonstration_of_a_simple_entangling_optical_gate_.txt | An ideal cz gate would take the four maximally en- tangled states to the four separable orthog- respectively. , and DA onal states: i | For the four input states in Fig. 3(e), the measured out- put density matrices are shown in Fig. 3(f). In fact, they are close to the four orthogonal separable states eiϕ2 3.07 ( | and ϕ2 = 0.32 as determined by a best fit. For ease of visualisation, we have rotated these states into the logical basis in Fig. 3(g). The average of the fidelities between all 5% combinations of the measured output states is 24 (ideally zero), demonstrating that the states are close to orthogonal. Their average tangle, ¯T = 0.04 0.05, and linear entropy, ¯SL = 0.42 0.07, indicates that they are unentangled, albeit somewhat mixed. This circuit is working quite well as a Bell-state analyser. |
Processed_Demonstration_of_a_simple_entangling_optical_gate_.txt | The average fidelity of the measured output states with the above separable states is F = 79 if we ana- lyzed the output of the circuit in this rotated basis, we would correctly identify the Bell-state with a probability of 79%. More directly, we can measure each of the separa- ble states for each Bell-state input by explicitly analysing in the rotated basis, which gives the directly measured truth table for the CZ gate when operated as a Bell-state analyser. The results are shown in Fig. 3(c), and the av- erage probability of success is 78 3%, in agreement with the tomography results. |
Processed_Demonstration_of_a_simple_entangling_optical_gate_.txt | It is interesting to note that whenever a postselected event occurs, the Bell measurement has effectively dis- criminated one of four input wave plate settings applied to a single input qubit. That is to say, two bits of classical information (representing the four wave plate settings) have been encoded into a single qubit. This is reminis- cent of quantum dense coding [24, 25, 26], although be- cause the Bell measurement is non-deterministic, a pro- tocol using this gate would be less efficient than ordinary classical communication. Nevertheless, this still demon- strates the power of entanglement for dense coding given a deterministic Bell analyser, such as can be constructed in principle using measurement-induced nonlinearity. |
Processed_Demonstration_of_a_simple_entangling_optical_gate_.txt | In summary, we have proposed and demonstrated a new architecture for entangling optical gates. The key advantage of this new gate architecture is its simplicity and suitability for scaling—it requires only one nonclas- sical mode matching condition, and no classical inter- ferometers. This is very promising for micro-optic and integrated-optic realisations of this gate, where extremely good mode matching can be expected. Finally, we have demonstrated the operation of this gate as a Bell-state analyser which has the advantage of higher success prob- ability and no ancillae compared to alternative recent demonstrations [9, 27]. |
Processed_Demonstration_of_a_simple_entangling_optical_gate_.txt | This work was supported by the Australian Research Council (ARC), the Queensland Government, and the US Advanced Research and Development Agency (ARDA). |
Processed_Demonstration_of_a_simple_entangling_optical_gate_.txt | RP acknowledges support from the Austrian Science Foundation (FWF). We wish to acknowledge Rohan Dal- ton for valuable discussions. [1] M. A. Nielsen and I. L. Chuang, Quantum Computation and Quantum Information (Cambridge University Press, Cambridge, 2000). |
Processed_Demonstration_of_a_simple_entangling_optical_gate_.txt | burn, Phys. Rev. A 65, 012314 (2002). [11] E. Knill, Phys. Rev. A 66, 052306 (2002). [12] T. B. Pittman, B. C. Jacobs, and J. D. Franson, Phys. |
