Description
The Universal Sentence Encoder encodes text into high-dimensional vectors that can be used for text classification, semantic similarity, clustering, and other natural language tasks.
The model is trained and optimized for greater-than-word length text, such as sentences, phrases, or short paragraphs. It is trained on a variety of data sources and a variety of tasks with the aim of dynamically accommodating a wide variety of natural language understanding tasks. The input is the variable-length text and the output is a 512-dimensional vector. The universal-sentence-encoder model has trained with a deep averaging network (DAN) encoder.
This model supports 16 languages (Arabic, Chinese-simplified, Chinese-traditional, English, French, German, Italian, Japanese, Korean, Dutch, Polish, Portuguese, Spanish, Thai, Turkish, Russian) text encoder.
The details are described in the paper “Multilingual Universal Sentence Encoder for Semantic Retrieval”.
Note: This model only works on Linux and macOS operating systems and is not compatible with Windows due to the incompatibility of the SentencePiece library.
How to use
embeddings = UniversalSentenceEncoder.pretrained("tfhub_use_multi", "xx") \
.setInputCols("document") \
.setOutputCol("sentence_embeddings")
val embeddings = UniversalSentenceEncoder.pretrained("tfhub_use_multi", "xx")
.setInputCols("document")
.setOutputCol("sentence_embeddings")
import nlu
text = ["I love NLP", "Me encanta usar SparkNLP"]
embeddings_df = nlu.load('xx.use.multi').predict(text, output_level='sentence')
embeddings_df
Results
It gives a 512-dimensional vector of the sentences
Model Information
| Model Name: | tfhub_use_multi |
| Compatibility: | Spark NLP 3.0.0+ |
| License: | Open Source |
| Edition: | Official |
| Input Labels: | [sentence] |
| Output Labels: | [sentence_embeddings] |
| Language: | xx |
Benchmarking
- We apply this model to the STS benchmark for semantic similarity.
STSBenchmark | dev | test |
-----------------------------------|--------|-------|
Correlation coefficient of Pearson | 0.829 | 0.809 |
- For semantic similarity retrieval, we evaluate the model on [Quora and AskUbuntu retrieval task.](https://arxiv.org/abs/1811.08008). Results are shown below:
Dataset | Quora | AskUbuntu | Average |
-----------------------|-------|-----------|---------|
Mean Average Precision | 89.2 | 39.9 | 64.6 |
- For the translation pair retrieval, we evaluate the model on the United Nation Parallel Corpus. Results are shown below:
Language Pair | en-es | en-fr | en-ru | en-zh |
---------------|--------|-------|-------|-------|
Precision@1 | 85.8 | 82.7 | 87.4 | 79.5 |