Most of the major databases on the semantic web have links to Princeton WordNet (PWN) synonym set (synset) identifiers, which differ for each PWN release, and are thus incompatible between versions. On the other hand, both PWN and the more recent Open English Wordnet (OEWN) provide permanent word sense identifiers (the sense keys), which can solve this interoperability problem. We present an algorithm that runs in linear time, to automatically derive a synset mapping between any pair of Wordnet versions that use PWN sense keys. This allows to update old WordNet links, and seamlessly interoperate with newer English Wordnet versions for which no prior mapping exists. By applying the proposed algorithm on the fly, at load time, we combine the Open Multilingual Wordnet (OMW 1.4, which uses old PWN 3.0 identifiers) with OEWN Edition 2021, and obtain almost perfect precision and recall. We compare the results of our approach using respectively synset offsets, versus the Collaborative InterLingual Index (CILI version 1.0) as synset identifiers, and find that the synset offsets perform better than CILI 1.0 in all cases, except a few ties.

We fit WordNet relations to word embeddings, using 3CosAvg and LRCos, two set-based methods for analogy resolution, and introduce 3CosWeight, a new, weighted variant of 3CosAvg. We test the performance of the resulting semantic vectors in lexicographic semantics tests, and show that none of the tested classifiers can learn symmetric relations like synonymy and antonymy, since the source and target words of these relations are the same set. By contrast, with the asymmetric relations (hyperonymy / hyponymy and meronymy), both 3CosAvg and LRCos clearly outperform the baseline in all cases, while 3CosWeight attained the best scores with hyponymy and meronymy, suggesting that this new method could provide a useful alternative to previous approaches.