PURPOSES : In this study, model-agnostic methods are applied for interpreting machine learning models, such as the feature global effect, the importance of a feature, the joint effects of features, and explaining individual predictions. METHODS : Model-agnostic global interpretation techniques, such as partial dependence plot (PDP), accumulated local effect (ALE), feature interaction (H-statistics), and permutation feature importance, were applied to describe the average behavior of a machine learning model. Moreover, local model-agnostic interpretation methods, individual conditional expectation curves (ICE), local surrogate models (LIME), and Shapley values were used to explain individual predictions. RESULTS : As global interpretations, PDP and ALE-Plot demonstrated the relationship between a feature and the prediction of a machine learning model, where the feature interaction estimated whether one feature depended on the other feature, and the permutation feature importance measured the importance of a feature. For local interpretations, ICE exhibited how changing a feature changes the interested instance’s prediction, LIME explained the relationship between a feature and the instance’s prediction by replacing the machine model with a locally interpretable model, and Shapley values presented how to fairly contribute to the instance’s prediction among the features. CONCLUSIONS : Model-agnostic methods contribute to understanding the general relationship between features and a prediction or debut a model from the global and/or local perspective, securing the reliability of the learning model.

Density and sound speed contrasts (g and h , respectively), and swimming angle were measured for sandfish (Arctoscopus japonicus) without swimbladder. The density contrast was measured by the volume displacement method while the sound speed contrast was measured by the acoustic measurements of travel time (time-of-flight method). The swimming angle was measured by dividing it into daytime, nighttime, daytime feeding and nighttime feeding. The g was 1.001 to 1.067 with an average (± standard deviation) of 1.032 (± 0.017), and the h was 1.007 to 1.022 with an average (± standard deviation) of 1.015 (± 0.003). The swimming angles (mean ± standard deviation) were 16.8 ± 10.3° during the daytime, 1.9 ± 12.3° during the nighttime, 30.2 ± 12.6° in the daytime feeding and 35.0 ± 13.2° in the nighttime feeding. These results will provide important parameters input to calculate theoretical scattering models for estimating the acoustic target strength of sandfish.

Rare earth elements, which are important components of motors, are in high demand and thus constantly get more expensive. This tendency is driven by the growth of the electric vehicle market, as well as environmental issues associated with rare-earth metal manufacturing. TC 298 of the ISO manages standardization in the areas of rare-earth recycling, measurement, and sustainability. Korea, a resource-poor country, is working on international standardization projects that focus on recycling and encouraging the domestic adoption of international standards. ITU-T has previously issued recommendations regarding the recycling of rare-earth metals from e-waste. ISO TC 298 expands on the previous recommendations and standards for promoting the recycling industry. Recycling-related rare earth standards and drafts covered by ISO TC 298, as well as Korea’s strategies, are reviewed and discussed in this article.

수처리 및 의약바이오 분야에서 유효물질 분리에 활용되고 있는 알루미나 중공사 분리막은 얇은 두께로 인해 취 급 및 적용시 쉽게 파괴되는 단점이 있기 때문에 분리막의 강도를 100 MPa 이상으로 향상시키기 위한 연구가 필요하다. 본 연구에서는 나노입자의 함량을 0, 1, 3, 5 wt%로 증가시켰을 때 제조된 중공사 분리막의 특성을 평가하였다. 그 결과, 나노입 자의 함량이 증가함에 따라 중공사 분리막의 강도는 79 MPa에서 115 MPa로 증가하였으며, 밀도는 1.76 g/m3에서 1.88 g/m3 으로 증가하였고 기공률과 평균기공크기는 각각 51%에서 48%로, 416 nm에서 352 nm로 감소한 것을 확인하였다. 스폰지구 조가 발달하고 스폰지구조의 기공크기가 향상된 알루미나 중공사 분리막은 100 MPa 이상으로 기계적 강도가 향상되었으며, 약 100000 GPU의 높은 질소 투과도 및 약 3000 L/m2h의 높은 물 투과도를 나타내었다. 따라서, γ-알루미나 나노입자를 소 결조제로 첨가하는 것은 α-알루미나 중공사 분리막의 기계적 강도를 효과적으로 증진시키고 높은 투과성능을 유지할 수 있 는 매우 유효한 방법임을 확인하였다.

본 연구에서는 용매로 침액되었을 시 전반적으로 더 균일한 상전이법 기반 평막 제조 시 주로 사용되는 부직포 지지체의 영향을 분석하였다. 도프용액의 점도가 낮을 경우 용액이 부직포층으로 쉽게 침투하여 불균일한 막이 형성되는 것 을 확인하였으며, 이를 방지하기 위해 부직포층을 유기용매로 침액하는 기법을 도입하였다. 부직포층이 유기 분리막이 생성 되는 것을 확인하였으며, 수투과 및 용매투과율도 향상하는 것을 알 수 있었다. 부직포 침액의 영향은 낮은 점도에서 확연하 게 나타났으며, 고분자용액의 점도가 높은 경우 침액 여부에 관계없이 동일한 성능을 얻을 수 있었다.