Event Abstract

Back to Event

An Ontology-Based Semantic Question Complexity Model and Its Applications in Neuroinformatics

Aref Eshghishargh1*, Simon Milton1, Gary F. Egan2, Andrew Lonie3, Scott Kolbe3, Neil E. Killeen3 and Jason M. Lohrey4
  • 1 University of Melbourne, Computing and Information Systems, Australia
  • 2 Monash University, Monash Biomedical Imaging (BMI), Australia
  • 3 University of Melbourne, Australia
  • 4 Arcitecta, Australia

Neuroscience is an important field of study because of the huge number of neurological disorders (Köhler et al., 2012) and the quest scientist have for solving brain mysteries (Akil et al., 2011). Enormous data-sets containing large images are commonly produced in this field (Ozyurt et al., 2010). Usually, data are gathered using different methods in different labs and data resources are scattered. Also, ordinary methods of searching might not be enough to resolve the semantics of scientific questions properly. Therefore, scientists are always looking for better tools to handle and search the data. Data structures and schemas such as ontologies (Gruber, 1993) have evolved to assist managing the bulk of information in this field and ontology-based (or enhanced) applications have been developed to assist neuroscientists in handling the information. Ontology-based applications have practiced different approaches (Gupta et al., 2010) including using ontologies for question answering both in restricted (closed) domains with the approach discussed in (Mollá and Vicedo, 2007) and open domains such as the approach used in PowerAqua (Lopez et al., 2012). There have been few efforts in question answering in neuroscience and a limited range of question types have been addressed in them. Current approaches do not use the full potential of ontologies, mostly use basic relationships such as ‘is_a’ and most of applications have limited capability of query expression (Lopez et al., 2012). For example, they can answer queries asking for “volume” or “inferior parietal lobule”, but are not able to understand and answer questions such as “what is the volume of inferior parietal lobule of the brain?”. Therefore, an approach which addresses question answering in a systematic manner and considers the complexity of questions and their structure seems necessary. This way, shortages of the field and the scope of the research will be clearer; also, more complex questions can be identified and answered. In this research, an ontology-based model is proposed which has the ability to track changes and helps in answering conditional questions. To build this model, sample questions sourced from experts and literature are analyzed, tokenized and then clustered until a state known as theoretical saturation (Bowen, 2008) is reached. The outcome of this process is 8 clusters which are called dimensions. These eight dimensions are: relationships such as partOf and subClassOf; concepts involved such as hippocampus and Precentral Gyrus; domain-specific phrases that specify scientific processes or attributes such as curvature, activation, volume and thickness; changes such as extra or time (temporal changes); summary or statistical phrases such as summary, total number; data resources such as ontologies or data-files; conditions such as connected and finally indeterminate phrases such as elderly. Relationships, concepts and domain-specific phrases are loaded from ontologies in neuroscience such as NIFSTD (Imam et al., 2011) and this, makes the model an ontology-based one. The model has some unique features such as the potential to cover temporal changes. The evaluation of the model was done via a different set of sample questions from a different expert team of neuroscientists. After proposing the model, its applications including capability of inferring information, question classification, and interface design are investigated. At the lowest level, dimensions can be individually used to infer or enquire information. For example, questions containing indeterminate phrases can be marked as ‘unanswerable’ as more information is needed to resolve them. This information can be either enquired using user interaction processes or guessed through applying methods such as fuzzy logic. Another example is that in many cases, the type of the answer can be known by looking at the dimensions present in the question. For example, when the value of the domain-specific phrase dimension is volume, it means that the question is looking for the volume which is a number. Using the model and techniques such as ontology-based query expansion (Bhogal et al., 2007), a range of question can be answered which were not answerable without the ontologies. A question classification or taxonomy can be shaped using model dimensions. Question classification is the task of assigning categories to questions (Yu et al., 2005). Individual or multiple dimensions can be used to build this classification. For example, considering the number of resources as the core dimension of the classification, level 1 would be questions using only one ontology as the resource. Level 2 would have two ontologies as the resource and possible ontology mapping would be needed to map the two ontologies. This goes on and at the highest level (level n) the classification would have n ontologies. The question answering system which is built in response to this question classification will be a mediator system that integrates resources in order to answer a question. Another use of the model is in interface design; also dimensions can be seen as frame semantics (Petruck, 1996) of complex questions. Dimensions can be fields of a keyword-based (concept based) interface (Müller et al., 2004) where they are categories (or drop down lists) and related values are loaded from ontologies, data resources and other pre-defined lists into them. Doing this, the user will be guided through posing a scientifically valid question and the cost of query translation will be omitted. TAMBIS (Stevens et al., 2003, 2000) is an example of keyword-based interface. It worth mentioning that the model can also be used to build faceted search interfaces (Hearst, 2006) or text-based ones (Müller et al., 2004).

