Coder Social home page Coder Social logo

tc-bot's Introduction

End-to-End Task-Completion Neural Dialogue Systems

参考 文献 End-to-End Task-Completion Neural Dialogue SystemsA User Simulator for Task-Completion Dialogues. 本文档描述了如何运行仿真和不同对话代理(基于规则,命令行,强化学习),更多代理和用户模拟设置方法在文献的Recipe章节。

内容

数据

所有的数据都存放在该文件下: ./src/deep_dialog/data

  • 电影知识库
    movie_kb.1k.p : 94%(for user_goals_first_turn_template_subsets.v1.p)
    movie_kb.v2.p : 36%(for user_goals_first_turn_template_subsets.v1.p)

  • 用户目标
    user_goals_first_turn_template.v2.p --- user goals extracted from the first user turn
    user_goals_first_turn_template.part.movie.v1.p --- a subset of user goals [Please use this one, the upper bound success rate on movie_kb.1k.json is 0.9765.]

  • NLG规则模板
    dia_act_nl_pairs.v6.json :用户模拟器和代理的一些预定义NLG规则模板

  • Intent分类
    dia_acts.txt

  • Slot分类
    slot_set.txt

参数

基础设置

--agt:代理Id
--usr: 用户(或模拟器)Id
--max_turn: 对话最大轮数
--episodes: 对话迭代次数
--slot_err_prob: slot错分概率
--slot_err_mode: slot错分为哪个mode
--intent_err_prob: intent错分概率

数据设置

--movie_kb_path:代理方面电影的kb路径
--goal_file_path: 用户目标路径

模型设置

--dqn_hidden_size: DQN代理隐藏层层数t
--batch_size: DQN训练的batch大小
--simulation_epoch_size: 每一次迭代,对话仿真次数
--warm_start: use rule policy to fill the experience replay buffer at the beginning
--warm_start_epochs: 热启动运行对话数量

运行设置

--run_mode: 0 (NL)运行模式; 1(Dia_Act)debug模式; 2(Dia_Act and NL)debug模式; 3(training或者predict)非运行模式
--act_level: 0(Dia_Act级别用户模拟器); 1(NL级别用户模拟器)
--auto_suggest: 0 (no auto_suggest); 1(auto_suggest)
--cmd_input_mode: 0(输入方式NL); 1(输入方式Dia_Act). (这个参数只针对代理模式为AgentCmd模式时设置)

其他

--write_model_dir:写入模型的目录
--trained_model_path: 训练RL代理模型的目录,也是预测时加载模型的目录.

--learning_phase: train/test/all, 默认是all。拆分用户目标集为训练集和测试集,不要全部拆分; 我们引入一些随机因子,We introduce some randomness at the first sampled user action, even for the same user goal, the generated dialogue might be different.

运行对话代理

主程序run.py (1)初始化 Agent、User、NLU、NLG、对话管理DialogManager、对话参数设置。
(2)run_episodes()生成每轮对话,如果agt == 9 warm_start_simulation(),迭代episode_over, reward = dialog_manager.next_turn(),如果agt=9,没有trained_model_path,则训练网络模型并保存。

RequestBasicsAgent代理

python run.py --agt 5(RequestBasicsAgent代理) --usr 1(使用模拟器) --max_turn 40
	      --episodes 150
	      --movie_kb_path ./deep_dialog/data/movie_kb.1k.p
	      --goal_file_path ./deep_dialog/data/user_goals_first_turn_template.part.movie.v1.p
	      --intent_err_prob 0.00
	      --slot_err_prob 0.00
	      --episodes 500
	      --act_level 0

AgentCmd代理

NL输入

python run.py --agt 0(AgentCmd代理) --usr 1(使用模拟器) --max_turn 40
	      --episodes 150
	      --movie_kb_path ./deep_dialog/data/movie_kb.1k.p
	      --goal_file_path ./deep_dialog/data/user_goals_first_turn_template.part.movie.v1.p
	      --intent_err_prob 0.00
	      --slot_err_prob 0.00
	      --episodes 500
	      --act_level 0(Dia_Act级别用户模拟器)
	      --run_mode 0(NL)运行模式
	      --cmd_input_mode 0

Dia_Act输入

python run.py --agt 0(AgentCmd代理) --usr 1(使用模拟器) --max_turn 40
	      --episodes 150
	      --movie_kb_path ./deep_dialog/data/movie_kb.1k.p 
	      --goal_file_path ./deep_dialog/data/user_goals_first_turn_template.part.movie.v1.p
	      --intent_err_prob 0.00
	      --slot_err_prob 0.00
	      --episodes 500
	      --act_level 0(Dia_Act级别用户模拟器)
	      --run_mode 0 (NL)运行模式
	      --cmd_input_mode 1

End2End RL代理(DQN代理)

