GYAN: A methodology for rule extraction from artificial neural networks

Artificial neural network (ANN) learning methods provide a robust and non-linear approach to approximating the target function for many classification, regression and clustering problems. ANNs have demonstrated good predictive performance in a wide variety of practical problems. However, there are strong arguments as to why ANNs are not sufficient for the general representation of knowledge. The arguments are the poor comprehensibility of the learned ANN, and the inability to represent explanation structures. The overall objective of this thesis is to address these issues by: (1) explanation of the decision process in ANNs in the form of symbolic rules (predicate rules with variables); and (2) provision of explanatory capability by mapping the general conceptual knowledge that is learned by the neural networks into a knowledge base to be used in a rule-based reasoning system. A multi-stage methodology GYAN is developed and evaluated for the task of extracting knowledge from the trained ANNs. The extracted knowledge is represented in the form of restricted first-order logic rules, and subsequently allows user interaction by interfacing with a knowledge based reasoner. The performance of GYAN is demonstrated using a number of real world and artificial data sets. The empirical results demonstrate that: (1) an equivalent symbolic interpretation is derived describing the overall behaviour of the ANN with high accuracy and fidelity, and (2) a concise explanation is given (in terms of rules, facts and predicates activated in a reasoning episode) as to why a particular instance is being classified into a certain category.

Multiple camera management using wide baseline matching

Camera calibration information is required in order for multiple camera networks to deliver more than the sum of many single camera systems. Methods exist for manually calibrating cameras with high accuracy. Manually calibrating networks with many cameras is, however, time consuming, expensive and impractical for networks that undergo frequent change. For this reason, automatic calibration techniques have been vigorously researched in recent years. Fully automatic calibration methods depend on the ability to automatically find point correspondences between overlapping views. In typical camera networks, cameras are placed far apart to maximise coverage. This is referred to as a wide base-line scenario. Finding sufficient correspondences for camera calibration in wide base-line scenarios presents a significant challenge. This thesis focuses on developing more effective and efficient techniques for finding correspondences in uncalibrated, wide baseline, multiple-camera scenarios. The project consists of two major areas of work. The first is the development of more effective and efficient view covariant local feature extractors. The second area involves finding methods to extract scene information using the information contained in a limited set of matched affine features. Several novel affine adaptation techniques for salient features have been developed. A method is presented for efficiently computing the discrete scale space primal sketch of local image features. A scale selection method was implemented that makes use of the primal sketch. The primal sketch-based scale selection method has several advantages over the existing methods. It allows greater freedom in how the scale space is sampled, enables more accurate scale selection, is more effective at combining different functions for spatial position and scale selection, and leads to greater computational efficiency. Existing affine adaptation methods make use of the second moment matrix to estimate the local affine shape of local image features. In this thesis, it is shown that the Hessian matrix can be used in a similar way to estimate local feature shape. The Hessian matrix is effective for estimating the shape of blob-like structures, but is less effective for corner structures. It is simpler to compute than the second moment matrix, leading to a significant reduction in computational cost. A wide baseline dense correspondence extraction system, called WiDense, is presented in this thesis. It allows the extraction of large numbers of additional accurate correspondences, given only a few initial putative correspondences. It consists of the following algorithms: An affine region alignment algorithm that ensures accurate alignment between matched features; A method for extracting more matches in the vicinity of a matched pair of affine features, using the alignment information contained in the match; An algorithm for extracting large numbers of highly accurate point correspondences from an aligned pair of feature regions. Experiments show that the correspondences generated by the WiDense system improves the success rate of computing the epipolar geometry of very widely separated views. This new method is successful in many cases where the features produced by the best wide baseline matching algorithms are insufficient for computing the scene geometry.

