Interpretable attention-based multi-encoder transformer based QSPR model for assessing toxicity and environmental impact of chemicals.

The rising demand from consumer goods and pharmaceutical industry is driving a fast expansion of newly developed chemicals. The conventional toxicity testing of unknown chemicals is expensive, time-consuming, and raises ethical concerns. The quantitative structure-property relationship (QSPR) is an efficient computational method because it saves time, resources, and animal experimentation. Advances in machine learning have improved chemical analysis in QSPR studies, but the real-world application of machine learning-based QSPR studies was limited by the unexplainable 'black box' feature of the machine learnings. In this study, multi-encoder structure-to-toxicity (S2T)-transformer based QSPR model was developed to estimate the properties of polychlorinated biphenyls (PCBs) and endocrine disrupting chemicals (EDCs). Simplified molecular input line entry systems (SMILES) and molecular descriptors calculated by the Dragon 6 software, were simultaneously considered as input of QSPR model. Furthermore, an attention-based framework is proposed to describe the relationship between the molecular structure and toxicity of hazardous chemicals. The S2T-transformer model achieved the highest R2 scores of 0.918, 0.856, and 0.907 for logarithm of octanol-water partition coefficient (Log KOW), octanol-air partition coefficient (Log KOA), and bioconcentration factor (Log BCF) estimation of PCBs, respectively. Moreover, the attention weights were able to properly interpret the lateral (meta, para) chlorination associated with PCBs toxicity and environmental impact.

Sludge bulking monitoring in industrial wastewater treatment plants through graphical methods: A dynamic graph embedding and Bayesian networks approach.

Sludge bulking is a prevalent issue in wastewater treatment plants (WWTPs) that negatively impacts effluent quality by hindering the normal functioning of treatment processes. To tackle this problem, we propose a novel graph-based monitoring framework that employs advanced graph-based techniques to detect and diagnose sludge bulking events. The proposed framework utilizes historical datasets under normal operating conditions to extract pertinent features and causal relationships between process variables. This enables operators to trigger alarms and diagnose the root cause of the bulking event. Sludge bulking detection is carried out using the dynamic graph embedding (DGE) method, which identifies similarities among process variables in both temporal and neighborhood dependencies. Consequently, the dynamic Bayesian network (DBN) computes the prior and posterior probabilities of a belief, updated at each time step. Variations in these probabilities indicate the potential root cause of the sludge bulking event. The results demonstrate that the DGE outperforms other linear and non-linear feature extraction methods, achieving a detection rate of 99%, zero false alarms, and less than one percent incorrect detections. Additionally, the DBN-based diagnostic method accurately identified the majority of sludge bulking root causes, primarily those resulting from sudden drops in COD concentration, with an accuracy of 98% an improvement of 11% over state-of-the-art techniques.

Smart solutions for urban health risk assessment: A PM2.5 monitoring system incorporating spatiotemporal long-short term graph convolutional network.

Current spatial-temporal early warning systems aim to predict outdoor air quality in urban areas either at short or long temporal horizons. These systems implemented architectures without considering the geographical distribution of each air quality monitoring station, increasing the uncertainty of the forecasting framework. This study developed an integrated spatiotemporal forecasting architecture incorporating an extensive air quality PM2.5 monitoring network and simultaneously forecasts PM2.5 concentrations at all locations, allowing the monitoring of the health risk associated with exposure to these levels. First, this study uses a graph convolutional layer to incorporate the spatial relationship of the neighboring stations at their current state with real-time measurements. Then, it is coupled to a deep learning temporal model to form the long- and short-term time-series graph convolutional network (LSTGraphNet) model, anticipating high pollutant concentration events. This work tested the proposed model with a case study of an existing ambient air quality monitoring network in South Korea. LSTGraphNet model showed prediction performances of PM2.5 at multiple monitoring stations with a mean absolute error (MAE) of 1.82 µg/m3, 4.46 µg/m3, and 4.87 µg/m3 for forecasting horizons of one, three, and 6 h ahead, respectively. Compared to conventional sequential models, this architecture was superior among the state-of-the-art baselines, where the MAE decreased to 41%, respectively. The results of the study showed that the proposed architecture was superior to conventional sequential models and could be used as a tool for decision-making in smart cities by revealing hotspots of higher and lower PM2.5 concentrations in the long term.

Inter-regional multimedia fate analysis of PAHs and potential risk assessment by integrating deep learning and climate change scenarios.

Polycyclic aromatic hydrocarbons (PAHs) are hazardous compounds associated with respiratory disease and lung cancer. Increasing fossil fuel consumption, which causes climate change, has accelerated the emissions of PAHs. However, potential risks by PAHs have not been predicted for Korea, and appropriate PAH regulations under climate change have yet to be developed. This study assesses the potential risks posed by PAHs using climate change scenarios based on deep learning, and a multimedia fugacity model was employed to describe the future fate of PAHs. The multimedia fugacity model describes the dynamics of sixteen PAHs by reflecting inter-regional meteorological transportation. A deep neural network predicts future environmental and economic conditions, and the potential risks posed by PAHs, in the year 2050, using a prediction model and climate change scenarios. The assessment indicates that cancer risks would increase by more than 50%, exceeding the lower risk threshold in the southern and western regions. A mix of strategies for developing PAH regulatory policies highlighted the necessity of increasing PAHs monitoring stations and controlling fossil fuel usage based on the domestic and global conditions under climate change scenarios.

Transfer learning driven sequential forecasting and ventilation control of PM2.5 associated health risk levels in underground public facilities.

Particulate matter with aerodynamic diameter less than 2.5 µm (PM2.5) has become a major public concern in closed indoor environments, such as subway stations. Forecasting platform PM2.5 concentrations is significant in developing early warning systems, and regulating ventilation systems to ensure commuter health. However, the performance of existing forecasting approaches relies on a considerable amount of historical sensor data, which is usually not available in practical situations due to hostile monitoring environments or newly installed equipment. Transfer learning (TL) provides a solution to the scant data problem, as it leverages the knowledge learned from well-measured subway stations to facilitate predictions on others. This paper presents a TL-based residual neural network framework for sequential forecast of health risk levels traced by subway platform PM2.5 levels. Experiments are conducted to investigate the potential of the proposed methodology under different data availability scenarios. The TL-framework outperforms the RNN structures with a determination coefficient (R2) improvement of 42.84%, and in comparison, to stand-alone models the prediction errors (RMSE) are reduced up to 40%. Additionally, the forecasted data by TL-framework under limited data scenario allowed the ventilation system to maintain IAQ at healthy levels, and reduced PM2.5 concentrations by 29.21% as compared to stand-alone network.