SEEE DIGIBOOK ON ENGINEERING & TECHNOLOGY, VOL. 01 (2), JUN 2020 PP. (45-56)
Abstract– The World Wide Web contains several forms of semantic relationships between entities. Relation extraction is a crucial stage in a variety of Web-related tasks, including information retrieval (IR), information extraction, and social network extraction. A supervised relation extraction system trained to extract a specific relation type (source relation) may not accurately extract a novel relation type (target relation) for which it has not been trained. However, it is costly to manually prepare training data for each new relation type that may be extracted. We offer a method for adapting an existing relation extraction system to automatically extract new relation types. Our suggested method consists of two stages: learning a lower dimensional projection between distinct relations and learning a relational classifier for the target relation type by sampling instances. First, in order to describe a semantic relationship between two things, we extract lexical and syntactic patterns from co-occurrence contexts. Then, a bipartite graph is constructed between relation-specific (RS) and relation-independent (RI) patterns. On the bipartite graph, spectral clustering is conducted to compute a lower-dimensional projection. Using a modest number of labelled cases, we then train a classifier for the target relation type. We provide a one-sided under sampling strategy to account for the dearth of target relation training examples. Using a data set including 2,000 instances of 20 distinct relation types, we test the proposed technique. Our experimental results indicate that the proposed strategy obtains a macro average F-score that is statistically significant at the 62.77 level. In addition, the suggested strategy outperforms various baselines and a weakly supervised relation extraction method previously proposed.
Index Terms – Cloud Computing; Encryption Homomorphic;
REFERENCE
[1] Bollegala, D., Matsuo, Y., & Ishizuka, M. (2011). Minimally supervised novel relation extraction using a latent relational mapping. IEEE Transactions on Knowledge and Data Engineering, 25(2), 419-432.
[2] Nayak, T., Majumder, N., Goyal, P., & Poria, S. (2020). Deep neural approaches to relation triplets extraction: A comprehensive survey. Cognitive Computation, 13(5), 1215-1232.
[3] Sapate, M. P., & Singh, S. T. Improved Method for Relation Adaptation to Extract the New Relations Efficiently from Relational Mapping.
[4] Pawar, S., Palshikar, G. K., & Bhattacharyya, P. (2017). Relation extraction: A survey. arXiv preprint arXiv:1712.05191.
[5] Nguyen, D. Q., & Verspoor, K. (2019, April). End-to-end neural relation extraction using deep biaffine attention. In European Conference on Information Retrieval (pp. 729-738). Springer, Cham.
[6] Zhang, X., Liu, T., Li, P., Jia, W., & Zhao, H. (2020). Robust neural relation extraction via multi-granularity noises reduction. IEEE Transactions on Knowledge and Data Engineering, 33(9), 3297-3310.
[7] Zhao, X., Yang, M., Qu, Q., Xu, R., & Li, J. (2020). Exploring Privileged Features for Relation Extraction with Contrastive Student-Teacher Learning. IEEE Transactions on Knowledge and Data Engineering.
[8] Li, J., Huang, G., Chen, J., & Wang, Y. (2019). Dual CNN for relation extraction with knowledge-based attention and word embeddings. Computational intelligence and neuroscience, 2019.
[9] Liu, K., & El-Gohary, N. (2017). Ontology-based semi-supervised conditional random fields for automated information extraction from bridge inspection reports. Automation in construction, 81, 313-327.
Venkataraman K, Dhanabal S
Department of Information Technology
Rathinam Technical Campus,
Coimbatore, Tamilnadu, India