The Moldable Inspector

Object inspectors are an essential category of tools that allow developers to comprehend the run-time of object-oriented systems. Traditional object inspectors favor a generic view that focuses on the low-level details of the state of single objects. Based on 16 interviews with software developers and a follow-up survey with 62 respondents we identified a need for object inspectors that support different high-level ways to visualize and explore objects, depending on both the object and the current developer need. We propose the Moldable Inspector, a novel inspector model that enables developers to adapt the inspection workflow to suit their immediate needs by making the inspection context explicit, providing multiple interchangeable domain-specific views for each object, and supporting a workflow that groups together multiple levels of connected objects. We show that the Moldable Inspector can address multiple kinds of development needs involving a wide range of objects.

Using space to represent categories: Insights from gaze position

We investigated the boundaries among imagery, memory, and perception by measuring gaze during retrieved versus imagined visual information. Eye fixations during recall were bound to the location at which a specific stimulus was encoded. However, eye position information generalized to novel objects of the same category that had not been seen before. For example, encoding an image of a dog in a specific location enhanced the likelihood of looking at the same location during subsequent mental imagery of other mammals. The results suggest that eye movements can also be launched by abstract representa- tions of categories and not exclusively by a single episode or a specific visual exemplar.

Hodgkin's lymphoma cell lines express a fusion protein encoded by intergenically spliced mRNA for the multilectin receptor DEC-205 (CD205) and a novel C-type lectin receptor DCL-1

Classic Hodgkin's lymphoma (HL) tissue contains a small population of morphologically distinct malignant cells called Hodgkin and Reed-Sternberg (HRS) cells, associated with the development of HL. Using 3'-rapid amplification of cDNA ends ( RACE) we identified an alternative mRNA for the DEC-205 multilectin receptor in the HRS cell line L428. Sequence analysis revealed that the mRNA encodes a fusion protein between DEC-205 and a novel C-type lectin DCL-1. Although the 7.5-kb DEC-205 and 4.2-kb DCL-1 mRNA were expressed independently in myeloid and B lymphoid cell lines, the DEC-205/DCL-1 fusion mRNA (9.5 kb) predominated in the HRS cell lines ( L428, KM-H2, and HDLM-2). The DEC-205 and DCL-1 genes comprising 35 and 6 exons, respectively, are juxtaposed on chromosome band 2q24 and separated by only 5.4 kb. We determined the DCL-1 transcription initiation site within the intervening sequence by 5'-RACE, confirming that DCL-1 is an independent gene. Two DEC-205/DCL-1 fusion mRNA variants may result from cotranscription of DEC-205 and DCL-1, followed by splicing DEC-205 exon 35 or 34-35 along with DCL-1 exon 1. The resulting reading frames encode the DEC-205 ectodomain plus the DCL-1 ectodomain, the transmembrane, and the cytoplasmic domain. Using DCL-1 cytoplasmic domain-specific polyclonal and DEC-205 monoclonal antibodies for immunoprecipitation/Western blot analysis, we showed that the fusion mRNA is translated into a DEC-205/DCL-1 fusion protein, expressed in the HRS cell lines. These results imply an unusual transcriptional control mechanism in HRS cells, which cotranscribe an mRNA containing DEC-205 and DCL-1 prior to generating the intergenically spliced mRNA to produce a DEC-205/DCL-1 fusion protein.

The application of cognitive orientation to daily occupational performance (CO-OP) with children 5-7 years with developmental coordination disorder

There has been an increase in the use of cognitive frameworks in occupational therapy with children with developmental coordination disorder (DCD). Investigations into the utility of one such cognitive approach, namely Cognitive Orientation to (daily) Occupational Performance (CO-OP), with children with DCD have shown the intervention to be effective with children over 7 years. However, there has been limited research into its utility with younger children. This paper presents two case studies to demonstrate the use of CO-OP with children aged 5-7 years. Two boys with DCD engaged in 10 sessions of CO-OP. These younger children were found to be able to use the global framework (Goal, Plan, Do, Check) to improve their task performance, to develop plans using domain-specific strategies and to engage in checking strategies. Issues relating to attention, motivation and goal setting are discussed in the context of the two case studies.

