390 research outputs found
Knowledge refinement for a formulation system.
No full textThis thesis describes the application of the knowledge refinement tool KRUST to the design system TFS, whose task is tablet formulation for a major pharmaceutical company. KRUST has already been successfully applied to a variety of classificatory problems, and a generic refinement framework is being developed. This thesis explores the differences in knowledge content and problem-solving steps for design rather than diagnosis systems, and how this affects the refinement process. It describes how novel components found in the design system were included within KRUST’s underlying knowledge model, and how KRUST’s refinement mechanisms were extended to apply to the design system by adding new operators to the existing tool-sets. Following this necessary adaptation of KRUST, new mechanisms were introduced whereby inductive learning from proofs of related examples is used to constrain and guide KRUST’s refinement generation. The concept of a generic refinement tool is introduced. In the course of the work described here, KRUST’s knowledge and operator representations have developed in a way that facilitate its future application to different shells. The successful application of KRUST to TFS is used to show that KRUST has grown nearer to being a truly generic tool, and provides evidence that the construction of such a tool is both feasible and desirable. Lastly, the role of knowledge refinement within software development is explored. Traditionally, refinement has been applied only to debugging, but the thesis shows how refinement can also play a role in software maintenance. In the course of its development, TFS has undergone both routine debugging, and also maintenance, when the formulation task was altered by a change in company policy. It was thus possible to test the extent to which KRUST was able to reproduce automatically the changes that were originally made manually to TFS, and hence to evaluate KRUST’s effectiveness in both debugging and maintenance roles
Knowledge driven approaches to e-learning recommendation.
No full textLearners often have difficulty finding and retrieving relevant learning materials to support their learning goals because of two main challenges. The vocabulary learners use to describe their goals is different from that used by domain experts in teaching materials. This challenge causes a semantic gap. Learners lack sufficient knowledge about the domain they are trying to learn about, so are unable to assemble effective keywords that identify what they wish to learn. This problem presents an intent gap. The work presented in this thesis focuses on addressing the semantic and intent gaps that learners face during an e-Learning recommendation task. The semantic gap is addressed by introducing a method that automatically creates background knowledge in the form of a set of rich learning-focused concepts related to the selected learning domain. The knowledge of teaching experts contained in e-Books is used as a guide to identify important domain concepts. The concepts represent important topics that learners should be interested in. An approach is developed which leverages the concept vocabulary for representing learning materials and this influences retrieval during the recommendation of new learning materials. The effectiveness of our approach is evaluated on a dataset of Machine Learning and Data Mining papers, and our approach outperforms benchmark methods. The results confirm that incorporating background knowledge into the representation of learning materials provides a shared vocabulary for experts and learners, and this enables the recommendation of relevant materials. We address the intent gap by developing an approach which leverages the background knowledge to identify important learning concepts that are employed for refining learners' queries. This approach enables us to automatically identify concepts that are similar to queries, and take advantage of distinctive concept terms for refining learners' queries. Using the refined query allows the search to focus on documents that contain topics which are relevant to the learner. An e-Learning recommender system is developed to evaluate the success of our approach using a collection of learner queries and a dataset of Machine Learning and Data Mining learning materials. Users with different levels of expertise are employed for the evaluation. Results from experts, competent users and beginners all showed that using our method produced documents that were consistently more relevant to learners than when the standard method was used. The results show the benefits in using our knowledge driven approaches to help learners find relevant learning materials
Employing multi-modal sensors for personalised smart home health monitoring.
