Trimmed KL divergence between Gaussian mixtures for robust unsupervised acoustic anomaly detection.
Nash M. BorgesGerard G. L. MeyerPublished in: INTERSPEECH (2009)
Keyphrases
- anomaly detection
- kl divergence
- gaussian mixture
- unsupervised learning
- expectation maximization
- kullback leibler divergence
- em algorithm
- information theoretic
- mixture model
- closed form
- gaussian mixture model
- mahalanobis distance
- covariance matrix
- probability density function
- exponential family
- supervised learning
- computer vision
- semi supervised
- density estimation
- gaussian distribution
- pattern recognition
- probability density
- mutual information
- active learning
- training data
- probabilistic latent semantic analysis
- image processing