Welfare Guarantees from Data

Authors: Darrell Hoy, Denis Nekipelov, Vasilis Syrgkanis

NeurIPS 2017 | Conference PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Experimental We apply our approach to datasets from a sponsored search auction system and find empirical results that are a significant improvement over bounds from worst-case analysis.
Researcher Affiliation Collaboration Darrell Hoy University of Maryland darrell.hoy@gmail.com Denis Nekipelov University of Virginia denis@virginia.edu Vasilis Syrgkanis Microsoft Research vasy@microsoft.com
Pseudocode No The paper does not contain any pseudocode or algorithm blocks.
Open Source Code No The paper does not provide any links to open-source code or state that code is available.
Open Datasets No The paper states: 'We applied our approach to a real world dataset from a sponsored search auction system' and 'We applied our analysis to the Bing Ads sponsored search auction system.' This is proprietary data and no public access information (link, citation, or repository) is provided.
Dataset Splits No The paper does not explicitly provide details about train/validation/test splits. It mentions using 'data generated by advertisers repeatedly participating in a sponsored search auction' and retrieving 'data of auctions for the phrase for the period of a week', but no specific splitting methodology for reproducibility.
Hardware Specification No The paper does not mention any specific hardware used for running the experiments (e.g., GPU models, CPU types, or cloud infrastructure specifications).
Software Dependencies No The paper does not list any software dependencies with specific version numbers. It mentions the 'Bing Ads sponsored search auction system' but not the software environment or libraries used for analysis.
Experiment Setup Yes The paper describes the experimental setup in detail, including the data source, the simulation of auctions for each phrase under alternative bids ('simulating the auctions for the week under any alternative bid an advertiser could submit (bids are multiples of cents)'). It also outlines how quantities like the average threshold (ˆTi) and expected revenue (d REV) were estimated, and how ˆT was computed ('computing the quantity ˆT = maxx X P i [n] ˆTi xi = i ˆTi γi αm 1(i)').