A semagram is a flexible structure for encoding the semantics of a given concept via a slot-filler structure.
This "semagram base” [1] is the result of a systematic analysis, unification and extension of semantic relations (or slots) derived from different resources:
- ANW dictionary [2]
- Visual Attributes [3]
- Property Norms [4]
This project is the first result of the collaboration between:
- Dr. Luigi Di Caro (University of Turin) and
- Prof. Roberto Navigli (University Sapienza of Rome).
Two PhD students contributed to the annotation and the semi-automatic development of the released resource:
- Valentina Leone (PhD student at the University of Turin)
- Giovanni Siragusa (PhD student at the University of Turin)
The semagram base is released under a CC-BY 4.0 license.
[1] Leone, V., Siragusa, G., Di Caro, L., Navigli, R. (2020). Building Semantic Grams of Human Knowledge. In Proc. of LREC (to appear).
[2] Moerdijk, F., Tiberius, C., and Niestadt, J. (2008). Accessing the anw dictionary. In Proc. of the workshop on Cognitive Aspects of the Lexicon, pages 18–24.
[3] Silberer, C., Ferrari, V., and Lapata, M. (2013). Models of semantic representation with visual attributes. In ACL(1), pages 572–582.
[4] Devereux, B. J., Tyler, L. K., Geertzen, J., and Randall, B. (2014). The cslb concept property norms. Behavior research methods, 46(4):1119–1127.