Learning the kernel matrix with semidefinite programming

Kernel-based learning algorithms work by embedding the data into a Euclidean space, and then searching for linear relations among the embedded data points. The embedding is performed implicitly, by specifying the inner products between each pair of points in the embedding space. This information is contained in the so-called kernel matrix, a symmetric and positive semidefinite matrix that encodes the relative positions of all points. Specifying this matrix amounts to specifying the geometry of the embedding space and inducing a notion of similarity in the input space - classical model selection problems in machine learning. In this paper we show how the kernel matrix can be learned from data via semidefinite programming (SDP) techniques. When applied to a kernel matrix associated with both training and test data this gives a powerful transductive algorithm -using the labeled part of the data one can learn an embedding also for the unlabeled part. The similarity between test points is inferred from training points and their labels. Importantly, these learning problems are convex, so we obtain a method for learning both the model class and the function without local minima. Furthermore, this approach leads directly to a convex method for learning the 2-norm soft margin parameter in support vector machines, solving an important open problem.

Learning the Kernel Matrix with Semi-Definite Programming

Kernel-based learning algorithms work by embedding the data into a Euclidean space, and then searching for linear relations among the embedded data points. The embedding is performed implicitly, by specifying the inner products between each pair of points in the embedding space. This information is contained in the so-called kernel matrix, a symmetric and positive definite matrix that encodes the relative positions of all points. Specifying this matrix amounts to specifying the geometry of the embedding space and inducing a notion of similarity in the input space -- classical model selection problems in machine learning. In this paper we show how the kernel matrix can be learned from data via semi-definite programming (SDP) techniques. When applied to a kernel matrix associated with both training and test data this gives a powerful transductive algorithm -- using the labelled part of the data one can learn an embedding also for the unlabelled part. The similarity between test points is inferred from training points and their labels. Importantly, these learning problems are convex, so we obtain a method for learning both the model class and the function without local minima. Furthermore, this approach leads directly to a convex method to learn the 2-norm soft margin parameter in support vector machines, solving another important open problem. Finally, the novel approach presented in the paper is supported by positive empirical results.

Street computing : towards an integrated open data Application Programming Interface (API) for cities

This special issue of the Journal of Urban Technology brings together five articles that are based on presentations given at the Street Computing workshop held on 24 November 2009 in Melbourne in conjunction with the Australian Computer-Human Interaction conference (OZCHI 2009). Our own article introduces the Street Computing vision and explores the potential, challenges and foundations of this research vision. In order to do so, we first look at the currently available sources of information and discuss their link to existing research efforts. Section 2 then introduces the notion of Street Computing and our research approach in more detail. Section 3 looks beyond the core concept itself and summarises related work in this field of interest.

A new approach to automatically producing schedules for cane railways

The scheduling of locomotive movements on cane railways has proven to be a very complex task. Various optimisation methods have been used over the years to try and produce an optimised schedule that eliminates or minimises bin supply delays to harvesters and the factory, while minimising the number of locomotives, locomotive shifts and cane bins, and also the cane age. This paper reports on a new attempt to develop an automatic scheduler using a mathematical model solved using mixed integer programming and constraint programming approaches and blocking parallel job shop scheduling fundamentals. The model solution has been explored using conventional constraint programming search techniques and found to produce a reasonable schedule for small-scale problems with up to nine harvesters. While more effort is required to complete the development of the full model with metaheuristic search techniques, the work completed to date gives confidence that the metaheuristic techniques will provide near optimal solutions in reasonable time.

