Asymptotic Risk of Bézier Simplex Fitting

Authors: Akinori Tanaka, Akiyoshi Sannai, Ken Kobayashi, Naoki Hamada2416-2424

AAAI 2020 | Conference PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Experimental Those results are verified numerically under small to moderate sample sizes. In this paper, we study the asymptotic risk of the two fitting methods of the B ezier simplex: the all-at-once fitting and the inductive skeleton fitting, and compare their performance with respect to the degree. We examine the empirical performances of the all-at-once fitting and the inductive skeleton fitting and verify the asymptotic risks derived in Section 3.1 over synthetic instances and multi-objective optimization instances.
Researcher Affiliation Collaboration Akinori Tanaka RIKEN AIP, Keio University akinori.tanaka@riken.jp Akiyoshi Sannai RIKEN AIP, Keio University akiyoshi.sannai@riken.jp Ken Kobayashi Fujitsu Laboratories LTD., RIKEN AIP, Tokyo Tech ken-kobayashi@fujitsu.com Naoki Hamada Fujitsu Laboratories LTD., RIKEN AIP hamada-naoki@fujitsu.com
Pseudocode No The paper does not contain any clearly labeled 'Pseudocode' or 'Algorithm' blocks or figures.
Open Source Code Yes The source code and library dependencies are provided in https://github.com/rafcc/aaai-20.1534.
Open Datasets Yes To investigate the relationship between the generalization performance and our theoretical risk, we provide two complementary instances of multi-objective optimization problems: a generalized location problem called MED (Harada, Sakuma, and Kobayashi 2006; Hamada et al. 2010) and a multi-objective hyper-parameter tuning of the group lasso (Yuan and Lin 2006) on the Birthwt dataset (Hosmer and Lemeshow 1989; Venables and Ripley 2002).
Dataset Splits No The paper mentions 'training points' and a 'test set', but it does not explicitly describe a 'validation set' or provide specific percentages/counts for a train/validation/test split.
Hardware Specification Yes Experiment programs were implemented in Python 3.7.1 and run on a Windows 7 PC with an Intel Core i7-4790CPU (3.60 GHz) and 16 GB RAM.
Software Dependencies No Experiment programs were implemented in Python 3.7.1 and run on a Windows 7 PC with an Intel Core i7-4790CPU (3.60 GHz) and 16 GB RAM. The source code and library dependencies are provided in https://github.com/rafcc/aaai-20.1534. While Python version is specified, the paper does not list other ancillary software dependencies with specific version numbers directly in the main text.
Experiment Setup Yes In this experiment, we estimated the B ezier simplex with degree D = 2 or 3, and compared the following three fitting methods: all-at-once the all-at-once fitting (Section 3.1); inductive skeleton (non-optimal) the inductive skeleton fitting (Section 3.2) with N (0) = = N (M 1) = N/M, which does not provide the optimal value of the risk shown in Table 1; inductive skeleton (optimal) the inductive skeleton fitting (Section 3.2) where N (0), . . . , N (M 1) are determined by minimizing the risk shown in Table 1 under the constraints M 1 m=0 N (m) = N and N (m) 0 (m = 0, . . . , M 1). We make them strongly convex by the following perturbation: f1 = f1 + ε x 2 , f2 = f2 + ε x 2 , f3 = f3 + ε x 2 where ε is an arbitrarily small positive number (we set ε = 10 4). We changed the size of the training set from N { 250, 500, 1000, 2000 }.