Differentiable Economics for Randomized Affine Maximizer Auctions

Authors: Michael Curry, Tuomas Sandholm, John Dickerson

IJCAI 2023 | Conference PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Experimental By using the gradient-based optimization tools of differentiable economics, we can now train lottery AMAs, competing with or outperforming prior approaches in revenue. We train a lottery AMA on this setting and find revenue competitive with both previous AMA approaches [Sandholm and Likhodedov, 2015; Tang and Sandholm, 2012] as well as the Regret Net neural network approach (which performs better but is not perfectly strategyproof) [Duetting et al., 2019].
Researcher Affiliation Collaboration Michael Curry1 , Tuomas Sandholm2 and John Dickerson3 1 University of Zurich, ETH AI Center 2 Carnegie Mellon University, Optimized Markets, Inc., Strategic Machine, Inc., Strategy Robot, Inc. 3 University of Maryland curry@ifi.uzh.ch, sandholm@cs.cmu.edu, johnd@umd.edu
Pseudocode Yes For diagrams and pseudocode for training, please see the supplemental material.
Open Source Code No The paper mentions 'pseudocode for training' being in the supplemental material, but does not provide an explicit statement or link to the open-source code for the methodology itself.
Open Datasets Yes Spherical distribution In order to demonstrate a revenue improvement by offering lotteries, we consider a particular valuation distribution which we refer to as the spherical distribution for lack of a better name this is a distribution on a number of discrete, random points, scaled and normalized according to the proof construction in [Briest et al., 2010]. 2-bidder, 2-item uniform setting We also consider a 2-bidder, 2-item additive auction where item values are independently distributed on U[0, 1].
Dataset Splits No The paper mentions training on 'fresh valuation samples' and testing on '100000 sampled valuations', but does not specify a distinct validation set or explicit percentages for training/validation/test splits.
Hardware Specification Yes We train all auctions for 9000 steps, with 215 fresh valuation samples per gradient update, on cluster nodes with 2080Ti GPUs.
Software Dependencies No During training, we use the softmax function as a differentiable surrogate for the max and argmax operations: that is, arg maxk f(ak) softmaxτ(f(a1), , f(a K)), a and maxk f(ak) softmaxτ(f(a1), , f(a K)), f(a) . As the softmax temperature parameter τ approaches 0, this approach recovers the true argmax. Using this soft version of the AMA definition, we directly compute the total payment and differentiate it with respect to the parameters via the Jax autograd system [Bradbury et al., 2018] along with Haiku [Hennigan et al., 2020] and optax [Hessel et al., 2020]. The software names are mentioned but no specific version numbers are provided.
Experiment Setup Yes For lottery AMAs, we allow 2048 or 4096 allocations. The inverse of the softmax temperature parameter is 100; we use an Adam optimizer with learning rate of 0.01. We train all auctions for 9000 steps, with 215 fresh valuation samples per gradient update