Fitness assessment of document model

Document classification is a supervised machine learning process, where predefined category labels are assigned to documents based on the hypothesis derived from training set of labelled documents. Documents cannot be directly interpreted by a computer system unless they have been modelled as a collection of computable features. Rogati and Yang [M. Rogati and Y. Yang, Resource selection for domain-specific cross-lingual IR, in SIGIR 2004: Proceedings of the 27th annual international conference on Research and Development in Information Retrieval, ACM Press, Sheffied: United Kingdom, pp. 154-161.] pointed out that the effectiveness of document classification system may vary in different domains. This implies that the quality of document model contributes to the effectiveness of document classification. Conventionally, model evaluation is accomplished by comparing the effectiveness scores of classifiers on model candidates. However, this kind of evaluation methods may encounter either under-fitting or over-fitting problems, because the effectiveness scores are restricted by the learning capacities of classifiers. We propose a model fitness evaluation method to determine whether a model is sufficient to distinguish positive and negative instances while still competent to provide satisfactory effectiveness with a small feature subset. Our experiments demonstrated how the fitness of models are assessed. The results of our work contribute to the researches of feature selection, dimensionality reduction and document classification.

A reference architecture for instructional educational software

Our extensive research has indicated that high-school teachers are reluctant to make use of existing instructional educational software (Pollard, 2005). Even software developed in a partnership between a teacher and a software engineer is unlikely to be adopted by teachers outside the partnership (Pollard, 2005). In this paper we address these issues directly by adopting a reusable architectural design for instructional educational software which allows easy customisation of software to meet the specific needs of individual teachers. By doing this we will facilitate more teachers regularly using instructional technology within their classrooms. Our domain-specific software architecture, Interface-Activities-Model, was designed specifically to facilitate individual customisation by redefining and restructuring what constitutes an object so that they can be readily reused or extended as required. The key to this architecture is the way in which the software is broken into small generic encapsulated components with minimal domain specific behaviour. The domain specific behaviour is decoupled from the interface and encapsulated in objects which relate to the instructional material through tasks and activities. The domain model is also broken into two distinct models - Application State Model and Domainspecific Data Model. This decoupling and distribution of control gives the software designer enormous flexibility in modifying components without affecting other sections of the design. This paper sets the context of this architecture, describes it in detail, and applies it to an actual application developed to teach high-school mathematical concepts.

Background and foreground knowledge in dynamic ontology construction

Ontologies have become a key component in the Semantic Web and Knowledge management. One accepted goal is to construct ontologies from a domain specific set of texts. An ontology reflects the background knowledge used in writing and reading a text. However, a text is an act of knowledge maintenance, in that it re-enforces the background assumptions, alters links and associations in the ontology, and adds new concepts. This means that background knowledge is rarely expressed in a machine interpretable manner. When it is, it is usually in the conceptual boundaries of the domain, e.g. in textbooks or when ideas are borrowed into other domains. We argue that a partial solution to this lies in searching external resources such as specialized glossaries and the internet. We show that a random selection of concept pairs from the Gene Ontology do not occur in a relevant corpus of texts from the journal Nature. In contrast, a significant proportion can be found on the internet. Thus, we conclude that sources external to the domain corpus are necessary for the automatic construction of ontologies.

