TaskBot是一个基于深度学习、机器学习、规则系统的任务驱动型多轮对话框架，欢迎star、contribute。100行代码，实现你的chatbot。

• 多轮跨场景语义继承
• 模块化组件，面向快速的2次定制化开发任务
• 多套替代方案，系统冷启动及热更新。

• Query Expandsion| dictionary based, similarity based, context based
• Spelling error correction
• segment

• Bag of Words:|tfidf
• Network|word2vec[4], glove[3], EMlo[2]
• Topic Model| LSI, LDA, LSA[5]

• Named entity recognization| BiRNN-CRF[6,7], Fuzzy&Keyword Match
• Entity Normalization|Rule, Keyword Match
• Intent Classification| FastText[8, 9], TextCNN, Multi-Round-LSTM, Pattern Match
• Emotion recognization

• Dialog state tracker
• Action Policy| High confidence, Rule, Reinforcement Learning
• record

• Retrieval based| Text similarity rank, Metric Learning, rule based, Keyword Match
• Knowlage Graph based[9, 10, 11]|
• Generation Model| seq2seq

[1] github.com/deepmipt/DeepPavlov

[2] EMol:Deep contextualized word embeddings

[3] GloVe: Global Vectors for Word Representation

[4] Word2vec: Distributed Representations of Words and Phrases and their Compositionality

[5] Topic Model: LDA, LSI, LSA

[6] BiLSTM-CRF: Empower Sequence Labeling with Task-Aware Neural Language Model

[7] A BiLSTM-CRF inplement with PyTorch

[8] FastText: Compressing text classification models

[9] Nickel, M., Murphy, K., Tresp, V., & Gabrilovich, E. A Review of Relational Machine Learning for Knowledge Graphs.

[10] Socher, R., Chen, D., Manning, C. D., & Ng, A. (2013). Reasoning with neural tensor networks for knowledge base completion. In Advances in Neural Information Processing Systems (pp. 926-934).

[11] Bordes, A., Usunier, N., Garcia-Duran, A., Weston, J., & Yakhnenko, O. (2013). Translating embeddings for modeling multi-relational data. In Advances in Neural Information Processing Systems (pp. 2787-2795).

TaskBot is licensed under the GNU GENERAL PUBLIC LICENSE VERSION 3.0