All Projects → zhpmatrix → Lic2019 Competition

zhpmatrix / Lic2019 Competition

2019语言与智能技术竞赛-基于知识图谱的主动聊天

Programming Languages

python
139335 projects - #7 most used programming language

Projects that are alternatives of or similar to Lic2019 Competition

SpaceFusion
NAACL'19: "Jointly Optimizing Diversity and Relevance in Neural Response Generation"
Stars: ✭ 73 (-33.03%)
Mutual labels:  chatbot, dialogue-systems
Nlp Projects
word2vec, sentence2vec, machine reading comprehension, dialog system, text classification, pretrained language model (i.e., XLNet, BERT, ELMo, GPT), sequence labeling, information retrieval, information extraction (i.e., entity, relation and event extraction), knowledge graph, text generation, network embedding
Stars: ✭ 360 (+230.28%)
Mutual labels:  knowledge-graph, dialogue-systems
Nlp4rec Papers
Paper list of NLP for recommender systems
Stars: ✭ 162 (+48.62%)
Mutual labels:  knowledge-graph, dialogue-systems
Meld
MELD: A Multimodal Multi-Party Dataset for Emotion Recognition in Conversation
Stars: ✭ 373 (+242.2%)
Mutual labels:  chatbot, dialogue-systems
Knowledge Graphs
A collection of research on knowledge graphs
Stars: ✭ 845 (+675.23%)
Mutual labels:  knowledge-graph, dialogue-systems
Convai Bot 1337
NIPS Conversational Intelligence Challenge 2017 Winner System: Skill-based Conversational Agent with Supervised Dialog Manager
Stars: ✭ 65 (-40.37%)
Mutual labels:  chatbot, dialogue-systems
Neuraldialog Cvae
Tensorflow Implementation of Knowledge-Guided CVAE for dialog generation ACL 2017. It is released by Tiancheng Zhao (Tony) from Dialog Research Center, LTI, CMU
Stars: ✭ 279 (+155.96%)
Mutual labels:  chatbot, dialogue-systems
Deeppavlov
An open source library for deep learning end-to-end dialog systems and chatbots.
Stars: ✭ 5,525 (+4968.81%)
Mutual labels:  chatbot, dialogue-systems
Chatbot cn
基于金融-司法领域(兼有闲聊性质)的聊天机器人,其中的主要模块有信息抽取、NLU、NLG、知识图谱等,并且利用Django整合了前端展示,目前已经封装了nlp和kg的restful接口
Stars: ✭ 791 (+625.69%)
Mutual labels:  knowledge-graph, dialogue-systems
Moel
MoEL: Mixture of Empathetic Listeners
Stars: ✭ 38 (-65.14%)
Mutual labels:  chatbot, dialogue-systems
Korean restaurant reservation
Implement korean restaurant reservation dialogue system based on hybrid code network.
Stars: ✭ 73 (-33.03%)
Mutual labels:  chatbot, dialogue-systems
Simple
SimplE Embedding for Link Prediction in Knowledge Graphs
Stars: ✭ 104 (-4.59%)
Mutual labels:  knowledge-graph
Economic audit knowledge graph
经济责任审计知识图谱:网络爬虫、关系抽取、领域词汇判定
Stars: ✭ 98 (-10.09%)
Mutual labels:  knowledge-graph
Query2box
Query2box: Reasoning over Knowledge Graphs in Vector Space Using Box Embeddings
Stars: ✭ 98 (-10.09%)
Mutual labels:  knowledge-graph
Kglab
Graph-Based Data Science: an abstraction layer in Python for building knowledge graphs, integrated with popular graph libraries – atop Pandas, RDFlib, pySHACL, RAPIDS, NetworkX, iGraph, PyVis, pslpython, pyarrow, etc.
Stars: ✭ 98 (-10.09%)
Mutual labels:  knowledge-graph
Chatbot
Русскоязычный чатбот
Stars: ✭ 106 (-2.75%)
Mutual labels:  chatbot
Kge Hake
The code of paper Learning Hierarchy-Aware Knowledge Graph Embeddings for Link Prediction. Zhanqiu Zhang, Jianyu Cai, Yongdong Zhang, Jie Wang. AAAI 2020.
Stars: ✭ 101 (-7.34%)
Mutual labels:  knowledge-graph
Botfuel Dialog
Botfuel SDK to build highly conversational chatbots
Stars: ✭ 96 (-11.93%)
Mutual labels:  chatbot
Whatsappapi
A simple API to integrate chatbots written in Javascript with WhatsApp Web 💬📲 (Store hotfix https://github.com/Zibri/WhatsAppWebApi/blob/master/zstore.js)
Stars: ✭ 94 (-13.76%)
Mutual labels:  chatbot
Scm Biz Suite
供应链中台系统基础版,集成零售管理, 电子商务, 供应链管理, 财务管理, 车队管理, 仓库管理, 人员管理, 产品管理, 订单管理, 会员管理, 连锁店管理, 加盟管理, 前端React/Ant Design, 后端Java Spring+自有开源框架,全面支持MySQL, PostgreSQL, 全面支持国产数据库南大通用GBase 8s,通过REST接口调用,前后端完全分离。
Stars: ✭ 1,310 (+1101.83%)
Mutual labels:  knowledge-graph

