• Corpus ID: 240158311

Neural, Neural Everywhere: Controlled Generation Meets Scaffolded, Structured Dialogue∗

@inproceedings{Chi2021NeuralNE,
  title={Neural, Neural Everywhere: Controlled Generation Meets Scaffolded, Structured Dialogue∗},
  author={Ethan A. Chi and Chetanya Rastogi and Alexander Iyabor and Hari Sowrirajan and Avanika Narayan and Ashwin Paranjape},
  year={2021}
}
In this paper, we present the second iteration of Chirpy Cardinal, an open-domain dialogue agent developed for the Alexa Prize SGC4 competition. Building on the success of the SGC3 Chirpy, we focus on improving conversational flexibility, initiative, and coherence. We introduce a variety of methods for controllable neural generation, ranging from prefix-based neural decoding over a symbolic scaffolding, to pure neural modules, to a novel hybrid infilling-based method that combines the best of… 
Neural-Augmented Retrieval for Open-Domain Dialogue Systems
TLDR
This work replaced ChirpyCardinal’s GloVe-based retrieval method with ColBERT, experimented with using BART or T5 for contextualized template infilling, and evaluated these schemes through the relevance retrieval output and quality of the infilled template.
Stylistic Response Generation by Controlling Personality Traits and Intent
TLDR
Automatic annotation schemes are employed to enrich the corpora with noisy estimates of personality and intent annotations, and the impact of using such features as control codes for response generation using automatic evaluation metrics, ablation studies and human judgement is assessed.
Empathetic Conversational Systems: A Review of Current Advances, Gaps, and Opportunities
TLDR
A critical review of this rapidly growing field by examining the current advances in four dimensions of conceptual empathy models and frameworks, the adopted empathy-related concepts, the datasets and algorithmic techniques developed, and the evaluation strategies.

References

SHOWING 1-10 OF 60 REFERENCES
Neural Generation Meets Real People: Towards Emotionally Engaging Mixed-Initiative Conversations
TLDR
This work presents Chirpy Cardinal, an open-domain dialogue agent, as a research platform for the 2019 Alexa Prize competition, capable of talking knowledgeably about a wide variety of topics, as well as chatting empathetically about ordinary life.
Wizard of Wikipedia: Knowledge-Powered Conversational agents
TLDR
The best performing dialogue models are able to conduct knowledgeable discussions on open-domain topics as evaluated by automatic metrics and human evaluations, while a new benchmark allows for measuring further improvements in this important research direction.
DIALOGPT : Large-Scale Generative Pre-training for Conversational Response Generation
TLDR
It is shown that conversational systems that leverage DialoGPT generate more relevant, contentful and context-consistent responses than strong baseline systems.
Advancing the State of the Art in Open Domain Dialog Systems through the Alexa Prize
TLDR
Several key open-ended problems such as conversational speech recognition, open domain natural language understanding, commonsense reasoning, statistical dialog management, and dialog evaluation are addressed.
Can You Put it All Together: Evaluating Conversational Agents’ Ability to Blend Skills
TLDR
This work investigates several ways to combine models trained towards isolated capabilities, ranging from simple model aggregation schemes that require minimal additional training, to various forms of multi-task training that encompass several skills at all training stages.
Recipes for Building an Open-Domain Chatbot
TLDR
Human evaluations show the best models outperform existing approaches in multi-turn dialogue on engagingness and humanness measurements, and the limitations of this work are discussed by analyzing failure cases of the models.
Human-like informative conversations: Better acknowledgements using conditional mutual information
TLDR
The results demonstrate that systems mimicking human conversational traits improve overall quality and more broadly illustrate the utility of linguistic principles in improving dialogue agents.
Hierarchical Neural Story Generation
TLDR
This work collects a large dataset of 300K human-written stories paired with writing prompts from an online forum that enables hierarchical story generation, where the model first generates a premise, and then transforms it into a passage of text.
Topical-Chat: Towards Knowledge-Grounded Open-Domain Conversations
TLDR
Topical-Chat is introduced, a knowledge-grounded humanhuman conversation dataset where the underlying knowledge spans 8 broad topics and conversation partners don’t have explicitly defined roles, to help further research in opendomain conversational AI.
Open-Domain Conversational Agents: Current Progress, Open Problems, and Future Directions
TLDR
The properties of continual learning, providing engaging content, and being well-behaved are discussed -- and how to measure success in providing them and their recommendations to the community are discussed.
...
...