And associative memory

Kohonen, T. (1988) Self-Organization and Associative Memory, Springer-Verlag, Berlin. 

Kohonen T. Self-organisation and associative memory. Berlin Springer-Verlag, 1989. 

T. Kohonen, Self Organization and Associated Memory. Springer-Verlag, Heidelberg, 1989. W.J. Meissen, J.R.M. Smits, L.M.C. Buydens and G. Kateman, Using artificial neural networks for solving chemical problems. II. Kohonen self-organizing feature maps and Hopfield networks. Chemom. Intell. Lab. Syst., 23 (1994) 267-291. 

Kohonen T (1984) Self-organization and associative memory. Springer, Berlin Heidelberg New York... 

T. Kohonen, S ef-Organisation and Associative Memory, Springer-VerLag, Berlin, 1987. 

Kohonen, T. Self-organization and Associative memory, 3rd Ed, 1989, Springer, Berlin. 

For reversible interactions, the linear dynamic range is determined by that portion of the coating-analyte sorption isotherm (discussed in Section 5.4) that lies between the LOD and the saturation limit. For irreversible interactions, the LDR will depend on the sorption/reaction kinetics and the coating capacity. For practical reasons, it is desirable to have the widest LDR possible, although inexpensive microprocessors and associated memory make correction for minor nonlinearities straightforward. For an example of a wide linear dynamic range, refer to Figure 5.13. 

The autosampler functions as a separate IEEE-488 device, distinct from the source system. It also Incorporates its own internal microcomputer and associated memory. It allows random access of any sample In the tray at any time during the analysis. Using the autosampler for sample introduction allows the analyst to perform other analysis-related tasks while an analysis is taking piace. Including the system controller, both monochromators, the RF system and the autosampler, the instrument assembly contains five separate microprocessors. 

Carpenter, G. 1989. Neural network models for pattern recognition and associative memory. Neural Networks, 2 243-258. 

Schweizer and collaborators have elaborated an extensive mode-coupling model of polymer dynamics [52-54]. The model does not make obvious assumptions about the nature of polymer motion or the presence or absence of particular long-lived dynamic structures, e.g., tubes it yields a set of generalized Langevin equations and associated memory functions. Somewhat realistic assumptions are made for the equilibrium structure of the solutions. Extensive calculations were made of the molecular weight dependences for probe diffusion in melts, often leading by calculation rather than assumption to power-law behaviors for various transport coefficients. However, as presented in the papers noted here, the model is applicable to melts rather than solutions Momentum variables have been completely suppressed, so there are no hydrodynamic interactions. Readers should recall that hydrodynamic interactions usually refer to interactions that are solvent-mediated. 

The methods for deriving effective potentials from known protein structures can be grouped into three main categories statistical approaches, optimization methods, and associative memory methods. All of these make an a priori assumption about the terms to be included in the force-field and then use an ensemble of known protein three-dimensional structures to derive the parameters associated with the different terms. 

Pattern classification and associative memory networks can be trained to distinguish patterns into separate classes and associate input-output pairs. 

---