M³ constructs multi-scale variation-aware training measures via Morton octrees, cutting physics-weighted errors up to 4.7× and beating denser random data under aggressive subsampling.
A survey on deep active learning: Recent advances and new frontiers
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
citation-role summary
background 1
citation-polarity summary
fields
cs.AI 1years
2026 1verdicts
CONDITIONAL 1roles
background 1polarities
background 1representative citing papers
citing papers explorer
-
M$^3$: Reframing Training Measures for Discretized Physical Simulations
M³ constructs multi-scale variation-aware training measures via Morton octrees, cutting physics-weighted errors up to 4.7× and beating denser random data under aggressive subsampling.