• Comparison of three schemes of intelligent customer service chat module


    In the current market of intelligent customer service and its NLP field, the chat module is a very important research field. At present, the industry’s solutions for the chat module are generally the following three solutions Based on fixed template We all know that the general solution for business Q&A is based on template Q&A. […]

  • Hengyuan cloud (gpushare)_ Beyond the model of pre training NLP


    Article source|Hengyuan cloud community Original address|Training an NLP model beyond pre training from zero Original author|Mathor Welcome welcome, warm welcomeCommunity leaders are back ~I can carry articles happily again! Get out of the way. I’m going to start moving. Body start:\This article is based on a paper on arXivNLP From Scratch Without Large-Scale Pretraining: A […]

  • NLP Essays (II)


    When the performance of AI on a single point task approaches or exceeds that of human beings, it will bring huge business opportunities to the industry. In terms of visual classification, retrieval, matching, target detection and other tasks, as the relevant algorithms become more and more accurate, the industry has also begun to try these […]

  • Brief introduction of common pre training language models


    Model structure evolution Based on the evolution direction and purpose, this paper combs some pre training language models that I am familiar with but not very familiar with: First, let’s take a look at the “full version of Bert”: RoBERTa: A Robustly Optimized BERT Pretraining Approach(2019)It can be regarded as the whole Bert. There are […]

  • Hengyuan cloud_ Byte beating mrasp pre training model is really fragrant


    Article source|A shared computing platform focusing on AI industry: Hengyuan cloud Original address|Mrasp pre training model Original author Jiao Hui It is estimated that the bosses are either taking exams or on holiday recently, so they don’t post in the communityI have no choice but to dig deep in the vast posts. Hey, it’s a […]

  • Natural language processing — Text vectorization (2)


    1、 Abstract This sharing is based on the continuation of the previous text vector method. In the last content, we mainly shared two methods of text vectorization: the word bag model representation method and the word vector method based on deep learning. Although the word bag model can simply express words as vectors, it will […]

  • 5-minute NLP series – Summary of 11 word embedding model


    TF-IDF, Word2Vec, GloVe, FastText, ELMO, CoVe, BERT, RoBERTa The function of word embedding in depth model is to provide input features for downstream tasks, such as sequence marking and text classification. In the past decade, many word embedding methods have been proposed. This article will make a complete summary of these word embedding models Context […]

  • Computer graduation project springboot + Vue JS new covid-19 news corpus event argument collaborative tagging system


    Suitable for the subject New covid-19 news corpus event argument collaborative tagging system Recommended reasons: Cool topic + social frontier hot spot + innovation to awesome! Development technology Front end technology: Vue js elementUI Back end technology: springboot JPA Database: MySQL functional module 1. User registration and login function. The default login is the administrator […]

  • Customize a robot to help you chat with TA


    Auto chat example Chat 1: User: are you there? BOT: in User: what are you doing? BOT: watch TV User: what TV are you watching BOT: vivid User: is it nice? BOT: very funny User: where are you? BOT: at home User: you’re the only one in the family? BOT: I like playing alone User: […]

  • Quickly build ner model, Amazon sagemaker ground truth and Amazon comprehensive give you tips


    Previously, we had an article on the difficulties in Chinese named entity recognition (NER) and the solutions of Amazon sagemaker. Interested children’s shoes are welcome to click here to review:Compared with English, what is the difficulty of Chinese named entity recognition? How to solve it。 In this article, we intend to talk about the difficulties […]

  • Natural language processing common resources note sharing


    Statement: personal notes, opinions, continuously updated The personal and standby natural language processing related resources collected by individuals are only for academic exchangeProject address:https://github.com/junchaoIU/ChineseNLP_ResourceNote Author:🍧 Wu, Junchao Open source Python Library project address brief introduction Jieba participle https://github.com/fxsjy/jieba Chinese Thesaurus Chinese information extraction tool https://github.com/fighting41love/cocoNLP Extract structured information from Chinese text data, such as time, […]

  • Using EasyWeChat and ChatterBot to build a official account of “automatic recovery robot”


    Since the revision of the official account list page, many people say that the role of the official account menu will be weakened. Moreover, for personal numbers, menu development cannot be operated in the development mode. So we simply “give up the menu” and make “automatic reply” to replace the menu function. To develop the […]