Evaluation of Unsupervised Compositional Representations [article]

Hanan Aldarmaki, Mona Diab
2018 arXiv   pre-print
We evaluated various compositional models, from bag-of-words representations to compositional RNN-based models, on several extrinsic supervised and unsupervised evaluation benchmarks. Our results confirm that weighted vector averaging can outperform context-sensitive models in most benchmarks, but structural features encoded in RNN models can also be useful in certain classification tasks. We analyzed some of the evaluation datasets to identify the aspects of meaning they measure and the
more » ... eristics of the various models that explain their performance variance.
arXiv:1806.04713v2 fatcat:dxb47tq22zfqdhesa52dr5yz2i