Candid covariance-free incremental principal component analysis

Weng, Juyang; Zhang, Yilu; Hwang, Wonjae

doi:10.1109/tpami.2003.1217609

Cited by 399 publications

(27 citation statements)

References 14 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…The initial setting for ϵ-greedy is 0.6, which decreases via multiplication with 0.995 after every episode, and is reset when a new module is created. The learning rates for IncSFA: for CCIPCA, a 1/ t learning rate is used, with amnesic parameter l = 2 (Weng et al, 2003), the MCA learning rate is a constant η MCA = 0.05. We collected results over 25 different runs.…”

Section: Experiments and Resultsmentioning

confidence: 99%

“…Specifically, we use IncSFA and the ROC method, respectively. Some details of these algorithms are described below, but more thorough descriptions can be found elsewhere (Guedalia et al, 1999; Weng et al, 2003; Peng and Yi, 2006; Zhang et al, 2005; Kompella et al, 2012b). …”

Section: Curious Dr Misfamentioning

confidence: 99%

“…IncSFA uses incremental algorithms for the two required PCAs. For the first, Candid Covariance-Free Incremental PCA (Zhang and Weng, 2001; Weng et al, 2003), is used, which can also reduce the dimensionality by only computing the K highest eigenvectors. For the second, Minor Components Analysis (MCA; Oja, 1992; Peng and Yi, 2006; Peng et al, 2007) updates the J slowest features.…”

Section: Curious Dr Misfamentioning

confidence: 99%

See 2 more Smart Citations

An intrinsic value system for developing multiple invariant representations with incremental slowness learning

et al. 2013

View full text Add to dashboard Cite

Curiosity Driven Modular Incremental Slow Feature Analysis (CD-MISFA;) is a recently introduced model of intrinsically-motivated invariance learning. Artificial curiosity enables the orderly formation of multiple stable sensory representations to simplify the agent's complex sensory input. We discuss computational properties of the CD-MISFA model itself as well as neurophysiological analogs fulfilling similar functional roles. CD-MISFA combines 1. unsupervised representation learning through the slowness principle, 2. generation of an intrinsic reward signal through learning progress of the developing features, and 3. balancing of exploration and exploitation to maximize learning progress and quickly learn multiple feature sets for perceptual simplification. Experimental results on synthetic observations and on the iCub robot show that the intrinsic value system is essential for representation learning. Representations are typically explored and learned in order from least to most costly, as predicted by the theory of curiosity.

show abstract

Section: Experiments and Resultsmentioning

confidence: 99%

Section: Curious Dr Misfamentioning

confidence: 99%

Section: Curious Dr Misfamentioning

confidence: 99%

See 1 more Smart Citation

An intrinsic value system for developing multiple invariant representations with incremental slowness learning

et al. 2013

View full text Add to dashboard Cite

show abstract

“…Here, we rely on the CCIPCA algorithm by Weng et al [39]. Several successful applications of CCIPCA can be demonstrated [43-45].…”

Section: Methodsmentioning

confidence: 99%

“…In the following, we briefly outline the CCIPCA iteration. For further details on CCIPCA, see [39]. A convergence proof is given in [46].…”

Section: Methodsmentioning

confidence: 99%

The looks of an odour - Visualising neural odour response patterns in real time

et al. 2013

View full text Add to dashboard Cite

BackgroundCalcium imaging in insects reveals the neural response to odours, both at the receptor level on the antenna and in the antennal lobe, the first stage of olfactory information processing in the brain. Changes of intracellular calcium concentration in response to odour presentations can be observed by employing calcium-sensitive, fluorescent dyes. The response pattern across all recorded units is characteristic for the odour.MethodPreviously, extraction of odour response patterns from calcium imaging movies was performed offline, after the experiment. We developed software to extract and to visualise odour response patterns in real time. An adaptive algorithm in combination with an implementation for the graphics processing unit enables fast processing of movie streams. Relying on correlations between pixels in the temporal domain, the calcium imaging movie can be segmented into regions that correspond to the neural units.ResultsWe applied our software to calcium imaging data recorded from the antennal lobe of the honeybee Apis mellifera and from the antenna of the fruit fly Drosophila melanogaster. Evaluation on reference data showed results comparable to those obtained by previous offline methods while computation time was significantly lower. Demonstrating practical applicability, we employed the software in a real-time experiment, performing segmentation of glomeruli - the functional units of the honeybee antennal lobe - and visualisation of glomerular activity patterns.ConclusionsReal-time visualisation of odour response patterns expands the experimental repertoire targeted at understanding information processing in the honeybee antennal lobe. In interactive experiments, glomeruli can be selected for manipulation based on their present or past activity, or based on their anatomical position. Apart from supporting neurobiology, the software allows for utilising the insect antenna as a chemosensor, e.g. to detect or to classify odours.

show abstract