Conversational AI Reading Materials.

Updated on Dec 18, 2020.

==> Tutorials

Updated on Dec 18, 2020

• Deep Generative Models in IJCAL 2018

• Conversational AI in ACL、SIGIR 2018

• Deep Learning for Dialogue Systems in COLING 2018

• Unified View of Deep Generative Models in ICML 2018

• DRL for Natural Language Processing in ACL 2018

• Deep+Learning+for+ChatBots in EMNLP 2018

• Transfer Learning for NLP released by Sebastian Ruder on Feb 26, 2019

• Zero-shot Learning for CV IN CVPR 2017

• Deep Adversarial Learning for NLP on NAACL, June 2, 2019

• Transfer Learning in NLP in NAACL, June 2, 2019

• Neural Approaches to Conversational AI, a recent survey for dialog systems

• The Bright Future of ACL/NLP in ACL, July 2019.

• Building Dialog System With Less Supervision in West Coast NLP workshop, Sep 7, 2019

• Knowledge-Aware Natural Language Understanding, the PhD thesis of Pradeep Dasigi in CMU, 2019.

• Towards scalable multi-domain conversational agents in the 4th International Workshop on Search-Oriented Conversational AI.

• The Big Bad NLP Database

• Deeper Conversational AI in NeurIPS, Dec 10, 2020

• Interactive Learning for Conversational Understanding by Gokhan Tur from Amazon Alexa AI, 2020 (Stanford NLP Seminar)

• GPT-3-Few-Shot Learning with a Giant Language Model by Melanie Subbiah from OpenAI, 2020 (Stanford NLP Seminar)

==> Deadlines

NIPS 2018 Workshops: Around Octomber 25, 2018

CVPR-2019: November 16, 2018 (including rebuttal)

NAACL-2019: Abastract: December 3, 2018, Full paper: December 10, 2018

KDD-2019: February 3, 2019

ICML-2019: January, 18 2019

IJCAI-2019: February 25, 2019

ACL-2019: March 4, 2019

EMNLP-2019: May 21, 2019

More can be found at the Conference Deadline

==> Books

Updated on Jan 28, 2020

• The Study of Language
• Semantics Third Edition
• Reinforcement Learning for Adaptive Dialogue Systems
• Neural Network Methods in Natural Language Processing

==> Datasets

Updated on Aug 17, 2020

Multilingual:

• Cross-lingual dataset for intent and slot detection, Cross-lingual dialog state tracking dataset. Papers: XLM, XLM-R, mBert, mBart

Intent Detection:

• BANKING77: A single domain intent dataset with 77 intents; CLINC150: A dataset with 15 domains and 150 intents in total, and this dataset is mainly designed for Out-of-Scope (OOS) predictions.

Updated on Sep. 13, 2018.

==> GAN

• SeqGAN - SeqGAN: Sequence Generative Adversarial Nets with Policy Gradient
• MaliGAN - Maximum-Likelihood Augmented Discrete Generative Adversarial Networks
• RankGAN - Adversarial ranking for language generation
• LeakGAN - Long Text Generation via Adversarial Training with Leaked Information
• TextGAN - Adversarial Feature Matching for Text Generation
• GSGAN - GANS for Sequences of Discrete Elements with the Gumbel-softmax Distribution

==> Reinforcement Learning

• (*****) Deep Reinforcement Learning For Sequence to Sequence Models
• Maximum Entropy Inverse Reinforcement Learning
• (*****) Towards Diverse Text Generation with Inverse Reinforcement Learning
• (*****) Reinforcement Learning and Control as Probabilistic Inference-Tutorial and Review
• (*****) Policy gradient methods for reinforcement learning with function approximation
• (*****) Proximal policy optimization algorithms

Change the on-policy gradient method to off-policy gradient method to improve the data efficiency. And use trust region optimization method to search for the better result.

• (*****) Continuous control with deep reinforcement learning

Use the deterministic policy instead of stochastic policy to deal with the continuous action space. The architecture of framework is Actor-Critic.

==> Transfer Learning/Meta-Learning

• (*****) Model-Agnostic Meta-Learning
• A spect-augmented Adversarial Networks for Domain Adaptation
• (*****) Natural Language to Structured Query Generation via Meta-Learning
• (*****) Learning a Prior over Intent via Meta-Inverse Reinforcement Learning

==> Multi-Agent Learning

• Multi-Agent Actor-Critic for Mixed Cooperative-Competitive Environments
• Multi-agent cooperation and the emergence of (natural) language
• (*****)Counterfactual Multi-Agent Policy Gradients (AAAI best paper) ==========

(Application)

==> (Textual) Dialog Generation

• A Hierarchical Latent Structure for Variational Conversation Modeling
• A Hierarchical Latent Variable Encoder-Decoder Model for Generating Dialogues
• (*****) Improving Variational Encoder-Decoders in Dialogue Generation
• (*****) DialogWAE- Multimodal Response Generation with Conditional Wasserstein Auto-Encoder
• Learning Discourse-level Diversity for Neural Dialog Models using Conditional Variational Autoencoders

==> VQA (Visual Question Answering)

• (**) Visual Question Answering- A Survey of Methods and Datasets
• (****) Visual Question Answering as a Meta Learning Task
• Cross-Dataset Adaptation for Visual Question Answering
• Joint Image Captioning and Question Answering
• Learning Answer Embeddings for Visual Question Answering
• Question Answering through Transfer Learning from Large Fine-grained Supervision Data

==> Visual Dialog

• (**) Visual Dialog
• (****) Are You Talking to Me-Reasoned Visual Dialog Generation through Adversarial Learning
• (****)Two can play this Game-Visual Dialog with Discriminative Question Generation
• Zero-Shot Dialog Generation with Cross-Domain Latent Actions
• Adversarial Learning of Task-Oriented Neural Dialog Models
• (*****)Best of Both Worlds-Transferring Knowledge from Discriminative Learning to a Generative Visual Dialog Model
• (*****)Learning Cooperative Visual Dialog Agents with Deep Reinforcement Learning
• (****)Mind Your Language-Learning Visually Grounded Dialog in a Multi-Agent Setting

==> Embodied Question Answering

• (*****)Embodied Question Answering (https://embodiedqa.org/)

==> Video Question Answering

• Open-Ended Long-form Video Question Answering via Adaptive Hierarchical Reinforced Networks.
• Multi-Turn Video Question Answering via Multi-Stream Hierarchical Attention Context Network.
• TVQA-Localized, Compositional Video Question Answering

==> Conversational Information Retrieval

• Dialog-based Interactive Image Retrieval （NIPS2018）
• Improving Search through A3C Reinforcement Learning based Conversational Agent

==> Video Dialog

• A New Dataset: Audio-Visual Scene-Aware Dialog
• End-to-End Audio Visual Scene-Aware Dialog using Multimodal Attention-Based Video Features
• Audio Visual Scene-Aware Dialog (AVSD) Challenge at DSTC7
• More details on the webpage