Development and application of machine learning models for prediction of soil available cadmium based on soil properties and climate features

Identifying the key influencing factors in soil available cadmium (Cd) is crucial for preventing the Cd accumulation in the food chain. However, current experimental methods and traditional prediction models for assessing available Cd are time-consuming and ineffective. In this study, machine learning (ML) models were developed to investigate the intricate interactions among soil properties, climate features, and available Cd, aiming to identify the key influencing factors. The optimal model was obtained through a combination of stratified sampling, Bayesian optimization, and 10-fold cross-validation. It was further explained through the utilization of permutation feature importance, 2D partial dependence plot, and 3D interaction plot. The findings revealed that pH, surface pressure, sensible heat net flux and organic matter content significantly influenced the Cd accumulation in the soil. By utilizing historical soil surveys and climate change data from China, this study predicted the spatial distribution trend of available Cd in the Chinese region, highlighting the primary areas with heightened Cd activity. These areas were primarily located in the eastern, southern, central, and northeastern China. This study introduces a novel methodology for comprehending the process of available Cd accumulation in soil. Furthermore, it provides recommendations and directions for the remediation and control of soil Cd pollution.

A machine learning-based predictive model of causality in orthopaedic medical malpractice cases in China

PURPOSE

To explore the feasibility and validity of machine learning models in determining causality in medical malpractice cases and to try to increase the scientificity and reliability of identification opinions.

METHODS

We collected 13,245 written judgments from PKULAW.COM, a public database. 963 cases were included after the initial screening. 21 medical and ten patient factors were selected as characteristic variables by summarising previous literature and cases. Random Forest, eXtreme Gradient Boosting (XGBoost) and Light Gradient Boosting Machine (LightGBM) were used to establish prediction models of causality for the two data sets, respectively. Finally, the optimal model is obtained by hyperparameter tuning of the six models.

RESULTS

We built three real data set models and three virtual data set models by three algorithms, and their confusion matrices differed. XGBoost performed best in the real data set, with a model accuracy of 66%. In the virtual data set, the performance of XGBoost and LightGBM was basically the same, and the model accuracy rate was 80%. The overall accuracy of external verification was 72.7%.

CONCLUSIONS

The optimal model of this study is expected to predict the causality accurately.

Perspectives on incorporating expert feedback into model updates

Machine learning (ML) practitioners are increasingly tasked with developing models that are aligned with non-technical experts' values and goals. However, there has been insufficient consideration of how practitioners should translate domain expertise into ML updates. In this review, we consider how to capture interactions between practitioners and experts systematically. We devise a taxonomy to match expert feedback types with practitioner updates. A practitioner may receive feedback from an expert at the observation or domain level and then convert this feedback into updates to the dataset, loss function, or parameter space. We review existing work from ML and human-computer interaction to describe this feedback-update taxonomy and highlight the insufficient consideration given to incorporating feedback from non-technical experts. We end with a set of open questions that naturally arise from our proposed taxonomy and subsequent survey.

An Application of Machine Learning to Logistics Performance Prediction: An Economics Attribute-Based of Collective Instance

In this work, a machine learning application was constructed to predict the logistics performance index based on economic attributes. The prediction procedure employs both linear and non-linear machine learning algorithms. The macroeconomic panel dataset is used in this investigation. Furthermore, it was combined with the microeconomic panel dataset obtained through the data envelopment analysis method for evaluating financial efficiency. The procedure was implemented in six ASEAN member countries. The non-linear algorithm of an artificial neural network performed best on the complex pattern of a collective instance of these six countries, followed by the penalized linear of the Ridge regression method. Due to the limited amount of training data for each country, the artificial neural network prediction procedure is only applicable to the datasets of Singapore, Malaysia, and the Philippines. Ridge regression fits the Indonesia, Thailand and Vietnam datasets. The results provide precise trend forecasting. Macroeconomic factors are driving up the logistics performance index in Vietnam in 2020. Malaysia logistics performance is influenced by the logistics business's financial efficiency. The results at the country level can be used to track, improve, and reform the country's short-term logistics and supply chain policies. This can bring significant gains in national logistics and supply chain capabilities, as well as support for global trade collaboration, all for the long-term development of the region.

Explainable machine learning for public policy: Use cases, gaps, and research directions

Explainability is highly desired in machine learning (ML) systems supporting high-stakes policy decisions in areas such as health, criminal justice, education, and employment. While the field of explainable ML has expanded in recent years, much of this work has not taken real-world needs into account. A majority of proposed methods are designed with generic explainability goals without well-defined use cases or intended end users and evaluated on simplified tasks, benchmark problems/datasets, or with proxy users (e.g., Amazon Mechanical Turk). We argue that these simplified evaluation settings do not capture the nuances and complexities of real-world applications. As a result, the applicability and effectiveness of this large body of theoretical and methodological work in real-world applications are unclear. In this work, we take steps toward addressing this gap for the domain of public policy. First, we identify the primary use cases of explainable ML within public policy problems. For each use case, we define the end users of explanations and the specific goals the explanations have to fulfill. Finally, we map existing work in explainable ML to these use cases, identify gaps in established capabilities, and propose research directions to fill those gaps to have a practical societal impact through ML. The contribution is (a) a methodology for explainable ML researchers to identify use cases and develop methods targeted at them and (b) using that methodology for the domain of public policy and giving an example for the researchers on developing explainable ML methods that result in real-world impact.

