ARC-AGI-2挑战中的多视角Transformer
Multi-Perspective Transformers in ARC-AGI-2 Challenge
AI总结:
本文提出使用TinyLM结合测试时训练和专家产品方法解决ARC-AGI-2视觉谜题,实现训练集96.1%和评估集21.7%的准确率。
摘要:
ARC-AGI-2是一个衡量机器从有限示例中泛化、解释符号意义并灵活应用规则的基准测试。本文讨论了使用TinyLM解决该挑战的方法,包括测试时训练(TTT)和专家产品(POE)的微调。模型在训练集上达到96.1%的准确率,在评估集上达到21.7%的准确率。
英文摘要:
ARC-AGI-2 is a benchmark of human-intuitive visual puzzles that measures a machine's ability to generalize from limited examples, interpret symbolic meaning, and flexibly apply rules in varying contexts. In this paper, we discuss our approach to solving the ARC-AGI-2 puzzles with TinyLM, with additional fine-tuning at test time, including Test-Time-Training (TTT) and Products of Experts (POE). Our model achieves 96.1% accuracy on the training set and 21.7% accuracy on the evaluation set.
机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。