analiza aiqbench
This arXiv preprint proposes a three-stage framework (Attribution-Mapping-Execution) for allocating value among training data, base models, fine-tuning, and prompts in generative AI pipelines. Its claims rest on comparisons with human reference judgments, and the author has withdrawn this version because the problem formulation, scope, and evaluation changed, so the results should be treated as provisional. Readers interested in data attribution and compensation for AI training inputs should watch for a revised version before drawing conclusions.
AME: A Multi-Type Contributor Attribution Framework in Generative AI Markets
arXiv:2606.16075v2 Announce Type: replace Abstract: Generative AI enables value creation through multi-stage collaboration among heterogeneous contributors, including training data, base models, fine-tuning behaviors, and prompts. However, how to fairly allocate the data value remains largely unexplored. This paper formulates multi-stage generative AI value allocation as a new research problem and identifies three core challenges: heterogeneous data contribution valuation, data rights mapping, and trustworthy execution. We propose AME (Attribution-Mapping-Execution) framework, a unified framework that integrates data contribution valuation, data rights mapping, and trustworthy execution into a single workflow. Experimental results demonstrate that AME framework achieves data value allocation outcomes more consistent with human reference judgments while maintaining low-cost trustworthy execution. Our work provides an initial foundation for value assessment and revenue allocation in generative AI data markets.
Abstract:Generative AI enables value creation through multi-stage collaboration among heterogeneous contributors, including training data, base models, fine-tuning behaviors, and prompts. However, how to fairly allocate the data value remains largely unexplored. This paper formulates multi-stage generative AI value allocation as a new research problem and identifies three core challenges: heterogeneous data contribution valuation, data rights mapping, and trustworthy execution. We propose AME (Attribution-Mapping-Execution) framework, a unified framework that integrates data contribution valuation, data rights mapping, and trustworthy execution into a single workflow. Experimental results demonstrate that AME framework achieves data value allocation outcomes more consistent with human reference judgments while maintaining low-cost trustworthy execution. Our work provides an initial foundation for value assessment and revenue allocation in generative AI data markets.
We have decided to withdraw this manuscript because the research has undergone substantial revisions in its problem formulation, scope, and evaluation. As a result, the current version no longer adequately represents the revised study. We therefore withdraw this earlier version to avoid potential misunderstandings regarding the research findings
Machine Learning (cs.LG); Computer Vision and Pattern Recognition (cs.CV)
arXiv:2606.16075 [cs.LG]
(or arXiv:2606.16075v2 [cs.LG] for this version)
Text integral sub licența deschisă a publicației · 186 cuvinte
Subiecte conexe
Priority-Aware Shapley Value
arXiv · 16 iun. 2026
AME: A Multi-Type Contributor Attribution Framework in Generative AI Markets
arXiv · 16 iun. 2026
Shapley Valuation of Finite-Copy Quantum Data Depends on Physical Access
arXiv · 7 sept. 2026
Digital Twins for Opinion Dynamics: A Generative LLM Framework for Social Networks
arXiv · 18 sept. 2026