A global taxonomy of interpretable AI: unifying the terminology for the technical and social sciences

Since its emergence in the 1960s, Artificial Intelligence (AI) has grown to conquer many technology products and their fields of application. Machine learning, as a major part of the current AI solutions, can learn from the data and through experience to reach high performance on various tasks. This growing success of AI algorithms has led to a need for interpretability to understand opaque models such as deep neural networks. Various requirements have been raised from different domains, together with numerous tools to debug, justify outcomes, and establish the safety, fairness and reliability of the models. This variety of tasks has led to inconsistencies in the terminology with, for instance, terms such as interpretable, explainable and transparent being often used interchangeably in methodology papers. These words, however, convey different meanings and are "weighted" differently across domains, for example in the technical and social sciences. In this paper, we propose an overarching terminology of interpretability of AI systems that can be referred to by the technical developers as much as by the social sciences community to pursue clarity and efficiency in the definition of regulations for ethical and reliable AI development. We show how our taxonomy and definition of interpretable AI differ from the ones in previous research and how they apply with high versatility to several domains and use cases, proposing a-highly needed-standard for the communication among interdisciplinary areas of AI.

Economic development, weather shocks and child marriage in South Asia: A machine learning approach

Globally, 21 percent of young women are married before their 18th birthday. Despite some progress in addressing child marriage, it remains a widespread practice, in particular in South Asia. While household predictors of child marriage have been studied extensively in the literature, the evidence base on macro-economic factors contributing to child marriage and models that predict where child marriage cases are most likely to occur remains limited. In this paper we aim to fill this gap and explore region-level indicators to predict the persistence of child marriage in four countries in South Asia, namely Bangladesh, India, Nepal and Pakistan. We apply machine learning techniques to child marriage data and develop a prediction model that relies largely on regional and local inputs such as droughts, floods, population growth and nightlight data to model the incidence of child marriages. We find that our gradient boosting model is able to identify a large proportion of the true child marriage cases and correctly classifies 77% of the true marriage cases, with a higher accuracy in Bangladesh (92% of the cases) and a lower accuracy in Nepal (70% of cases). In addition, all countries contain in their top 10 variables for classification nighttime light growth, a shock index of drought over the previous and the last two years and the regional level of education, suggesting that income shocks, regional economic activity and regional education levels play a significant role in predicting child marriage. Given the accuracy of the model to predict child marriage, our model is a valuable tool to support policy design in countries where household-level data remains limited.

Crowd Anomaly Detection via Spatial Constraints and Meaningful Perturbation

Crowd anomaly detection is a practical and challenging problem to computer vision and VideoGIS due to abnormal events’ rare and diverse nature. Consequently, traditional methods rely on low-level reconstruction in a single image space, easily affected by unimportant pixels or sudden variations. In addition, real-time detection for crowd anomaly detection is challenging, and localization of anomalies requires other supervision. We present a new detection approach to learn spatiotemporal features with the spatial constraints of a still dynamic image. First, a lightweight spatiotemporal autoencoder has been proposed, capable of real-time image reconstruction. Second, we offer a dynamic network to obtain a compact representation of video frames in motion, reducing false-positive anomaly alerts by spatial constraints. In addition, we adopt the perturbation visual interpretation method for anomaly visualization and localization to improve the credibility of the results. In experiments, our results provide competitive performance across various scenarios. Besides, our approach can process 52.9–63.4 fps in anomaly detection, making it practical for crowd anomaly detection in video surveillance.

Achieving Equity with Predictive Policing Algorithms: A Social Safety Net Perspective

Whereas using artificial intelligence (AI) to predict natural hazards is promising, applying a predictive policing algorithm (PPA) to predict human threats to others continues to be debated. Whereas PPAs were reported to be initially successful in Germany and Japan, the killing of Black Americans by police in the US has sparked a call to dismantle AI in law enforcement. However, although PPAs may statistically associate suspects with economically disadvantaged classes and ethnic minorities, the targeted groups they aim to protect are often vulnerable populations as well (e.g., victims of human trafficking, kidnapping, domestic violence, or drug abuse). Thus, determining how to enhance the benefits of PPA while reducing bias through better management is important. In this paper, we propose a policy schema to address this issue. First, after clarifying relevant concepts, we examine major criticisms of PPAs and argue that some of them should be addressed. If banning AI or making it taboo is an unrealistic solution, we must learn from our errors to improve AI. We next identify additional challenges of PPAs and offer recommendations from a policy viewpoint. We conclude that the employment of PPAs should be merged into broader governance of the social safety net and audited publicly by parliament and civic society so that the unjust social structure that breeds bias can be revised.

Mitigating belief projection in explainable artificial intelligence via Bayesian teaching

State-of-the-art deep-learning systems use decision rules that are challenging for humans to model. Explainable AI (XAI) attempts to improve human understanding but rarely accounts for how people typically reason about unfamiliar agents. We propose explicitly modelling the human explainee via Bayesian teaching, which evaluates explanations by how much they shift explainees' inferences toward a desired goal. We assess Bayesian teaching in a binary image classification task across a variety of contexts. Absent intervention, participants predict that the AI's classifications will match their own, but explanations generated by Bayesian teaching improve their ability to predict the AI's judgements by moving them away from this prior belief. Bayesian teaching further allows each case to be broken down into sub-examples (here saliency maps). These sub-examples complement whole examples by improving error detection for familiar categories, whereas whole examples help predict correct AI judgements of unfamiliar cases.