• Task-oriented (in closed domain) vs Non-task-oriented (i.e., chat bots [8], in open domain) [1]
• Generative vs retrieval-based
• Question answering tells us how to do and dialogue system tells us the steps to do it successfully.

• ASR -> NLU -> DM -> NLG

• NLU: maps the utterance into semantic slots

  • Domain identification: text classification
  • User domain specific intent detection: text classification (We can incorporate preceding text information during perform classification, see [7])
  • Slot filling: sequence labeling
    • Parameters of the INTENT
    • Since slots are intent-specific, we can Joint learning of intent identification and slot filling
  • Industrial
    • To avoid OOV
      • Character-based classification and sequence labeling
      • Fasttext sub n-gram
      • Combine CNN or LSTM on character embedding and word embedding
    • New intent or new domain
      • Zero-shot learning

• DM: Dialogue Management

  • Dialogue State Tracking: estimates the goal at every turn. A probability distribution over each slot for each turn.

  • Policy Learning: generates next action

    • Supervised learning
    • Reinforcement learning
      • Entity: agent and environment
      • Rule: agent takes action and environment gives reward and state
      • Policy: a rule that the agent should follow to select actions given the current state

  • Models

    • Rule-based is popular in production system [8]
      • Frame based
      • Finite state machine
        • Node is action
        • Edge is slot status
      • Agenda based
    • Markov Decision Process
    • Neural network

• NLG: convert action to utterance

  • Action - template
  • Hybrid model [8]
    • If the user dialogue actions can be found in the predefined sentence templates, the **template-based NLG ** is applied; otherwise, the utterance is generated by the model-based NLG

• End-to-End_Task-Completion_Neural_Dialogue_Systems, A_User_Simulator_for_Task-Completion_Dialogues
• see report of Dr.Che

• Given pre-defined responses, retrieval based model predict one response given current input context [2]
• Single/Multi turn response matching
• Two steps [3]:
  • retrieval top-k response candidates by directly matching
  • reranking and give best candidate by incorporing context
• Multi-turn
  • Sequential matching network, and other state-of-the-art models
• AIML for pattern matching

• Generative new responses from scratch [2]
• Multi-turn
  • HRED transfered from query suggestion, see industrail application
  • VHRED
  • Building End-To-End Dialogue Systems Using Generative Hierarchical Neural Network Models
  • Context-sensitive ACL 2018
• Avoid generic responses
  • Mutual information
• Cs224s slides

• Alime

• How to use question to increase Chatbot's interactivity ?

• Minlie Huang
• Rui Yan
• Nan Duan

• Ubuntu Dialogue Corpus v2.0
• A New Multi-Turn, Multi-Domain, Task-Oriented Dialogue Dataset
• Dialog State Tracking Challenge
• Douban Conversation Corpus
• 2017第六届全国社会媒体处理大会之中文人机对话技术评测

• 阿里巴巴高级算法专家李永彬：小蜜智能对话开发平台的技术探索与场景实践
  • Few-shot learning (小样本学习) for intent classification
    • Few-shot learning = meta learning + supervised learning)
    • For solving any-intent few-sample problem
    • Few-Shot text classification with induction network
      • Capsules + dynamic routing
  • DM
    • RL
  • User simulator
    • Testing dialogue pipeline

• wxpy
• wxBot
• WeRoBot: weChat bot
• Awesome-Chatbot gives a list of public repositories about chatbot

• Baidu UNIT
• IP梦工厂
• 图灵机器人
• 小i机器人
• 海知智能
• XiaoIce
• Pandorabots
• Wit.ai
• Rasa
  • Rasa NLU
    • Intent classification
    • Entity recognition
  • Rasa Core
  • Rasa UI
  • tensorflow_embedding in NLU from StarSpace: Embed All The Things!
  • Chinese chatbot example
• DialogFlow
• ChatterBot
• Rokid
• DeepPavlov

• Rasa is a framework, we can define new component to replace official one
• Steps:
  • Generate data with Chatito
  • Train rasa-nlu, define new component if necessary
    • Increase nlu accuracy with button attribute

[1] A Survey on Dialogue Systems: Recent Advances and New Frontiers
[2] Deep Learning for Chatbots part 1, part 2
[3] https://c.m.163.com/news/l/180148.html?spssid=fba792c95ad60299db5435a91da37e10&spsw=1&spss=other.
[4] A Network-based End-to-End Trainable Task-oriented Dialogue System, and implementation, Ali Xiaomi
[5] End-to-end LSTM-based dialog control optimized with supervised and reinforcement learning, Hybrid Code Networks: practical and efficient end-to-end dialog control with supervised and reinforcement learning, and implementation
[6] Learning End-to-End Goal-Oriented Dialog from Facebook, and implementation
[7] Sequential Short-Text Classification with Recurrent and Convolutional Neural Networks
[8] “Chitty-Chitty-Chat Bot”: Deep Learning for Conversational AI
[9] End-to-End Task-Completion Neural Dialogue Systems, implementation