arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 1702.08359stat.MLcs.LG
收藏

动态词嵌入

Dynamic Word Embeddings

Disney Research(迪士尼研究院)

Robert Bamler, Stephan Mandt

摘要:

我们提出了一个概率语言模型的时间戳的文本数据,跟踪随着时间的推移,单个单词的语义演变。该模型通过嵌入空间中的潜在轨迹来表示单词和上下文。在每个时刻,从word 2 vec的概率版本推断嵌入向量[Mikolov等人,2013年]。这些嵌入向量通过潜在扩散过程在时间上连接。我们描述了两种可扩展的变分推断算法--skip-gram平滑和skip-gram过滤--使我们能够在所有时间联合训练模型;从而在所有数据上学习,同时允许单词和上下文向量漂移。在三个不同语料库上的实验结果表明,我们的动态模型推断出的单词嵌入轨迹比基于在时间片上单独训练的静态模型的竞争方法更具可解释性,并导致更高的预测可能性。

英文摘要:

We present a probabilistic language model for time-stamped text data which tracks the semantic evolution of individual words over time. The model represents words and contexts by latent trajectories in an embedding space. At each moment in time, the embedding vectors are inferred from a probabilistic version of word2vec [Mikolov et al., 2013]. These embedding vectors are connected in time through a latent diffusion process. We describe two scalable variational inference algorithms--skip-gram smoothing and skip-gram filtering--that allow us to train the model jointly over all times; thus learning on all data while simultaneously allowing word and context vectors to drift. Experimental results on three different corpora demonstrate that our dynamic model infers word embedding trajectories that are more interpretable and lead to higher predictive likelihoods than competing methods that are based on static models trained separately on time slices.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