Research map / Learning theory and algorithms / Prediction and decision losses
Efficient Optimal Selection for Composited Advertising Creatives with Tree Structure
Read the paperCheck your idea against 108,967 papers
AI-extracted summary. Written by AI from the paper's title only (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| Use ingredient-tree Thompson sampling with dynamic-programming search instead of enumerating creatives | AES · Real-world dataset | search time: 729.70 $\pm$ 3.09 -> 165.04 $\pm$ 0.07 | Helped |
| Use ingredient-tree Thompson sampling with dynamic-programming search instead of MVT hill climbing | AES · Real-world dataset | search time: 656.96 $\pm$ 2.45 -> 165.04 $\pm$ 0.07 | Helped |
| Model ingredient interactions with a tree rather than selecting each ingredient independently | AES · Synthetic data | overall CTR: about 8.0% | Helped |
| Use the ingredient-tree model instead of context-free Egreedy or UCB | AES · Synthetic data | overall CTR | Helped |
| Explore sampled model weights and select creatives by dynamic programming rather than exploring the whole creative space | AES · Synthetic data | overall CTR | Helped |
| Use tree-based ingredient interactions rather than context-free or simple independent-ingredient selection | AES · Real-world dataset | overall CTR | Helped |
| Include both vertex and edge weights rather than using edge weights alone (ablation) | AES · Real-world dataset | overall CTR | Helped |
| Include both vertex and edge weights rather than using vertex weights alone (ablation) | AES · Real-world dataset | overall CTR | Helped |
| Constrain ingredient interactions to the tree structure rather than model all possible interactions (ablation) | AES · Real-world dataset | CTR estimator performance | No clear effect |
| Use dynamic programming rather than hill climbing for optimal creative selection | AES · Real-world dataset | overall CTR | Helped |
Closest papers by meaning
- Parallel Ranking of Ads and Creatives in Real-Time Advertising Systems (AAAI 2024 · similarity 0.88)
- Design Your Ad: Personalized Advertising Image and Text Generation with Unified Autoregressive Models (CVPR 2026 · similarity 0.85)
- A Combinatorial-Bandit Algorithm for the Online Joint Bid/Budget Optimization of Pay-per-Click Advertising Campaigns (AAAI 2018 · similarity 0.83)
- Online Conversion Rate Prediction via Multi-Interval Screening and Synthesizing under Delayed Feedback (AAAI 2024 · similarity 0.81)
- Percentile Risk-Constrained Budget Pacing for Guaranteed Display Advertising in Online Optimization (AAAI 2024 · similarity 0.81)
- GenAuction: A Generative Auction for Online Advertising (AAAI 2025 · similarity 0.81)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.