Diffauction enables fair auction driven image ads with buyer click prediction
DiffAuction: An Auction-Driven Framework for Generative Image Advertising
Computer Vision and Pattern Recognition
Summary
Showing multiple advertisers fairly in a single AI-generated image is hard and slow with current methods. The authors created DiffAuction, a system that quickly adjusts how much space each advertiser gets in the image. They also introduce a way to estimate how likely people are to click on each ad part using AI models that look at the image and understand user interest. Finally, they design a fair auction system that decides which advertisers win and how much they pay, aiming for the platform to earn more money. Their tests show their system controls ad shares better and increases auction revenue.
What this means in practice
- •For digital marketing teams: Generate images combining multiple advertisers while fairly allocating space and estimating click interest for ad pricing.$Commercial implications: Enables platforms to auction shared generative image space to advertisers with payment rules based on predicted user clicks.
- •For ad platform engineers: Implement auctions that select generated image compositions maximizing expected revenue with truthful advertiser bidding incentives.
Authors
Yijie Su, Ruiyang Si, Tonghan Wang
Abstract
Diffusion models enable multiple advertisers to compete for exposure within a shared generated image. Yet existing approaches are computationally costly when generating images under varying target-share configurations, and it remains unclear how the resulting outcomes should be evaluated and priced. To address these gaps, we propose DiffAuction, a framework integrating controllable generation, click-intent evaluation, and auction design. First, our training-free Residual Share Calibration (RSC) decomposes joint-attention updates into product-specific residuals and rescales them according to target shares. We further introduce Share-Conditioned Prompting (SCP) to adapt the viewpoint to each desired allocation. Second, we develop an evaluator that uses a VLM to extract visual evidence and an LLM to assess Click-Through Intent (CTI), a proxy for click propensity, along four dimensions: Product Prominence, Brand Recognizability, Scene Integration, and Contextual Value Communication. We validate its assessments against judgments from 10 trained annotators on 1000 images. Third, we formulate this advertising setting as an auction over jointly feasible generation configurations. Each configuration has a public expected CTI vector, while advertisers hold configuration-specific private values per unit of CTI that map click intent into commercial value. We adopt a learnable affine maximizer auction (AMA), which jointly selects a shared generation configuration and determines advertiser-specific payments to maximize expected platform revenue. The resulting allocation and payment rules guarantee dominant-strategy incentive compatibility (DSIC) and individual rationality (IR). Experiments show that our generation method strengthens the monotonic relationship between target shares and CTI relative to baselines, while the learned AMA achieves higher platform revenue in the resulting auction setting.