Machine Learning for Synthetic Data Generation: a Review

Lu, Yingzhou; Wang, Huazheng; Wei, Wüchang

doi:10.48550/arxiv.2302.04062

Cited by 19 publications

(27 citation statements)

References 88 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…It is the favored method for increasing a training dataset's amount and diversity to improve DL accuracy. 29,30 It entails that an actual product was not captured for the image dataset but was created by the model based on patterns and features learned from real images. 31 This technology is employed in numerous other industries, including healthcare, 32 commerce, manufacturing, agriculture, 33 and more.…”

Section: Literature Reviewmentioning

confidence: 99%

Intelligent IoT framework with GAN‐synthesized images for enhanced defect detection in manufacturing

Aramkul,

Sugunnasil

2023

Computational Intelligence

View full text Add to dashboard Cite

The manufacturing industry is always exploring techniques to optimize processes, increase product quality, and more accurately identify defects. The technique of deep learning is the strategy that will be used to handle the issues presented. However, the challenge of using AI in this domain is the small and imbalanced dataset for training affected by the severe shortage of defective data. Moreover, data acquisition requires significant labor, time, and resources. In response to these demands, this research presents an intelligent internet of things (IoT) framework enriched by generative adversarial network (GAN). This framework was developed in response to the needs outlined above. The framework applies the IoT for real‐time data collection and communication, while GAN are utilized to synthesize high‐fidelity images of manufacturing defects. The quality of the GAN‐synthesized image is quantified by the average FID score of 8.312 for non‐defective images and 7.459 for defective images. As evidenced by the similarity between the distributions of synthetic and real images, the proposed generative model can generate visually authentic and high‐fidelity images. As demonstrated by the results of the defect detection experiments, the accuracy can be enhanced to the maximum of 96.5% by integrating the GAN‐synthesized images with the real images. Concurrently, this integration reduces the occurrence of false alarms.

show abstract

Section: Literature Reviewmentioning

confidence: 99%

Intelligent IoT framework with GAN‐synthesized images for enhanced defect detection in manufacturing

Aramkul,

Sugunnasil

2023

Computational Intelligence

View full text Add to dashboard Cite

show abstract

“…Machine learning algorithms often require a substantial amount of labeled data for effective training, but collecting such data in real-world experiments can be time-consuming and costly, which is particularly relevant in atomic BEC experiments. Recently, to tackle the issue of obtaining a costly labeled training set for machine learning algorithms, the use of synthetic data has been proposed [46][47][48]. This is especially attractive as the GPE can generate synthetic BEC images for CNN training with parameters that are similar to those of the experimental setup.…”

Section: Introductionmentioning

confidence: 99%

Vortex detection in atomic Bose–Einstein condensates using neural networks trained on synthetic images

Kim,

Kwon,

Rabga

et al. 2023

Mach. Learn.: Sci. Technol.

View full text Add to dashboard Cite

Quantum vortices in atomic Bose-Einstein condensates (BECs) are topological defects characterized by quantized circulation of particles around them. In experimental studies, vortices are commonly detected by time-of-flight imaging, where their density-depleted cores are enlarged. In this work, we describe a machine learning-based method for detecting vortices in experimental BEC images, particularly focusing on turbulent condensates containing irregularly distributed vortices. Our approach employs a convolutional neural network (CNN) trained solely on synthetic simulated images, eliminating the need for manual labeling of the vortex positions as ground truth. We find that the CNN achieves accurate vortex detection in real experimental images, thereby facilitating analysis of large experimental datasets without being constrained by specific experimental conditions. This novel approach represents a significant advancement in studying quantum vortex dynamics and streamlines the analysis process in the investigation of turbulent BECs.

show abstract

“…But besides the exploration of these proof of concepts in this eld, research about other use cases is still ongoing. Our use case focuses on the generation of AI narratives, which is still a relatively novel research eld (Lu et al 2023).…”

Section: Introductionmentioning

confidence: 99%

The Aluminum Standard: Using Generative Artificial Intelligence Tools to Synthesize and Annotate Non-Structured Patient Data

Ochoa,

Mustafa,

Weil

et al. 2023

Preprint

View full text Add to dashboard Cite

Background Medical narratives are fundamental to the correct identification of a patient's health condition. This is not only because it describes the patient’s situation. It also contains relevant information about the patient’s context and health state evolution. Narratives are usually vague and cannot be categorized easily. On the other hand, once the patient’s situation is correctly identified based on a narrative, it is then possible to map the patient’s situation into precise classification schemas and ontologies that are machine-readable. To this end, language models can be trained to read and extract elements from these narratives. However, the main problem is the lack of data for model identification and model training in languages other than English. Alternative available data, like MIMIC (Johnson et al. 2016) is written in English and for specific patient conditions like intensive care. Thus, when model training is required for other types of patients, like oncology (and not intensive care), this could lead to bias. To facilitate clinical narrative models training, a method for creating high-quality synthetic narratives is needed. Method We devised workflows based on generative AI methods to synthesize narratives in the German Language. Since we required highly realistic narratives, we generated prompts, written with high-quality medical terminology, asking for clinical narratives containing both a main and co-disease. The frequency of distribution of both the main and co-disease was extracted from the hospital’s structured data, such that the synthetic narratives reflect the disease distribution among the patient’s cohort. In order to validate the quality of the synthetic narratives, we annotated them to train a Named Entity Recognition (NER) algorithm. According to our assumptions, the validation of this system implies that the synthesized data used for its training are of acceptable quality. Result We report precision, recall and F1 score for the NER model while also considering metrics that take into account both exact and partial entity matches. We obtained a precision of 0.851 for Entity Type match metric, with a F1 score of 0.188. Conclusion Despite its inherent limitations, this technology can accelerate model identification and training. By using this approach, data can be interoperable across languages and regions without compromising data safety.

show abstract

Machine Learning for Synthetic Data Generation: a Review

Cited by 19 publications

References 88 publications

Intelligent IoT framework with GAN‐synthesized images for enhanced defect detection in manufacturing

Intelligent IoT framework with GAN‐synthesized images for enhanced defect detection in manufacturing

Vortex detection in atomic Bose–Einstein condensates using neural networks trained on synthetic images

The Aluminum Standard: Using Generative Artificial Intelligence Tools to Synthesize and Annotate Non-Structured Patient Data

Contact Info

Product

Resources

About