TCN: Table Convolutional Network for Web Table Interpretation

Relational Web tables, embedded in HTML or published on data platforms, have become an important resource for many applications, including question answering or entity augmentation. To utilize the data, we require some understanding of what the tables are about. Previous research on recovering Web table semantics has largely focused on simple tables, which only describe a single semantic concept. However, there is also a significant number of de-normalized multi-concept tables on the Web. Treating these as single-concept tables results in many incorrect relations being extracted. In this paper, we propose a normalization approach to decompose multi-concept tables into smaller single-concept tables. First, we identify columns that represent keys or identifiers of entities. Then, we utilize the table schema as well as intrinsic data correlations to identify concept boundaries and split the tables accordingly. Experimental results on real Web tables show that our approach is feasible and effectively identifies semantic concepts.

We present a method based on header paths for efficient and complete extraction of labeled data from tables meant for humans. Although many table configurations yield to the proposed syntactic analysis, some require access to semantic knowledge. Clicking on one or two critical cells per table, through a simple interface, is sufficient to resolve most of these problem tables. Header paths, a purely syntactic representation of visual tables, can be transformed ("factored") into existing representations of structured data such as category trees, relational tables, and RDF triples. From a random sample of 200 web tables from ten large statistical web sites, we generated 376 relational tables and 34,110 subject-predicate-object RDF triples.

The growing amount of semi-structured and unstructured data on tourism Web sites with heterogeneous designs requires information extraction (IE) mechanisms, to create, for instance, tourism portals. In order to build semantic eTourism environments, the acquisition of room prices is of particular interest. Room prices and related information often appear in tabular structures, which still challenge Web information extraction techniques. In this paper, we begin by identifying various price table patterns which are characterized by the position of a number of features that determine a room price. We then describe an extended ontology model for tourism prices. Finally, we present TAINEX, a plug-in for functional and structural analysis and data interpretation of price tables, which extends the existing prototype TourIE, a rule/ontology-based information extraction system for Web sites with heterogeneous designs.

Several processes have been advanced in the literature for reverse engineering of relational databases. The inputs to these processes are relational tables and available contextual information. The output is a model of the underlying logical intent, apart from the implementation artifacts. Most of the existing processes for database reverse engineering are inadequate; they assume too high a quality of input information. The authors of these processes are skilled database designers and they are overly optimistic about the state-of-the-art, as practiced. This paper catalogs odd aspects of relational database designs that we have encountered over the past several years. Many of these database designs are from commercial software products.

We consider the problem of extracting structured records from semi-structured web pages with no human supervision required for each target web site. Previous work on this problem has either required significant human effort for each target site or used brittle heuristics to identify semantic data types. Our method only requires annotation for a few pages from a few sites in the target domain. Thus, after a tiny investment of human effort, our method allows automatic extraction from potentially thousands of other sites within the same domain. Our approach extends previous methods for detecting data fields in semi-structured web pages by matching those fields to domain schema columns using robust models of data values and contexts. Annotating 2---5 pages for 4---6 web sites yields an extraction accuracy of 83.8% on job offer sites and 91.1% on vacation rental sites. These results significantly outperform a baseline approach.

Traditionally, information extraction from web tables has focused on small, more or less homogeneous corpora, often based on assumptions about the use of tags. A multitude of different HTML implementations of web tables make these approaches difficult to scale. In this paper, we approach the problem of domain-independent information extraction from web tables by shifting our attention from the tree-based representation of webpages to a variation of the two-dimensional visual box model used by web browsers to display the information on the screen. The there by obtained topological and style information allows us to fill the gap created by missing domain-specific knowledge about content and table templates. We believe that, in a future step, this approach can become the basis for a new way of large-scale knowledge acquisition from the current Visual Web.

Automatic information extraction from semi-structured Web pages by pattern discovery

To extract information automatically from semi-structured web pages, this paper puts forward a method named IESS for discovering the record model based on DOM and Maximal Similar Sub Tree, to identify records automatically and correctly when there are some differences in expression models of records that belong to the same type. Furthermore, the system can extract information from result pages of paper searching websites automatically. The experiments made through with some common paper searching websites have demonstrated that this system has high efficiency and accuracy.