Sensitivity of the NMR density matrix to pulse sequence parameters : a simplified analytic approach

We present a formalism for the analysis of sensitivity of nuclear magnetic resonance pulse sequences to variations of pulse sequence parameters, such as radiofrequency pulses, gradient pulses or evolution delays. The formalism enables the calculation of compact, analytic expressions for the derivatives of the density matrix and the observed signal with respect to the parameters varied. The analysis is based on two constructs computed in the course of modified density-matrix simulations: the error interrogation operators and error commutators. The approach presented is consequently named the Error Commutator Formalism (ECF). It is used to evaluate the sensitivity of the density matrix to parameter variation based on the simulations carried out for the ideal parameters, obviating the need for finite-difference calculations of signal errors. The ECF analysis therefore carries a computational cost comparable to a single density-matrix or product-operator simulation. Its application is illustrated using a number of examples from basic NMR spectroscopy. We show that the strength of the ECF is its ability to provide analytic insights into the propagation of errors through pulse sequences and the behaviour of signal errors under phase cycling. Furthermore, the approach is algorithmic and easily amenable to implementation in the form of a programming code. It is envisaged that it could be incorporated into standard NMR product-operator simulation packages.

A new analytic approach for power system fault diagnosis employing the temporal information of alarm messages

Traditional analytic models for power system fault diagnosis are usually formulated as an unconstrained 0–1 integer programming problem. The key issue of the models is to seek the fault hypothesis that minimizes the discrepancy between the actual and the expected states of the concerned protective relays and circuit breakers. The temporal information of alarm messages has not been well utilized in these methods, and as a result, the diagnosis results may be not unique and hence indefinite, especially when complicated and multiple faults occur. In order to solve this problem, this paper presents a novel analytic model employing the temporal information of alarm messages along with the concept of related path. The temporal relationship among the actions of protective relays and circuit breakers, and the different protection configurations in a modern power system can be reasonably represented by the developed model, and therefore, the diagnosed results will be more definite under different circumstances of faults. Finally, an actual power system fault was served to verify the proposed method.

On the reliability of classifying programming tasks using a Neo-Piagetian theory of cognitive development

Recent research has proposed Neo-Piagetian theory as a useful way of describing the cognitive development of novice programmers. Neo-Piagetian theory may also be a useful way to classify materials used in learning and assessment. If Neo-Piagetian coding of learning resources is to be useful then it is important that practitioners can learn it and apply it reliably. We describe the design of an interactive web-based tutorial for Neo-Piagetian categorization of assessment tasks. We also report an evaluation of the tutorial's effectiveness, in which twenty computer science educators participated. The average classification accuracy of the participants on each of the three Neo-Piagetian stages were 85%, 71% and 78%. Participants also rated their agreement with the expert classifications, and indicated high agreement (91%, 83% and 91% across the three Neo-Piagetian stages). Self-rated confidence in applying Neo-Piagetian theory to classifying programming questions before and after the tutorial were 29% and 75% respectively. Our key contribution is the demonstration of the feasibility of the Neo-Piagetian approach to classifying assessment materials, by demonstrating that it is learnable and can be applied reliably by a group of educators. Our tutorial is freely available as a community resource.

A hierarchical decomposition approach for coordinated dispatch of plug-in electric vehicles

Plug-in electric vehicles (PEVs) are increasingly popular in the global trend of energy saving and environmental protection. However, the uncoordinated charging of numerous PEVs can produce significant negative impacts on the secure and economic operation of the power system concerned. In this context, a hierarchical decomposition approach is presented to coordinate the charging/discharging behaviors of PEVs. The major objective of the upper-level model is to minimize the total cost of system operation by jointly dispatching generators and electric vehicle aggregators (EVAs). On the other hand, the lower-level model aims at strictly following the dispatching instructions from the upper-level decision-maker by designing appropriate charging/discharging strategies for each individual PEV in a specified dispatching period. Two highly efficient commercial solvers, namely AMPL/IPOPT and AMPL/CPLEX, respectively, are used to solve the developed hierarchical decomposition model. Finally, a modified IEEE 118-bus testing system including 6 EVAs is employed to demonstrate the performance of the developed model and method.

