analiza aiqbench
This paper introduces VERDI, a framework designed to improve the transferability of optimization strategies across different foundation world models by rigorously distinguishing between simple retrieval and actual knowledge transfer. By validating candidate strategies on target-side verifiers before reuse and evolving diagnostic representations based on contradictions, the method significantly reduces search costs and GPU usage while minimizing negative transfer in embodied intelligence tasks.
VERDI:Verified Contradictions for Continual World-Model Optimization
arXiv:2608.09537v2 Announce Type: replace Abstract: Foundation world models have made remarkable progress in planning, simulation, and embodied intelligence. However, optimizing a pretrained world model toward a user-specified objective remains difficult: each campaign typically rediscovers optimization strategies from scratch, and the resulting knowledge rarely transfers to the next model. Existing research agents automate the optimization loop but treat successful strategies as directly reusable recipes, without principled safeguards for when transfer is appropriate. We argue instead that retrieval is not transfer: a strategy validated on one model is at best an optimization hypothesis for another, and becomes transferable knowledge only after target-side experimental valida- tion. Guided by this principle, we propose VERDI , a continual framework for evidence-licensed world model optimization. VERDI characterizes each world model through shared inference-time probes to construct an Optimization Fin- gerprint, retrieves relevant prior experience as ranked hypotheses, and validates every candidate under a frozen target-side verifier before admitting it as reusable evidence; contradictions among nearby fingerprints further trigger probe evolution, continually refining the diagnostic representation itself. Experiments on Ctrl-World, the Cosmos family, and RoboCoin show that VERDI reduces search cost by 68%, GPU cost by 69%, and negative transfer from 0.34 to 0.06, while predicting transfer outcomes with 83% sign accuracy.
Abstract:Foundation world models have made remarkable progress in planning, simulation, and embodied intelligence. However, optimizing a pretrained world model toward a user-specified objective remains difficult: each campaign typically rediscovers optimization strategies from scratch, and the resulting knowledge rarely transfers to the next model. Existing research agents automate the optimization loop but treat successful strategies as directly reusable recipes, without principled safeguards for when transfer is appropriate. We argue instead that retrieval is not transfer: a strategy validated on one model is at best an optimization hypothesis for another, and becomes transferable knowledge only after target-side experimental valida- tion. Guided by this principle, we propose VERDI , a continual framework for evidence-licensed world model optimization. VERDI characterizes each world model through shared inference-time probes to construct an Optimization Fin- gerprint, retrieves relevant prior experience as ranked hypotheses, and validates every candidate under a frozen target-side verifier before admitting it as reusable evidence; contradictions among nearby fingerprints further trigger probe evolution, continually refining the diagnostic representation itself. Experiments on Ctrl-World, the Cosmos family, and RoboCoin show that VERDI reduces search cost by 68%, GPU cost by 69%, and negative transfer from 0.34 to 0.06, while predicting transfer outcomes with 83% sign accuracy.
Artificial Intelligence (cs.AI)
arXiv:2608.09537 [cs.AI]
(or arXiv:2608.09537v2 [cs.AI] for this version)
https://doi.org/10.48550/arXiv.2608.09537 Focus to learn more arXiv-issued DOI via DataCite
arXivLabs is a framework that allows collaborators to develop and share new arXiv features directly on our website.
Both individuals and organizations that work with arXivLabs have embraced and accepted our values of openness, community, excellence, and user data privacy. arXiv is committed to these values and only works with partners that adhere to them.
Have an idea for a project that will add value for arXiv's community? Learn more about arXivLabs.
Text integral sub licența deschisă a publicației · 292 cuvinte
Subiecte conexe
A game theory for foundation models shows new paths to rational cooperation through similarity inference
arXiv · 5 aug. 2026
Unified Learning of Temporal Task Structure and Action Timing for Bimanual Robot Manipulation
arXiv · 22 sept. 2026
ReDiSC: A Reparameterized Masked Diffusion Model for Scalable Node Classification with Structured Predictions
arXiv · 30 iul. 2026
Beyond the Golden Teacher: Enhancing Graph Learning through LLM-GNN Co-teaching
arXiv · 12 iun. 2026