Web: http://arxiv.org/abs/2208.03309

Sept. 19, 2022, 1:12 a.m. | Wenxiao Wang, Alexander Levine, Soheil Feizi

cs.LG updates on arXiv.org arxiv.org

Data poisoning considers an adversary that distorts the training set of
machine learning algorithms for malicious purposes. In this work, we bring to
light one conjecture regarding the fundamentals of data poisoning, which we
call the Lethal Dose Conjecture. The conjecture states: If $n$ clean training
samples are needed for accurate predictions, then in a size-$N$ training set,
only $\Theta(N/n)$ poisoned samples can be tolerated while ensuring accuracy.
Theoretically, we verify this conjecture in multiple cases. We also offer a …

arxiv data

More from arxiv.org / cs.LG updates on arXiv.org

Machine Learning Product Manager (Canada, Remote)

@ FreshBooks | Canada

Data Engineer

@ Amazon.com | Irvine, California, USA

Senior Autonomy Behavior II, Performance Assessment Engineer

@ Cruise LLC | San Francisco, CA

Senior Data Analytics Engineer

@ Intercom | Dublin, Ireland

Data Analyst Intern

@ ADDX | Singapore

Data Science Analyst - Consumer

@ Yelp | London, England, United Kingdom

Senior Data Analyst - Python+Hadoop

@ Capco | India - Bengaluru

DevOps Engineer, Data Team

@ SingleStore | Hyderabad, India

Software Engineer (Machine Learning, AI Platform)

@ Phaidra | Remote

Sr. UI/UX Designer - Artificial Intelligence (ID:1213)

@ Truelogic Software | Remote, anywhere in LATAM

Analytics Engineer

@ carwow | London, England, United Kingdom

HRIS Data Analyst

@ SecurityScorecard | Remote