FrameNet-assisted Noun Compound Interpretation

Findings of the Association for Computational Linguistics: ACL-IJCNLP 2021, pages 2901–2911 August 1–6, 2021. ©2021 Association for Computational Linguistics 2901 FrameNet-assisted Noun Compound Interpretation Girishkumar Ponkiya†, Diptesh Kanojia‡, Pushpak Bhattacharyya† †Indian Institute of Technology Bombay, Mumbai ‡Centre for Translation Studies, University of Surrey, United Kingdom {girishp,diptesh,pb}@cse.iitb.ac.in Girish K Palshikar TCS Research, Tata Consultancy Services, Pune gk.palshikar@tcs.com Abstract Given a noun compound (NC), we address the problem of predicting the appropriate semantic label linking the constituents of the NC. This problem is called Noun Compound Interpretation (NCI). We use FrameNet as a semantic label repository. For example, given the noun compound (board approval), we pre- dict the frame (DENY OR GRANT PERMISSION, as per FrameNet) as appropriate and the semantic role of the modifier word (AUTHORITY) as the semantic label linking board and approval; the resulting label is DENY OR GRANT PERMISSION:AUTHORITY. Our semantic label repository is very large (≈ 11k labels) compared to the NC data available for training (approx 1900). Thus, learning in this case, especially for unseen semantic la- bels, is hard. We propose to solve this prob- lem by predicting semantic labels in a contin- uous label embedding space, which is novel. This embedding space is created by learning label embeddings using the FrameNet data. The embeddings are then used to train two separate models – one for predicting Frames and the other for FEs. As the label embed- ding space captures the semantics of the labels, using these embeddings enables generalizing well on unseen labels, thus achieving zero-shot learning. Our preliminary investigations show that the proposed approach performs well for unseen labels, achieving 5% and 2% points im- provements over baselines for the frame and FE prediction, respectively. The study shows the promise of the use of continuous space embeddings for noun compound interpretation and points to the need for further investigation. 1 Introduction A noun compound is a sequence of two or more nouns that act as a single entity with well-defined meaning (e.g., paper submission, colon cancer, etc.). Semantic relations between the component nouns are implicit. For instance, the information that ‘it is a juice made from orange’ is hidden in orange juice. Uncovering this semantic relation is called the problem of Noun Compound Inter- pretation (NCI). NCI needs ML, as the task faces the challenge of ambiguity, and disambiguation by rules is well nigh impossible because of multifari- ous complex underlying language phenomena. The proposition of storing NCs and doing table lookup for interpretation is also impractical due to a large number of NCs and the challenge of high produc- tivity (new nouns and NCs get created frequently, e.g., corona vaccine is a relatively new NC). Often, the exact relation, sentiment, etc. are also governed by contextual pragmatics. For instance, the sentiment towards tax money depends on who the beneficiary is, which again depends on the pred- icate. The predicate give could indicate negative sentiment (for the tax-payer), whereas the predicate receive would indicate positive sentiment (for the government). Due to such instances, …

Content truncated to 500 words