Ontologies, taxonomies, thesauri:learning from texts

The use of ontologies as representations of knowledge is widespread but their construction, until recently, has been entirely manual. We argue in this paper for the use of text corpora and automated natural language processing methods for the construction of ontologies. We delineate the challenges and present criteria for the selection of appropriate methods. We distinguish three ma jor steps in ontology building: associating terms, constructing hierarchies and labelling relations. A number of methods are presented for these purposes but we conclude that the issue of data-sparsity still is a ma jor challenge. We argue for the use of resources external tot he domain specific corpus.

miniDVMS v1.8 : A user manual(The data visualization and modeling system)

Today, the data available to tackle many scientific challenges is vast in quantity and diverse in nature. The exploration of heterogeneous information spaces requires suitable mining algorithms as well as effective visual interfaces. miniDVMS v1.8 provides a flexible visual data mining framework which combines advanced projection algorithms developed in the machine learning domain and visual techniques developed in the information visualisation domain. The advantage of this interface is that the user is directly involved in the data mining process. Principled projection methods, such as generative topographic mapping (GTM) and hierarchical GTM (HGTM), are integrated with powerful visual techniques, such as magnification factors, directional curvatures, parallel coordinates, and user interaction facilities, to provide this integrated visual data mining framework. The software also supports conventional visualisation techniques such as principal component analysis (PCA), Neuroscale, and PhiVis. This user manual gives an overview of the purpose of the software tool, highlights some of the issues to be taken care while creating a new model, and provides information about how to install and use the tool. The user manual does not require the readers to have familiarity with the algorithms it implements. Basic computing skills are enough to operate the software.

Multiple sclerosis: natural history, diagnosis and changes in the brain

This article describes the general symptoms, diagnosis and changes in the nervous system in multiple sclerosis. A second article will describe the specific visual symptoms which are believed to be characteristic of the disease.

Probabilistic methods for drug discovery

This thesis introduces a flexible visual data exploration framework which combines advanced projection algorithms from the machine learning domain with visual representation techniques developed in the information visualisation domain to help a user to explore and understand effectively large multi-dimensional datasets. The advantage of such a framework to other techniques currently available to the domain experts is that the user is directly involved in the data mining process and advanced machine learning algorithms are employed for better projection. A hierarchical visualisation model guided by a domain expert allows them to obtain an informed segmentation of the input space. Two other components of this thesis exploit properties of these principled probabilistic projection algorithms to develop a guided mixture of local experts algorithm which provides robust prediction and a model to estimate feature saliency simultaneously with the training of a projection algorithm.Local models are useful since a single global model cannot capture the full variability of a heterogeneous data space such as the chemical space. Probabilistic hierarchical visualisation techniques provide an effective soft segmentation of an input space by a visualisation hierarchy whose leaf nodes represent different regions of the input space. We use this soft segmentation to develop a guided mixture of local experts (GME) algorithm which is appropriate for the heterogeneous datasets found in chemoinformatics problems. Moreover, in this approach the domain experts are more involved in the model development process which is suitable for an intuition and domain knowledge driven task such as drug discovery. We also derive a generative topographic mapping (GTM) based data visualisation approach which estimates feature saliency simultaneously with the training of a visualisation model.

Investigating the effects of antiepileptic drugs on the electrophysiology of vision

The principal aim of this work was to examine the effects of antiepileptic drugs (AEDs) on vision. Vigabatrin acts by increasing GABA at brain inhibitory synapses by irreversibly binding to GABA-transaminase. Remacemide is a novel non-competitive NMDA receptor antagonist and fast sodium channel inhibitor that results in the inhibition of the NMDA receptors located in the neuronal membrane calcium channels increasing glutamate in the brain. Vigabatrin has been shown to cause a specific pattern of visual field loss, as one in three adults taking vigabatrin have shown a bilateral concentric constriction. Remacemide has unknown effects on vision. The majority of studies of the effects of AEDs on vision have not included the paediatric population due to difficulties assessing visual field function using standard perimetry testing. Evidently an alternative test is required to establish and monitor visual field problems associated with AEDs both in children and in adults who cannot comply with perimetry. In order to test paediatric patients exposed to vigabatrin, a field-specific visual evoked potential was developed. Other tests performed on patients taking either vigabatrin or remacemide were electroretinograms, electro-oculograms, multifocal VEPs and perimetry. Comparing these tests to perimetry results from vigabatrin patients the field specific VEP was found to have a high sensitivity and specificity, as did the 30Hz flicker amplitude. The modified VEP was also found to provide useful results in vigabatrin patients. Remacemide did not produce a similar visual field loss to vigabatrin although macular vision was affected. The field specific VEP is a useful method for detecting vigabatrin associated visual field loss that is well tolerated by young children. This technique combined with the ERG under light adapted (30Hz flicker) condition is presently the superior method for detecting vigabatrin-attributed peripheral field defects present in children below the developmental age of 9. The effects of AEDs on vision should be monitored carefully and the use of multifocal stimulation allows for specific areas of the retina and visual pathway to be monitored.

