Kanamori 2010. “Statistical outlier detection using direct density ratio estimation” • 使えるデータ:正常データと正常と異常が混じったテストデータ. • 半教師あり異常検知ともいえる? • Hidoらの論文では線形モデルによって密度比を推定していた. Inlier-based Outlier Detection 19
We expect that D3RE is applicable to other applications, such as • 因果推論; • 変分ベイズ法; • 敵対的生成ネットワーク. n 推薦システムにも応用できて,すでにThe Web conferenceに採択. n ノンパラメトリック操作変数を,深層密度比推定を用いて行った. • 近日中にarXivに公開. 結論 23
K., and Schölkopf, B. Covariate shift by kernel mean matching. Dataset Shift in Machine Learning, 131-160 (2009), 01 2009. n Hastie, T., Tibshirani, R., and Friedman, J. The elements of statistical learning: data mining, inference and prediction. Springer, 2001. n Kanamori, T., Hido, S., and Sugiyama, M. A least-squares approach to direct importance estimation. Journal of Machine Learning Research, 10(Jul.):1391–1445, 2009. n Kiryo, R., Niu, G., du Plessis, M. C., and Sugiyama, M. Positive-unlabeled learning with non-negative risk estimator. In NeurIPS, 2017. n Sugiyama, M., Suzuki, T., and Kanamori, T. Density ratio matching under the bregman divergence: A unified frame-work of density ratio estimation. Annals of the Institute of Statistical Mathematics, 64, 10 2011b Reference 24