boilerbot
Official Open-Source Implementation of BoilerBot: A Reliable Task-Oriented Chatbot Enhanced with Large Language Models.
Science Score: 44.0%
This score indicates how likely this project is to be science-related based on various indicators:
-
✓CITATION.cff file
Found CITATION.cff file -
✓codemeta.json file
Found codemeta.json file -
✓.zenodo.json file
Found .zenodo.json file -
○DOI references
-
○Academic publication links
-
○Committers with academic emails
-
○Institutional organization owner
-
○JOSS paper metadata
-
○Scientific vocabulary similarity
Low similarity (12.9%) to scientific vocabulary
Keywords
Repository
Official Open-Source Implementation of BoilerBot: A Reliable Task-Oriented Chatbot Enhanced with Large Language Models.
Basic Info
- Host: GitHub
- Owner: jinensetpal
- License: apache-2.0
- Language: Python
- Default Branch: main
- Homepage: https://www.amazon.science/alexa-prize/proceedings/boilerbot-a-reliable-task-oriented-chatbot-enhanced-with-large-language-models
- Size: 387 KB
Statistics
- Stars: 7
- Watchers: 1
- Forks: 0
- Open Issues: 0
- Releases: 0
Topics
Metadata Files
README.md
BoilerBot: A Reliable Task-Oriented Chatbot Enhanced with Large Language Models
This is an official re-implementation of the task-oriented conversational chatbot Purdue's BoilerBot team developed for the Alexa Prize TaskBot Challenge 2.
Architecture
Setup
Ensure ports 8001-8003 & 27017 are unused. Then, from the root of the repository, run:
bash
docker compose up
If you want to use cache from your host machine, you will have to change paths under volumes in compose.yml if XDG_CACHE_HOME != /home/<user> or if you are on Windows / OSX.
Running
I'm working on developing a web frontend using APL Viewhost to render the APLs. In the meantime, the deployment is only available as an API. You can send POST requests to http://localhost:8001/chat with query: str, uid: str|None, apl:true|false, or visit http://localhost:8001/docs and press 'Try it out' for a friendlier UI.
The security key for requests to the GPU server is insecure_placeholder.
UID
The UID enables conversation persistence. To start a new conversation, leave the field blank. To continue a conversation, use the UID returned from the previous request.
Changes from the Original Deployment
Our TaskBot was developed using Amazon's Cobot (conversational bot) framework. The source for the framework, alongside Amazon's datasets and APIs are closed source and couldn't be a part of our public implementation. Therefore: 1. The Cobot framework was rewritten using an API base instead of the Lambda function that was originally used. 2. Task-relevant off-the-shelf models were taken from HuggingFace or manually developed to replace Amazon APIs. 3. Community datasets (WikiHow-Dataset, RecipeNLG) were used as the task databases. 4. LLM modules are not functional. Will post an update as soon as I find the time and compute.
Owing to the less robust neural approaches, noisy datasets, and a framework re-implementation, some amount of reliability (e.g. intent classification) and efficiency (e.g. sequential NLU) of the system is lost. It is also why some design decisions are hacky hotfixes. If you run into a bug, please submit an issue.
Citation
If you use our work, you can cite it using:
bibtex
@inproceedings{boilerbot,
title={BoilerBot: A Reliable Task-Oriented Chatbot Enhanced with Large Language Models},
author={Hu, Yifei and Setpal, Jinen and Zhang, Damin and Zietek, Jacob and Lambert, Jack and Gonzalez, Rey Alex and Rayz, Julia},
booktitle={2nd Proceedings of Alexa Prize TaskBot Challenge (Alexa Prize 2023)},
year={2023},
url = {https://www.amazon.science/alexa-prize/proceedings/boilerbot-a-reliable-task-oriented-chatbot-enhanced-with-large-language-models}
}
Owner
- Name: Jinen Setpal
- Login: jinensetpal
- Kind: user
- Location: Purdue University
- Company: @DAGsHub
- Website: jinensetpal.github.io
- Twitter: 48bitmachine
- Repositories: 8
- Profile: https://github.com/jinensetpal
Research in Interpretable Network Optimization
Citation (CITATION.cff)
cff-version: 1.2.0
title: >-
BoilerBot: A Reliable Task-Oriented Chatbot Enhanced with
Large Language Models
message: >-
If you use this software, please cite it using the
metadata from this file.
