SDLGASMay 27, 2021

Evaluation of concept drift adaptation for acoustic scene classifier based on Kernel Density Drift Detection and Combine Merge Gaussian Mixture Model

arXiv:2105.13220v12 citations
Originality Synthesis-oriented
AI Analysis

This work addresses concept drift adaptation for acoustic scene classifiers, which is an incremental improvement in domain-specific applications.

The paper tackled concept drift adaptation in acoustic scene classification by evaluating Kernel Density Drift Detection and Combine Merge Gaussian Mixture Model, finding that different drift types require specific hyperparameter configurations to optimize adaptation frequency and avoid overfitting, with CMGMM pruning improving model performance.

Based on the experimental results, all concepts drift types have their respective hyperparameter configurations. Simple and gradual concept drift have similar pattern which requires a smaller α value than recurring concept drift because, in this type of drift, a new concept appear continuously, so it needs a high-frequency model adaptation. However, in recurring concepts, the new concept may repeat in the future, so the lower frequency adaptation is better. Furthermore, high-frequency model adaptation could lead to an overfitting problem. Implementing CMGMM component pruning mechanism help to control the number of the active component and improve model performance.

Code Implementations1 repo
Foundations

The foundational work for this paper's niche, ranked by how specifically the neighbourhood builds on it — not by global fame.

Your Notes