No full textSmart home systems are employed worldwide for a variety of automated monitoring tasks. FITsense is a system that performs personalised smart home health monitoring using sensor data. In this thesis, we expand upon this system by identifying the limits of health monitoring using simple IoT sensors, and establishing deployable solutions for new rich sensing technologies. The FITsense system collects data from FitHomes and generates behavioural insights for health monitoring. To allow the system to expand to arbitrary home layouts, sensing applications must be delivered while relying on sparse "ground truth" data. An enhanced data representation was tested for improving activity recognition performance by encoding observed temporal dependencies. Experiments showed an improvement in activity recognition accuracy over baseline data representations with standard classifiers. Channel State Information (CSI) was chosen as our rich sensing technology for its ambient nature and potential deployability. We developed a novel Python toolkit, called CSIKit, to handle various CSI software implementations, including automatic detection for off-the-shelf CSI formats. Previous researchers proposed a method to address AGC effects on COTS CSI hardware, which we tested and found to improve correlation with a baseline without AGC. This implementation was included in the public release of CSIKit. Two sensing applications were delivered using CSIKit to demonstrate its functionality. Our statistical approach to motion detection with CSI data showed a 32% increase in accuracy over an infrared sensor-based solution using data from 2 unique environments. We also demonstrated the first CSI activity recognition application on a Raspberry Pi 4, which achieved an accuracy of 92% with 11 activity classes. An application was then trained to support movement detection using data from all COTS CSI hardware. This was combined with our signal divider implementation to compare CSI wireless and sensing performance characteristics. The IWL5300 exhibited the most consistent wireless performance, while the ESP32 was found to produce viable CSI data for sensing applications. This establishes the ESP32 as a low-cost high-value hardware solution for CSI sensing. To complete this work, an in-home study was performed using real-world sensor data. An ESP32-based CSI sensor was developed to be integrated into our IoT network. This sensor was tested in a FitHome environment to identify how the data from our existing simple sensors could aid sensor development. We performed an experiment to demonstrate that annotations for CSI data could be gathered with infrared motion sensors. Results showed that our new CSI sensor collected real-world data of similar utility to that collected manually in a controlled environment
Complexity modelling for case knowledge maintenance in case-based reasoning.
No full textCase-based reasoning solves new problems by re-using the solutions of previously solved similar problems and is popular because many of the knowledge engineering demands of conventional knowledge-based systems are removed. The content of the case knowledge container is critical to the performance of case-based classification systems. However, the knowledge engineer is given little support in the selection of suitable techniques to maintain and monitor the case base. This research investigates the coverage, competence and problem-solving capacity of case knowledge with the aim of developing techniques to model and maintain the case base. We present a novel technique that creates a model of the case base by measuring the uncertainty in local areas of the problem space based on the local mix of solutions present. The model provides an insight into the structure of a case base by means of a complexity profile that can assist maintenance decision-making and provide a benchmark to assess future changes to the case base. The distribution of cases in the case base is critical to the performance of a case-based reasoning system. We argue that classification boundaries represent important regions of the problem space and develop two complexity-guided algorithms which use boundary identification techniques to actively discover cases close to boundaries. We introduce a complexity-guided redundancy reduction algorithm which uses a case complexity threshold to retain cases close to boundaries and delete cases that form single class clusters. The algorithm offers control over the balance between maintaining competence and reducing case base size. The performance of a case-based reasoning system relies on the integrity of its case base but in real life applications the available data invariably contains erroneous, noisy cases. Automated removal of these noisy cases can improve system accuracy. In addition, error rates can often be reduced by removing cases to give smoother decision boundaries between classes. We show that the optimal level of boundary smoothing is domain dependent and, therefore, our approach to error reduction reacts to the characteristics of the domain by setting an appropriate level of smoothing. We introduce a novel algorithm which identifies and removes both noisy and boundary cases with the aid of a local distance ratio. A prototype interface has been developed that shows how the modelling and maintenance approaches can be used in practice in an interactive manner. The interface allows the knowledge engineer to make informed maintenance choices without the need for extensive evaluation effort while, at the same time, retaining control over the process. One of the strengths of our approach is in applying a consistent, integrated method to case base maintenance to provide a transparent process that gives a degree of explanation
Self adapting websites: mining user access logs.