Street computing : towards an integrated open data Application Programming Interface (API) for cities

This special issue of the Journal of Urban Technology brings together five articles that are based on presentations given at the Street Computing Workshop held on 24 November 2009 in Melbourne in conjunction with the Australian Computer- Human Interaction conference (OZCHI 2009). Our own article introduces the Street Computing vision and explores the potential, challenges, and foundations of this research trajectory. In order to do so, we first look at the currently available sources of information and discuss their link to existing research efforts. Section 2 then introduces the notion of Street Computing and our research approach in more detail. Section 3 looks beyond the core concept itself and summarizes related work in this field of interest. We conclude by introducing the papers that have been contributed to this special issue.

An integrated approach for earthwork allocation, sequencing and routing

Planning techniques for large scale earthworks have been considered in this article. To improve these activities a “block theoretic” approach was developed that provides an integrated solution consisting of an allocation of cuts to fills and a sequence of cuts and fills over time. It considers the constantly changing terrain by computing haulage routes dynamically. Consequently more realistic haulage costs are used in the decision making process. A digraph is utilised to describe the terrain surface which has been partitioned into uniform grids. It reflects the true state of the terrain, and is altered after each cut and fill. A shortest path algorithm is successively applied to calculate the cost of each haul, and these costs are summed over the entire sequence, to provide a total cost of haulage. To solve this integrated optimisation problem a variety of solution techniques were applied, including constructive algorithms, meta-heuristics and parallel programming. The extensive numerical investigations have successfully shown the applicability of our approach to real sized earthwork problems.

Profiling : an application assignment approach for green data centers

In the past few years, there has been a steady increase in the attention, importance and focus of green initiatives related to data centers. While various energy aware measures have been developed for data centers, the requirement of improving the performance efficiency of application assignment at the same time has yet to be fulfilled. For instance, many energy aware measures applied to data centers maintain a trade-off between energy consumption and Quality of Service (QoS). To address this problem, this paper presents a novel concept of profiling to facilitate offline optimization for a deterministic application assignment to virtual machines. Then, a profile-based model is established for obtaining near-optimal allocations of applications to virtual machines with consideration of three major objectives: energy cost, CPU utilization efficiency and application completion time. From this model, a profile-based and scalable matching algorithm is developed to solve the profile-based model. The assignment efficiency of our algorithm is then compared with that of the Hungarian algorithm, which does not scale well though giving the optimal solution.

Integrated approach to optimize open-pit mine block sequencing

Critical stage in open-pit mining is to determine the optimal extraction sequence of blocks, which has significant impacts on mining profitability. In this paper, a more comprehensive block sequencing optimisation model is developed for the open-pit mines. In the model, material characteristics of blocks, grade control, excavator and block sequencing are investigated and integrated to maximise the short-term benefit of mining. Several case studies are modeled and solved by CPLEX MIP and CP engines. Numerical investigations are presented to illustrate and validate the proposed methodology.

Curating in uncharted territories: An examination of a programming model & communicative platform for curation in non-traditional art spaces

This thesis articulates and examines public engagement programming in an emerging, non¬-traditional site. As a practice-led research project, the creative work proposes a site responsive, engagement centric, agile model for curatorial programming that developed out of the dynamic, new media/digital, curatorial practice at QUT's Creative Industries Precinct. The model and its accompanying exegetical framework, Curating in Uncharted Territories, offer a theoretically informed approach to programming, delivering and reporting for curatorial practices in a non¬-traditional sites of public engagement. The research provides the foundation for full development of the model and the basis for further research.

Mathematics, programming, and STEM

Learning mathematics is a complex and dynamic process. In this paper, the authors adopt a semiotic framework (Yeh & Nason, 2004) and highlight programming as one of the main aspects of the semiosis or meaning-making for the learning of mathematics. During a 10-week teaching experiment, mathematical meaning-making was enriched when primary students wrote Logo programs to create 3D virtual worlds. The analysis of results found deep learning in mathematics, as well as in technology and engineering areas. This prompted a rethinking about the nature of learning mathematics and a need to employ and examine a more holistic learning approach for the learning in science, technology, engineering, and mathematics (STEM) areas.