Incremental Ensemble Gaussian Processes

Research output: Contribution to journalArticlepeer-review

Abstract

Belonging to the family of Bayesian nonparametrics, Gaussian process (GP) based approaches have well-documented merits not only in learning over a rich class of nonlinear functions, but also quantifying the associated uncertainty. However, most GP methods rely on a single preselected kernel function, which may fall short in characterizing data samples that arrive sequentially in time-critical applications. To enable online kernel adaptation, the present work advocates an incremental ensemble (IE-) GP framework, where an EGP meta-learner employs an ensemble of GP learners, each having a unique kernel belonging to a prescribed kernel dictionary. With each GP expert leveraging the random feature-based approximation to perform online prediction and model update with scalability, the EGP meta-learner capitalizes on data-adaptive weights to synthesize the per-expert predictions. Further, the novel IE-GP is generalized to accommodate time-varying functions by modeling structured dynamics at the EGP meta-learner and within each GP learner. To benchmark the performance of IE-GP and its dynamic variant in the case where the modeling assumptions are violated, rigorous performance analysis has been conducted via the notion of regret. Also, online unsupervised learning is explored under the novel IE-GP framework. Synthetic and real data tests demonstrate the effectiveness of the proposed schemes.

Original languageEnglish (US)
JournalIEEE Transactions on Pattern Analysis and Machine Intelligence
DOIs
StateAccepted/In press - 2022

Bibliographical note

Publisher Copyright:
IEEE

Keywords

  • Benchmark testing
  • Dimensionality reduction
  • ensemble learning
  • Gaussian processes
  • Kernel
  • online prediction
  • Radio frequency
  • random features
  • regret analysis
  • Scalability
  • Task analysis
  • Training

PubMed: MeSH publication types

  • Journal Article

Fingerprint

Dive into the research topics of 'Incremental Ensemble Gaussian Processes'. Together they form a unique fingerprint.

Cite this