No full textThe Web can be regarded as a large repository of diversified information in the form of millions of websites distributed across the globe. However, the ever increasing number of websites in the Web has made it extremely difficult for users to find the right informa- tion that satisfies their current needs. In order to address this problem, many researchers explored Web Mining as a way of developing intelligent websites, which could present the information available in a website in a more meaningful way by relating it to a users need. Web Mining applies data mining techniques on web usage, web content or web structure data to discover useful knowledge such as topical relations between documents, users access patterns and website usage statistics. This knowledge is then used to develop intelligent websites that can personalise the content of a website based on a users prefer- ence. However, existing intelligent websites are too focussed on filtering the information available in a website to match a users need, ignoring the true source of users problems in the Web. The majority of problems faced by users in the Web today, can be reduced to issues related to a websites design. All too often, users needs change rapidly but the websites remain static and existing intelligent websites such as customisation, personalisa- tion and recommender systems only provide temporary solutions to this problem. An idea introduced to address this limitation is the development of adaptive websites. Adaptive websites are sites that automatically change their organisation and presentation based on users access patterns. Shortcutting is a sophisticated method used to change the organi- sation of a website. It involves connecting two documents that were previously unlinked in a website by adding a new hyperlink between them based on correlations in users visits. Existing methods tend to minimize the number of clicks required to find a target document by providing a shortcut between the initial and target documents in a users navigational path. This approach assumes the sequence of intermediate documents appearing in the path is insignificant to a users information need and bypasses them. In this work, we explore the idea of adaptive websites and present our approach to it using wayposts to address the above mentioned limitation. Wayposts are intermediate documents in a users path which may contain information significant to a users need that could lead him to his intended target document. Our work identifies such wayposts from frequently travelled users paths and suggests them as potential navigational shortcuts, which could be used to improve a websites organisation
Integrating content and semantic representations for music recommendation.
No full textMusic recommender systems are used by millions of people every day to discover new and exciting music. Central to making recommendations is the representation of each track, which may be used to calculate similarity. Content representations capture the musical and texture facets of each track, and semantic representations describe social and cultural information provided by listeners. This thesis is motivated by an analysis of the strengths and weaknesses of both content and semantic representations. Content representations can be available for all tracks in a collection, but provide poor recommendation quality. Semantic representations suffer from the cold-start problem and are not available for all tracks, but provide good recommendation quality when a strong representation is available. These observations highlight the need to integrate both content and semantic representations, and use the strengths of each to improve music recommendation quality and discovery. A bridge of the gap between content and semantic representations is achieved in this thesis through hybrid representation. Content texture representations are examined, and a new music-inspired texture representation is defined. This content is integrated with semantic tags directly, and through a mid-level pseudo-tag representation. The effect of these approaches is to increase the high quality discovery of tracks, and to allow users to uncover interesting novel recommendations. The challenge of evaluating music recommendations when many tracks are undertagged is addressed. Implicit and explicit feedback provided by users is exploited to define a new ground truth similarity measure, which accurately reflects how different recommendation methods perform. A user study is conducted to evaluate both this measure, and the performance of integrated representations for recommending strong novel recommendations
