Biased human decisions have consequential impacts across various domains, yielding unfair treatment of individuals and resulting in suboptimal outcomes for organizations and society. In recognition of this fact, organizations regularly design and deploy interventions aimed at mitigating these biases. However, measuring human decision biases remains an important but elusive task. Organizations are frequently concerned with mistaken decisions disproportionately affecting one group. In practice, however, this is typically not possible to assess due to the scarcity of a gold standard: a label that indicates what the correct decision would have been. In this work, we propose a machine learning-based framework to assess bias in human-generated decisions when gold standard labels are scarce. We provide theoretical guarantees and empirical evidence demonstrating the superiority of our method over existing alternatives. This proposed methodology establishes a foundation for transparency in human decision-making, carrying substantial implications for managerial duties, and offering potential for alleviating algorithmic biases when human decisions are used as labels to train algorithms.

In this study, we presented two innovative methods, which are Threshold-Based Derivative (TBD) and Adaptive Derivative Peak Detection(ADPD), that enhance the accuracy of Learning models for blood glucose estimation using Mid-Infrared (MIR) spectroscopy. In these presented methods, we have enhanced the model's accuracy by integrating absorbance data and its differentiation with critical points. Blood samples were characterized with Fourier Transform Infrared (FTIR) spectroscopy and advanced preprocessing steps. The learning models were Ridge Regression and Support Vector Regression(SVR) using Leave-One-out Cross-Validation. Results exhibited that TBD and ADPD significantly outperform basic used methods. For SVR, the TBD increased the r2 score by around 27%, and ADPD increased it by around 10%. these Ridge Regression values were between 36% and 24%. In addition, Results demonstrate that TBD and ADPD significantly outperform conventional methods, achieving lower error rates and improved clinical accuracy, validated through Clarke and Parkes Error Grid Analysis.

Determining company similarity is a vital task in finance, underpinning hedging, risk management, portfolio diversification, and more. Practitioners often rely on sector and industry classifications to gauge similarity, such as SIC-codes and GICS-codes - the former being used by the U.S. Securities and Exchange Commission (SEC), and the latter widely used by the investment community. Since these classifications can lack granularity and often need to be updated, using clusters of embeddings of company descriptions has been proposed as a potential alternative, but the lack of interpretability in token embeddings poses a significant barrier to adoption in high-stakes contexts. Sparse Autoencoders (SAEs) have shown promise in enhancing the interpretability of Large Language Models (LLMs) by decomposing LLM activations into interpretable features. We apply SAEs to company descriptions, obtaining meaningful clusters of equities in the process. We benchmark SAE features against SIC-codes, Major Group codes, and Embeddings. Our results demonstrate that SAE features not only replicate but often surpass sector classifications and embeddings in capturing fundamental company characteristics. This is evidenced by their superior performance in correlating monthly returns - a proxy for similarity - and generating higher Sharpe ratio co-integration strategies, which underscores deeper fundamental similarities among companies.

Machine learning has been widely adopted in biomedical research, fueled by the increasing availability of data. However, integrating datasets across institutions is challenging due to legal restrictions and data governance complexities. Federated learning allows the direct, privacy preserving training of machine learning models using geographically distributed datasets, but faces the challenge of how to appropriately control for covariate effects. The naive implementation of conventional covariate control methods in federated learning scenarios is often impractical due to the substantial communication costs, particularly with high-dimensional data. To address this issue, we introduce dsLassoCov, a machine learning approach designed to control for covariate effects and allow an efficient training in federated learning. In biomedical analysis, this allow the biomarker selection against the confounding effects. Using simulated data, we demonstrate that dsLassoCov can efficiently and effectively manage confounding effects during model training. In our real-world data analysis, we replicated a large-scale Exposome analysis using data from six geographically distinct databases, achieving results consistent with previous studies. By resolving the challenge of covariate control, our proposed approach can accelerate the application of federated learning in large-scale biomedical studies.

Credit card fraud detection remains a critical challenge in financial security, with machine learning models like XGBoost(eXtreme gradient boosting) emerging as powerful tools for identifying fraudulent transactions. However, the inherent class imbalance in credit card transaction datasets poses significant challenges for model performance. Although sampling techniques are commonly used to address this imbalance, their implementation sometimes precedes the train-test split, potentially introducing data leakage. This study presents a comparative analysis of XGBoost's performance in credit card fraud detection under three scenarios: Firstly without any imbalance handling techniques, secondly with sampling techniques applied only to the training set after the train-test split, and third with sampling techniques applied before the train-test split. We utilized a dataset from Kaggle of 284,807 credit card transactions, containing 0.172\% fraudulent cases, to evaluate these approaches. Our findings show that although sampling strategies enhance model performance, the reliability of results is greatly impacted by when they are applied. Due to a data leakage issue that frequently occurs in machine learning models during the sampling phase, XGBoost models trained on data where sampling was applied prior to the train-test split may have displayed artificially inflated performance metrics. Surprisingly, models trained with sampling techniques applied solely to the training set demonstrated significantly lower results than those with pre-split sampling, all the while preserving the integrity of the evaluation process.

