Incremental learning for interactive e-mail filtering

In this article, we propose a framework, namely, Prediction-Learning-Distillation (PLD) for interactive document classification and distilling misclassified documents. Whenever a user points out misclassified documents, the PLD learns from the mistakes and identifies the same mistakes from all other classified documents. The PLD then enforces this learning for future classifications. If the classifier fails to accept relevant documents or reject irrelevant documents on certain categories, then PLD will assign those documents as new positive/negative training instances. The classifier can then strengthen its weakness by learning from these new training instances. Our experiments’ results have demonstrated that the proposed algorithm can learn from user-identified misclassified documents, and then distil the rest successfully.

Can myocardial contrast echo provide incremental benefit to exercise echo? Implications of the intensity and duration of hyperaemia

Background. Stress myocardial contrast echo (MCE) is technically challenging with exercise (Ex) because of cardiacmovementandshort duration ofhyperemia.Vasodilators solve these limitations, but are less potent for inducing abnormal wall motion (WM). We sought whether a combined dipyridamole (DI; 0.56 mg/kg i.v. 4 min) and Ex stress protocol would enable MCE to provide incremental benefit toWManalysis for detection of CAD. Methods. Standard echo images were followed by real time MCE at rest and following stress in 85 pts, 70 undergoing quantitative coronary angiography and 15 low risk pts.WMAfrom standard and LVopacification images, and then myocardial perfusion were assessed sequentially in a blinded fashion. A subgroup of 13 pts also underwent Ex alone, to assess the contribution of DI to quantitative myocardial flow reserve (MFR). Results. Significant (>50%) stenoses were present in 43 pts, involving 69 territories. Addition of MCE improved SE sensitivity for detection of CAD (91% versus 74%, P = 0.02) and better appreciation of disease extent (87% versus 65%territories, P=0.003), with a non-significant reduction in specificity. In 55 territories subtended by a significant stenosis, but with no resting WM abnormality, ability to identify ischemia was also significantly increased by MCE (82% versus 60%, P = 0.002). MFR was less with Ex alone than with DIEx stress (2.4 ± 1.6 versus 4.0 ± 1.9, P = 0.05), suggesting prolongation of hyperaemia with DI may be essential to the results. Conclusions. Dipyridamole-exercise MCE adds significant incremental benefit to standard SE, with improved diagnostic sensitivity and more accurate estimation of extent of CAD.

Incremental model transformation for the evolution of model-driven systems

Model transformations are an integral part of model-driven development. Incremental updates are a key execution scenario for transformations in model-based systems, and are especially important for the evolution of such systems. This paper presents a strategy for the incremental maintenance of declarative, rule-based transformation executions. The strategy involves recording dependencies of the transformation execution on information from source models and from the transformation definition. Changes to the source models or the transformation itself can then be directly mapped to their effects on transformation execution, allowing changes to target models to be computed efficiently. This particular approach has many benefits. It supports changes to both source models and transformation definitions, it can be applied to incomplete transformation executions, and a priori knowledge of volatility can be used to further increase the efficiency of change propagation.

Playing in continuous spaces: Some analysis and extension of population-based incremental learning

As an alternative to traditional evolutionary algorithms (EAs), population-based incremental learning (PBIL) maintains a probabilistic model of the best individual(s). Originally, PBIL was applied in binary search spaces. Recently, some work has been done to extend it to continuous spaces. In this paper, we review two such extensions of PBIL. An improved version of the PBIL based on Gaussian model is proposed that combines two main features: a new updating rule that takes into account all the individuals and their fitness values and a self-adaptive learning rate parameter. Furthermore, a new continuous PBIL employing a histogram probabilistic model is proposed. Some experiments results are presented that highlight the features of the new algorithms.

Incremental derivation of abstraction relations for data refinements

Data refinements are refinement steps in which a program’s local data structures are changed. Data refinement proof obligations require the software designer to find an abstraction relation that relates the states of the original and new program. In this paper we describe an algorithm that helps a designer find an abstraction relation for a proposed refinement. Given sufficient time and space, the algorithm can find a minimal abstraction relation, and thus show that the refinement holds. As it executes, the algorithm displays mappings that cannot be in any abstraction relation. When the algorithm is not given sufficient resources to terminate, these mappings can help the designer find a suitable abstraction relation. The same algorithm can be used to test an abstraction relation supplied by the designer.