没有NLU和NLG模块训练End2End RL代理(NLU模块模拟噪声)

python run.py --agt 9(DQN代理) --usr 1(使用模拟器) --max_turn 40
	      --movie_kb_path ./deep_dialog/data/movie_kb.1k.p
	      --dqn_hidden_size 80
	      --experience_replay_pool_size 1000
	      --episodes 500
	      --simulation_epoch_size 100
	      --write_model_dir ./deep_dialog/checkpoints/rl_agent/
	      --run_mode 3  (training或者predict)非运行模式
	      --act_level 0(Dia_Act级别用户模拟器)
	      --slot_err_prob 0.00
	      --intent_err_prob 0.00
	      --batch_size 16
	      --goal_file_path ./deep_dialog/data/user_goals_first_turn_template.part.movie.v1.p
	      --warm_start 1
	      --warm_start_epochs 120

有NLU和NLG模块训练End2End RL代理

python run.py --agt 9 (DQN) --usr 1(使用模拟器) --max_turn 40
	      --movie_kb_path ./deep_dialog/data/movie_kb.1k.p
	      --dqn_hidden_size 80
	      --experience_replay_pool_size 1000
	      --episodes 500
	      --simulation_epoch_size 100
	      --write_model_dir ./deep_dialog/checkpoints/rl_agent/
	      --run_mode 3  (training或者predict)非运行模式
	      --act_level 1(NL级别用户模拟器)
	      --slot_err_prob 0.00
	      --intent_err_prob 0.00
	      --batch_size 16
	      --goal_file_path ./deep_dialog/data/user_goals_first_turn_template.part.movie.v1.p
	      --warm_start 1
	      --warm_start_epochs 120

基于N轮对话测试Rl代理:

python run.py --agt 9 (DQN)--usr 1(使用模拟器) --max_turn 40
	      --movie_kb_path ./deep_dialog/data/movie_kb.1k.p
	      --dqn_hidden_size 80(DQN隐藏层层数)
	      --experience_replay_pool_size 1000
	      --episodes 300 
	      --simulation_epoch_size 100
	      --write_model_dir ./deep_dialog/checkpoints/rl_agent/
	      --slot_err_prob 0.00
	      --intent_err_prob 0.00
	      --batch_size 16
	      --goal_file_path ./deep_dialog/data/user_goals_first_turn_template.part.movie.v1.p
	      --trained_model_path ./deep_dialog/checkpoints/rl_agent/noe2e/agt_9_478_500_0.98000.p
	      --run_mode 3(training或者predict)非运行模式

Evaluation

为了评估代理的性能,三个重要指标:成功率、平均价值和平均轮数。

  1. 画学习曲线 python draw_learning_curve.py --result_file ./deep_dialog/checkpoints/rl_agent/noe2e/agt_9_performance_records.json
  2. 在Excel表格中画学习曲线

Reference

主要参考文献

@inproceedings{li2017end,
  title={End-to-End Task-Completion Neural Dialogue Systems},
  author={Li, Xuijun and Chen, Yun-Nung and Li, Lihong and Gao, Jianfeng and Celikyilmaz, Asli},
  booktitle={Proceedings of The 8th International Joint Conference on Natural Language Processing},
  year={2017}
}

@article{li2016user,
  title={A User Simulator for Task-Completion Dialogues},
  author={Li, Xiujun and Lipton, Zachary C and Dhingra, Bhuwan and Li, Lihong and Gao, Jianfeng and Chen, Yun-Nung},
  journal={arXiv preprint arXiv:1612.05688},
  year={2016}
}

tc-bot's People

Contributors

xiul-msr avatar xiaoqian19940510 avatar yvchen avatar xjli avatar shangyusu avatar

Watchers

 avatar  avatar

Recommend Projects

  • React photo React

    A declarative, efficient, and flexible JavaScript library for building user interfaces.

  • Vue.js photo Vue.js

    🖖 Vue.js is a progressive, incrementally-adoptable JavaScript framework for building UI on the web.

  • Typescript photo Typescript

    TypeScript is a superset of JavaScript that compiles to clean JavaScript output.

  • TensorFlow photo TensorFlow

    An Open Source Machine Learning Framework for Everyone

  • Django photo Django

    The Web framework for perfectionists with deadlines.

  • D3 photo D3

    Bring data to life with SVG, Canvas and HTML. 📊📈🎉

Recommend Topics

  • javascript

    JavaScript (JS) is a lightweight interpreted programming language with first-class functions.

  • web

    Some thing interesting about web. New door for the world.

  • server

    A server is a program made to process requests and deliver data to clients.

  • Machine learning

    Machine learning is a way of modeling and interpreting data that allows a piece of software to respond intelligently.

  • Game

    Some thing interesting about game, make everyone happy.

Recommend Org

  • Facebook photo Facebook

    We are working to build community through open source technology. NB: members must have two-factor auth.

  • Microsoft photo Microsoft

    Open source projects and samples from Microsoft.

  • Google photo Google

    Google ❤️ Open Source for everyone.

  • D3 photo D3

    Data-Driven Documents codes.