Machine Learning in an Auction Environment
Authors: Patrick Hummel, R. Preston McAfee
JMLR 2016 | Venue PDF | Archive PDF | Plain Text | LLM Run Details
| Reproducibility Variable | Result | LLM Response |
|---|---|---|
| Research Type | Experimental | Finally, we empirically verify these findings through simulations and illustrate that adding active learning in the auction environment we consider only changes overall efficiency by a few hundredths of a percentage point. ... We now seek to empirically verify that the benefit that can be obtained from active exploration is indeed quite small by conducting simulations under some empirically realistic scenarios. |
| Researcher Affiliation | Industry | Patrick Hummel EMAIL 1600 Amphitheatre Parkway Google Inc. Mountain View, CA 94043, USA. R. Preston Mc Afee EMAIL One Microsoft Way Microsoft Corp. Redmond, WA 98052, USA |
| Pseudocode | No | The paper describes mathematical formulations and theoretical analysis of the problem, but does not include any explicitly labeled pseudocode or algorithm blocks. |
| Open Source Code | No | The paper does not contain any explicit statements about releasing source code for the methodology, nor does it provide any links to code repositories. |
| Open Datasets | No | The paper conducts simulations based on chosen distributions and parameters (e.g., beta distribution for click-through rate, lognormal distribution for competing bids), which are informed by empirical evidence from other studies. However, it does not use or provide concrete access information (link, DOI, repository, or specific citation to a named dataset) for a publicly available dataset that the authors themselves used or created for their experiments. |
| Dataset Splits | No | The paper describes simulations with specified parameters and scenarios, not experiments on a dataset with explicit training, validation, or test splits. Therefore, information about dataset splits is not applicable or provided. |
| Hardware Specification | No | The paper describes the simulation setup, including parameters and distributions, but does not provide any specific details about the hardware (e.g., GPU models, CPU types, or cloud computing resources) used to run these simulations. |
| Software Dependencies | No | The paper does not specify any software libraries, frameworks, or programming languages with version numbers that were used to conduct the simulations or implement the methodology. |
| Experiment Setup | Yes | Throughout we assume that payoffs are discounted at a rate of δ = 0.9995 and that there are T = 10000 time periods. ... we assume that the CPC bidder s clickthrough rate is initially drawn from a beta distribution with parameters α and β. ... We thus consider values of α and β satisfying (α, β) = (10, 1000) and (20, 2000). ... we assume throughout that the CPM bidder s bid is drawn from a lognormal distribution with parameters µ and σ2. We use a value of σ2 = log(2) ... We thus consider a range of values of µ from 4.25 ... to 3.5 |