CogALex-V Shared Task: ROOT18
This work addresses a challenge in computational linguistics for researchers and practitioners by evaluating distributional methods in semantic relation classification, but it is incremental as it builds on existing shared tasks and highlights limitations rather than introducing a novel solution.
The paper tackled the problem of classifying word pairs as semantically related or unrelated and then categorizing related pairs into specific semantic relations (synonymy, antonymy, hypernymy, meronymy) using unsupervised distributional measures as features in a classifier called ROOT18, which performed solidly on the first subtask but poorly on the second subtask, indicating that distributional measures are insufficient for discriminating between multiple semantic relations simultaneously.
In this paper, we describe ROOT 18, a classifier using the scores of several unsupervised distributional measures as features to discriminate between semantically related and unrelated words, and then to classify the related pairs according to their semantic relation (i.e. synonymy, antonymy, hypernymy, part-whole meronymy). Our classifier participated in the CogALex-V Shared Task, showing a solid performance on the first subtask, but a poor performance on the second subtask. The low scores reported on the second subtask suggest that distributional measures are not sufficient to discriminate between multiple semantic relations at once.