Change Detection Adam Tsakalidis and Maria Liakata EMNLP2020 (ArXiv: “Autoencoding Word Representations through Time for Semantic Change Detection”) 論文紹介
t=0 と t=i を alignment(PROCR):Hamilton2016[3] ii. t=0 と t=i について、意味が変化しない k 単語を元に alignment (PROCRk):Tsakalidis2019[1] iii. (ii) を隣合う2つの時期について行い、全時期で共通する k 単語で alignment(PROCRkt):Tsakalidis2019[1] - 最初と最後の2点間のみを使用 i. Random Forest(RF):t=0 から t=i を予測 ii. 提案手法の再構築モデル(LSTMr):t=0, t=i を再構築 iii. 提案手法の予測モデル(LSTMf):t=0 から t=i を予測 16
Barbara McGillivray. Mining the UK Web Archive for Semantic Change Detection, RANLP2019. [link] [2] Alex Rosenfeld, Katrin Erk. Deep Neural Models of Semantic Shift, ACL2018. [link] [3] William L. Hamilton, Jure Leskovec, Dan Jurafsky. Diachronic Word Embeddings Reveal Statistical Laws of Semantic Change, ACL2016. [link] [4] Philippa Shoemark, Farhana Ferdousi Liza, Dong Nguyen, Scott Hale, Barbara McGillivray. Room to Glo: A Systematic Comparison of Semantic Change Detection Approaches with Word Embeddings, EMNLP2019. [link] 23