To extract information automatically from semi-structured web pages, this paper puts forward a method named IESS for discovering the record model based on DOM and Maximal Similar Sub Tree, to identify records automatically and correctly when there are some differences in expression models of records that belong to the same type. To test the performance of the method, a scientific literature statistical analysis system is designed. The practice shows that users can quickly understand the distribution of papers in their retrieving field and grasp the importance with the help of the system.

Business information extraction from semi-structured webpages

This paper proposes an adaptive bag-of-phrases (BoP) algorithm for mobile scene recognition based on bag-of-words approach. Conventional BoW methods do not consider the dependence and pairwise relationship among different codewords. However, these contextual relations between pairwise codewords play an important role for users to recognize an image. In light of this problem, this paper proposes an effective BoP technique to integrate both the spatial and contextual information between visual words for scene recognition. It first uses hierarchical k-means algorithm to construct a universal codebook for all categories. The contextual (dependence) relationship between pairwise words is then mined for each category based on the mutual information they contain. Subsequently, a visual phrase vocabulary is constructed which is then used to generate a BoP histogram through a proposed quantization method. Finally, support vector machine (SVM) is used to train these histograms into a classifier. Experimental results on the Scene 15 dataset show that the proposed method is effective for mobile scene recognition.

This study investigates how subordinate-supervisor knowledge dissimilarity (relational demography perspective) and unit-level knowledge diversity, as resulted from structural change, influence individual job satisfaction after the change. We focus on two dimensions of individual knowledge, namely educational background (explicit knowledge) and organizational tenure (tacit knowledge). We develop hypotheses about the negative impact of dyadic knowledge dissimilarity and positive impact of unit knowledge diversity on change satisfaction. Moreover, we advance the hypothesis that the implementation lag, i.e. the time elapsed from the structural change, moderates the positive relationship between unit knowledge diversity and change satisfaction. Data were collected through mixed methods. Quantitative analyses document that individual job satisfaction after the change is negatively affected by functional background dissimilarity between subordinates and supervisors, and that the positive effect of tenure diversity at the unit level on change satisfaction is stronger when the implementation lag is longer. Our qualitative findings illuminate on how units with higher diversity develop new, high-quality routines and processes over time. Contributions to theory and practice are discussed.

In Federated Learning (FL), data communication among clients is denied. However, it is difficult to learn from the decentralized client data, which is under-sampled, especially for segmentation tasks that need to extract enough contextual semantic information. Existing FL studies always average client models to one global model in segmentation tasks while neglecting the diverse knowledge extracted by the models. To maintain and utilize the diverse knowledge, we propose a novel training paradigm called Federated Learning with Z-average and Cross-teaching (FedZaCt) to deal with segmentation tasks. From the model parameters’ aspect, the Z-average method constructs individual client models, which maintain diverse knowledge from multiple client data. From the model distillation aspect, the Cross-teaching method transfers the other client models’ knowledge to supervise the local client model. In particular, FedZaCt does not have the global model during the training process. After training, all client models are aggregated into the global model by averaging all client model parameters. The proposed methods are applied to two medical image segmentation datasets including our private aortic dataset and a public HAM10000 dataset. Experimental results demonstrate that our methods can achieve higher Intersection over Union values and Dice scores.

In the fast-paced world of product design, businesses seek a competitive edge by swiftly addressing user requirements and developing precise solutions. Depending on diverse knowledge, conceptual design makes enterprises invest significantly in knowledge management. Following the philosophy of decision-based design, a knowledge graph-aided decision guidance method is proposed to streamline and enhance knowledge utilisation in product conceptual design. Firstly, the decision-making process in product conceptual design is modelled using the Concept-Decision-Knowledge (CDK) model, yielding the meta-model of CDK (mCDK). A data-augmented BERT-BiLSTM-CRF model is adopted to extract information from diverse data sources, forming a CDK knowledge graph (CDK-KG). A decision case for employing problem-solving knowledge configuration is generated to resolve specific design decision problems when new requirements arise. Validation of the method is demonstrated through a case study on the launch vehicle conceptual design of the first and second-stage separation system. The results indicate that the proposed method can automatically extract information from data resources to construct a knowledge graph. Furthermore, it can provide decision cases for design requirements through knowledge configuration, supporting decision-making.

HA-GCEN: Hyperedge-abundant graph convolutional enhanced network for hate speech detection