On April 20, Francisco M. López successfully defended his doctoral thesis "Learning Active Perception through Abstraction". He conducted his research in the group of FIAS Senior Fellow Jochen Triesch, with support from The Adaptive Mind (TAM) excellence cluster.

In his work, López developed computational models to better understand how humans, in particular infants, learn to perceive and interact with the world. By putting together aspects methods from variational inference, reinforcement learning, and virtual embodiments, his doctoral thesis bridged developmental science and artificial intelligence (AI).  A central contribution of his PhD was the contribution to the development of MIMo, a multimodal infant model that enables faster-than-real-time simulations of sensorimotor development. MIMo has become an increasingly popular research platform within the developmental AI community. López's work on a growing version of MIMo was recognized with the BabyBot Paper Award at the 2025 IEEE International Conference on Development and Learning. The main project of his PhD was the introduction of the Active Variational Efficient Coding (AVEC) theory, an extension of the Efficient Coding Hypothesis to the frameworks of variational inference and active perception. López showed how AVEC can explain fundamental aspects of sensorimotor coordination, such as the self-calibration of binocular vision, allowing MIMo to learn vergence and saccades without supervision. Beyond active perception, López also explored aspects of perceptual and aesthetic preferences. He was able to reproduce several behavioral results, including the attractiveness of average faces and biases toward symmetry and familiarity. López will continue working as a postdoctoral researcher in the group of Jochen Triesch at FIAS while also serving as Visiting Fellow at the University of New South Wales, Australia.

Publications: 

López, F. M., Shi, B.E, & Triesch, J. (in press). Efficient Coding in Active Perception: A Developmental Perspective on Autonomous Control. Advances in Child Development and Behavior, 70. 

López, F. M., Kanazawa, H., Fiala, O., Balashov, Y., Marcel, V., Rustler, L., Lenz, M., Kim, D., Kuniyoshi, Y., Triesch, J., & Hoffmann, M. (under review). Simulating Infant First‑Person Sensorimotor Experience via Motion Retargeting from Babies to Humanoids. Submitted to 2026 IEEE International Conference on Development and Learning. 

López, F. M., Ernst, M. R., Cruz, F., Hoffmann, M., & Triesch, J. (under review). Infant Spontaneous Movement Noise Improves Exploration in Deep RL. Submitted to 2026 IEEE International Conference on Development and Learning. 

Philipp, L., López, F. M., & Triesch, J. (under review). Embodiment Shapes Rolling Behavior in a Multimodal Infant Model. Submitted to 2026 IEEE International Conference on Development and Learning. 

López, F. M., Lenz, M., Fedozzi, M. G., Aubret, A., & Triesch, J. (2025). MIMo grows! Simulating body and sensory development in a multimodal infant model. In 2025 IEEE International Conference on Development and Learning. 

Faßbender, L., Falck, J., López, F. M., Shing, Y. L., Triesch, J., & Schwarzer, G. (2025). A comparison of force adaptation in toddlers and adults during a drawer opening task. Scientific Reports, 15(1), 3699.

López, F. M., & Triesch, J. (2025). Hierarchical Residuals Exploit Brain‑Inspired Compositionality. In 2025 European Symposium on Artificial Neural Networks, Computational Intelligence and Machine Learning (ESANN). 

Mattern, D., Schumacher, P., López, F. M., Raabe, M. C., Ernst, M. R., Aubret, A. & Triesch, J. (2024). MIMo: A Multimodal Infant Model for Studying Cognitive Development. IEEE Transactions on Cognitive and Developmental Systems, 16(4), 1291‑1301. 

López, F. M., Shi, B. E., & Triesch, J. (2024) Prioritizing Compression Explains Human Perceptual Preferences. In Intrinsically‑Motivated and Open‑Ended Learning Workshop @ NeurIPS2024. 

López, F. M., Raabe, M. C., Shi, B. E., & Triesch, J. (2024). Self‑Calibrating Saccade‑Vergence Interactions. In 2024 IEEE International Conference on Development and Learning (ICDL) (pp. 1‑7). IEEE. 

Ernst, M. R., López, F. M., Aubret, A., Fleming, R. W., & Triesch, J. (2024). Self‑Supervised Learning of Color Constancy. In 2023 IEEE International Conference on Development and Learning (ICDL) (pp. 1‑7). IEEE. 

López, F. M., Shi, B. E., & Triesch, J. (2023). Eye‑hand coordination develops from active multimodal compression. In 2023 IEEE International Conference on Development and Learning (ICDL) (pp. 437‑442). IEEE.