Topological Data Analysis of Decision Boundaries with Application to Model Selection
We propose the labeled \v{C}ech complex, the plain labeled Vietoris-Rips complex, and the locally scaled labeled Vietoris-Rips complex to perform persistent homology inference of decision boundaries in classification tasks. We provide theoretical conditions and analysis for recovering the homology of a decision boundary from samples. Our main objective is quantification of deep neural network complexity to enable matching of datasets to pre-trained models; we report results for experiments using MNIST, FashionMNIST, and CIFAR10.
Code (1)
Tasks
General ClassificationModel SelectionTopological Data AnalysisSimilar Papers 제목 키워드 기반
Some Geometrical and Topological Properties of DNNs' Decision Boundaries
Geometry and topology of decision regions are closely related with classification performance and robustness against adversarial attacks. In this paper, we use differential geometry to theoretically explore the geometric…
Topological Data Analysis for Neural Network Analysis: A Comprehensive Survey
This survey provides a comprehensive exploration of applications of Topological Data Analysis (TDA) within neural network analysis. Using TDA tools such as persistent homology and Mapper, we delve into the intricate stru…
ArticlesModel SelectionSurveyTopological Data AnalysisSpatial database implementation of fuzzy region connection calculus for analysing the relationship of diseases
Analyzing huge amounts of spatial data plays an important role in many emerging analysis and decision-making domains such as healthcare, urban planning, agriculture and so on. For extracting meaningful knowledge from geo…
Decision MakingFinding the Homology of Decision Boundaries with Active Learning
Accurately and efficiently characterizing the decision boundary of classifiers is important for problems related to model selection and meta-learning. Inspired by topological data analysis, the characterization of decisi…
Active LearningMeta-LearningModel SelectionTopological Data AnalysisOn The Topological Expressive Power of Neural Networks
We propose a topological description of neural network expressive power. We adopt the topology of the space of decision boundaries realized by a neural architecture as a measure of its intrinsic expressive power. By sa…