Abstract: Generative learning has emerged as a powerful paradigm in machine learning, enabling systems to create new data samples, model complex distributions, and uncover hidden structures within datasets. Variational AutoEncoders (VAEs) are central to this progress, offering a probabilistic framework for generative modeling. Yet, as datasets grow in scale and complexity, classical approaches face limitations in efficiency and expressiveness.
In this talk, Dr. Bapi Chatterjee and Dhiraj Madan from IBM will present how quantum computing can enhance generative learning by augmenting VAEs with quantum-inspired techniques. The session will explore how quantum-enhanced VAEs leverage the principles of quantum mechanics to expand representational capacity, improve sampling efficiency, and tackle challenges that are computationally intensive for classical systems.
Key themes will include:
Fundamentals of generative learning and the role of VAEs.
How quantum-enhanced architectures can improve learning and inference.
Applications in domains such as drug discovery, materials science, and complex optimization.
Challenges in integrating quantum computing with machine learning pipelines, including scalability and hardware constraints.
The presentation will highlight both theoretical insights and practical demonstrations, showing how quantum-enhanced VAEs can redefine the boundaries of generative AI. By combining the strengths of machine learning and quantum computing, this approach opens new pathways for solving problems that were previously intractable, pointing toward a future where generative models are more powerful, efficient, and impactful.