Iterative NLP Query Refinement for Enhancing Domain-Specific Information Retrieval: A Case Study in Career Services

The exploding number of computational models produced by Systems Biologists over the last years is an invitation to structure and exploit this new wealth of information. Researchers would like to trace models relevant to specific scientific questions, to explore their biological content, to align and combine them, and to match them with experimental data. To automate these processes, it is essential to consider semantic annotations, which describe their biological meaning. As a prerequisite for a wide range of computational methods, we propose general and flexible similarity measures for Systems Biology models computed from semantic annotations. By using these measures and a large extensible ontology, we implement a platform that can retrieve, cluster, and align Systems Biology models and experimental data sets. At present, its major application is the search for relevant models in the BioModels Database, starting from initial models, data sets, or lists of biological concepts. Beyond similarity searches, the representation of models by semantic feature vectors may pave the way for visualisation, exploration, and statistical analysis of large collections of models and corresponding data.

Exploiting the category structure of Wikipedia for entity ranking

To overcome the semantic gap in the retrieval process, the user relevance feedback frame work is added to the current image retrieval system. Relevance feedback method iteratively refines and updates the retrieved result by learning the user-labeled examples to further improves the overall performance. In the proposed scheme feature descriptor derived from the EDBTC compressed data stream. Firstly, a color image is decomposed using EDBTC scheme to produce two new image representations, namely color quantizer and bitmap image by using Improved LBG-VQ (ILBG)algorithm. Two image feature descriptors called Color Histogram Feature (CHF), and Bit Pattern Feature(BPF) can be subsequently generated from the EDBTC color quantizer and its corresponding bitmap image respectively without performing the decoding process. The similarity degree between two images is simply measured with the similarity distance score of their feature descriptor. Second, we presented a relevance feedback (RF) framework for effective image retrieval by using a support vector machine (SVM). Effectiveness of EDBTC feature descriptor is quantitatively examined and compared in the RGB color space as well as in HSI color channel. Extensive experiments shows that there is a significant improvement in retrieval performance with the proposed method using SVMRF compared with the retrieval performance without RF.

Notwithstanding recent advances in machine vision, video activity recognition from multiple cameras still remains a challenging task as many real-world interactions cannot be automatically recognised for many reasons, such as partial occlusion or coverage black-spots. In this paper we propose a new technique that infers the unseen relationship between two individuals captured by different cameras and use it to retrieve relevant video clips if there is a likely interaction between the two individuals. We introduce a human object interaction (HOI) model integrating the causal relationship between the humans and the objects. For this we first extract the key frames and generate the labels or annotations using the state-of-the-art image captioning models. Next, we extract SVO (subject, verb, object) triples and encode the descriptions into a vector form for HOI inference using the Stanford CoreNLP parser. In order to calculate the HOI co-existence and the possible causality score we use transfer entropy. From our experimentation, we found that integrating casual relations into the content indexing process and using transfer entropy to calculate the causality score leads to improvement in retrieval performance.

Bayesian belief networks for IR

We will present a new algorithm for improving the retrieval performance using query expansion, based on a hierarchy of clusters. In order to create this hierarchical data structure, a clustering algorithm is executed multiple times with different initial conditions. With the aid of this hierarchical data structure, we have achieved significant improvement in retrieval performance over previously known methods in terms of both recall and precision. In our experiments with Japanese patent data, we have employed a co-clustering algorithm as a clustering method.

Query specific graph-based query reformulation using UMLS for clinical information access.

We present an approach for robust shape retrieval from data-bases containing articulated 3D shapes. We represent each shape by the eigenvectors of an appropriately defined affinity matrix, obtaining a spectral embedding. Retrieval is then performed on these embeddings using global shape descriptors. Transformation into the spectral domain normalizes the shapes against articulation (bending), rigid-body transformations, and uniform scaling. Experimentally, we show absolute improvement in retrieval performance when conventional shape descriptors are used in the spectral domain on the McGill database of articulated 3D shapes. We also propose a simple eigenvalue-based descriptor, which is easily computed and performs comparably against the best known shape descriptors applied to the original shapes.

Pseudo-Relevance Feedback (PRF) assumes that the top-ranking n documents of the initial retrieval are relevant and extracts expansion terms from them. In this work, we introduce the notion of pseudo-irrelevant documents, i.e. high-scoring documents outside of top n that are highly unlikely to be relevant. We show how pseudo-irrelevant documents can be used to extract better expansion terms from the top-ranking n documents: good expansion terms are those which discriminate the top-ranking n documents from the pseudo-irrelevant documents. Our approach gives substantial improvements in retrieval performance over Model-based Feedback on several test collections.

