Use XLING Universal Sentence Encoder

[samhavens]

New Universal Sentence Encoder module supports the 8 most common languages, English, Chinese, Japanese, Korean, Spanish, French, German, Italian. It should be an almost-drop-in replacement[0] for the current USE model: https://aihub.cloud.google.com/u/0/p/products%2F1c473955-f951-42cc-9191-df4e6dce0ceb

[0] Will also need to install sentencepiece and tf-sentencepiece