Latest Research
The Shavit Lab has significantly advanced the understanding of scalable algorithms and data structures essential for multi-core processors. In recent years, the lab has expanded into the burgeoning field of connectomics, where we apply our computational expertise to unravel the complexities of neural networks in the brain. By developing sophisticated algorithms for processing and analyzing large-scale neural data, our lab aims to reconstruct the intricate connectivity patterns of the brain with high precision. Additionally, the lab studies techniques to make large machine learning models more efficient by leveraging sparsity, and how to ensure AI safety. Our interdisciplinary approach aims to advance both fields of neuroscience and artificial intelligence.
News
- January 22, 2025: Our work Wasserstein Distances, Neuronal Entanglement, and Sparsity has been accepted into ICLR 2025 as a Spotlight Presentation!
- December 23, 2024: Our work Presynaptic input synchrony at scale has been accepted into COSYNE 2025!
- December 3, 2024: Our work Jailbreak Defense in a Narrow Domain: Limitations of Existing Methods and a New Transcript-Classifier Approach has been accepted into NeurIPS 2024 Workshop on New Frontiers in Adversarial Machine Learning and Workshop on Socially Responsible Language Modelling Research!
- October 16, 2024: Our work Structure Matters: Deciphering Neural Network's Properties from its Structure has been accepted into NeurIPS 2024 Workshop on Symmetry and Geometry in Neural Representations!
- September 5, 2024: Introducing a new manuscript: On the Complexity of Neural Computation in Superposition.
- May 29, 2024: Introducing two new manuscripts: A connectomics-driven analysis reveals novel characterization of border regions in mouse visual cortex and Sparse Expansion and Neuronal Disentanglement.
- May 28, 2024: We celebrated the end of the year with a delicious barbecue dinner!