一. 任务形式化

这个赛道的目标是设立基于知识图谱的主动聊天任务。知识驱动对话-官网地址

输入:

对话目标g,其中g=START->TOPIC_A->TOPIC_B;表示机器从冷启动状态主动聊到话题A,然后聊到话题B。意味着在该任务中,由机器主动的引导对话;在该任务中,具体的话题包括电影和娱乐人物主体。

相关知识信息M,其中M=f1,f2,...,fn. 包括三类,分别是:话题A的知识信息,话题B的知识信息,话题A和话题B的关联信息。在该任务中,具体的相关知识信息包括电影票房,导演和评价等,以SPO形式表示。也就是(Subject, Predicate, Object),即(实体一,谓词,实体二)。

当前对话序列H=u1,u2,...u(t-1)

输出:

机器回复ut.

二.数据介绍(见官方网站)

三.评价方法

自动评估指标和人工评估指标结合。自动评估指标考虑三个层面的度量,分别是字级别(F1-score),词级别(BLEU)和回复多样性(DISTINCT)。关于回复多样性,还是基于词的计算,不过考察的是生成词的另外一个维度。在参考2中作者这样写道:

distinct-1 and distinct-2 are respectively the number of distinct unigrams and bigrams divided by total number of generated words

四.一般流程

参考PyTorch官方提供的tutorial(见参考4),从seq2seq的角度解决问题的方法是,将多轮对话拆分成平行句。例如,针对当前对话序列H=u1,u2,...u(t-1),可以拆分成t-2组样本,分别是:u1->u2;u2->u3;...;u(t-2)->u(t-1);但是,这样的划分方式存在明显的问题是:句子之间的平滑。这应该是一个问题,但是还没有深入思考过。

五.想法实现

目前,重构了PyTorch官方的Chatbot的Tutorial代码,将各个模块解耦出来,顺带发现了一个Bug。在此基础上,准备实现一个baseline,代码地址

实现思路上,主要参考论文《A Knowledge-Grounded Neural Conversation Model》,包括两个主要部分:

第一:如何将知识图谱的信息嵌入到模型中;

第二: memory机制的实现;

补充:

IJCAI16的文章《Neural Generative Question Answering》的3.2节提出了两种方式用于输入和KG的交互:

第一: Bilinear Model

第二: CNN-based Matching Model

参考:

0.www2018,《DKN: Deep Knowledge-Aware Network for News Recommendation》

1.第六届全国社会媒体处理大会-SMP2017中文人机对话技术评测(ECDT)

包含两个任务:用户意图领域分类和特定域任务型人机对话在线评测

2.《A Diversity-Promoting Objective Function for Neural Conversation Models》

3.《A Persona-Based Neural Conversation Model》

4.chatbot tutorial with pytorch

5.2018JDDC的对话比赛亚军方案

6.机器如何猜你所想?阿里小蜜预测平台解密

7.拆开阿里小蜜的内核,看智能人机交互的实现逻辑

8.一文揭秘!自底向上构建知识图谱全过程

9.为电商而生的知识图谱,如何感应用户需求?

10.健康知识图谱,阿里工程师如何实现?

Note that the project description data, including the texts, logos, images, and/or trademarks, for each open source project belongs to its rightful owner. If you wish to add or remove any projects, please contact us at [email protected].