Tuesday 9:10-12:10 (CS Building R102)
Event Date Description Course Material
Lecture 02/21/2017 Introduction
Modular Dialogue System
Slides: [Course Logistics] [Modular Dialogue System]
Special 02/28/2017 Break
Milestone 03/03/2017 Milestone 0 Due
Special 03/07/2017 Break
Special 03/14/2017 Break
Lecture 03/21/2017 Ontology/Database and Language Understanding Suggested Readings:
1. [LSTM for Utterance Classification]
2. [RNN for Slot Filling]
3. [Joint Frame Semantic Parsing]
Slides: [Ontology] [Language Understanding]
Lecture 03/28/2017 Dialogue State Tracking Slides: [Dialogue State Tracking]
Special 04/04/2017 Break
Milestone 04/08/2017 Milestone 1 Due Slides: [Milestone 1]
Lecture 04/11/2017 Dialogue Policy Optimization 1 / User Simulation Suggested Readings:
1. [User Simulator]
Slides: [Policy Optimization 1] [User Simulation]
Guest Lecture 04/18/2017 Automatic Speech Recognition
Guest Lecture by Prof. Hung-Yi Lee
Slides: []
Lecture 04/25/2017 Dialogue Policy Optimization 2 Suggested Readings:
1. [DQN Based Dialogue Agent]
2. [RL Based Dialogue Agent]
3. [Policy Gradient Based Dialogue Agent]
Slides: [Policy Optimization 2]
Lecture 05/02/2017 Natural Language Generation
Microsoft API Tutorial by Eric
Suggested Readings:
1. [Semantically-Conditioned LSTM-Based NLG]
2. [Neural Conversational Model]
Slides: [Natural Language Generation] [Course Schedule]
Milestone 05/06/2017 Milestone 2 Due Slides: [Milestone 2]
Special 05/09/2017 TA Discussion
Presentation 05/16/2017 Progress Presentation & Demonstration
Lecture 05/23/2017 Natural Language Generation
Slides: [Natural Language Generation] [Evaluation]
Special 05/30/2017 Break
Milestone 06/03/2017 Milestone 3 Due (21:00) Slides: [Milestone 3]
Lecture 06/06/2017 Multimodality, Robustness & Flexibility
Guest Talk by NVIDIA
Slides: [Multimodality] [System Improvement] [NVIDIA Talk]
Lecture 06/13/2017 Recent Trends & Final Discussion Suggested Readings:
1. [End-to-End Joint NLU and DM]
2. [Multi-Turn Language Understanding]
3. [K-SAN: Knowledge-Guided Structural Attention Network]
4. [MUSE: Modularizaing Unsupervised Sense Embedding]
5. [App Behavior Pattern for LU]
6. [LU Impact Analysis]
Slides: [Joint NLU+DM] [Multi-Turn NLU] [Contextual NLU+DM]
Presentation 06/20/2017 Final Poster Presentation Slides: [Poster Exhibition Tips] [poster template]