GAGA: Deciphering Age-path of Generalized Self-paced Regularizer

Authors: Xingyu Qu, Diyang Li, Xiaohan Zhao, Bin Gu

NeurIPS 2022 | Conference PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Experimental We demonstrate the performance of GAGA on real-world datasets, and find considerable speedup between our algorithm and competing baselines.
Researcher Affiliation Academia 1 Mohamed bin Zayed University of Artificial Intelligence 2 Nanjing University of Information Science & Technology
Pseudocode Yes Algorithm 1 Alternate Convex Search (ACS) and Algorithm 2 Generalized Age-path Algorithm (GAGA)
Open Source Code Yes Did you include the code, data, and instructions needed to reproduce the main experimental results (either in the supplemental material or as a URL)? [Yes]
Open Datasets Yes Dataset Source Samples Dimensions Task mfeat-pixel UCI [43] 2000 240 C pendigits UCI 3498 16 hiva agnostic Open ML 4230 1620. UCI [43], Open ML [44].
Dataset Splits No No explicit specification of training/validation/test dataset splits (e.g., percentages, counts, or explicit cross-validation strategy) in the main body of the paper.
Hardware Specification Yes All experiments are run on a machine with Intel(R) Core(TM) i7-2600 CPU @ 3.40GHz, 16 GB RAM and NVIDIA GeForce GTX 1080Ti.
Software Dependencies Yes The codes are implemented in Python 3.8 based on the Scikit-learn (version 0.24) library [57] and Matplotlib (version 3.4.3) for visualization. For comparison, ACS is implemented via the open source Geatpy (version 2.7.0) framework [58]. The ODE solver is based on SciPy (version 1.7.3) using LSODE routine [54, 55].
Experiment Setup Yes For SVM, we use the Gaussian kernel K(x1, x2) = exp( γκ x1 x2 2). For linear SP-regularizer, λmin and λmax are set as 1e-4 and 500 respectively. The trade-off parameters C for SVM and α for Lasso are searched from {1e-3, 1e-2, . . . , 1e2} and {1e-5, 1e-4, . . . , 1e-1} respectively. The kernel coefficient γκ is searched from {0.1, 0.2, . . . , 2}.