Bezier Trajectory Matching replaces SGD trajectories with quadratic Bezier surrogates to create a structured lower-rank supervision signal for dataset condensation, matching or exceeding standard trajectory matching on clinical data especially in low-prevalence and low-budget settings.
The model consists of a single residual temporal block with 64 channels, kernel size 9, dilation 1, BatchNorm, PReLU activations, and dropout (0.75)
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.LG 1years
2026 1verdicts
UNVERDICTED 1representative citing papers
citing papers explorer
-
Geometric Characterisation and Structured Trajectory Surrogates for Clinical Dataset Condensation
Bezier Trajectory Matching replaces SGD trajectories with quadratic Bezier surrogates to create a structured lower-rank supervision signal for dataset condensation, matching or exceeding standard trajectory matching on clinical data especially in low-prevalence and low-budget settings.