Enhancing agrobacterium-mediated transformation efficiency of sugarcane : progress towards an efficient, genotype-independent method

Several key issues need to be resolved before an efficient and reproducible Agrobacterium-mediated sugarcane transformation method can be developed for a wider range of sugarcane cultivars. These include loss of morphogenetic potential in sugarcane cells after Agrobacterium-mediated transformation, effect of exposure to abiotic stresses during in vitro selection, and most importantly the hypersensitive cell death response of sugarcane (and other nonhost plants) to Agrobacterium tumefaciens. Eight sugarcane cultivars (Q117, Q151, Q177, Q200, Q208, KQ228, QS94-2329, and QS94-2174) were evaluated for loss of morphogenetic potential in response to the age of the culture, exposure to Agrobacterium strains, and exposure to abiotic stresses during selection. Corresponding changes in the polyamine profiles of these cultures were also assessed. Strategies were then designed to minimize the negative effects of these factors on the cell survival and callus proliferation following Agrobacterium-mediated transformation. Some of these strategies, including the use of cell death protector genes and regulation of intracellular polyamine levels, will be discussed.

Essays on irrigation development, farm production and unaccounted costs : theory and empirical evidence

A review of the literature related to issues involved in irrigation induced agricultural development (IIAD) reveals that: (1) the magnitude, sensitivity and distribution of social welfare of IIAD is not fully analysed; (2) the impacts of excessive pesticide use on farmers’ health are not adequately explained; (3) no analysis estimates the relationship between farm level efficiency and overuse of agro-chemical inputs under imperfect markets; and (4) the method of incorporating groundwater extraction costs is misleading. This PhD thesis investigates these issues by using primary data, along with secondary data from Sri Lanka. The overall findings of the thesis can be summarised as follows. First, the thesis demonstrates that Sri Lanka has gained a positive welfare change as a result of introducing new irrigation technology. The change in the consumer surplus is Rs.48,236 million, while the change in the producer surplus is Rs. 14,274 millions between 1970 and 2006. The results also show that the long run benefits and costs of IIAD depend critically on the magnitude of the expansion of the irrigated area, as well as the competition faced by traditional farmers (agricultural crowding out effects). The traditional sector’s ability to compete with the modern sector depends on productivity improvements, reducing production costs and future structural changes (spillover effects). Second, the thesis findings on pesticides used for agriculture show that, on average, a farmer incurs a cost of approximately Rs. 590 to 800 per month during a typical cultivation period due to exposure to pesticides. It is shown that the value of average loss in earnings per farmer for the ‘hospitalised’ sample is Rs. 475 per month, while it is approximately Rs. 345 per month for the ‘general’ farmers group during a typical cultivation season. However, the average willingness to pay (WTP) to avoid exposure to pesticides is approximately Rs. 950 and Rs. 620 for ‘hospitalised’ and ‘general’ farmers’ samples respectively. The estimated percentage contribution for WTP due to health costs, lost earnings, mitigating expenditure, and disutility are 29, 50, 5 and 16 per cent respectively for hospitalised farmers, while they are 32, 55, 8 and 5 per cent respectively for ‘general’ farmers. It is also shown that given market imperfections for most agricultural inputs, farmers are overusing pesticides with the expectation of higher future returns. This has led to an increase in inefficiency in farming practices which is not understood by the farmers. Third, it is found that various groundwater depletion studies in the economics literature have provided misleading optimal water extraction quantity levels. This is due to a failure to incorporate all production costs in the relevant models. It is only by incorporating quality changes to quantity deterioration, that it is possible to derive socially optimal levels. Empirical results clearly show that the benefits per hectare per month considering both the avoidance costs of deepening agro-wells by five feet from the existing average, as well as the avoidance costs of maintaining the water salinity level at 1.8 (mmhos/Cm), is approximately Rs. 4,350 for farmers in the Anuradhapura district and Rs. 5,600 for farmers in the Matale district.

Automatic domain ontology extraction for context-sensitive opinion mining

Automated analysis of the sentiments presented in online consumer feedbacks can facilitate both organizations’ business strategy development and individual consumers’ comparison shopping. Nevertheless, existing opinion mining methods either adopt a context-free sentiment classification approach or rely on a large number of manually annotated training examples to perform context sensitive sentiment classification. Guided by the design science research methodology, we illustrate the design, development, and evaluation of a novel fuzzy domain ontology based contextsensitive opinion mining system. Our novel ontology extraction mechanism underpinned by a variant of Kullback-Leibler divergence can automatically acquire contextual sentiment knowledge across various product domains to improve the sentiment analysis processes. Evaluated based on a benchmark dataset and real consumer reviews collected from Amazon.com, our system shows remarkable performance improvement over the context-free baseline.