AI Paper Atlas

Research map / Learning theory and algorithms / Combinatorial and search optimization

Asymptotic Risk of Bézier Simplex Fitting

Akinori Tanaka, Akiyoshi Sannai, Ken Kobayashi, Naoki Hamada

AAAI 2020 · Learning theory and algorithms › Combinatorial and search optimization · approach: Globally optimal geometric optimization · 6 citations (OpenAlex)

Read the paperCheck your idea against 137,217 papers

AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
Use optimal skeleton-wise sample allocation instead of all-at-once fittingBézier simplex fitting, degree 2 · Simplex dimensions M=2–8asymptotic risk: 3.0 / N -> 2.91421 / N; 6.0 / N -> 5.49632 / N; 10.0 / N -> 8.66660 / N; 15.0 / N -> 11.9936 / N; 21.0 / N -> 15.1663 / Improved
Use optimal skeleton-wise sample allocation instead of all-at-once fittingBézier simplex fitting, degree 3 · Simplex dimension M=2asymptotic risk: 4.0 / N -> 3.72726 / NImproved
Use optimal skeleton-wise sample allocation instead of all-at-once fittingBézier simplex fitting, degree 3 · Simplex dimensions M=3–8asymptotic risk: 10.0 / N -> 10.6472 / N; 20.0 / N -> 23.8821 / N; 35.0 / N -> 44.7548 / N; 56.0 / N -> 73.1387 / N; 84.0 / N -> 107.570 Hurt
Use optimal rather than equal skeleton sample allocationInductive skeleton fitting · Synthetic instances, D=2 and D=3MSEImproved
Use inductive skeleton fitting rather than all-at-once fittingLocation problem · Location problem, D=2, N=50 and 100MSE: 2.855e-04 $\pm$ 2.114e-05 -> 2.691e-04 $\pm$ 8.541e-06; 2.660e-04 $\pm$ 1.227e-05 -> 2.608e-04 $\pm$ 5.946e-06Improved
Use inductive skeleton fitting rather than all-at-once fittingLocation problem · Location problem, D=3, N=50 and 100MSE: 3.596e-04 $\pm$ 7.935e-05 -> 3.269e-04 $\pm$ 3.969e-05; 2.810e-04 $\pm$ 1.569e-05 -> 2.796e-04 $\pm$ 1.478e-05No clear effect
Use all-at-once rather than inductive skeleton fittingGroup lasso · Group lasso, D=2 and D=3, N=50 and 100MSE: 1.041e-04 $\pm$ 1.614e-05 vs 4.966e-04 $\pm$ 1.848e-05; 8.949e-05 $\pm$ 6.083e-06 vs 5.020e-04 $\pm$ 1.276e-05; 4.354e-0Improved

Closest papers by meaning

Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy