BROCCOLI: overlapping and outlier-robust biclustering through proximal stochastic gradient descent

Data Mining and Knowledge Discovery - Matrix tri-factorization subject to binary constraints is a versatile and powerful framework for the simultaneous clustering of observations and features, also...     

Recurrent kernel machines: computing with infinite echo state networks

Echo state networks (ESNs) are large, random recurrent neural networks with a single trained linear readout layer. Despite the untrained nature of the recurrent weights, they are capable of performing universal computations on temporal input data, which makes them interesting for both theoretical research and practical applications. The key to their success lies in the fact that the network computes a broad set of nonlinear, spatiotemporal mappings of the input data, on which linear regression or classification can easily be performed. One could consider the reservoir as a spatiotemporal kernel, in which the mapping to a high-dimensional space is computed explicitly. In this letter, we build on this idea and extend the concept of ESNs to infinite-sized recurrent neural networks, which can be considered recursive kernels that subsequently can be used to create recursive support vector machines. We present the theoretical framework, provide several practical examples of recursive kernels, and apply them to typical temporal tasks.     

The Impact of Software Growth on the Electronics Industry

Software growth has significantly impacted the computer industry and is changing many other electronics industries. It is not a question of whether intellectual property will be sold in the form of software, but which companies will do it successfully. The actions electronics companies take today will ultimately determine how large a role they will play in a world increasingly dominated by software     

Temperature-dependent effects of cadmium on Daphnia magna: accumulation versus sensitivity

Standard toxicity tests are performed at one constant, optimal temperature (usually 20 degrees C), while in the field variable and suboptimal temperatures may occur. Lack of knowledge on the interactions between chemicals and temperature hampers the extrapolation of laboratory toxicity data to ecosystems. Therefore, the aim of this study was to analyze the effects of temperature on cadmium toxicity to the waterflea Daphnia magna and to address possible processes responsible for temperature-dependent toxicity. This was investigated by performing standard toxicity tests with D. magna under a wide temperature range. Thermal effects on accumulation kinetics were determined by estimating uptake and elimination rates from accumulation experiments. To study temperature dependency of the intrinsic sensitivity of the daphnids to cadmium, the DEBtox model was used to estimate internal threshold concentrations (ITCs) and killing rates from the toxicity and accumulation data. The results revealed that increasing temperature lowered the ITC and increased the killing rate and the uptake rate of the metal. Enhanced sensitivity of D. magna was shown to be the primary factor for temperature-dependent toxicity. Since temperature has such a major impact on toxicity, a temperature correction may be necessary when translating toxicity data from the laboratory to the field.     

Measuring picogram per liter concentrations of freely dissolved parent and alkyl PAHs (PAH-34), using passive sampling with polyoxymethylene

Passive sampling with nondepletive sorbents is receiving increasing interest because of its potential to measure freely dissolved concentrations of hydrophobic organic compounds (HOCs) at very low concentrations, as well as its potential for both laboratory use and field deployment. However, consistent approaches have yet to be developed for the majority of HOCs of environmental and regulatory interest. In the present study, a passive sampling method was developed which allows the freely dissolved concentrations of 18 parent and 16 groups of alkyl polycyclic aromatic hydrocarbons (PAHs) on the U.S. Environmental Protection Agency (USEPA)'s "PAH-34" target compound list to be measured. Commercially available 76-μm-thick polyoxymethylene (POM) was placed in sediment/water slurries and exposed for up to 126 days, with 28 days found to be sufficient to obtain equilibrium among the sediment, water, and POM phases for the target 2- to 6-ring PAHs. The POM/water partition coefficients (K(POM)) necessary to calculate freely dissolved concentrations for parent PAHs were determined in two separate laboratories (one using pure standards, and the other using coal tar/petroleum-contaminated sediments) and agreed very well. Since the so-called "16" alkyl PAHs on the PAH-34 list actually include several hundreds of isomers for which no standards exist, sediments impacted by coal tar, or spiked with a coal tar/petroleum nonaqueous phase liquid (NAPL) were also used to measure K(POM) values for each alkyl PAH cluster. The log K(POM) values ranged from ca. 3.0 to 6.2 for 2- to 6-ring parent PAHs, and correlated well with SPARC octanol/water coefficients (K(OW)) (correlation coefficient of r(2) = 0.986). However, log K(POM) values for alkyl PAHs deviated increasingly from SPARC log K(OW) values with increasing degree of alkylation. A simple empirical model that incorporates the number of carbon atoms in a PAH gave a better fit to the experimental log K(POM) values, and was used to estimate log K(POM) for alkyl PAHs that could not be directly measured. Detection limits (as freely dissolved concentrations) ranged from ca. 1 part per trillion (ng/L) for the 2-ring PAH naphthalene, down to <1 pg/L (part per quadrillion) for the 5- and 6-ring PAHs. Sorption isotherms were linear (r(2) > 0.99) over at least 4 orders of magnitude.     

Advance warning for loss of separation in an industrial distillation column

This paper describes the development of a data‐driven advance warning system for the onset of loss of separation in an industrial distillation column. The system would enable preventive actions to avoid several hours of bad operation and subsequent recovery of the process. Data of more than 2 years of process operation were used to identify and validate various monitoring systems based on both static principal component analysis (PCA) and dynamic PCA. Despite the presence of autocorrelation in the data, only minor differences in advance warning were observed between PCA and dynamic PCA. The developed system provides warnings for 35% to 45% of the observed periods of bad column operation, with respective advance warning times of 16 and 6 minutes. It proves a valuable additional tool to monitor the operation of the distillation column and avoid losses of product, with the potential of reducing bad operation (and the associated costs) by up to 45% and substantially improving overall process reliability.     

Evaluating changes in driver behaviour: A risk profiling approach

New road safety strategies continue to be devised by researchers and policy makers with pay-as-you-drive (PAYD) schemes gaining increasing attention. However, empirically measuring the effectiveness of these strategies is challenging due to the influence of the road environment and other factors external to the driver. The analysis presented here applies Temporal and Spatial Identifiers to control for the road environment and Driver Behaviour Profiles to provide a common measure of driving behaviour based on the risk of a casualty crash for assessing the effectiveness of a PAYD scheme on reducing driving risks. The results show that in many cases personalised feedback alone is sufficient to induce significant changes, but the largest reductions in risk are observed when drivers are also awarded a financial incentive to change behaviour. Importantly, the more frequent the exposure to the speeding information, the greater the magnitude of the change. However, the changes are disproportionately associated with those that were already safer drivers in the baseline period suggesting that some drivers may be predisposed to changing their behaviour. These results suggest that it would be beneficial to provide real-time or daily feedback on speeding behaviour in conjunction with a financial reward scheme, potentially as a component of insurance premiums.     

Survey and the assimilation of a modernist narrative in urbanism

The presence of civic and regional survey as a locus of discussion in early British planning texts is closely associated with the involvement of the Scottish biologist planner Patrick Geddes in the early planning movement. This paper reads the Geddesian survey idea and its assimilation into planning discourse as the expression of a modern apology of planning. Geddes's understanding of survey places science and society in a relationship in which society is not merely determined by the progressive development of scientific insights. Instead, survey is understood as a process in which a growing awareness of the urban environment encourages an informed and self-conscious citizenry to take charge of its own future. In the Geddesian survey-project the citizen appears as the modern subject, caught in an ambiguous relationship to the environment. He or she seems to be fully determined by the environment, while at the same time able to shape that environment freely and independently. The effort to ground planning in scientific survey inserts the planning process within a logic of environmental determinism, while attempting to safeguard a place for planning as an emancipated act of self-determination. Planning appears both as a process determined by 'nature' as well as the quintessential expression of the human con trol over 'nature'. Survey, presented as preliminary to planning, is understood as a thought vehicle which opens up a space for the development of the modern plan in its full reformist breath of both physical reform and social reconstruction.