著者

Kentaro Kanamori Takuya Takagi Ken Kobayashi Hiroki Arimura

出版者

The Japanese Society for Artificial Intelligence

雑誌

人工知能学会論文誌 (ISSN:13460714)

巻号頁・発行日

vol.36, no.6, pp.C-L44_1-12, 2021-11-01 (Released:2021-11-01)

参考文献数

51

---

Post-hoc explanation methods for machine learning models have been widely used to support decision-making. Counterfactual Explanation (CE), also known as Actionable Recourse, is one of the post-hoc explanation methods that provides a perturbation vector that alters the prediction result obtained from a classifier. Users can directly interpret the perturbation as an “action” to obtain their desired decision results. However, actions extracted by existing methods often become unrealistic for users because they do not adequately consider the characteristics corresponding to the data distribution, such as feature-correlations and outlier risk. To suggest an executable action for users, we propose a new framework of CE, which we refer to as Distribution-Aware Counterfactual Explanation (DACE), that extracts a realistic action by evaluating its reality on the empirical data distribution. Here, the key idea is to define a new cost function based on the Mahalanobis distance and the local outlier factor. Then, we propose a mixed-integer linear optimization approach to extracting an optimal action by minimizing the defined cost function. Experiments conducted on real datasets demonstrate the effectiveness of the proposed method compared with existing CE methods.

著者

Umihiko Kaneko Daisuke Hachinohe Ken Kobayashi Hidemasa Shitan Keijiro Mitsube Azusa Furugen Takeshi Kawamura Ryuji Koshima Tsutomu Fujita

出版者

International Heart Journal Association

雑誌

International Heart Journal (ISSN:13492365)

巻号頁・発行日

vol.61, no.5, pp.1059-1069, 2020-09-29 (Released:2020-09-29)

参考文献数

27

被引用文献数

8 7

---

Because of its rigidity and non-steerability, the presence of a horizontal aortic root poses a major anatomical issue during transcatheter aortic valve replacement (TAVR) with Evolut self-expanding valve. Previous studies have elucidated the difficulties of coaxial implantation of the self-expanding valve in patients with horizontal aorta, often resulting in increased complications and a lower device success rate. To date, most patients with extremely horizontal aorta (aortic root angle ≥ 70°) have been excluded from major TAVR clinical trials. Therefore, available data on TAVR with Evolut in this challenging anatomy are limited, and standardized treatment strategies and clinical results remain unknown. Herein, we report a clinical case series of TAVR with Evolut in extremely horizontal aorta. Among seven patients (aged 80-92 years; STS score, 12.6% ± 7.9%) who underwent TAVR with Evolut system, aortic root angle ranged from 71° to 83° (mean, 75.1°± 4.5°). All patients achieved device success with dedicated strategies and were clinically stable at 3-month follow-up. None of the patients had more than mild paravalvular leakage (PVL) at any point during follow-up.Complications in three patients included complete atrioventricular block requiring a permanent pacemaker implantation, cerebral infarction because of atrial fibrillation 3 days after TAVR, and cardiac tamponade requiring pericardiocentesis. In this case series, Evolut self-expanding TAVR in extremely horizontal aorta was effective and feasible with a high device success rate. Based on anatomical features, some dedicated strategies majorly contribute to the success of this procedure. Large-scale multicenter studies are required to confirm our findings.