Novius koebelei Olliff in Craw 1892
No full text<i>Novius koebelei</i> Olliff in Craw <p>(Figs 19–24, 48–59, 64, 76–80, 90, 96–99)</p> <p> <i>Novius Koebelei</i> Olliff in Craw, 1892: 14. TL: Australia, introduced to U.S.A.; Coquillett, 1893: 20; Lea, 1902: 493.</p> <p> <i>Rodolia koebelei</i>: Korschefsky, 1931: 101; Gordon, 1972: 26; Ślipiński, 2007: 143.</p> <p> <i>Novius limbatus</i> Blackburn, 1895: 254. TL: Queensland, near Cairns. <b>Syn. nov.</b></p> <p> <i>Rodolia limbata</i>: Ślipiński, 2007: 143.</p> <p> <i>Rodolia blackburni</i> Ukrainsky, 2009: 285 (replacement name for <i>Novius limbata</i> Blackburn, 1895, not Motschulsky, 1866).</p> <p> <i>Novius tridens</i> Lea, 1902: 492. TL: Queensland, near Cairns. <b>Syn. nov.</b></p> <p> <i>Rodolia tridens</i>: Ślipiński, 2007: 143.</p> <p> <i>Novius simplicipennis</i> Blackburn, 1895: 253. TL: Queensland, Toowoomba. <b>Syn. nov.</b></p> <p> <i>Rodolia simplicipennis</i>: Ślipiński, 2007: 143.</p> <p> <i>Novius discoidalis</i> Blackburn, 1895: 253. TL: Queensland, near Toowoomba. <b>Syn. nov.</b></p> <p> <i>Rodolia discoidalis</i>: Ślipiński, 2007: 143.</p> <p> <i>Novius tripustulatus</i> Blackburn, 1895: 254. TL: Queensland, near Cairns. <b>Syn. nov.</b></p> <p> <i>Rodolia tripustulata</i>: Ślipiński, 2007: 143.</p> <p> <i>Novius ruber</i> Blackburn, 1889a: 148. TL: New South Wales, Mulwala. <b>Syn. nov.</b></p> <p> <i>Rodolia rubra</i>: Ślipiński, 2007: 143.</p> <p> <b> <i>Diagnosis</i>.</b> <i>Novius koebelei</i> can only be diagnosed by the details of the male genitalia, in particular by the presence of the small apical barb at the apex of the penis guide.</p> <p> <b> <i>Description</i>.</b> Length 2.7–3.5 mm. Body oval, widest near middle, 1.2–1.3 times longer than wide. Color pattern variable. Head, pronotum and scutellum usually uniformly dark. Elytra of typical form orange or red often with darker sutural stripe posteriorly and small lateral dark spot near midlength of elytron situated near lateral margin and continuing as darker stripe along lateral margin posteriorly. Melanic forms have elytra with black areas of various sizes or, rarely entirely black (Figs 48–59, 64, 76–80, 90). Interocular distance in frontal view 1.6–2.0 times eye width. Male genitalia (Figs 19–23, 108–111): parameres slender, not expanded apically; penis guide slightly longer than parameres, stout and narrowing apically with small apical barb. Penis bent and pointed apically. Female genitalia as in Figs 96–98.</p> <p> <b> <i>Type material</i>.</b> <i>Novius Koebelei</i> Olliff in Craw: Lectotype, here designated, the specimen illustrated in Craw, 1892: frontispiece, plate 1, fig 3, “ <i>Novius Koebelei</i>, Olliff.: male enlarged.”; <i>Novius limbatus</i> Blackburn: Lectotype, here designated (BMNH), “ Australia Blackburn Coll. B.M. 1910-236/ <i>Novius limbatus</i> Blackb. / T 5938, Qu.”; <i>Novius tridens</i> Lea: Holotype male: “10408 <i>Novius tridens</i> Lea, N.S. Wales, TYPE, S.A. Museum”; <i>Novius simplicipennis</i> Blackburn: Lectotype, here designated (BMNH), “ Australia Blackburn Coll. B.M. 1910-236/ <i>Novius simplicipennis</i> Blackb. / T 4164, Qu”; <i>Novius discoidalis</i> Blackburn: Lectotype male, here designated (BMNH), “ Australia Blackburn Coll. B.M. 1910-236/ <i>Novius discoidalis</i> Blackb. / T 5936, Qu.”; <i>Novius tripustulatus</i> Blackburn: Lectotype female, here designated (BMNH), “ Australia Blackburn Coll. B.M. 1910-236/ <i>Novius tripustulatus</i> Blackb. / T 5937, Qu.”; <i>Novius ruber</i> Blackburn: Lectotype female, here designated (BMNH), Blackburn: “ Australia Blackburn Coll. B.M. 1910- 236/ <i>Novius ruber</i> Blackb. / T 2963.”