An incremental tri-partite approach to ontology learning

In this paper we present a new approach to ontology learning. Its basis lies in a dynamic and iterative view of knowledge acquisition for ontologies. The Abraxas approach is founded on three resources, a set of texts, a set of learning patterns and a set of ontological triples, each of which must remain in equilibrium. As events occur which disturb this equilibrium various actions are triggered to re-establish a balance between the resources. Such events include acquisition of a further text from external resources such as the Web or the addition of ontological triples to the ontology. We develop the concept of a knowledge gap between the coverage of an ontology and the corpus of texts as a measure triggering actions. We present an overview of the algorithm and its functionalities.

Testing incremental politics : the case of North Sea oil

In the quest to secure the much vaunted benefits of North Sea oil, highly non-incremental technologies have been adopted. Nowhere is this more the case than with the early fields of the central and northern North Sea. By focusing on the inflexible nature of North Sea hardware, in such fields, this thesis examines the problems that this sort of technology might pose for policy making. More particularly, the following issues are raised. First, the implications of non-incremental technical change for the successful conduct of oil policy is raised. Here, the focus is on the micro-economic performance of the first generation of North Sea oil fields and the manner in which this relates to government policy. Secondly, the question is posed as to whether there were more flexible, perhaps more incremental policy alternatives open to the decision makers. Conclusions drawn relate to the degree to which non-incremental shifts in policy permit decision makers to achieve their objectives at relatively low cost. To discover cases where non-incremental policy making has led to success in this way, would be to falsify the thesis that decision makers are best served by employing incremental politics as an approach to complex problem solving.

Incremental nature of anterior eye grading scales determined by objective image analysis

Aim: To use previously validated image analysis techniques to determine the incremental nature of printed subjective anterior eye grading scales. Methods: A purpose designed computer program was written to detect edges using a 3 × 3 kernal and to extract colour planes in the selected area of an image. Annunziato and Efron pictorial, and CCLRU and Vistakon-Synoptik photographic grades of bulbar hyperaemia, palpebral hyperaemia roughness, and corneal staining were analysed. Results: The increments of the grading scales were best described by a quadratic rather than a linear function. Edge detection and colour extraction image analysis for bulbar hyperaemia (r2 = 0.35-0.99), palpebral hyperaemia (r2 = 0.71-0.99), palpebral roughness (r2 = 0.30-0.94), and corneal staining (r2 = 0.57-0.99) correlated well with scale grades, although the increments varied in magnitude and direction between different scales. Repeated image analysis measures had a 95% confidence interval of between 0.02 (colour extraction) and 0.10 (edge detection) scale units (on a 0-4 scale). Conclusion: The printed grading scales were more sensitive for grading features of low severity, but grades were not comparable between grading scales. Palpebral hyperaemia and staining grading is complicated by the variable presentations possible. Image analysis techniques are 6-35 times more repeatable than subjective grading, with a sensitivity of 1.2-2.8% of the scale.

Incremental tensor biased discriminant analysis:a new color-based visual tracking method

Most existing color-based tracking algorithms utilize the statistical color information of the object as the tracking clues, without maintaining the spatial structure within a single chromatic image. Recently, the researches on the multilinear algebra provide the possibility to hold the spatial structural relationship in a representation of the image ensembles. In this paper, a third-order color tensor is constructed to represent the object to be tracked. Considering the influence of the environment changing on the tracking, the biased discriminant analysis (BDA) is extended to the tensor biased discriminant analysis (TBDA) for distinguishing the object from the background. At the same time, an incremental scheme for the TBDA is developed for the tensor biased discriminant subspace online learning, which can be used to adapt to the appearance variant of both the object and background. The experimental results show that the proposed method can track objects precisely undergoing large pose, scale and lighting changes, as well as partial occlusion. © 2009 Elsevier B.V.