Abstract
One of the major tasks involved in the development of a knowledge-based problem solver is domain knowledge representation and reasoning. In this paper, we address this task for a knowledge-based kinematics word problem solver that solves problems from the domain of kinematics automatically, where solving each problem involves identifying the value of a specific unknown quantity referred to, within the problem. Knowledge about kinematics domain is captured at two levels: quantitative level and a more abstract qualitative level. We leverage OWL (Web Ontology Language) and RDF (Resource Description Framework) rules to represent both qualitative and quantitative knowledge of the domain in a single framework. We build an ontology, wherein we identify a fixed number of classes and properties that provide a vocabulary to formally represent a domain qualitatively and quantitatively. We then define the kinematics domain in terms of these classes and properties using RDF rules and OWL axioms. This is then used as a knowledge base (KB) to a kinematics problem solver. The input to this solver is represented as an RDF graph, called the problem scenario graph. Inference based on the OWL axioms and RDF rules in the KB adds knowledge, that is required to solve the problem, to the problem scenario graph. The knowledge enriched problem scenario graph is then used by an external reasoner to infer the value of the unknown quantity in the problem. We created a dataset of around 100 problems from the domain to provide a qualitative analysis of the solver by describing the various failure modes with examples.
Similar content being viewed by others
Notes
These are axioms of the form: if the current value of a parameter is [0,0] and the value of the parameter is increasing, then its next possible value is (0, inf). Since the landmark values in a parameter’s quantity space are in ascending order, an increase (decrease) in the parameter will result in shifting right (left) from the current value in the quantity space.
It may be noted that angle of projection being 0 always does not mean straight line motion—for example, a ball projected horizontally from a height has motion in vertical and horizontal dimensions.
References
Abraham, S.S., Khemani, D.: Hybrid of qualitative and quantitative knowledge models for solving physics word problems. In: FLAIRS Conference, pp. 510–515 (2016)
Abraham, S.S., Sundaram, S.S.: Combining qualitative and quantitative reasoning for solving kinematics word problems. In: The Thirty-First International Flairs Conference (2018)
Bobrow, D., Falkenhainer, B., Farquhar, A., Fikes, R., Forbus, K., Gruber, T., Iwasaki, Y., Kuipers, B., et al.: A compositional modeling language. In: Proceedings of the 10th International Workshop on Qualitative Reasoning about Physical Systems, pp. 12–21 (1996)
Carroll, J.J., Dickinson, I., Dollin, C., Reynolds, D., Seaborne, A., Wilkinson, K.: Jena: implementing the semantic web recommendations. In: Proceedings of the 13th International World Wide Web Conference on Alternate Track Papers and Posters, pp. 74–83. ACM (2004)
Clark, P., Chaw, S.Y., Barker, K., Chaudhri, V., Harrison, P., Fan, J., John, B., Porter, B., Spaulding, A., Thompson, J., et al.: Capturing and answering questions posed to a knowledge-based system. In: Proceedings of the 4th International Conference on Knowledge Capture, pp. 63–70. ACM (2007)
De Kleer, J.: Qualitative and Quantitative Knowledge in Classical Mechanics. MIT, Artificial Intelligence Laboratory, London (1975)
De Kleer, J.: Multiple representations of knowledge in a mechanics problem-solver. In: Readings in Qualitative Reasoning About Physical Systems, pp. 40–45. Elsevier, London (1990)
Forbus, K.D.: Qualitative reasoning about physical processes. IJCAI 1981, 326–330 (1981)
Glimm, B., Horrocks, I., Motik, B., Stoilos, G., Wang, Z.: Hermit: an owl 2 reasoner. J. Autom. Reason. 53(3), 245–269 (2014)
Huang, D., Shi, S., Lin, C.Y., Yin, J.: Learning fine-grained expressions to solve math word problems. In: Proceedings of the 2017 Conference on Empirical Methods in Natural Language Processing, pp. 805–814 (2017)
Kuipers, B.: Qualitative simulation. Artif. Intell. 29(3), 289–338 (1986)
Kushman, N., Artzi, Y., Zettlemoyer, L., Barzilay, R.: Learning to automatically solve algebra word problems. In: Proceedings of the 52nd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers), Vol. 1, pp. 271–281 (2014)
Minsky, M.: Minsky’s frame system theory. In: TINLAP’75: Proceedings of the 1975 workshop on Theoretical issues in natural language processing, pp. 104–116 (1975)
Mitra, A., Baral, C.: Learning to use formulas to solve simple arithmetic problems. In: Proceedings of the 54th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers), Vol. 1, pp. 2144–2153 (2016)
Pisan, Y.: An integrated architecture for engineering problemsolving. Tech. rep., Northwestern Univ Evanston Il Dept of Computerscience (1998)
Pisan, Y., Bachmann, A.: Using qualitative reasoning to solve dynamic problems. In: Proceedings of the 12th International Workshop on Qualitative Reasoning, pp. 167–173 (1998)
Upadhyay, S., Chang, M.W., Chang, K.W., Yih, W.t.: Learning from explicit and implicit supervision jointly for algebra word problems. In: Proceedings of the 2016 Conference on Empirical Methods in Natural Language Processing, pp. 297–306 (2016)
Wang, H., Noy, N., Rector, A., Musen, M., Redmond, T., Rubin, D., Tu, S., Tudorache, T., Drummond, N., Horridge, M., et al.: Frames and owl side by side. In: Presentation Abstracts, p. 54. Citeseer (2006)
Zhang, D., Wang, L., Xu, N., Dai, B.T., Shen, H.T.: The gap of semantic parsing: a survey on automatic math word problem solvers. arXiv preprint arXiv:1808.07290 (2018)
Zhou, L., Dai, S., Chen, L.: Learn to solve algebra word problems using quadratic programming. In: Proceedings of the 2015 Conference on Empirical Methods in Natural Language Processing, pp. 817–822 (2015)
Author information
Authors and Affiliations
Corresponding author
Additional information
Publisher's Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
About this article
Cite this article
Abraham, S.S., Sundaram, S.S. An Ontology-Based Kinematics Problem Solver Using Qualitative and Quantitative Knowledge. New Gener. Comput. 37, 551–584 (2019). https://doi.org/10.1007/s00354-019-00067-x
Received:
Accepted:
Published:
Issue Date:
DOI: https://doi.org/10.1007/s00354-019-00067-x