Marie-Jean Meurs

Semantic text mining support for lignocellulose research

BackgroundBiofuels produced from biomass are considered to be promising sustainable alternatives to fossil fuels. The conversion of lignocellulose into fermentable sugars for biofuels production requires the use of enzyme cocktails that can efficiently and economically hydrolyze lignocellulosic biomass. As many fungi naturally break down lignocellulose, the identification and characterization of the enzymes involved is a key challenge in the research and development of biomass-derived products and fuels. One approach to meeting this challenge is to mine the rapidly-expanding repertoire of microbial genomes for enzymes with the appropriate catalytic properties.ResultsSemantic technologies, including natural language processing, ontologies, semantic Web services and Web-based collaboration tools, promise to support users in handling complex data, thereby facilitating knowledge-intensive tasks. An ongoing challenge is to select the appropriate technologies and combine them in a coherent system that brings measurable improvements to the users. We present our ongoing development of a semantic infrastructure in support of genomics-based lignocellulose research. Part of this effort is the automated curation of knowledge from information on fungal enzymes that is available in the literature and genome resources.ConclusionsWorking closely with fungal biology researchers who manually curate the existing literature, we developed ontological natural language processing pipelines integrated in a Web-based interface to assist them in two main tasks: mining the literature for relevant knowledge, and at the same time providing rich and semantically linked information.

An approach to controlling user models and personalization effects in recommender systems

Personalization nowadays is a commodity in a broad spectrum of computer systems. Examples range from online shops recommending products identified based on the users previous purchases to web search engines sorting search hits based on the user browsing history. The aim of such adaptive behavior is to help users to find relevant content easier and faster. However, there are a number of negative aspects of this behavior. Adaptive systems have been criticized for violating the usability principles of direct manipulation systems, namely controllability, predictability, transparency, and unobtrusiveness. In this paper, we propose an approach to controlling adaptive behavior in recommender systems. It allows users to get an overview of personalization effects, view the user profile that is used for personalization, and adjust the profile and personalization effects to their needs and preferences. We present this approach using an example of a personalized portal for biochemical literature, whose users are biochemists, biologists and genomicists. Also, we report on a user study evaluating the impacts of controllable personalization on the usefulness, usability, user satisfaction, transparency, and trustworthiness of personalized systems.

mycoCLAP, the database for characterized lignocellulose-active proteins of fungal origin: resource and text mining curation support

Enzymes active on components of lignocellulosic biomass are used for industrial applications ranging from food processing to biofuels production. These include a diverse array of glycoside hydrolases, carbohydrate esterases, polysaccharide lyases and oxidoreductases. Fungi are prolific producers of these enzymes, spurring fungal genome sequencing efforts to identify and catalogue the genes that encode them. To facilitate the functional annotation of these genes, biochemical data on over 800 fungal lignocellulose-degrading enzymes have been collected from the literature and organized into the searchable database, mycoCLAP (http://mycoclap.fungalgenomics.ca). First implemented in 2011, and updated as described here, mycoCLAP is capable of ranking search results according to closest biochemically characterized homologues: this improves the quality of the annotation, and significantly decreases the time required to annotate novel sequences. The database is freely available to the scientific community, as are the open source applications based on natural language processing developed to support the manual curation of mycoCLAP. Database URL: http://mycoclap.fungalgenomics.ca

Graph decomposition approaches for terminology graphs

We propose a graph-based decomposition methodology of a network of document features represented by a terminology graph. The graph is automatically extracted from raw data based on Natural Language Processing techniques implemented in the TermWatch system. These graphs are Small Worlds. Based on clique minimal separators and the associated graph of atoms: a subgraph without clique separator, we show that the terminology graph can be divided into a central kernel which is a single atom and a periphery made of small atoms. Moreover, the central kernel can be separated based on small optimal minimal separators.

Natural Language Processing for Semantic Assistance in Web Portals

Web portals are a major class of web-based content management systems. They can provide users with a single point of access to a multitude of content sources and applications. However, further analysis of content brokered through a portal is not supported by current portal systems, leaving it to their users to deal with information overload. We present the first work examining the integration of natural language processing into web portals to provide users with semantic assistance in analyzing and interpreting content. This integration is based on the portal standard JSR286 and open source NLP frameworks. Two application scenarios, news analysis and biocuration, highlight the feasibility and usefulness of our approach.

Data Sampling and Supervised Learning for HIV Literature Screening

This paper presents a supervised learning approach to support the screening of HIV literature. The manual screening of biomedical literature is an important task in the process of systematic reviews. Researchers and curators have the very demanding, time-consuming, and error-prone task of manually identifying documents that should be included in a systematic review concerning a specific problem. We developed a supervised learning approach to support screening tasks, by automatically flagging potentially relevant documents from a list retrieved by a literature database search. To overcome the main issues associated with the automatic literature screening task, we evaluated the use of data sampling, feature combinations, and feature selection methods, generating a total of 105 classification models. The models yielding the best results were composed of a Logistic Model Trees classifier, a fairly balanced training set, and feature combination of Bag-Of-Words and MeSH terms. According to our results, the system correctly labels the great majority of relevant documents, making it usable to support HIV systematic reviews to allow researchers to assess a greater number of documents in less time.

Using Collaborative Tagging for Text Classification: From Text Classification to Opinion Mining

Numerous initiatives have allowed users to share knowledge or opinions using collaborative platforms. In most cases, the users provide a textual description of their knowledge, following very limited or no constraints. Here, we tackle the classification of documents written in such an environment. As a use case, our study is made in the context of text mining evaluation campaign material, related to the classification of cooking recipes tagged by users from a collaborative website. This context makes some of the corpus specificities difficult to model for machine-learning-based systems and keyword or lexical-based systems. In particular, different authors might have different opinions on how to classify a given document. The systems presented hereafter were submitted to the D´Efi Fouille de Textes 2013 evaluation campaign, where they obtained the best overall results, ranking first on task 1 and second on task 2. In this paper, we explain our approach for building relevant and effective systems dealing with such a corpus.

Semantic composition process in a speech understanding system

A knowledge representation formalism for SLU is introduced. It is used for incremental and partially automated annotation of the Media corpus in terms of semantic structures. An automatic interpretation process is described for composing semantic structures from basic semantic constituents using patterns involving constituents and words. The process has procedures for obtaining semantic compositions and for generating frame hypotheses by inference. This process is evaluated on a dialogue corpus manually annotated at the word and semantic constituent levels.

Automatic Triage of Mental Health Online Forum Posts: CLPsych 2016 System Description.

This paper presents a system capable of performing automatic triage of forum posts from ReachOut.com, a mental health online forum. The system assigns to each post a tag that indicates how urgently moderator attention is needed. The evaluation is based on experiments conducted on the CLPsych 2016 task, and the system is released as an open-source software.

Mutual Disambiguation for Entity Linking

The disambiguation algorithm presented in this paper is implemented in SemLinker, an entity linking system. First, named entities are linked to candidate Wikipedia pages by a generic annotation engine. Then, the algorithm re-ranks candidate links according to mutual relations between all the named entities found in the document. The evaluation is based on experiments conducted on the test corpus of the TAC-KBP 2012 entity linking task.