Audio Captioning RAG via Generative Pair-to-Pair Retrieval with Refined Knowledge Base

Đã lưu trong:
Chi tiết về thư mục
Xuất bản năm:arXiv.org (Dec 19, 2024), p. n/a
Tác giả chính: Choi Changin
Tác giả khác: Lim Sungjun, Rhee Wonjong
Được phát hành:
Cornell University Library, arXiv.org
Những chủ đề:
Truy cập trực tuyến:Citation/Abstract
Full text outside of ProQuest
Các nhãn: Thêm thẻ
Không có thẻ, Là người đầu tiên thẻ bản ghi này!

MARC

LEADER 00000nab a2200000uu 4500
001 3117168571
003 UK-CbPIL
022 |a 2331-8422 
035 |a 3117168571 
045 0 |b d20241219 
100 1 |a Choi Changin 
245 1 |a Audio Captioning RAG via Generative Pair-to-Pair Retrieval with Refined Knowledge Base 
260 |b Cornell University Library, arXiv.org  |c Dec 19, 2024 
513 |a Working Paper 
520 3 |a Recent advances in audio understanding tasks leverage the reasoning capabilities of LLMs. However, adapting LLMs to learn audio concepts requires massive training data and substantial computational resources. To address these challenges, Retrieval-Augmented Generation (RAG) retrieves audio-text pairs from a knowledge base (KB) and augments them with query audio to generate accurate textual responses. In RAG, the relevance of the retrieved information plays a crucial role in effectively processing the input. In this paper, we analyze how different retrieval methods and knowledge bases impact the relevance of audio-text pairs and the performance of audio captioning with RAG. We propose generative pair-to-pair retrieval, which uses the generated caption as a text query to accurately find relevant audio-text pairs to the query audio, thereby improving the relevance and accuracy of retrieved information. Additionally, we refine the large-scale knowledge base to retain only audio-text pairs that align with the contextualized intents. Our approach achieves state-of-the-art results on benchmarks including AudioCaps, Clotho, and Auto-ACD, with detailed ablation studies validating the effectiveness of our retrieval and KB construction methods. 
653 |a Impact analysis 
653 |a Audio data 
653 |a Knowledge bases (artificial intelligence) 
653 |a Queries 
653 |a Information retrieval 
653 |a Query processing 
653 |a Ablation 
700 1 |a Lim Sungjun 
700 1 |a Rhee Wonjong 
773 0 |t arXiv.org  |g (Dec 19, 2024), p. n/a 
786 0 |d ProQuest  |t Engineering Database 
856 4 1 |3 Citation/Abstract  |u https://www.proquest.com/docview/3117168571/abstract/embedded/ZKJTFFSVAI7CB62C?source=fedsrch 
856 4 0 |3 Full text outside of ProQuest  |u http://arxiv.org/abs/2410.10913