BlenderBot 3: a deployed conversational agent that continually learns to responsibly engage

Kurt Shuster, Jing Xu, Mojtaba Komeili, Da Ju, Eric Michael Smith +13 more
8/5/2022
cs.CLcs.AI

No official code found. Best implementation: facebookresearch/ParlAI (10,618 stars, MIT, last commit 2023-11-03).

Abstract

We present BlenderBot 3, a 175B parameter dialogue model capable of open-domain conversation with access to the internet and a long-term memory, and having been trained on a large number of user defined tasks. We release both the model weights and code, and have also deployed the model on a public web page to interact with organic users. This technical report describes how the model was built (architecture, model and training scheme), and details of its deployment, including safety mechanisms. Human evaluations show its superiority to existing open-domain dialogue agents, including its predecessors (Roller et al., 2021; Komeili et al., 2022). Finally, we detail our plan for continual learning using the data collected from deployment, which will also be publicly released. The goal of this research program is thus to enable the community to study ever-improving responsible agents that learn through interaction.

View on arXivView PDF

Code Implementations(1)

facebookresearch/ParlAICommunity100%
MIT

A framework for training and evaluating AI models on a variety of openly available dialogue datasets.

10,6182,043C++, CSSApr 24, 20172 years agoMIT

Similar Papers

Dialogue Agents 101: A Beginner's Guide to Critical Ingredients for Designing Effective Conversational Systems

84% match2023

LiveChat: A Large-Scale Personalized Dialogue Dataset Automatically Constructed from Live Streaming

83% match1 repo382023

DialogStudio: Towards Richest and Most Diverse Unified Dataset Collection for Conversational AI

82% match1 repo5272023

Behavior Modeling for Training-free Building of Private Domain Multi Agent System

82% match2025

Conversation Chronicles: Towards Diverse Temporal and Relational Dynamics in Multi-Session Conversations

82% match1 repo1092023

Improving Open Language Models by Learning from Organic Interactions

82% match2023

Cite this paper

@article{shuster2022blenderbot,
  title  = {BlenderBot 3: a deployed conversational agent that continually learns to responsibly engage},
  author = {Kurt Shuster and Jing Xu and Mojtaba Komeili and Da Ju and Eric Michael Smith and Stephen Roller and Megan Ung and Moya Chen and Kushal Arora and Joshua Lane and Morteza Behrooz and William Ngan and Spencer Poff and Naman Goyal and Arthur Szlam and Y-Lan Boureau and Melanie Kambadur and Jason Weston},
  year   = {2022},
  eprint = {2208.03188},
  archivePrefix = {arXiv},
  url    = {http://arxiv.org/abs/2208.03188v3}
}

Add a badge to your README

Shows whether official code is linked for this paper.

Code of Paper badge
[![Code of Paper](https://codeofpaper.com/badge/2208.03188.svg)](https://codeofpaper.com/paper/2208.03188)

Discussion