NLP Wiki

This page is read only. You can view the source, but not change it. Ask your administrator if you think this is wrong.
====== Task Oriented Dialog ======

===== Papers =====
  * [[https://arxiv.org/pdf/2109.14739.pdf|Su et al 2021 - Multi-Task Pre-Training for Plug-and-Play Task-Oriented Dialogue System]] Introduces PPTOD model
  * Discovering Latent Structure
    * [[https://aclanthology.org/N19-1178.pdf|Shi et al 2019 - Unsupervised Dialog Structure Learning]]
    * [[https://arxiv.org/pdf/2009.08115.pdf|Zhang et al 2020 - A Probabilistic End-To-End Task-Oriented Dialog Model with Latent Belief States towards Semi-Supervised Learning]]

===== Few-Shot Task Oriented Dialog =====
  * **Schema-Guided Dialog (SGD)**:[[https://arxiv.org/pdf/1909.05855.pdf|2019 - Towards Scalable Multi-domain Conversational Agents: The Schema-Guided Dialogue Dataset]]

===== Zero Shot Task Oriented Dialog =====
  * [[https://arxiv.org/pdf/2212.09939.pdf|Zhao et al 2022 - ANYTOD: A Programmable Task-Oriented Dialog System]]

===== Prompting and In-Context Learning =====
See also Response Generation, below.

  * [[https://arxiv.org/pdf/2110.08118.pdf|Madotto et al 2021 - Few-Shot Bot: Prompt-Based Learning for Dialogue Systems]]
  * [[https://arxiv.org/pdf/2203.08568.pdf|Hu et al 2022 - In-Context Learning for Few-Shot Dialogue State Tracking]]
  * [[https://arxiv.org/pdf/2307.01453.pdf|King & Flanigan 2023 - Diverse Retrieval-Augmented In-Context Learning for Dialogue State Tracking]]

===== Semi-supervised Methods =====
  * [[https://aclanthology.org/2022.aacl-short.37.pdf|Zhang et al 2022 - CSS: Combining Self-training and Self-supervised Learning for Few-shot Dialogue State Tracking]]

===== Domain Adaptation =====
  * [[https://ieeexplore.ieee.org/document/9016052?denied=|Yu et al 2020 - DAFA: Dialog System Domain Adaptation With a Filter and an Amplifier]]

===== Dialog State Tracking =====
"Dialog state tracking (DST) extracts users’ goals in task-oriented dialog systems, where dialog states are often represented in terms of a set of slot-value pairs" (from [[https://arxiv.org/pdf/2106.00291v1.pdf|Dai 2021]]).

  * MultiWoz
    * [[https://github.com/budzianowski/multiwoz|MultiWOZ]] Also a data to text generation dataset.
    * Paper v1.0: [[https://arxiv.org/pdf/1807.06517.pdf|Ramadan et al 2018 - Large-Scale Multi-Domain Belief Tracking with Knowledge Sharing]].
    * Paper v2.0: [[https://arxiv.org/pdf/1810.00278.pdf|Budzianowski et al 2018 - MultiWOZ - A Large-Scale Multi-Domain Wizard-of-Oz Dataset for Task-Oriented Dialogue Modelling]]
    * See also [[https://paperswithcode.com/sota/multi-domain-dialogue-state-tracking-on-1|Multi-domain Dialogue State Tracking on MULTIWOZ 2.1]]
    * STAR: [[https://arxiv.org/pdf/2101.09374.pdf|Ye et al 2021 - Slot Self-Attentive Dialogue State Tracking]] SOTA as of 2022
    * [[https://arxiv.org/pdf/2104.00773.pdf|Ye et al 2021 - MultiWOZ 2.4: A Multi-Domain Task-Oriented Dialogue Dataset with Essential Annotation Corrections to Improve State Tracking Evaluation]] Contains a comparison of SOTA methods on MultiWoz 2.1 and 2.4.
  * [[https://arxiv.org/pdf/2106.00291.pdf|Dai et al 2021 - Preview, Attend and Review: Schema-Aware Curriculum Learning for Multi-Domain Dialog State Tracking]]

===== Response Generation =====
  * **With LLMs**
    * [[https://arxiv.org/pdf/2110.08118.pdf|Madotto et al 2021 - Few-Shot Bot: Prompt-Based Learning for Dialogue Systems]]
    * [[https://arxiv.org/pdf/2302.11520.pdf|Li et al 2023 - Guiding Large Language Models via Directional Stimulus Prompting]]

===== Related Pages =====
  * [[Chatbots]]
  * [[Dialog]]