Unverified50% confidenceCautionExact time
M-plicits checkpoints are individually fitted independent surfaces, with each shape trained separately; fitting a new shape requires retraining, and the method does not possess cross-shape generalization capability.
1
Sources
50%
Confidence
Long-term
Relevance
10/5/2026
First Seen
Sources
Related Entities
Related Claims
UnverifiedNeural representations trained by overfitting to a single shape are called overfitting-based or per-shape fitting methods, while methods with generalization capability are called feed-forward methods (e.g., DeepSDF, OccNet).64% similarUnverifiedGPEvac adopts a permutation-invariant scoring mechanism, allowing a single trained policy to be directly transferred to building layouts with varying topologies and scales without retraining.61% similarUnverifiedAs a per-scene-trained method, FactorSplat still requires training a separate model for each scan.60% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/977093API
curl https://kongchang.com/api/v1/knowledge/claims/977093MCP
get_claim(id=977093)