Notice: The reproducibility variables underlying each score are classified using an automated LLM-based pipeline, validated against a manually labeled dataset. LLM-based classification introduces uncertainty and potential bias; scores should be interpreted as estimates. Full accuracy metrics and methodology are described in Coakley et alK. L. Coakley, T. Snelleman, H. Hoos, and O. E. Gundersen, "The embrace of open science: An analysis of a decade of AI research and 56 800 conference papers," Under Review, 2026..
Beyond Last-Click: An Optimal Mechanism for Ad Attribution
Authors: Nan An, Weian Li, Qi Qi, Changyuan Yu, Liang Zhang
NeurIPS 2025 | Venue PDF | LLM Run Details
| Reproducibility Variable | Result | LLM Response |
|---|---|---|
| Research Type | Experimental | Finally, numerical experiments are conducted to show that PVM consistently outperforms LCM in terms of attribution accuracy and fairness. |
| Researcher Affiliation | Collaboration | Nan An Gaoling School of Artificial Intelligence Renmin University of China Beijing, China EMAIL Weian Li School of Software Shandong University Jinan, China EMAIL Qi Qi Gaoling School of Artificial Intelligence Renmin University of China Beijing, China EMAIL Changyuan Yu Baidu Inc. Beijing, China EMAIL Liang Zhang Gaoling School of Artificial Intelligence Renmin University of China Beijing, China EMAIL |
| Pseudocode | No | The paper defines the mechanisms (LCM and PVM) using mathematical definitions and descriptive text (Definition 2 and Definition 3) but does not include structured pseudocode or an algorithm block. |
| Open Source Code | No | The numerical experiments are based on data that include sensitive information from both users and advertising platforms. Due to privacy concerns and contractual restrictions, we are unable to release the raw data. |
| Open Datasets | No | The numerical experiments are based on data that include sensitive information from both users and advertising platforms. Due to privacy concerns and contractual restrictions, we are unable to release the raw data. However, although we cannot disclose the data, we provide a description of its general distribution in supplement materials to assist in the reproducibility of our results, along with a detailed explanation of the simulation process and experimental setup in supplement materials. |
| Dataset Splits | No | The paper states, "Each configuration was evaluated using 5 * 10^4 simulated user paths, repeated over 10 independent runs," which describes the simulation volume, but it does not specify traditional training/test/validation splits for a dataset. |
| Hardware Specification | No | The paper states, "As introduced in supplement materials," regarding compute resources. However, the provided text does not contain the supplement materials, and therefore no specific hardware details are mentioned in the main paper text. |
| Software Dependencies | No | The paper does not mention any specific software dependencies with version numbers. |
| Experiment Setup | Yes | Each configuration was evaluated using 5 * 10^4 simulated user paths, repeated over 10 independent runs. |