Simple Mechanisms for Welfare Maximization in Rich Advertising Auctions

Authors: Gagan Aggarwal, Kshipra Bhawalkar, Aranyak Mehta, Divyarthi Mohan, Alexandros Psomas

NeurIPS 2022 | Conference PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Experimental Finally, we experimentally test our algorithms on real-world data. In this section we present some empirical results for our truthful mechanisms.
Researcher Affiliation Collaboration Gagan Aggarwal Google Research gagana@google.com Kshipra Bhawalkar Google Research kshipra@google.com Aranyak Mehta Google Research aranyak@google.com Divyarthi Mohan Tel Aviv University divyarthim@tau.ac.il Alexandros Psomas Purdue University apsomas@cs.purdue.edu
Pseudocode Yes Algorithm 1 (ALGI). First, run ALGB. Let Wi be the space allotted to advertiser i. Second, post-process to allocate the ad j with maximum value that fits in Wi, i.e. j 2 argmaxwij Wibij. Any remaining space is left unallocated.
Open Source Code No Did you include the code, data, and instructions needed to reproduce the main experimental results (either in the supplemental material or as a URL)? [No] The data is proprietary.
Open Datasets No We evaluated our algorithms from real world data obtained from a large search engine. The data consists of a sample of approximately 11000 queries, selected to have at least 6 advertisers each. The assets are proprietary
Dataset Splits No The paper states it uses a sample of real-world data but does not provide specific training, validation, or test dataset splits.
Hardware Specification No We ran our algorithms on a single machine. This statement does not provide specific hardware details such as CPU/GPU models or memory.
Software Dependencies No The paper does not provide specific software dependencies with version numbers.
Experiment Setup Yes We use 500 as the space limit as that is larger than the space of any individual rich ad.