type: inproceedings
authors:
- given-names: Yifei
family-names: Hu
email: hu381@purdue.edu
affiliation: Purdue University
- given-names: Jinen
family-names: Setpal
email: jsetpal@cs.purdue.edu
affiliation: Purdue University
orcid: 'https://orcid.org/0000-0002-8959-0451'
- given-names: Damin
family-names: Zhang
email: zhan4060@purdue.edu
affiliation: Purdue University
- given-names: Jacob
family-names: Zietek
email: jzietek@cs.purdue.edu
affiliation: Purdue University
- given-names: Jack
family-names: Lambert
email: lamberjb@purdue.edu
affiliation: Purdue University
- given-names: Alex
family-names: Gonzalez
email: gonza736@purdue.edu
affiliation: Purdue University
- given-names: Julia
family-names: Rayz
email: jtaylor1@purdue.edu
affiliation: Purdue University
identifiers:
- type: url
value: >-
https://www.amazon.science/alexa-prize/proceedings/boilerbot-a-reliable-task-oriented-chatbot-enhanced-with-large-language-models
description: Publication
- type: url
value: >-
https://assets.amazon.science/8c/03/80c814a749f58e73a1aeda2ff282/boilerbot-tb2-final-2023.pdf
description: PDF
repository-code: 'https://dagshub.com/jinensetpal/boilerbot/'
abstract: >-
This paper outlines the design and deployment of
BoilerBot: a task-oriented multi- modal conversational
agent developed for the Alexa Prize TaskBot 2 competition.
BoilerBot features flexible response generation,
leveraging large language models (LLMs) to enable
adaptable user experiences. We discuss our novel
contributions towards advancing state-of-the-art
task-oriented conversational agents, highlighting
user-facing challenges and proposing fault-tolerant,
iterative solutions for carefully guided workflows that
enable Alexa users to maximize BoilerBot’s functionality.
license: Apache-2.0
GitHub Events
Total
- Watch event: 1
Last Year
- Watch event: 1
Committers
Last synced: over 1 year ago
Top Committers
| Name | Commits | |
|---|---|---|
| Jinen Setpal | j****8@g****m | 3 |
Issues and Pull Requests
Last synced: 12 months ago
All Time
- Total issues: 0
- Total pull requests: 0
- Average time to close issues: N/A
- Average time to close pull requests: N/A
- Total issue authors: 0
- Total pull request authors: 0
- Average comments per issue: 0
- Average comments per pull request: 0
- Merged pull requests: 0
- Bot issues: 0
- Bot pull requests: 0
Past Year
- Issues: 0
- Pull requests: 0
- Average time to close issues: N/A
- Average time to close pull requests: N/A
- Issue authors: 0
- Pull request authors: 0
- Average comments per issue: 0
- Average comments per pull request: 0
- Merged pull requests: 0
- Bot issues: 0
- Bot pull requests: 0
Top Authors
Issue Authors
Pull Request Authors
Top Labels
Issue Labels
Pull Request Labels
Dependencies
- python 3.11-slim build
- python 3.11-slim build
- python 3.11-slim build
- Requests ==2.31.0
- attrs ==23.2.0
- fastapi ==0.85.1
- injector ==0.20.1
- numpy ==1.25.1
- pymongo ==4.6.1
- sentence_transformers ==2.2.2
- thefuzz ==0.20.0
- transformers ==4.30.2
- uvicorn ==0.23.0
- Flask ==3.0.0
- gunicorn ==21.2.0
- Requests ==2.31.0
- chromadb ==0.3.29
- fastapi ==0.85.1
- numerizer ==0.2.3
- numpy ==1.25.1
- pandas ==2.0.3
- pydantic ==1.10.11
- scikit_learn ==1.3.0
- sentence_transformers ==2.2.2
- spacy ==3.6.0
- torch ==2.0.1
- transformers ==4.30.2
- uvicorn ==0.23.0