With the advent of GDPR, the domain of explainable AI and model interpretability has gained added impetus. Methods to extract and communicate visibility into decision-making models have become legal requirement. Two specific types of explanations, contrastive and counterfactual have been identified as suitable for human understanding. In this paper, we propose a model agnostic method and its systemic implementation to generate these explanations using shapely additive explanations (SHAP). We discuss a generative pipeline to create contrastive explanations and use it to further to generate counterfactual datapoints. This pipeline is tested and discussed on the IRIS, Wine Quality & Mobile Features dataset. Analysis of the results obtained follows.
This study assessed exposure by the roadside to highly toxic particle-bound polycyclic aromatic hydrocarbons (PPAHs) that are known to adsorb preferentially on fine particles, aerodynamic diameter (d ≤ 1 μm). The real-time air quality measurements were conducted in March, April, and May 2015 in Kanpur at two busy roadside locations: one outside IIT Kanpur main gate, IG, and another by a residential area, M3. The locations show varying land use type and traffic density. Higher averaged daily concentrations of PM, PM, and PM were observed at IG (PM 700-800 μg/m) owing to nature and high density of traffic, and occurrence of biomass burning nearby. Statistically significant relation (R > 90%, p < 0.05) between PM and PM highlights the influence of mobile sources on particle load at IG. IG, the busier location, had higher daily averaged concentration of aggregate PPAHs (104 ng/m) than M3 which is located near a residential area (38 ng/m). In contrast, the higher average daily value of PC/DC ratio (mass per unit surface area of PPAHs on nanoparticles) at M3 (4.87 ng/mm) than at IG (4.08 ng/mm) suggests that PAHs of greater mass occur on particles at M3. Finer particles are known to adsorb pollutants of a larger mass that are likely to be more toxic in case of PAHs suggest that ambient air at M3 has more toxicity potential. However, this inference is not based on chemical analyses, and chemical characteristics must also be taken into account for the detailed assessment of health risk. The multiple path dosimetry model (MPPD-v3.04) reveals that the 99.02% of PM inhaled, 77.01% of PM and 34.54% of PM are deposited in the outermost (head) region of the human respiratory tract.
Hyperbolic cooling towers have become the design standard for all natural-draft cooling towers because of their structural strength and minimum usage of material. The hyperbolic shape is particularly suited to cooling tower construction as the wide base provides a large space for the water and cooling system. As the tower widens out at the top, it supports the turbulent mixing as the heated air makes contact with the atmospheric air. Hyperbolic cooling tower is a tall structure with shells subjected to dead load and wind load. In absence of ground motion, wind becomes the major factor. In this study, 2 major models are studied with I and V column support. Each model is further divided into 2 models, i.e. one with SHELL element and another with SOLID element. All models were modelled and analyzed in ANSYS. The wind loads on these cooling tower have been calculated in the form of pressure by using the circumferentially distributed design wind pressure coefficients as given in IS: 11504 - 1985 code along with the design wind pressures at different levels as per IS: 875 (Part 3) - 1987 code. The analysis has been carried out using 8 noded shell element (SHELL281), 8 noded solid element (SOLID185) and 20 noded solid element (SOLID 186).
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.