ICICLE: A semantic-based retrieval system for WWW images

In this paper, we present ICICLE (Image ChainNet and Incremental Clustering Engine), a prototype system that we have developed to efficiently and effectively retrieve WWW images based on image semantics. ICICLE has two distinguishing features. First, it employs a novel image representation model called Weight ChainNet to capture the semantics of the image content. A new formula, called list space model, for computing semantic similarities is also introduced. Second, to speed up retrieval, ICICLE employs an incremental clustering mechanism, ICC (Incremental Clustering on ChainNet), to cluster images with similar semantics into the same partition. Each cluster has a summary representative and all clusters' representatives are further summarized into a balanced and full binary tree structure. We conducted an extensive performance study to evaluate ICICLE. Compared with some recently proposed methods, our results show that ICICLE provides better recall and precision. Our clustering technique ICC facilitates speedy retrieval of images without sacrificing recall and precision significantly.

Evaluation of unsupervised semantic mapping of natural language with Leximancer concept mapping

The Leximancer system is a relatively new method for transforming lexical co-occurrence information from natural language into semantic patterns in an unsupervised manner. It employs two stages of co-occurrence information extraction-semantic and relational-using a different algorithm for each stage. The algorithms used are statistical, but they employ nonlinear dynamics and machine learning. This article is an attempt to validate the output of Leximancer, using a set of evaluation criteria taken from content analysis that are appropriate for knowledge discovery tasks.

Resolution and characterisation of multiple isoforms of bovine kappa-casein by 2-DE following a reversible cysteine-tagging enrichment strategy

Visualisation of multiple isoforms of kappa-casein on 2-D gels is restricted by the abundant alpha- and beta-caseins that not only limit gel loading but also migrate to similar regions as the more acidic kappa-casein isoforms. To overcome this problem, we took advantage of the absence of cysteine residues in alpha(S1)- and beta-casein by devising an affinity enrichment procedure based on reversible biotinylation of cysteine residues. Affinity capture of cysteine-containing proteins on avidin allowed the removal of the vast majority of alpha(S1)- and beta-casein, and on subsequent 2-D gel analysis 16 gel spots were identified as kappa-casein by PMF. Further analysis of the C-terminal tryptic peptide along with structural predictions based on mobility on the 2-D gel allowed us to assign identities to each spot in terms of genetic variant (A or B), phosphorylation status (1, 2 or 3) and glycosylation status (from 0 to 6). Eight isoforms of the A and B variants with the same PTMs were observed. When the casein fraction of milk from a single cow, homozygous for the B variant of kappa-casein, was used as the starting material, 17 isoforms from 13 gel spots were characterised. Analysis of isoforms of low abundance proved challenging due to the low amount of material that could be extracted from the gels as well as the lability of the PTMs during MS analysis. However, we were able to identify a previously unrecognised site, T-166, that could be phosphorylated or glycosylated. Despite many decades of analysis of milk proteins, the reasons for this high level of heterogeneity are still not clear.

The influence of dopamine on semantic activation in Parkinson's disease: Evidence from a multipriming task

Research has suggested that semantic processing deficits in Parkinson's disease (PD) are related to striatal dopamine deficiency. As an investigation of the influence of dopamine on semantic activation in PD, 7 participants with PD performed a lexical-decision task when on and off levodopa medication. Seven healthy controls matched to the participants with PD in terms of sex, age, and education also participated in the study. By use of a multipriming paradigm, whereby 2 prime words were presented prior to the target word, semantic priming effects were measured across stimulus onset asynchronies (SOAs) of 250 Ins and 1,200 Ins. The results revealed a similar pattern of priming across SOAs for the control group and the PD participants on medication. In contrast, within-group comparisons revealed that automatic semantic activation was compromised in PD participants when off medication. The implications of these results for the neuromodulatory influence of dopamine on semantic processing in PD are discussed.

Classic identity negative priming involves accessing semantic representations in the left anterior temporal cortex

Classic identity negative priming (NP) refers to the finding that when an object is ignored, subsequent naming responses to it are slower than when it has not been previously ignored (Tipper, S.P., 1985. The negative priming effect: inhibitory priming by ignored objects. Q. J. Exp. Psychol. 37A, 571-590). It is unclear whether this phenomenon arises due to the involvement of abstract semantic representations that the ignored object accesses automatically. Contemporary connectionist models propose a key role for the anterior temporal cortex in the representation of abstract semantic knowledge (e.g., McClelland, J.L., Rogers, T.T., 2003. The parallel distributed processing approach to semantic cognition. Nat. Rev. Neurosci. 4, 310-322), suggesting that this region should be involved during performance of the classic identity NP task if it involves semantic access. Using high-field (4 T) event-related functional magnetic resonance imaging, we observed increased BOLD responses in the left anterolateral temporal cortex including the temporal pole that was directly related to the magnitude of each individual's NP effect, supporting a semantic locus. Additional signal increases were observed in the supplementary eye fields (SEF) and left inferior parietal lobule (IPL). (c) 2006 Elsevier Inc. All rights reserved.

A Correlation Analysis on LSA and HAL Semantic Space Models

In this paper, we compare a well-known semantic spacemodel, Latent Semantic Analysis (LSA) with another model, Hyperspace Analogue to Language (HAL) which is widely used in different area, especially in automatic query refinement. We conduct this comparative analysis to prove our hypothesis that with respect to ability of extracting the lexical information from a corpus of text, LSA is quite similar to HAL. We regard HAL and LSA as black boxes. Through a Pearsonrsquos correlation analysis to the outputs of these two black boxes, we conclude that LSA highly co-relates with HAL and thus there is a justification that LSA and HAL can potentially play a similar role in the area of facilitating automatic query refinement. This paper evaluates LSA in a new application area and contributes an effective way to compare different semantic space models.

Semantic caching for multiresolution spatial query processing in mobile environments

Spatial data are particularly useful in mobile environments. However, due to the low bandwidth of most wireless networks, developing large spatial database applications becomes a challenging process. In this paper, we provide the first attempt to combine two important techniques, multiresolution spatial data structure and semantic caching, towards efficient spatial query processing in mobile environments. Based on the study of the characteristics of multiresolution spatial data (MSD) and multiresolution spatial query, we propose a new semantic caching model called Multiresolution Semantic Caching (MSC) for caching MSD in mobile environments. MSC enriches the traditional three-category query processing in semantic cache to five categories, thus improving the performance in three ways: 1) a reduction in the amount and complexity of the remainder queries; 2) the redundant transmission of spatial data already residing in a cache is avoided; 3) a provision for satisfactory answers before 100% query results have been transmitted to the client side. Our extensive experiments on a very large and complex real spatial database show that MSC outperforms the traditional semantic caching models significantly

Semantic caching for web-based spatial applications

Client-side caching of spatial data is an important yet very much under investigated issue. Effective caching of vector spatial data has the potential to greatly improve the performance of spatial applications in the Web and wireless environments. In this paper, we study the problem of semantic spatial caching, focusing on effective organization of spatial data and spatial query trimming to take advantage of cached data. Semantic caching for spatial data is a much more complex problem than semantic caching for aspatial data. Several novel ideas are proposed in this paper for spatial applications. A number of typical spatial application scenarios are used to generate spatial query sequences. An extensive experimental performance study is conducted based on these scenarios using real spatial data. We demonstrate a significant performance improvement using our ideas.