Ignacio Serna
Independent Researcher
I am an independent researcher funded by a Coefficient Giving (formerly Open Philanthropy) grant to study the geometry of learned representations in the joint space of vision encoders with other modalities. My work probes the limits of linear structure in large-scale neural networks and asks how faithfully concepts are captured by linear directions — sitting at the intersection of computer vision, representation learning, and interpretability.
Affiliations
- Visitor at the Technical Safety & Governance Lab, University of Oxford, working with Fazl Barez
- Visiting Researcher at the Max Planck Institute for Human Development (Center for Humans & Machines)
- Visiting Faculty at IMDEA Software
Teaching
- Machine Learning Foundations at IE University, as Adjunct Professor (spring semester)
Background
PhD in Computer Science (2024, Autonomous University of Madrid), including a visiting stay at Caltech (California Institute of Technology) with Professor Pietro Perona . Previously postdoc at the Max Planck Institute with Iyad Rahwan. B.S. in Mathematics and a B.S. in Computer Science (2018), and an M.S. in Artificial Intelligence (2020).