arXivDaily arXiv每日学术速递 周一至周五更新
arXiv周末暂无论文更新,休息一下吧,周末愉快~~
arXiv 2602.16008cs.SDcs.AIcs.CLcs.LG
收藏

MAEB:大规模音频嵌入基准

MAEB: Massive Audio Embedding Benchmark

Carleton University(卡尔顿大学);Durham University(杜伦大学);Stanford University(斯坦福大学);Aarhus University(奥胡斯大学);Indian Institute of Technology, Kharagpur(印度理工学院,卡里格普);Harvard University(哈佛大学)
Adnan El Assadi, Isaac Chung, Chenghao Xiao, Roman Solomatin, Animesh Jha, Rahul Chand, Silky Singh, Kaitlyn Wang, Ali Sartaz Khan, Marc Moussa Nasser, Sufen Fo… 展开作者

Adnan El Assadi, Isaac Chung, Chenghao Xiao, Roman Solomatin, Animesh Jha, Rahul Chand, Silky Singh, Kaitlyn Wang, Ali Sartaz Khan, Marc Moussa Nasser, Sufen Fong, Pengfei He, Alan Xiao, Ayush Sunil Munot, Aditya Shrivastava, Artem Gazizov, Niklas Muennighoff, Kenneth Enevoldsen

AI总结:

MAEB是一个涵盖多语言音频任务的基准,揭示了不同模型在音频与语言任务上的性能差异,展示了音频编码器在跨模态任务中的表现关联性。

摘要:

我们介绍了大规模音频嵌入基准(MAEB),一个涵盖语音、音乐、环境声音以及跨模态音频-文本推理的30项任务的大型基准,覆盖100多种语言。我们评估了50多种模型,发现没有单一模型在所有任务上占据主导:对比音频-文本模型在环境声音分类(如ESC50)中表现优异,但在多语言语音任务(如SIB-FLEURS)中得分接近随机;而语音预训练模型则表现出相反的模式。所有模型在聚类任务上仍然面临挑战,即使最佳表现的模型也只能取得有限的成果。我们观察到,擅长声学理解的模型在语言任务上表现较差,反之亦然。我们还发现,音频编码器在MAEB上的表现与其在音频大语言模型中的表现高度相关。MAEB源自MAEB+任务集。MAEB旨在在减少评估成本的同时保持任务多样性,并整合到MTEB生态系统中,实现文本、图像和音频模态的统一评估。我们发布了MAEB以及所有98项任务,以及代码和排行榜,详见https://github.com/embeddings-benchmark/mteb。

英文摘要:

We introduce the Massive Audio Embedding Benchmark (MAEB), a large-scale benchmark covering 30 tasks across speech, music, environmental sounds, and cross-modal audio-text reasoning in 100+ languages. We evaluate 50+ models and find that no single model dominates across all tasks: contrastive audio-text models excel at environmental sound classification (e.g., ESC50) but score near random on multilingual speech tasks (e.g., SIB-FLEURS), while speech-pretrained models show the opposite pattern. Clustering remains challenging for all models, with even the best-performing model achieving only modest results. We observe that models excelling on acoustic understanding often perform poorly on linguistic tasks, and vice versa. We also show that the performance of audio encoders on MAEB correlates highly with their performance when used in audio large language models. MAEB is derived from MAEB+, a collection of 98 tasks. MAEB is designed to maintain task diversity while reducing evaluation cost, and it integrates into the MTEB ecosystem for unified evaluation across text, image, and audio modalities. We release MAEB and all 98 tasks along with code and a leaderboard at https://github.com/embeddings-benchmark/mteb.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