Released Resources

1) FLORS: the state-of-the-art online POS tagger in domain adaptation scenario. It performs online representation learning (i.e., domain adaptation is performed by incrementally adapting word representations to the new domain). (EMNLP'2015)

2) Phrase embeddings: the first version in the world. It contains continuous linguistic phrases like "eye disease'' as well as discontinuous phrases like "turn...on'' etc. (ICLR'2013, ACL'2014)

3) Word Meta-Embeddings: A version of word embeddings which were derived by combining several pre-trained word embedding versions, such as word2vec, GloVe, C&W etc. (ACL'2016)

4) Attentive Convolution: An attention-based convolution system. Input: two sentences; output: a highly-expressive representation; task: reason the sentence relations, e.g., paraphrasing, textual entailment etc. (TACL 2018)

5) Perspectrum: A benchmark that studies substantiated perspective discovery where, given a claim, a system is expected to discover a diverse set of well-corroborated perspectives that take a stance with respect to the claim. (NAACL'2019)