AI Paper Atlas

Research map / Adversarial robustness and security / Attacks on text and graphs

VCG under Sybil (False-Name) Attacks - A Bayesian Analysis

Yotam Gafni, Ron Lavi, Moshe Tennenholtz

AAAI 2020 · Adversarial robustness and security › Attacks on text and graphs · approach: Membership inference and inversion

Read the paperCheck your idea against 137,217 papers

AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
Increase probability q of facing 1-type (single-item) bidders from below to at least 2/3 for n=2VCG with false-name bids · 2 items, n=2, any Ftruthful BNE existence: q >= \frac{2}{3}Improved
Set q to 1 (all adversaries are 1-type)VCG with false-name bids · any F and nattack benefit: q=1Improved
Add a third bidder (n>2) instead of n=2VCG with false-name bids · general distributions Fglobal granularity threshold q*: \frac{2}{3} -> 1Hurt
Increase number of bidders nVCG with false-name bids, beta distributions beta=1 · fully verified beta distribution casesgranularity threshold q*Hurt
Increase alpha of beta(alpha,1) distributionVCG with false-name bids · beta distributions beta=1granularity threshold q*Improved
Restrict per-item distribution to uniform for any n>=3VCG with false-name bids · uniform [0,1]q* for 1-type and split attack: <= \frac{1}{2}Improved
Use different per-item distributions F1 and F2 for 1-type and 2-type biddersVCG with false-name bids, n=2 · constructed counter-exampletruthful BNE existence: expected gain of attack >= 0.25(1-q)Hurt
Replace split attack by attack (1,1),(1,0.2) for a 1-type bidder in a two-point distribution, n=3VCG false-name attack · F: 1 w.p. 0.5, 0.1 w.p. 0.5q*: 0.0845241 -> 0.225403Hurt
Replace 1-type adversary (1,0.15) by 2-type adversary (2,0.15) against attack(0.1,0.1)1-type attacker (1,1) · Example in Discussionattacker utility loss vs truthful: 0.15 -> 0.7Hurt
Change adversary from 2-item demand to 1-item demand in intro examplebidder 2 type (2,0.5) false-name attack · intro examplepayment change from attacking: 0.8 -> 0.6Hurt
Vary q for 2-type split attack on uniform: threshold monotone in qsplit attack · uniformmonotonicity of benefit in qImproved

Closest papers by meaning

Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy