Goker Erdogan

Goker Erdogan
name surname [at] google.com

I am a Senior Research Scientist at Google DeepMind. I am interested in understanding how machines can acquire rich, flexible representations of the world around them from little supervision and use these to solve a variety of challenging tasks efficiently. My research is largely motivated by the belief that the flexibility and efficiency of human intelligence calls for strong inductive biases informed by the structure of the real world.
I am passionate about taking insights from how we humans learn so much about the world from so little supervision and endowing machines with this remarkable ability.

Previously, I was a research scientist at Cogitai building AI/machine learning solutions to make continual learning practical in a broad range of applications.

Between 2012-2017, I was a graduate student in Dept. of Brain and Cognitive Sciences at University of Rochester and a member of Computational Cognition and Perception lab. My research focus was shape perception, and I used behavioral, neuroimaging and computational methods to understand the nature of the representations and algorithms involved in shape perception. [thesis][cv]

Publications

Gemini Team Google (Erdogan G., contributor) (2024). Gemini 1.5: Unlocking multimodal understanding across millions of tokens of context. pdf

Bica I., Ilic A., Bauer M., Erdogan G., Bošnjak M., Kaplanis C., Gritsenko A. A., Minderer M., Blundell C., Pascanu R., & Mitrovic J. (2024). Improving fine-grained understanding in image-text pre-training. In Forty-first International Conference on Machine Learning. pdf

Kabra R., Zoran D., Erdogan G., Matthey L., Creswell A., Botvinick M., Lerchner A., & Burgess C. P. (2021). SIMONe: View-Invariant, Temporally-Abstracted Object Representations via Unsupervised Video Decomposition. In Advances in Neural Information Processing Systems. pdf

Erdogan G., Jacobs R. A. (2017) Visual Shape Perception as Bayesian Inference of 3D Object-centered Shape Representations. Psychological Review. webpage pdf code supplementary materials

Erdogan G., Jacobs R. A. (2016) A 3D shape Inference Model Matches Human Visual Object Similarity Judgments Better Than Deep Convolutional Neural Networks. In Papafragou, A., Grodner, D., Mirman, D., & Trueswell, J.C. (Eds.) Proceedings of the 38th Annual Conference of the Cognitive Science Society. Austin, TX: Cognitive Science Society. webpage code supplementary materials

Erdogan G., Chen, Q., Garcea F. E., Mahon B. Z., Jacobs R. A. (2016) Multisensory Part-Based Representations of Objects in Human Lateral Occipital Complex. Journal of Cognitive Neuroscience. Vol. 28, No. 6, Pages 869-881. webpage pdf

Erdogan G., Yildirim I., Jacobs R. A. (2015) From Sensory Signals to Modality-Independent Conceptual Representations: A Probabilistic Language of Thought Approach. PLoS Comput Biol 11(11): e1004610. doi: 10.1371/journal.pcbi.1004610 webpage pdf

Erdogan G., Yildirim I., Jacobs R. A. (2015). An Analysis-by-synthesis Approach to Multisensory Object Shape Perception. Multimodal Machine Learning Workshop. NIPS 2015. webpage

Erdogan G., Yildirim I., Jacobs R. A. (2014). Transfer of Object Shape Knowledge across Visual and Haptic Modalities. In P. Bello, M. Guarini, M. McShane, & B. Scassellati (Eds.), Proceedings of the 36th Annual Conference of the Cognitive Science Society. Austin, TX: Cognitive Science Society. webpage


Teaching

November 2018 - Introduction to Machine Learning lecture notes
Spring 2014 - BCS111 Foundations of Cognitive Science
Spring 2015 - BCS153 Cognition
Fall 2015 - BCS183 Animal Minds

Posts