基於音樂周邊資訊之音樂自動標籤系統優化

Improving Music Auto-Tagging System by Exploiting Music Context

林奕勳(Yi-Hsun Lin)

指導教授 ： 陳宏銘

國立臺灣大學/電機資訊學院/電信工程學研究所/碩士(2019年)

https://doi.org/10.6342/NTU201800616

摘要

音樂自動標籤的成效取決於訓練資料的品質。然而實際上，在人工標註的訓練資料中，歌曲與標籤的連結可能有誤（假陽性）或是有所遺漏（假陰性）。在此論文中，我們提出了「代價敏感的標籤傳遞學習法」，來改善自動標籤系統。首先，我們利用音樂周邊資訊來篩選出彼此相似的歌曲，並在它們之間傳遞標籤。然後，我們再把傳遞的標籤和原始的標籤一起用來最佳化自動標籤模型。另外，為了提升抗噪能力，我們把代價敏感的機制結合進模型的損失函數，以調整陽性連結相對於陰性連結的權重。接著，將所提出的方法用於訓練三個自動標籤模型，以檢測其成效。這三個模型分別為：CNN、CRNN和SampleCNN。其中，我們採用百萬歌曲資料集（Million Song Dataset）作為訓練資料，並使用四種不同的音樂周邊資訊：歌手、歌單、標籤以及聆聽者來衡量歌曲的相似性。實驗結果顯示：一、所提出的方法能夠成功的提升這三個模型的效能。二、代價敏感的損失函數有助於減少遺漏標籤的影響。三、歌手資訊比其他三種周邊資訊更適合做為標籤傳遞的媒介。

關鍵字

音樂自動標籤 ； 音樂檢索系統 ； 標籤傳遞 ； 音樂周邊資訊 ； 代價敏感學習

並列摘要

The performance of music auto-tagging depends on the quality of training data. In practice, the links between songs and tags in the manually labeled training data can be incorrect (false positive) or missing (false negative). In this paper, we propose a cost-sensitive tag propagation learning method to improve auto-tagging. Specifically, we exploit music context to determine similar songs and propagate tags between them. Both propagated tags and original tags are used to optimize the auto-tagging models, and cost-sensitivity is incorporated into the loss function to enhance the robustness by adjusting the weight of relevant (positive) links with respect to irrelevant (negative) links. The proposed method is tested on three auto-tagging models: 2D-CNN, CRNN, and SampleCNN. The Million Song Dataset is used for training, and four music contexts, artist, playlist, tag, and listener, are used for song similarity measurement. The experimental results show 1) The proposed method can successfully improve the performance of the three auto-tagging models, 2) The cost-sensitive loss function helps reduce the impact of missing tags, and 3) The artist music context is more powerful for tag propagation than the other three music contexts.

並列關鍵字

music auto-tagging ； music information retrieval ； tag propagation ； cost-sensitive learning ； music contextual information

參考文獻

[1] O. Kanishcheva and G. Angelova, “A pipeline approach to image auto-tagging refinement,” in Proc. BCI, Craiova, Romania, 2015: ACM, p. 9.

[2] T. Bertin-Mahieux, D. Eck, F. Maillet, and P. Lamere, “Autotagger: A model for predicting social tags from acoustic features on large music databases,” Journal of New Music Research, vol. 37, no. 2, pp. 115–135, 2008.

[3] E. Law, B. Settles, and T. Mitchell, “Learning to tag from open vocabulary labels,” in Proc. ECML PKDD, Barcelona, Spain, 2010: Springer, pp. 211–226.

[4] P. Lamere, “Social tagging and music information retrieval,” Journal of New Music Research, vol. 37, no. 2, pp. 101–114, 2008.

[5] Y.-H. Yang, D. Bogdanov, P. Herrera, and M. Sordo, “Music retagging using label propagation and robust principal component analysis,” in Proc. WWW, Lyon, France, 2012: ACM, pp. 869–876.