Data Balancing Strategies: A Survey of Resampling and Augmentation Methods
Authors:
Behnam Yousefimehr,
Mehdi Ghatee,
Mohammad Amin Seifi,
Javad Fazli,
Sajed Tavakoli,
Zahra Rafei,
Shervin Ghaffari,
Abolfazl Nikahd,
Mahdi Razi Gandomani,
Alireza Orouji,
Ramtin Mahmoudi Kashani,
Sarina Heshmati,
Negin Sadat Mousavi
Abstract:
Imbalanced data poses a significant obstacle in machine learning, as an unequal distribution of class labels often results in skewed predictions and diminished model accuracy. To mitigate this problem, various resampling strategies have been developed, encompassing both oversampling and undersampling techniques aimed at modifying class proportions. Conventional oversampling approaches like SMOTE e…
▽ More
Imbalanced data poses a significant obstacle in machine learning, as an unequal distribution of class labels often results in skewed predictions and diminished model accuracy. To mitigate this problem, various resampling strategies have been developed, encompassing both oversampling and undersampling techniques aimed at modifying class proportions. Conventional oversampling approaches like SMOTE enhance the representation of the minority class, whereas undersampling methods focus on trimming down the majority class. Advances in deep learning have facilitated the creation of more complex solutions, such as Generative Adversarial Networks (GANs) and Variational Autoencoders (VAEs), which are capable of producing high-quality synthetic examples. This paper reviews a broad spectrum of data balancing methods, classifying them into categories including synthetic oversampling, adaptive techniques, generative models, ensemble-based strategies, hybrid approaches, undersampling, and neighbor-based methods. Furthermore, it highlights current developments in resampling techniques and discusses practical implementations and case studies that validate their effectiveness. The paper concludes by offering perspectives on potential directions for future exploration in this domain.
△ Less
Submitted 17 May, 2025;
originally announced May 2025.
State of Security Awareness in the AM Industry: 2020 Survey
Authors:
Mark Yampolskiy,
Paul Bates,
Mohsen Seifi,
Nima Shamsaei
Abstract:
Security of Additive Manufacturing (AM) gets increased attention due to the growing proliferation and adoption of AM in a variety of applications and business models. However, there is a significant disconnect between AM community focused on manufacturing and AM Security community focused on securing this highly computerized manufacturing technology. To bridge this gap, we surveyed the America Mak…
▽ More
Security of Additive Manufacturing (AM) gets increased attention due to the growing proliferation and adoption of AM in a variety of applications and business models. However, there is a significant disconnect between AM community focused on manufacturing and AM Security community focused on securing this highly computerized manufacturing technology. To bridge this gap, we surveyed the America Makes AM community, asking in total eleven AM security-related questions aiming to discover the existing concerns, posture, and expectations. The first set of questions aimed to discover how many of these organizations use AM, outsource AM, or provide AM as a service. Then we asked about biggest security concerns as well as about assessment of who the potential adversaries might be and their motivation for attack. We then proceeded with questions on any experienced security incidents, if any security risk assessment was conducted, and if the participants' organizations were partnering with external experts to secure AM. Lastly, we asked whether security measures are implemented at all and, if yes, whether they fall under the general cyber-security category. Out of 69 participants affiliated with commercial industry, agencies, and academia, 53 have completed the entire survey. This paper presents the results of this survey, as well as provides our assessment of the AM Security posture. The answers are a mixture of what we could label as expected, "shocking but not surprising," and completely unexpected. Assuming that the provided answers are somewhat representative to the current state of the AM industry, we conclude that the industry is not ready to prevent or detect AM-specific attacks that have been demonstrated in the research literature.
△ Less
Submitted 7 September, 2022;
originally announced September 2022.
