Keynote Speakers

Bio: Jamal Toutouh is an Assistant Professor at the University of Málaga (Spain). Previously, he was a Postdoctoral Fellow at the Massachusetts Institute of Technology (MIT) in the USA. He earned his Ph.D. in Computer Engineering at the University of Malaga. During his doctoral studies, he devised Machine Learning and Optimization methods inspired by Nature to address hard-to-solve optimization problems that arise from Smart Mobility. The thesis was awarded three awards, including the 2018 Best Spanish Ph.D. Thesis in Smart Cities. 

While he continues contributing to the Smart Cities domain with high-impact solutions for society, Jamal has become a Generative Machine Learning enthusiast. Jamal's most significant ongoing research focuses on coupling gradient-free and gradient-based distributed optimization methods to train Generative Adversarial Networks (GAN). Most of these training methods and strategies are included in Lipizzaner, an open-source framework to train GANs. This novel research line provides more efficient and effective training, which uses fewer data to compute accurate generative models with lower complexity. The results of his research have been presented at top conferences and published in high-impact journals. In addition, he has been invited to give talks, seminars, courses, and tutorials in universities and R&D departments of companies around the world.   

Abstract: Generative Machine Learning (GML) has emerged as a transformative field, revolutionizing various domains by enabling machines to learn and generate complex and realistic data. GML has opened doors to unprecedented possibilities, from generating realistic images, writing meaningful poems, synthesizing human-like speech, and creating compelling music compositions to designing novel molecules. Despite the success of GML, effective training of generative models remains an open question. During the talk, we will delve into the challenges associated with training generative models and the recent advancements that have paved the way for more efficient and effective training. We will discuss techniques such as adversarial training, which have significantly improved the stability and quality of generated samples. Additionally, we will explore strategies for tackling challenges such as mode collapse, training instability, and scalability, which are crucial for large-scale GML applications. Furthermore, the talk will shed light on developments in model optimization and parallelization strategies that have accelerated the training process and enabled the generation of high-quality samples within reasonable computational times. We will also discuss approaches for improving sample diversity and incorporating domain-specific constraints to achieve more controllable and meaningful generative outputs.

Bio: Dr. Manchón has been a leading voice within the AI industry for more than 25 years: as an entrepreneur, driving her AI startup through a successful exit, and as an executive at companies such as Intel, Amazon, Roku and now as the Senior Director of Engineering, Research Strategy at Google AI. Pilar’s main areas of expertise range from Generative AI, Conversational AI and Responsible AI to Cognitive and Ambient Computing, Human-Computer Interaction, UX, UXR, Robotics and Language Technologies in general. 

She is originally from Spain and holds a degree in Linguistics from the University of Seville, an MSc. in Cognitive Science and Natural Language from Edinburgh University and a European PhD in Computational Linguistics on Intelligent Dialog Systems. Her postgraduate research and entrepreneurial endeavors took her to Stanford University, MIT and Singularity University.

Pilar is in the Board of Directors and/or Board of Advisors at several companies, Universities, NGOs, and VCs. She is an investor, a lecturer and a prolific public speaker. 

In terms of public service, Dr. Manchón is a member of the AI Strategy Advisory Council of the Government of Spain, and also a member of the AI Strategy Advisory Council of the Government of Andalusia. Her contributions to the field of AI, the Latinx Community, DEI, Women leadership and Entrepreneurship have been broadly acknowledged in the US, Spain, LATAM and more recently the UN, with multiple awards and recognitions.

In her current role, Pilar architects moonshots in Generative AI and leads teams of Research Scientists, Engineers and UX Researchers across multiple projects. She is a Generative AI ambassador for Google C-level engagements, and the Executive Sponsor of initiatives such as AI Centers of Excellence and AI Chairs with several Universities. She co-sponsors the Women in Leadership group and serves on multiple steering committees for Google’s strategy in Research, University and Government Relations.