AAAST LLC is a company that excels in technology, innovation, and business strategy while consistently
delivering value to its customers, employees, and stakeholders. Success in IT can be measured in several
ways, including financial performance, industry influence, technological advancements, and customer
satisfaction.
What is PM GenAI?
PM GenAI stands for Principal Value Generative AI and is a novel artificial intelligence algorithm capable of revolutionizing numerous areas of science and technology. Dr. Philip de Melo, a renowned American data scientist has developed the fundamentals of PM GenAI that is considered one of the major breakthroughs in data science.
Unlike existing solutions, this approach introduces a fundamentally different mechanism by generation of synthetic data based on observational data sets, allowing for the first time to significantly improve accuracy, scalability, and cost-effectiveness of current algorithms. It allows to significantly boost the accuracy of current technologies used in healthcare, manufacturing, and finance and suggests significant advantages over traditional methods.
PM GenAI in healthcare
AI has great potential in healthcare and public health, but it also comes with several shortcomings, including:
- Bias in Data – AI models are only as good as the data they are trained on. If the data is biased or unrepresentative, the AI can produce inaccurate or unfair results, especially for underrepresented populations.
- Lack of Explainability – Many AI models, particularly deep learning ones, function as “black boxes,” making it difficult to understand how they arrive at certain conclusions, which is a concern for medical decision-making.
PM GenAI in healthcare generates synthetic data, integrates it with observational data and rejects low probability prediction.
Demonstration of PM GenAI in healthcare (Classification)
Application of PM GenAI to real diabetes data collected from over 750 patients using traditional machine learning methods for binary classification. Logistic regression and support vector machines show higher accuracy among other methods with the following classification report:
| Results (traditional ML) | Results ( PM GenAI) |
| Accuracy = 73.17% Precision = 50.00% Recall = 68.18% | Accuracy ≈ 94.51% Precision ≈ 88.90% Recall ≈ 90.92% |
Demonstration of PM GenAI in production of electric appliances
Similar results are received applying the new technology to a manufacturing problem: prediction of revenues.
| Results (traditional ML) | Results ( PM GenAI) |
| Accuracy = 69.24% Precision = 50.00% Recall = 63.20% | Accuracy ≈ 94.22% Precision ≈ 89.30% Recall ≈ 92.39% |
