mirror of
https://github.com/zoo-apps/papers.git
synced 2026-11-03 14:59:55 +00:00
zeekayandHanzo Dev
dced1dcfb9
Bylines, paper authorship, correspondence addresses and the founder blurb under each post named an individual. They name the organisation that published the work instead: the brand on posts and pages, the research byline on papers, and research@ in place of a personal address. Both are what these repositories already used elsewhere — nothing new was coined. Co-authors are untouched. Where a paper named someone else alongside, they stay named; only the one byline changed. Bibliography entries, running heads, PDF metadata and affiliations move with it, so a citation still resolves to whoever the cited paper names. Sample and fixture data uses the placeholder name these tests already use elsewhere, so no assertion changed meaning. Co-authored-by: Hanzo Dev <[email protected]>
79 lines
3.2 KiB
TeX
79 lines
3.2 KiB
TeX
\documentclass[11pt,letterpaper]{article}
|
|
\usepackage{shared/zoocover}
|
|
|
|
% Packages
|
|
\usepackage[utf8]{inputenc}
|
|
\usepackage[T1]{fontenc}
|
|
\usepackage{amsmath,amssymb,amsthm}
|
|
\usepackage{algorithm}
|
|
\usepackage{algorithmic}
|
|
\usepackage{graphicx}
|
|
\usepackage{hyperref}
|
|
\usepackage{xcolor}
|
|
\usepackage{booktabs}
|
|
\usepackage{tikz}
|
|
\usepackage{pgfplots}
|
|
\pgfplotsset{compat=1.18}
|
|
\usetikzlibrary{shapes,arrows,positioning,calc}
|
|
\usepackage{listings}
|
|
\usepackage{caption}
|
|
\usepackage{subcaption}
|
|
% geometry loaded by zoocover
|
|
|
|
% Theorem environments
|
|
\newtheorem{theorem}{Theorem}[section]
|
|
\newtheorem{lemma}[theorem]{Lemma}
|
|
\newtheorem{proposition}[theorem]{Proposition}
|
|
\newtheorem{corollary}[theorem]{Corollary}
|
|
\newtheorem{definition}{Definition}[section]
|
|
|
|
% Custom commands
|
|
\newcommand{\Zoo}{\textsc{Zoo}}
|
|
\newcommand{\Lux}{\textsc{Lux}}
|
|
\newcommand{\FHE}{\textsc{fhe}}
|
|
\newcommand{\TFHE}{\textsc{tfhe}}
|
|
\newcommand{\CKKS}{\textsc{ckks}}
|
|
\newcommand{\TEE}{\textsc{tee}}
|
|
|
|
% Hyperref setup
|
|
\hypersetup{
|
|
colorlinks=true,
|
|
linkcolor=blue,
|
|
citecolor=blue,
|
|
urlcolor=blue
|
|
}
|
|
|
|
\title{\textbf{Zoo FHE: Privacy-Preserving AI Training and Inference on the Zoo Network}}
|
|
\author{
|
|
Antje Worring, Zoo Labs Foundation Research\thanks{[email protected]}\\
|
|
\textit{Zoo Labs Foundation}\\
|
|
\texttt{[email protected]}
|
|
}
|
|
|
|
\date{August 2024}
|
|
|
|
\begin{document}
|
|
|
|
\zoocoverpage
|
|
|
|
\begin{abstract}
|
|
We present \textbf{Zoo FHE}, a framework for privacy-preserving AI training and inference on the Zoo Network (chain ID 200200), a conservation-focused L2 on Lux Network. Zoo FHE combines two fully homomorphic encryption schemes---\TFHE{} for boolean/integer operations and \CKKS{} for approximate fixed-point arithmetic---to support encrypted model weights, encrypted inference, and encrypted training gradients. The system is exposed to smart contracts via an EVM precompile at address \texttt{0x0700}, enabling on-chain privacy-preserving computation without off-chain trusted parties. We introduce threshold decryption for model outputs using a $(t, n)$-threshold scheme where decryption requires cooperation of $t$ committee members from a set of $n$. We evaluate Zoo FHE on three conservation-critical applications: private wildlife monitoring (encrypted camera trap classification), confidential medical AI for veterinary diagnostics, and encrypted genomics for species conservation. Our benchmarks on Zoo testnet show encrypted inference latency of 340ms for a ResNet-18 classifier using \CKKS{}, and encrypted gradient aggregation throughput of 1,200 updates/second using \TFHE{}, both executing within the EVM precompile with gas costs under 500,000 per operation.
|
|
|
|
\textbf{Keywords}: fully homomorphic encryption, privacy-preserving AI, Zoo Network, encrypted inference, conservation AI
|
|
\end{abstract}
|
|
|
|
\input{sections/01-introduction}
|
|
\input{sections/02-background}
|
|
\input{sections/03-zoo-fhe-architecture}
|
|
\input{sections/04-encrypted-ai-primitives}
|
|
\input{sections/05-threshold-decryption}
|
|
\input{sections/06-conservation-applications}
|
|
\input{sections/07-performance-evaluation}
|
|
\input{sections/08-security-analysis}
|
|
\input{sections/09-related-work}
|
|
\input{sections/10-future-work}
|
|
\input{sections/11-conclusion}
|
|
\input{sections/99-acknowledgments}
|
|
|
|
\end{document}
|