Processed_Second_moment_of_the_pion's_distribution_amplitude.txt | We present preliminary results from the QCDSF/UKQCD collaborations for the second moment of the pion’s distribution amplitude with two flavours of dynamical fermions. We use nonperturbatively determined renormal- isation coefficients to convert our results to the MS scheme at 5 GeV . Employing a linear chiral extrapolation from our large pion masses > 550 MeV, we find hξ2 i = 0.281(28), leading to a value of a2 = 0.236(82) for the second Gegenbauer moment. |
Processed_Second_moment_of_the_pion's_distribution_amplitude.txt | the pion, φπ(x, µ2), contains information on how the pion’s longitudinal momentum is divided between its quark and anti-quark constituents when probed in diffractive dijet production at E-791 [1] and exclusive pion photoproduction at CLEO [2]. |
Processed_Second_moment_of_the_pion's_distribution_amplitude.txt | where z2 = 0 is a vector along the light cone, x is the fraction of the pion’s longitudinal momentum, p, carried by the quark (¯x = 1 − x for the anti- quark), fπ is the pion decay constant, and µ is the factorisation scale. On a Euclidean lattice, we are not able to compute matrix elements of bilocal operators such as ¯d(0)γµγ5u(z) in Eq. (1), ∗Talk presented by J.M. Zanotti at Light Cone 2005, Cairns, Australia. |
Processed_Second_moment_of_the_pion's_distribution_amplitude.txt | for the pion, hence the first nontrivial moment that we are able to calculate is hξ2i. |
Processed_Second_moment_of_the_pion's_distribution_amplitude.txt | Although the first lattice calculations of hξ2i appeared almost 20 years ago [3], there has been surprisingly little activity in this area in recent times [4,5,6] to complement other theoretical in- vestigations, e.g. [7,8,9,10,11,12,13,14,15,16,17, 18]. The current state-of-the-art lattice calcu- [6] who lation comes from Del Debbio et al. performed a simulation in quenched QCD and renormalised their results perturbatively to the MS scheme at µ = 2.67 GeV, hξ2iMS(µ = 2.67 GeV) = 0.280(49)+0.030 −0.013 . In these proceedings we present preliminary re- sults from the QCDSF/UKQCD collaborations for hξ2iMS in two flavour lattice QCD. These re- sults complement our preliminary results on the pion form factor [19]. |
Processed_Second_moment_of_the_pion's_distribution_amplitude.txt | The H(4) representation on the lattice leads us to consider two operators which we call Oa and Ob [20], e.g. |
Processed_Second_moment_of_the_pion's_distribution_amplitude.txt | Table 1 Lattice parameters: Gauge coupling β, sea quark lattice hopping parameter κsea, spacing and pion mass. |
Processed_Second_moment_of_the_pion's_distribution_amplitude.txt | Although the mixing coefficient for Oa, ∂∂ is not yet known, we expect that it is small and hence we anticipate that the contribution to hξ2i from Oa, ∂∂ will be small. Hence, for the rest of the work presented here, we will consider only the contribution from Oa. In a forthcoming publica- tion, we will attempt to address all mixing issues associated with both operators Oa and Ob. |
Processed_Second_moment_of_the_pion's_distribution_amplitude.txt | Figure 1. Ratio defined in Eq. (9), using Oa from Eq. (5), on a 243 × 48 lattice with β = 5.40 and κsea = 0.13560. |
Processed_Second_moment_of_the_pion's_distribution_amplitude.txt | and up to four different kappa values per beta we have generated O(2000 − 8000) trajectories. Lat- tice spacings and spatial volumes vary between 0.075-0.123 fm and (1.5-2.2 fm)3 respectively. A summary of the parameter space spanned by our dynamical configurations can be found in Table 1. We set the scale via the force parameter, with r0 = 0.5 fm. |
Processed_Second_moment_of_the_pion's_distribution_amplitude.txt | Correlation functions are calculated on config- urations taken at a distance of 10 trajectories us- ing 4 different locations of the fermion source. We use binning to obtain an effective distance of 20 trajectories. The size of the bins has little ef- fect on the error, which indicates residual auto- correlations are small. |