Over the recent years, Shapley value (SV), a solution concept from cooperative game theory, has found numerous applications in data analytics (DA). This paper provides the first comprehensive study of SV used throughout the DA workflow, which involves three main steps: data fabric, data exploration, and result reporting. We summarize existing versatile forms of SV used in these steps by a unified definition and clarify the essential functionalities that SV can provide for data scientists. We categorize the arts in this field based on the technical challenges they tackled, which include computation efficiency, approximation error, privacy preservation, and appropriate interpretations. We discuss these challenges and analyze the corresponding solutions. We also implement SVBench, the first open-sourced benchmark for developing SV applications, and conduct experiments on six DA tasks to validate our analysis and discussions. Based on the qualitative and quantitative results, we identify the limitations of current efforts for applying SV to DA and highlight the directions of future research and engineering.

Software defect prediction is a critical aspect of software quality assurance, as it enables early identification and mitigation of defects, thereby reducing the cost and impact of software failures. Over the past few years, quantum computing has risen as an exciting technology capable of transforming multiple domains; Quantum Machine Learning (QML) is one of them. QML algorithms harness the power of quantum computing to solve complex problems with better efficiency and effectiveness than their classical counterparts. However, research into its application in software engineering to predict software defects still needs to be explored. In this study, we worked to fill the research gap by comparing the performance of three QML and five classical machine learning (CML) algorithms on the 20 software defect datasets. Our investigation reports the comparative scenarios of QML vs. CML algorithms and identifies the better-performing and consistent algorithms to predict software defects. We also highlight the challenges and future directions of employing QML algorithms in real software defect datasets based on the experience we faced while performing this investigation. The findings of this study can help practitioners and researchers further progress in this research domain by making software systems reliable and bug-free.

Through the rational compatibility of herbal medicines, TCM achieves synergistic effects and toxicity reduction (1). This core principle has demonstrated significant advantages in treating complex diseases. For example, Lianhua Qingwen Capsules have shown remarkable efficacy in alleviating symptoms and reducing hospitalization time for COVID-19 patients (2, 3). Additionally, PHY906, developed by Yale University based on the traditional Huangqin Decoction, has improved colorectal cancer treatment outcomes as a chemotherapy adjuvant (4). In recent years, the rapid development of artificial intelligence (AI) has introduced novel opportunities for investigating the complex mechanisms underlying TCM (5, 6). AI's exceptional data processing capabilities, particularly in multi-dimensional data analysis and complex relationship modeling, are transforming traditional medicine from experience-driven to datadriven paradigms (7-9). Notably, Graph Artificial Intelligence (GraphAI) offers a unique toolkit for exploring complex network-structured data by integrating knowledge graphs, graph computation, and graph neural networks (GNNs) (10, 11). The core challenges of TCM compatibility--complex interactions involving multiple components, targets, and pathways-- align closely with GraphAI's strengths in handling intricate relationships (12-14).

Federated learning enables decentralized model training without sharing raw data, preserving data privacy. However, its vulnerability towards critical security threats, such as gradient inversion and model poisoning by malicious clients, remain unresolved. Existing solutions often address these issues separately, sacrificing either system robustness or model accuracy. This work introduces Tazza, a secure and efficient federated learning framework that simultaneously addresses both challenges. By leveraging the permutation equivariance and invariance properties of neural networks via weight shuffling and shuffled model validation, Tazza enhances resilience against diverse poisoning attacks, while ensuring data confidentiality and high model accuracy. Comprehensive evaluations on various datasets and embedded platforms show that Tazza achieves robust defense with up to 6.7x improved computational efficiency compared to alternative schemes, without compromising performance.

Multimodal learning has recently gained significant popularity, demonstrating impressive performance across various zero-shot classification tasks and a range of perceptive and generative applications. Models such as Contrastive Language-Image Pretraining (CLIP) are designed to bridge different modalities, such as images and text, by learning a shared representation space through contrastive learning. Despite their success, the working mechanisms underlying multimodal learning are not yet well understood. Notably, these models often exhibit a modality gap, where different modalities occupy distinct regions within the shared representation space. In this work, we conduct an in-depth analysis of the emergence of modality gap by characterizing the gradient flow learning dynamics. Specifically, we identify the critical roles of mismatched data pairs and a learnable temperature parameter in causing and perpetuating the modality gap during training. Furthermore, our theoretical insights are validated through experiments on practical CLIP models. These findings provide principled guidance for mitigating the modality gap, including strategies such as appropriate temperature scheduling and modality swapping. Additionally, we demonstrate that closing the modality gap leads to improved performance on tasks such as image-text retrieval.