STACKFEED: Structured Textual Actor-Critic Knowledge Base Editing with FeedBack
Fuente:
arXiv
Salvato in:
| Autori principali: | , , , , , , , , , |
|---|---|
| Natura: | Preprint |
| Pubblicazione: |
2024
|
| Soggetti: | |
| Accesso online: | |
| Tags: |
Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
|
| _version_ | 1866911244794986496 |
|---|---|
| author | Kirtania, Shashank Gupta, Naman Gupta, Priyanshu Kariya, Krishna Gulwani, Sumit Iyer, Arun Parthasarathy, Suresh Radhakrishna, Arjun Rajamani, Sriram K. Soares, Gustavo |
| author_facet | Kirtania, Shashank Gupta, Naman Gupta, Priyanshu Kariya, Krishna Gulwani, Sumit Iyer, Arun Parthasarathy, Suresh Radhakrishna, Arjun Rajamani, Sriram K. Soares, Gustavo |
| contents | Large Language Models (LLMs) often generate incorrect or outdated information, especially in low-resource settings or when dealing with private data. To address this, Retrieval-Augmented Generation (RAG) uses external knowledge bases (KBs), but these can also suffer from inaccuracies. We introduce STACKFEED, a novel Structured Textual Actor-Critic Knowledge base editing with FEEDback approach that iteratively refines the KB based on expert feedback using a multi-actor, centralized critic reinforcement learning framework. STACKFEED defines a ReACT actor agent on each document to perform structured edits based on document specific targeted instructions. Experimental results showcase that STACKFEED significantly improves KB quality and performance of the RAG system. We evaluate STACKFEED on low-resource programming problems, modified python packaged and factual question-answering tasks. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2410_10584 |
| institution | arXiv |
| publishDate | 2024 |
| record_format | arxiv |
| spellingShingle | STACKFEED: Structured Textual Actor-Critic Knowledge Base Editing with FeedBack Kirtania, Shashank Gupta, Naman Gupta, Priyanshu Kariya, Krishna Gulwani, Sumit Iyer, Arun Parthasarathy, Suresh Radhakrishna, Arjun Rajamani, Sriram K. Soares, Gustavo Artificial Intelligence Machine Learning Multiagent Systems Large Language Models (LLMs) often generate incorrect or outdated information, especially in low-resource settings or when dealing with private data. To address this, Retrieval-Augmented Generation (RAG) uses external knowledge bases (KBs), but these can also suffer from inaccuracies. We introduce STACKFEED, a novel Structured Textual Actor-Critic Knowledge base editing with FEEDback approach that iteratively refines the KB based on expert feedback using a multi-actor, centralized critic reinforcement learning framework. STACKFEED defines a ReACT actor agent on each document to perform structured edits based on document specific targeted instructions. Experimental results showcase that STACKFEED significantly improves KB quality and performance of the RAG system. We evaluate STACKFEED on low-resource programming problems, modified python packaged and factual question-answering tasks. |
| title | STACKFEED: Structured Textual Actor-Critic Knowledge Base Editing with FeedBack |
| topic | Artificial Intelligence Machine Learning Multiagent Systems |
| url | https://arxiv.org/abs/2410.10584 |