Processed_Second_moment_of_the_pion's_distribution_amplitude.txt | We calculate the average of matrix elements computed with three choices of pion momenta ~p0 = (p, p, 0), ~p1 = (p, 0, p), ~p2 = (0, p, p), with the indices of the operators (Eq. (5)) chosen ac- cordingly. |
Processed_Second_moment_of_the_pion's_distribution_amplitude.txt | where i and j are spatial indices, and O4 is the operator given in Eq. (4) with no derivatives and µ0 = 4. |
Processed_Second_moment_of_the_pion's_distribution_amplitude.txt | Figure 1 shows a typical example of the ratio in Eq. (9) where we clearly observe two plateaus for t < LT /2 and t > LT /2, where LT is the time extent of the lattice. After extracting Ra from the plateaus, we use Eq. (9) to extract hξ2i. |
Processed_Second_moment_of_the_pion's_distribution_amplitude.txt | Figure 3. Chiral extrapolation at constant β for β = 5.40 (top) and β = 5.29 (bottom) for Oa from Eq. (5). |
Processed_Second_moment_of_the_pion's_distribution_amplitude.txt | For each of our datasets, we extract a value for hξ2ibare from Eq. (9) and renormalise using Eq. (12). Figure 2 shows these results plotted as a function of m2 π. Here we observe that the results are approximately constant as we vary the pion mass. |
Processed_Second_moment_of_the_pion's_distribution_amplitude.txt | In order to obtain a result in the continuum and chiral limits, we first extrapolate our results at constant β to the chiral limit. In Fig. 3 we dis- play the chiral extrapolations for β = 5.40 (top) and 5.29 (bottom), while Fig. 4 contains the cor- responding extrapolations for β = 5.25 (top) and 5.20 (bottom). These results exhibit only a mild dependence on the quark mass and their values in the chiral limit agree within errors. |
Processed_Second_moment_of_the_pion's_distribution_amplitude.txt | Figure 4. Chiral extrapolation at constant β for β = 5.25 (top) and β = 5.20 (bottom) for Oa from Eq. (5). |
Processed_Second_moment_of_the_pion's_distribution_amplitude.txt | we observe that even though our operators are not O(a)-improved, we find a negligible depen- dence on the lattice spacing, at least when com- pared to the statistical errors. Hence we take the result at our smallest lattice spacing (largest β) as our result in the continuum limit. |
Processed_Second_moment_of_the_pion's_distribution_amplitude.txt | Figure 5. Results for each value of β in the chiral limit as a function of a2 for Oa from Eq. (5). |
Processed_Second_moment_of_the_pion's_distribution_amplitude.txt | In the asymptotic limit, µ2 → ∞, all a2n(µ2) = 0, for n > 0. |
Processed_Second_moment_of_the_pion's_distribution_amplitude.txt | Analyses of CLEO data for Fπγγ∗ constrain the pion distribution amplitude by calculating a re- lationship between the first two Gegenbauer mo- ments (see eg., [15,16]), together with upper and lower bounds on their respective values. |
Processed_Second_moment_of_the_pion's_distribution_amplitude.txt | which allows us to extract a2, but not a4. In order to place a constraint on a4, we would need to cal- culate the fourth moment of the pion’s distribu- tion amplitude which requires using an operator involving four covariant derivatives. |
Processed_Second_moment_of_the_pion's_distribution_amplitude.txt | which is larger than the values a2(1 GeV) = 0.07(1) [17] and a2(1 GeV) = 0.19(19) [18], but within the usual constraints 0 ≤ a2(1 GeV) ≤ 0.3 commonly quoted in the literature, even when we consider that our value will increase slightly when we run our result to a smaller scale of µ = 1 GeV. |
Processed_Second_moment_of_the_pion's_distribution_amplitude.txt | We have presented a preliminary result for the second moment of the pion’s distribution am- plitude calculated on lattices generated by the QCDSF/UKQCD collaboration with two flavours of dynamical fermions. We use nonperturbatively determined renormalisation coefficients to con- vert our result to the MS scheme at 5 GeV2. We find hξ2i = 0.281(28), which is very close to an earlier lattice result and larger than the asymp- totic value. |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.