References

Akil, H., Martone, M. E., and Van Essen, D. C. (2011). Challenges and opportunities in mining neuroscience data. Science 331, 708–712. doi:10.1126/science.1199305.

Bhogal, J., Macfarlane, a., and Smith, P. (2007). A review of ontology based query expansion. Inf. Process. Manag. 43, 866–886. doi:10.1016/j.ipm.2006.09.003.

Bowen, G. a. (2008). Naturalistic inquiry and the saturation concept: a research note. Qual. Res. 8, 137–152. doi:10.1177/1468794107085301.

Gruber, T. R. (1993). A translation approach to portable ontology specifications. Knowl. Acquis. 5, 199–220.

Gupta, A., Condit, C., and Qian, X. (2010). BioDB: An ontology-enhanced information system for heterogeneous biological information. Data Knowl. Eng. 69, 1084–1102. doi:10.1016/j.datak.2010.07.003.

Hearst, M. (2006). Design recommendations for hierarchical faceted search interfaces. ACM SIGIR Work. Facet. search, 5–9.

Imam, F. T., Larson, S. D., Grethe, J. S., Gupta, A., Bandrowski, A., and Martone, M. E. (2011). NIFSTD and NeuroLex: Comprehensive Neuroscience Ontology Development Based on Multiple Biomedical Ontologies and Community Involvement.

Köhler, S., Doelken, S. C., Rath, A., Aymé, S., and Robinson, P. N. (2012). Ontological phenotype standards for neurogenetics. Hum. Mutat. 33, 1333–1339. doi:10.1002/humu.22112.

Lopez, V., Fernández, M., Motta, E., and Stieler, N. (2012). Poweraqua: Supporting users in querying and exploring the semantic web. Semant. Web.

Mollá, D., and Vicedo, J. L. (2007). Question answering in restricted domains: An overview. Comput. Linguist. 33, 41–61.

Müller, H.-M., Kenny, E. E., and Sternberg, P. W. (2004). Textpresso: an ontology-based information retrieval and extraction system for biological literature. PLoS Biol. 2, e309.

Ozyurt, I., Keator, D., Wei, D., Fennema-Notestine, C., Pease, K., Bockholt, J., and Grethe, J. (2010). Federated Web-accessible Clinical Data Management within an Extensible NeuroImaging Database. Neuroinformatics 8, 231–249. doi:10.1007/s12021-010-9078-6.

Petruck, M. R. (1996). Frame semantics. Handbook of pragmatics, 1-13.

Stevens, R., Baker, P., Bechhofer, S., Ng, G., Jacoby, A., Paton, N. W., Goble, C. A., and Brass, A. (2000). TAMBIS: transparent access to multiple bioinformatics information sources. Bioinformatics 16, 184–186.

Stevens, R., Goble, C., Paton, N. W., Bechhofer, S., Ng, G., Baker, P., and Brass, A. (2003). Complex query formulation over diverse information sources in TAMBIS. Bioinforma. Manag. Sci. Data. Morgan Kaufmann.

Yu, H., Sable, C., & Zhu, H. R. (2005). Classifying medical questions based on an evidence taxonomy. In Proceedings of the AAAI 2005 workshop on question answering in restricted domains.

Keywords: Ontology based modeling and reasoning, question complexity, interfaces, Question classification, question answering, question dimensions

Conference: Neuroinformatics 2015, Cairns, Australia, 20 Aug - 22 Aug, 2015.

Presentation Type: Poster, to be considered for oral presentation

Topic: General neuroinformatics

Citation: Eshghishargh A, Milton S, Egan GF, Lonie A, Kolbe S, Killeen NE and Lohrey JM (2015). An Ontology-Based Semantic Question Complexity Model and Its Applications in Neuroinformatics. Front. Neurosci. Conference Abstract: Neuroinformatics 2015. doi: 10.3389/conf.fnins.2015.91.00015

Copyright: The abstracts in this collection have not been subject to any Frontiers peer review or checks, and are not endorsed by Frontiers. They are made available through the Frontiers publishing platform as a service to conference organizers and presenters.

The copyright in the individual abstracts is owned by the author of each abstract or his/her employer unless otherwise stated.

Each abstract, as well as the collection of abstracts, are published under a Creative Commons CC-BY 4.0 (attribution) licence (https://creativecommons.org/licenses/by/4.0/) and may thus be reproduced, translated, adapted and be the subject of derivative works provided the authors and Frontiers are attributed.

For Frontiers’ terms and conditions please see https://www.frontiersin.org/legal/terms-and-conditions.

Received: 31 May 2015; Published Online: 05 Aug 2015.

* Correspondence: Mr. Aref Eshghishargh, University of Melbourne, Computing and Information Systems, Melbourne, Australia, aref.cs@gmail.com