On Calibrated Predictions for Auction Selection Mechanisms

نویسندگان

  • H. Brendan McMahan
  • Omkar Muralidharan
چکیده

Calibration is a basic property for prediction systems, and algorithms for achieving it are well-studied in both statistics and machine learning. In many applications, however, the predictions are used to make decisions that select which observations are made. This makes calibration difficult, as adjusting predictions to achieve calibration changes future data. We focus on click-through-rate (CTR) prediction for search ad auctions. Here, CTR predictions are used by an auction that determines which ads are shown, and we want to maximize the value generated by the auction. We show that certain natural notions of calibration can be impossible to achieve, depending on the details of the auction. We also show that it can be impossible to maximize auction efficiency while using calibrated predictions. Finally, we give conditions under which calibration is achievable and simultaneously maximizes auction efficiency: roughly speaking, bids and queries must not contain information about CTRs that is not already captured by the predictions.

برای دانلود رایگان متن کامل این مقاله و بیش از 32 میلیون مقاله دیگر ابتدا ثبت نام کنید

ثبت نام

اگر عضو سایت هستید لطفا وارد حساب کاربری خود شوید

منابع مشابه

A Comparative Study of Multi-Attribute Continuous Double Auction Mechanisms

Auctions have been as a competitive method of buying and selling valuable or rare items for a long time. Single-sided auctions in which participants negotiate on a single attribute (e.g. price) are very popular. Double auctions and negotiation on multiple attributes create more advantages compared to single-sided and single-attribute auctions. Nonetheless, this adds the complexity of the auctio...

متن کامل

Auction Resource Allocation Mechanisms in Grids of Heterogeneous Computers

This paper examines economic resource allocation through a number of auction types for a grid of ewaste computers. It examines the time to complete tasks and the energy usage of completing the tasks on a grid. A model of a simulated grid is developed and used to evaluate the resource allocation mechanisms. The model is an agent-based simulation where by user agents submit tasks to node agents t...

متن کامل

Selecting Efficient Service-providers in Electric Power Distribution Industry Using Combinatorial Reverse Auction

In this paper, a combinatorial reverse auction mechanism is proposed for selecting the most efficient service-providers for resolving sustained power interruptions in multiple regions of an electric power distribution company’s responsibility area. Through this mechanism, supplying the required service in each region is assigned to only one potential service-provider considering two criteria in...

متن کامل

How Well Do Selection Models Perform? Assessing the Acuracy of Art Auction Pre-sale Estimates.

Art auction catalogs provide a pre-sale prediction interval for the price each item is expected to fetch. When the owner consigns art work to the auction house, a reserve price is agreed upon, which is not announced to the bidders. If the highest bid does not reach it, the item is brought in. Since only the prices of the sold items are published, analysts only have a biased sample to examine du...

متن کامل

A Multi-attribute Reverse Auction Framework Under Uncertainty to the Procurement of Relief Items

One of the main activities of humanitarian logistics is to provide relief items for survivors in case of a disaster. To facilitate the procurement operation, this paper proposes a bidding framework for supplier selection and optimal allocation of relief items. The proposed auction process is divided into the announcement construction, bid construction and bid evaluation phases. In the announcem...

متن کامل

ذخیره در منابع من


  با ذخیره ی این منبع در منابع من، دسترسی به آن را برای استفاده های بعدی آسان تر کنید

عنوان ژورنال:
  • CoRR

دوره abs/1211.3955  شماره 

صفحات  -

تاریخ انتشار 2012