</p> <p> <b> <i>Other specimens examined</i>. New South Wales</b> : Sydney, R.C.L. Perkins, 1942-95 (4, BMNH); Exp.937, 8.9.59, Frank Wilson, Sydney, viii.1959, C.I.E. Coll. No. 16712, ANIC (5, ANIC). <b>Queensland:</b> Tambourine Mts, 11-17.v.1935 <b>(</b> 2, BMNH); Brisbane, 28.28S 153.01E, ix.1992, V. Brancatini (2, BMNH); MUS. VIC. ENT-1041, Goodna, x.1920, F.E. Wilson (1, VM); 19km S of Bundaberg, Pine Ck, 9.v.1975, H. Frauca, (1, VM); Rockhampton, 10-11.iii.1965, Exp. Dr. J. Balogh, (1, ANIC); Caloundra, iv.1965, L. Simpson, captured on <i>Acelypha</i> sp. (4, ANIC); Brisbane, 27.28S 153.01E, viii.1992, V. Brancatini, (LPL9521) Pred. of <i>Icerya seychellarum</i> on <i>Livistona chinensis</i> (1, ANIC); Kenmore, 27.30.7S 152.56.2E, 9.xi.1993, V. Brancatini, <i>Rodolia koebelei</i> ? LPL9532 (KE15), predator of <i>Icerya aegyptiaca</i> on decoy <i>Ficus benjamina,</i> (12, ANIC); Brisbane, 27.28S 153.01E, 28.v.1992, V. Brancatini, <i>Rodolia</i> sp? (LPL9521), Pred. of <i>Icerya seychellarum</i> on <i>Livistona chinensis,</i> (4, ANIC); Long Pocket, 27.30.6S 152.59.7E, v.1993, V. Brancatini, <i>Rodolia koebelei</i> ? LPL9532 (LP108), predator of <i>Icerya aegyptiaca</i> on decoy <i>Ficus benjamina</i>, (9, ANIC); Indooroopilly, 27.30.0S 152.58.4E, 7.x.1994, O. Fahey & V. Branvatini, <i>Rodolia koebelei</i> ? LPL9533 (IP24), predator of <i>Icerya aegyptiaca</i> on decoy <i>Ficus benjamina</i> (6, ANIC); Indooroopilly, 27.30.0S 152.58.4E, 1.vi.1994, V. Bran +catini & O. Fahey, <i>Rodolia koebelei</i> ? LPL9533 (IP31), predator of <i>Icerya aegyptiaca</i> on <i>Livistona chinensis,</i> (8, ANIC); Indooroopilly, 27.30.0S 152.58.4E, 30.viii.1994, V. Branvatini, <i>Rodolia koebelei</i> ? LPL 9533 (IP22), predator of <i>Icerya aegyptiaca</i> on <i>Livistona chinensis</i> (5, ANIC); Kenmore, 27.37.3S 152.56.0E, ix.1994, V. Brancatini, <i>Rodolia koebelei,</i> predator of <i>Icerya seychellarum</i> on <i>Michelia figo</i> (1, ANIC). <b>Northern Territory:</b> Darwin NT, 12.27S 130.50E, x.1992, V. Brancatini (18, BMNH); Darwin NT, 19.iv.1991, V. Brancatini (8, BMNH); Berrimah, Darwin, 16.iv.1991, V. Brancatini (2, BMNH); Alawa, Darwin, 22.iv.1991, V. Brancatini (6, BMNH); Darwin, 19.iv.1991, V. Brancatini, <i>Rodolia</i> sp. (LPL9507), ex <i>Icerya aegyptiaca</i> on <i>Nandina domestica</i>, Brit. Mus. (M.H.), 1993-94 (6, BMNH); 12.46S 132.39E, 12km NNW of Mt. Cahill, 25.x.1972, at light, E. Britton (1, ANIC); Darwin, 12.27S 130.50E, x.1992, V. Brancatini, <i>Rodolia</i> sp. (LPL9507), Lab. Cult. Reared on <i>Icerya aegyptiaca</i> on <i>Ficus benjamina</i> (4, ANIC); Darwin, 12.27S 130.55E, x./ xi. 1993, V. Brancatini, LPL 9507, Laboratory culture reared on <i>Icerya aegyptiaca</i> on <i>Ficus benjamina</i> (45, ANIC).</p> <p> <b> <i>Distribution</i>.</b> Australia (Fig. 24), introduced to the United States (California) and subsequently to many countries worldwide.</p> <p> <b> <i>Remarks</i>.</b> <i>Novius koebelei</i> has been described or mentioned many times in the literature, usually with reference to Olliff (1895) who was originally cited as the author of this species. Coquillett (1893) had previously described the larval and egg stages of this species, and was subsequently cited as the author by Gordon (1972, 1985). Gordon (1972) designated a fourth instar larva as the neotype of <i>Novius koebelei</i> Coquillett but this action is invalid (based on an individual that is insufficient to ensure recognition), and unnecessary because the first description of this species appeared in 1892 in the report by A. Craw who illustrated the adult and larvae accompanied by the name <i>Novius Koebelei</i>, Olliff. According to the ICZN Art 74.4, we here designate the specimen illustrated in fig. 3 of that paper as the lectotype of <i>Novius koebelei</i> Olliff in Craw (1892).</p>Published as part of <i>Pang, Hong, Tang, Xue-Fei, Booth, Roger G., Vandenberg, Natalia, Forrester, Juanita, Mchugh, Joseph & Ślipiński, Adam, 2020, Revision Of The Australian Coccinellidae (Coleoptera). Genus Novius Mulsant Of Tribe Noviini, pp. 1-24 in Annales Zoologici 70 (1)</i> on pages 8-11, DOI: 10.3161/00034541ANZ2020.70.1.001, <a href="http://zenodo.org/record/3776582">http://zenodo.org/record/3776582</a>
Representing and reasoning about concrete domains with inference fusion.
