Architectures of neuronal circuits

L Luo - Science, 2021‏ - science.org
BACKGROUND The human brain contains about 100 billion neurons, each of which makes
thousands of synaptic connections. Although individual neurons can themselves be …

Mechanistic Interpretability for AI Safety--A Review

L Bereska, E Gavves - arxiv preprint arxiv:2404.14082, 2024‏ - arxiv.org
Understanding AI systems' inner workings is critical for ensuring value alignment and safety.
This review explores mechanistic interpretability: reverse engineering the computational …

The tumor microenvironment shows a hierarchy of cell-cell interactions dominated by fibroblasts

S Mayer, T Milo, A Isaacson, C Halperin… - Nature …, 2023‏ - nature.com
The tumor microenvironment (TME) is comprised of non-malignant cells that interact with
each other and with cancer cells, critically impacting cancer biology. The TME is complex …

Finding neurons in a haystack: Case studies with sparse probing

W Gurnee, N Nanda, M Pauly, K Harvey… - arxiv preprint arxiv …, 2023‏ - arxiv.org
Despite rapid adoption and deployment of large language models (LLMs), the internal
computations of these models remain opaque and poorly understood. In this work, we seek …

Explainability in graph neural networks: A taxonomic survey

H Yuan, H Yu, S Gui, S Ji - IEEE transactions on pattern …, 2022‏ - ieeexplore.ieee.org
Deep learning methods are achieving ever-increasing performance on many artificial
intelligence tasks. A major limitation of deep models is that they are not amenable to …

On explainability of graph neural networks via subgraph explorations

H Yuan, H Yu, J Wang, K Li, S Ji - … conference on machine …, 2021‏ - proceedings.mlr.press
We consider the problem of explaining the predictions of graph neural networks (GNNs),
which otherwise are considered as black boxes. Existing methods invariably focus on …

Xgnn: Towards model-level explanations of graph neural networks

H Yuan, J Tang, X Hu, S Ji - Proceedings of the 26th ACM SIGKDD …, 2020‏ - dl.acm.org
Graphs neural networks (GNNs) learn node features by aggregating and combining
neighbor information, which have achieved promising performance on many graph tasks …

Predictive coding: a theoretical and experimental review

B Millidge, A Seth, CL Buckley - arxiv preprint arxiv:2107.12979, 2021‏ - arxiv.org
Predictive coding offers a potentially unifying account of cortical function--postulating that the
core function of the brain is to minimize prediction errors with respect to a generative model …

[HTML][HTML] Zoom in: An introduction to circuits

C Olah, N Cammarata, L Schubert, G Goh, M Petrov… - Distill, 2020‏ - distill.pub
Many important transition points in the history of science have been moments when science
“zoomed in.” At these points, we develop a visualization or tool that allows us to see the …

[HTML][HTML] How to build the virtual cell with artificial intelligence: Priorities and opportunities

C Bunne, Y Roohani, Y Rosen, A Gupta, X Zhang… - Cell, 2024‏ - cell.com
Cells are essential to understanding health and disease, yet traditional models fall short of
modeling and simulating their function and behavior. Advances in AI and omics offer …