Search

Presents a two-stage approach that is effective for performing fast clustering. Training algorithm and competitive neural network characterization; Gravitation neural network (GNN) clustering that fully takes advantage of the quantization preprocess; Characterization of the interesting properties of the GNN.

Considers the task of solving the independent component analysis problem x = As given observation x, with a constraint of nonnegativity of the source random vector s. Prewhitening and axis rotations; Nonnegative principal component analysis; Axis pair rotation; Simulation results; Application to analysis of musical audio.

The problem of finding all cycles in the exponentially growing state space of synchronous Boolean networks was studied in the paper by C. Farrow, J. Heidel, J. Maloney, and J. R. Scalar, "Equations for synchronous Boolean networks with biological applications," IEEE Trans. Neural Net- works, vol. 15, no. 2, pp. 348-354 Mar. 2004. No efficient algorithm was given to solve the problem. We show that even the determination of the number of fixed points (cycles of length 1) for monotone Boolean networks and the determination of the existence of fixed points for general Boolean networks are both strong NP-complete. [ABSTRACT FROM AUTHOR]