Research
A-RSI adapts task refresh timing for zero-data multimodal self-improvement
A-RSI refreshes its task pool when training reward variation and accuracy signals decline, without using validation or test scores. On an eleven-benchmark suite, the performance average rose from 57.77 to 61.05 using 800 fresh tasks.
Every outlet that ran it 1 article
Hugging Face PapersResearchLead
A-RSI: Adaptive Task Refresh for Zero-Data Multimodal Self-Improvement