A deep-learning framework parametrizes dual variables with flow conservation to produce revenue upper bounds and near-optimality certificates for multi-item multi-bidder DSIC auctions, with a lifting construction that extends discrete certificates to continuous uniform and general distributions.
Constantinos Daskalakis, Alan Deckelbaum, and Christos Tzamos
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.GT 1years
2026 1verdicts
UNVERDICTED 1representative citing papers
citing papers explorer
-
Duality for Optimal Multi-Item, Multi-Bidder Auction Design: Revenue Certificates through Deep Learning
A deep-learning framework parametrizes dual variables with flow conservation to produce revenue upper bounds and near-optimality certificates for multi-item multi-bidder DSIC auctions, with a lifting construction that extends discrete certificates to continuous uniform and general distributions.