textanalysis   785

« earlier    

Universal Sentence Encoder
We present models for encoding sentences into embedding vectors that specifically target transfer learning to other NLP tasks. The models are efficient and result in accurate performance on diverse transfer tasks. Two variants of the encoding models allow for trade-offs between accuracy and compute resources. For both variants, we investigate and report the relationship between model complexity, resource consumption, the availability of transfer task training data, and task performance. Comparisons are made with baselines that use word level transfer learning via pretrained word embeddings as well as baselines do not use any transfer learning. We find that transfer learning using sentence embeddings tends to outperform word level transfer. With transfer learning via sentence embeddings, we observe surprisingly good performance with minimal amounts of supervised training data for a transfer task. We obtain encouraging results on Word Embedding Association Tests (WEAT) targeted at detecting model bias. Our pre-trained sentence encoding models are made freely available for download and on TF Hub.
nlp  embeddings  wordembedding  research  paper  AI  textanalysis  machinecomprehension  machinelearning 
11 weeks ago by sachaa

« earlier    

related tags

3dprinting  602  ai  algorithmicbias  algorithmics  analysis  annotator  api  automate  automation  badstats  bcu  bias  bigdata  cheatsheet  classification  cnn  collaboration  computervision  conference  constitution  conversion  converter  corenlp  counterclaim  dataanalysis  datamining  datascience  dataset  dataviz  dates  ddj  deeplearning  development  dh  digital-humanities  digitalhumanities  dj  docker  ebook  edusocmedia  email  embeddings  emotion  entityextraction  extraction  extractor  facebook  film  fuzzymatching  fuzzystringmatching  gender  generative  google  graphs  history  howto  imageprocessing  imagerecognition  inls201  jobsearch  knowledgeextraction  knowledgegraphs  knowledgemining  language  law  learning  legaldocs  levenstein  linkeddata  machinecomprehension  machinelearning  medium  moodtech  nlp  odyssey  opensource  paper  partialmatching  people  poetry  politics  popculture  postagging  predictions  predictive_modelling  presentation  py  python  q&a  r  readability  recommendationengine  reconciliation  reference  reinforcementlearning  research  review  rnn  rstats  ruby  sarcasm  scrape  scraper  scraping  search  semanticanalysis  semanticweb  sentiment  sentimentanalysis  server  slack  socialmedia  sociology  spacy  stringdistance  stringmatching  tagger  tagging  task  teachingimages  tensorflow  text  textannotation  textdata  textgeneration  textmining  textprocessing  texttospeech  textviz  tools  topic-modeling  topicmodelling  training  tutorial  twitter  ux  video  visualization  vocabulary  washpo  word2vec  wordembedding  words  worstpractice  writing 

Copy this bookmark:



description:


tags: