Publication
Theoretical Issues in Deep Networks. (2019).
CBMM Memo 100 v1 (1.71 MB)
CBMM Memo 100 v3 (8/25/2019) (1.31 MB)
CBMM Memo 100 v4 (11/19/2019) (1008.23 KB)
Theoretical issues in deep networks. Proceedings of the National Academy of Sciences 201907369 (2020). doi:10.1073/pnas.1907369117
PNASlast.pdf (915.3 KB)
Theory II: Landscape of the Empirical Risk in Deep Learning. (2017).
CBMM Memo 066_1703.09833v2.pdf (5.56 MB)
Notes on Hierarchical Splines, DCLNs and i-theory. (2015).
CBMM Memo 037 (1.83 MB)
Loss landscape: SGD has a better view. (2020).
CBMM-Memo-107.pdf (1.03 MB)
Typos and small edits, ver11 (955.08 KB)
Small edits, corrected Hessian for spurious case (337.19 KB)
Evolving Images for Visual Neurons Using a Deep Generative Network Reveals Coding Principles and Neuronal Preferences. Cell 177, 1009 (2019).
Author's last draft (20.26 MB)
Infants’ Categorization of Social Actions. Cognitive Development Society (CDS) (2015).
Using fNIRS to Map Functional Specificity in the Infant Brain: An fROI Approach. (2015).
SRCD2015_NIRS_poster.pdf (2.14 MB)
Simultaneous whole-animal 3D imaging of neuronal activity using light-field microscopy. Nature Methods 11, 727 - 730 (2014).
NOPA: Neurally-guided Online Probabilistic Assistance for Building Socially Intelligent Home Assistants. arXiv (2023). at <https://arxiv.org/abs/2301.05223>
NOPA: Neurally-guided Online Probabilistic Assistance for Building Socially Intelligent Home Assistants. 2023 IEEE International Conference on Robotics and Automation (ICRA) (2023). doi:10.1109/ICRA48891.2023.10161352
The inferior temporal cortex is a potential cortical precursor of orthographic processing in untrained monkeys. Nature Communications 11, (2020).
s41467-020-17714-3.pdf (25.01 MB)
An Optimal Structured Zeroth-order Algorithm for Non-smooth Optimization. 37th Conference on Neural Information Processing Systems (NeurIPS 2023) (2023). at <https://proceedings.neurips.cc/paper_files/paper/2023/file/7429f4c1b267cf619f28c4d4f1532f99-Paper-Conference.pdf>
Understanding the Role of Recurrent Connections in Assembly Calculus. (2022).
CBMM-Memo-137.pdf (1.49 MB)
Feature learning in deep classifiers through Intermediate Neural Collapse. (2023).
Feature_Learning_memo.pdf (2.16 MB)
For interpolating kernel machines, the minimum norm ERM solution is the most stable. (2020).
CBMM_Memo_108.pdf (1015.14 KB)
Better bound (without inequalities!) (1.03 MB)
For interpolating kernel machines, minimizing the norm of the ERM solution maximizes stability. Analysis and Applications 21, 193 - 215 (2023).
Biologically Inspired Mechanisms for Adversarial Robustness. (2020).
CBMM_Memo_110.pdf (3.14 MB)
Scene-Domain Active Part Models for Object Representation. IEEE International Conference on Computer Vision (ICCV) 2497 - 2505 (2015). doi:10.1109/ICCV.2015.287
Ren_ICCV15.pdf (3.37 MB)
]