Generative AI Techniques

This course is designed to provide a comprehensive understanding of generative AI techniques and their potential applications in Production Control and Operations (P&OC). Participants will learn about various generative models, such as Generative Adversarial Networks (GANs), Variational Autoencoders (VAEs), and Transformers, and how they can be used to generate realistic synthetic data, enhance data augmentation, and solve domain-specific problems. The course covers the theoretical foundations of generative models, as well as practical implementation and evaluation strategies tailored to P&OC.

• Understand the fundamental concepts and theory behind generative AI models

• Implement and train GANs, VAEs, and Transformer-based generative models using deep learning frameworks

• Generate realistic synthetic data for various P&OC use cases, such as production line simulations, inventory demand patterns, and supply chain scenarios

• Apply generative models for data augmentation to improve the performance of supervised learning tasks in P&OC

• Utilize generative AI techniques for domain-specific applications, such as anomaly detection, predictive maintenance, and production scheduling optimization

1. Introduction to Generative AI in P&OC

• Overview of generative models and their applications in Production Control and Operations

• Probability distributions and sampling techniques

• Generative vs. discriminative models

• Hands-on exercises: Implementing basic generative models (e.g., Gaussian Mixture Models, Hidden Markov Models)

2. Generative Adversarial Networks (GANs) for P&OC

• GAN architecture and training process

• Variants of GANs (e.g., DCGAN, WGAN, CycleGAN)

• Evaluation metrics for GANs (e.g., Inception Score, Fréchet Inception Distance)

• Hands-on exercises: Training a GAN to generate synthetic production line data or inventory demand patterns

3. Variational Autoencoders (VAEs) in P&OC

• Variational inference and latent variable models

• VAE architecture and training process

• Conditional VAEs and their applications in P&OC

• Hands-on exercises: Implementing a VAE for generating synthetic supply chain scenarios

4. Generative Transformer Models and Applications in P&OC

• Transformer-based generative models (e.g., GPT, BERT)

• Fine-tuning generative Transformers for P&OC text data (e.g., work orders, quality reports)

• Domain-specific applications of generative AI in Production Control and Operations (e.g., anomaly detection, predictive maintenance, production scheduling optimization)

• Hands-on exercises: Fine-tuning a generative Transformer model for generating synthetic work orders or quality reports

• Strong understanding of deep learning concepts and architectures (e.g., CNN, RNN, Transformers)

• Proficiency in programming with Python and deep learning frameworks (e.g., TensorFlow, PyTorch)

• Familiarity with probability theory and statistical concepts

• Knowledge of production control and operations management principles