Semantic approaches present an efficient, detailed and easily understandable representation of knowledge from documents. Al-Quran contains a vast amount of knowledge that needs appropriate knowledge extraction. A semantic based approach can help in designing an efficient and explainable knowledge representation model for Al-Quran. This research aims to propose a semantic-graph knowledge representation model for verses of Al-Quran based on word dependencies. These features are used in the proposed knowledge representation model allowing the semantic graph matching to improve Al-Quran search applications' accuracy. The proposed knowledge representation model is essentially a formalism for generating a semantic graph representation of Quranic verses, which can be applied for knowledge base construction for other applications such as information retrieval system. A set of rules called Semantic Dependency Triple Rules are defined to be mapped into the semantic graph representing the verse's logic. The rules translate word dependencies and other NLP metadata into a triple form that holds logical information. The proposed model has been tested with English translation of Al-Quran on a document retrieval prototype The basic system has been enhanced with anaphoric pronouns correction, which has shown improvement in retrieval performance. The results have been compared with a closely related system and evaluated on the accuracy of the document retrieval in Precision, Recall and F-score measurements. The proposed model has achieved 65%, 60% and 62.4% for the measurements, respectively. It has also improved the overall accuracy of previous system by 43.8%.

The Bag-of-visual-Words (BovW) model is one of the most popular visual content representation methods for large-scale content-based video retrieval. The visual words are quantized according to a visual vocabulary, which is generated by a visual features clustering process (e.g. K-means, GMM, etc). In principle, two types of errors can occur in the quantization process. They are referred to as the UnderQuantize and OverQuantize problems. The former causes ambiguities and often leads to false visual content matches, while the latter generates synonyms and may lead to missing true matches. Unlike most state-of-the-art research that concentrated on enhancing the BovW model by disambiguating the visual words, in this paper, we aim to address the OverQuantize problem by incorporating the similarity of spatial-temporal contexts associated to pair-wise visual words. The visual words with similar context and appearance are assumed to be synonyms. These synonyms in the initial visual vocabulary are then merged to rebuild a more compact and descriptive vocabulary. Our approach was evaluated on the TRECVID2002 and CC_WEB_VIDEO datasets for two typical Query-By-Example (QBE) video retrieval applications. Experimental results demonstrated substantial improvements in retrieval performance over the initial visual vocabulary generated by the BovW model. We also show that our approach can be utilized in combination with the state-of-the-art disambiguation method to further improve the performance of the QBE video retrieval.

Stemming is used to produce stem or root of words. The process is vital to different research fields such as text mining, sentiment analysis, and text categorization, etc. Several techniques have been proposed to stemming Arabic text and among them, Khoja and light-10 stemmers are the most widely used. In this paper, we propose and evaluate two different stemming techniques to Arabic that are based on light stemming techniques. The new stemmers are compared to best reported light stemmer, which is light-10. Results and experiments, which were conducted using standard collections, reveal that The proposed stemmers yield 5.13% and 13.1% improvement in retrieval performance over light 10 with 0.369 average precision and 0.397, respectively and the improvement is statistically significant.

Because of the high impact of high-tech digital crime upon our society, it is necessary to develop effective Information Retrieval (IR) tools to support digital forensic investigations. In this paper, we propose an IR system for digital forensics that targets emails. Our system incorporates WordNet (i.e. a domain independent ontology for the vocabulary) into an Extended Boolean Model (EBM) by applying query expansion techniques. Structured Boolean queries in Backus-Naur Form (BNF) are utilized to assist investigators in effectively expressing their information requirements. We compare the performance of our system on several email datasets with a traditional Boolean IR system built upon the Lucene keyword-only model. Experimental results show that our system yields a promising improvement in retrieval performance without the requirement of very accurate query keywords to retrieve the most relevant emails.

State-of-the-art content sharing platforms often require users to assign tags to pieces of media in order to make them easily retrievable. Since this task is sometimes perceived as tedious or boring, annotations can be sparse. Commenting on the other hand is a frequently used means of expressing user opinion towards shared media items. This work makes use of time series analyses in order to infer potential tags and indexing terms for audio-visual content from user comments. In this way, we mitigate the vocabulary gap between queries and document descriptors. Additionally, we show how large-scale encyclopaedias such as Wikipedia can aid the task of tag prediction by serving as surrogates for high-coverage natural language vocabulary lists. Our evaluation is conducted on a corpus of several million real-world user comments from the popular video sharing platform YouTube, and demonstrates significant improvements in retrieval performance.

Video producers, in telling a news story, tend to repeat important visual and speech material multiple times in adjacent shots, thus creating a certain level of redundancy. We describe this phenomenon, and use it to develop a framework to incorporate redundancy for cross-channel retrieval of visual items using speech. Testing our models in a series of retrieval experiments, we find that incorporating the fact that information occurs redundantly into cross-channel retrieval leads to significant improvements in retrieval performance.