Self-training from labeled features for sentiment analysis

Sentiment analysis concerns about automatically identifying sentiment or opinion expressed in a given piece of text. Most prior work either use prior lexical knowledge defined as sentiment polarity of words or view the task as a text classification problem and rely on labeled corpora to train a sentiment classifier. While lexicon-based approaches do not adapt well to different domains, corpus-based approaches require expensive manual annotation effort. In this paper, we propose a novel framework where an initial classifier is learned by incorporating prior information extracted from an existing sentiment lexicon with preferences on expectations of sentiment labels of those lexicon words being expressed using generalized expectation criteria. Documents classified with high confidence are then used as pseudo-labeled examples for automatical domain-specific feature acquisition. The word-class distributions of such self-learned features are estimated from the pseudo-labeled examples and are used to train another classifier by constraining the model's predictions on unlabeled instances. Experiments on both the movie-review data and the multi-domain sentiment dataset show that our approach attains comparable or better performance than existing weakly-supervised sentiment classification methods despite using no labeled documents.

Learning sentiment classification model from labeled features

We propose a novel framework where an initial classifier is learned by incorporating prior information extracted from an existing sentiment lexicon. Preferences on expectations of sentiment labels of those lexicon words are expressed using generalized expectation criteria. Documents classified with high confidence are then used as pseudo-labeled examples for automatical domain-specific feature acquisition. The word-class distributions of such self-learned features are estimated from the pseudo-labeled examples and are used to train another classifier by constraining the model's predictions on unlabeled instances. Experiments on both the movie review data and the multi-domain sentiment dataset show that our approach attains comparable or better performance than exiting weakly-supervised sentiment classification methods despite using no labeled documents.

Case based design of knitwear

In the developed world we are surrounded by man-made objects, but most people give little thought to the complex processes needed for their design. The design of hand knitting is complex because much of the domain knowledge is tacit. The objective of this thesis is to devise a methodology to help designers to work within design constraints, whilst facilitating creativity. A hybrid solution including computer aided design (CAD) and case based reasoning (CBR) is proposed. The CAD system creates designs using domain-specific rules and these designs are employed for initial seeding of the case base and the management of constraints. CBR reuses the designer's previous experience. The key aspects in the CBR system are measuring the similarity of cases and adapting past solutions to the current problem. Similarity is measured by asking the user to rank the importance of features; the ranks are then used to calculate weights for an algorithm which compares the specifications of designs. A novel adaptation operator called rule difference replay (RDR) is created. When the specifications to a new design is presented, the CAD program uses it to construct a design constituting an approximate solution. The most similar design from the case-base is then retrieved and RDR replays the changes previously made to the retrieved design on the new solution. A measure of solution similarity that can validate subjective success scores is created. Specification similarity can be used as a guide whether to invoke CBR, in a hybrid CAD-CBR system. If the newly resulted design is suffciently similar to a previous design, then CBR is invoked; otherwise CAD is used. The application of RDR to knitwear design has demonstrated the flexibility to overcome deficiencies in rules that try to automate creativity, and has the potential to be applied to other domains such as interior design.