No full textDescription Logic (DL)-based concept modelling formalisms provide a powerful means to represent and reason about taxonomic knowledge. They benefit from unambiguous semantics which make possible the automatic classification of concept definitions. However, the inevitable "trade-off" between the expressive power and the computational complexity prevents DLs from widely being applied to model problem domains featured by heterogeneous knowledge (e.g, mechanical domains). Current DL-based inferential systems adopt a tableau-based algorithm to enable automated reasoning, which, to some extent, can be considered as a constraint system that relies on concept constructors for specifying restrictions on the defining concepts. Such facts present a reasonable inspiration to envisage a hybrid approach extending existing DL-based inferential engines with expressive and deductive powers of concrete constraints and, in the meantime, leaving the original systems intact—without modifying their underlying inference algorithms. This idea is presented as a generic schema, the inference fusion framework, which dynamically fuses homogeneous reasoners to provide a heterogeneous inference. The fusion process is facilitated by fragmenting and rescheduling reasonings with regard to concrete knowledge to constraint solvers and updating DL-based inferential systems with the corresponding feedback. A mapping mechanism is responsible for the communication between the DL-based reasoning and the constraint-based one. In this thesis, a hybrid modelling language, DL(D)/S, is proposed to demonstrate the applicability of inference fusion. DL(D)/S extends DLs with two extensions that are the Knowledge Base (KB) global constraints defined on hybrid role successors, hybrid concepts, and the numeric constraints on role cardinality variables. A series of examples are used to give an in-depth explanation of the hybrid reasoning regarding DL(D)/S. The CONstrained COncept Reasoning system (CONCOR) system is proposed as the inference fusion-based hybrid reasoning system. CONCOR takes full advantage of existing DL-based inferential systems and constraint solvers. Input hybrid concepts are fragmented, normalised and distributed to different inferential engines if appropriate. With the help of CONCOR, DL-based systems can take the results of the constraint reasoning "blindly" and upgrade the conceptual hierarchical structures accordingly. DL(D)/S is presented as an extension of the expressive DL-based language, ALCN. However, since the inference fusion framework and the CONCOR system do not explicitly depend on any particular DL-based systems for the taxonomic reasoning, other DLs could have been used instead, as long as they meet the basic expressive requirements of the hybrid reasoning. Evaluation of CONCOR with real-world examples presents promising results. Concept taxonomy is correctly updated based on the inferential results of concrete constraints. Evaluation also supports the announced advantages of the CONCOR system architecture. The work presented in this thesis is only an experiment of the inference fusion approach. Further validations and evaluations of this idea and the practical hybrid reasoning systems still require more efforts to be made in the next stage
Eastern Fables
No full textSix fables from the Panchatantra, each with an illustration. The Gardener and the Bear and The Crane and the Craw-Fish are familiar to me. Good examples of the Panchatantra's more rambling style.This is a hardbound book (hard cover)#24 of 250F.E.I
- …
