pnnl/projection_ntk

RESOURCE

Abstract

This is a repository for work conducted in our research paper we would like to make open source. The main features are that we forked another open source library and added functionality that enabled us to compute neural-kernel objects using random vector projection software for both computer vision and large language text models. The new features are specifically code that enables the calculation of objects we define and refer to as the projection pseudo neural tangent kernel and the projection trace neural tangent kernelÍž both of which are approximations of the empirical neural tangent kernel. The software that computes the projections themselves are not part of the invention, but are necessarily tied to our software and licensed via MIT license, and BSD 2 license.
Developers:
Georgiev, Kristian [1] Engel, Andrew [2] Park, Sam [3] Leclerc, Guillaume Ilyas, Andrew Cohen-Wang, Benjamin Tanadi, Guspan Vendrow, Joshua Chiang, Tony [2] Choudhury, Sutanay [2]
  1. @MIT
  2. Pacific Northwest National Laboratory (PNNL), Richland, WA (United States)
  3. MIT CSAIL
Release Date:
2023-11-13
Project Type:
Open Source, Publicly Available Repository
Software Type:
Scientific
Licenses:
MIT License
BSD 2-clause "Simplified" License
Sponsoring Org.:
Code ID:
115804
Site Accession Number:
Battelle IPID 32904-E
Research Org.:
Pacific Northwest National Laboratory (PNNL), Richland, WA (United States)
Country of Origin:
United States

RESOURCE

Citation Formats

Georgiev, Kristian, Engel, Andrew, Park, Sam, Leclerc, Guillaume, Ilyas, Andrew, Cohen-Wang, Benjamin, Tanadi, Guspan, Vendrow, Joshua, Chiang, Tony, and Choudhury, Sutanay. pnnl/projection_ntk. Computer Software. https://github.com/pnnl/projection_ntk. USDOE. 13 Nov. 2023. Web. doi:10.11578/dc.20231113.1.
Georgiev, Kristian, Engel, Andrew, Park, Sam, Leclerc, Guillaume, Ilyas, Andrew, Cohen-Wang, Benjamin, Tanadi, Guspan, Vendrow, Joshua, Chiang, Tony, & Choudhury, Sutanay. (2023, November 13). pnnl/projection_ntk. [Computer software]. https://github.com/pnnl/projection_ntk. https://doi.org/10.11578/dc.20231113.1.
Georgiev, Kristian, Engel, Andrew, Park, Sam, Leclerc, Guillaume, Ilyas, Andrew, Cohen-Wang, Benjamin, Tanadi, Guspan, Vendrow, Joshua, Chiang, Tony, and Choudhury, Sutanay. "pnnl/projection_ntk." Computer software. November 13, 2023. https://github.com/pnnl/projection_ntk. https://doi.org/10.11578/dc.20231113.1.
@misc{ doecode_115804,
title = {pnnl/projection_ntk},
author = {Georgiev, Kristian and Engel, Andrew and Park, Sam and Leclerc, Guillaume and Ilyas, Andrew and Cohen-Wang, Benjamin and Tanadi, Guspan and Vendrow, Joshua and Chiang, Tony and Choudhury, Sutanay},
abstractNote = {This is a repository for work conducted in our research paper we would like to make open source. The main features are that we forked another open source library and added functionality that enabled us to compute neural-kernel objects using random vector projection software for both computer vision and large language text models. The new features are specifically code that enables the calculation of objects we define and refer to as the projection pseudo neural tangent kernel and the projection trace neural tangent kernelÍž both of which are approximations of the empirical neural tangent kernel. The software that computes the projections themselves are not part of the invention, but are necessarily tied to our software and licensed via MIT license, and BSD 2 license.},
doi = {10.11578/dc.20231113.1},
url = {https://doi.org/10.11578/dc.20231113.1},
howpublished = {[Computer Software] \url{https://doi.org/10.11578/dc.20231113.1}},
year = {2023},
month = {nov}
}