State encoding

There are many ways to model the machine states of a finite state machine. Described here are some of the most common ones. The 

MooreFSM module described earlier is used as an example in describing these encodings. 

---

Data often need to be transformed into a format that will help the neural network learn. This transformation is called encoding and is a very important step. If the neural network is given data in a poorly encoded state, it may not be able to learn the features of each class. For example, you could take each dinosaur picture and encode it as a series of black-and-white pictures. You could then snbmit the individual black-and-white pixel values into the neural network and let it decide what features and shapes to choose from. For some problems, this is fine. However, better results would be obtained if the network were given all the lines and shapes of the dinosaur from the picture and then trained on these shapes. This shape-based encoding might yield better results than a pixel-based encoding. 

Figure Bl.14.9. Imaging pulse sequence including flow and/or diflfiision encoding. Gradient pulses before and after the inversion pulse are supplemented in any of the spatial dimensions of the standard spin-echo imaging sequence. Motion weighting is achieved by switching a strong gradient pulse pair G, (see solid black line). The steady-state distribution of flow (coherent motion) as well as diffusion (spatially...

Flow which fluctuates with time, such as pulsating flow in arteries, is more difficult to experimentally quantify than steady-state motion because phase encoding of spatial coordinate(s) and/or velocity requires the acquisition of a series of transients. Then a different velocity is detected in each transient. Hence the phase-twist caused by the motion in the presence of magnetic field gradients varies from transient to transient. However if the motion is periodic, e.g., v(r,t)=VQsin (n t +( )q] with a spatially varying amplitude Vq=Vq(/-), a pulsation frequency co =co (r) and an arbitrary phase ( )q, the phase modulation of the acquired data set is described as follows ... 

The forces in a protein molecule are modeled by the gradient of the potential energy V(s, x) in dependence on a vector s encoding the amino acid sequence of the molecule and a vector x containing the Cartesian coordinates of all essential atoms of a molecule. In an equilibrium state x, the forces (s, x) vanish, so x is stationary and for stability reasons we must have a local minimizer. The most stable equilibrium state of a molecule is usually the... 

A challenging task in material science as well as in pharmaceutical research is to custom tailor a compound s properties. George S. Hammond stated that the most fundamental and lasting objective of synthesis is not production of new compounds, but production of properties (Norris Award Lecture, 1968). The molecular structure of an organic or inorganic compound determines its properties. Nevertheless, methods for the direct prediction of a compound s properties based on its molecular structure are usually not available (Figure 8-1). Therefore, the establishment of Quantitative Structure-Property Relationships (QSPRs) and Quantitative Structure-Activity Relationships (QSARs) uses an indirect approach in order to tackle this problem. In the first step, numerical descriptors encoding information about the molecular structure are calculated for a set of compounds. Secondly, statistical and artificial neural network models are used to predict the property or activity of interest based on these descriptors or a suitable subset. 

Von Neumann was able to construct a self-reproducing UTM embedded within a 29-state/5-cell neighborhood two-dimensional cellular automaton, composed of several tens of thousands of cells. It was, to say the least, an enormously complex machine . Its set of 29 states consist largely of various logical building blocks (AND and OR gates, for example), several types of transmission lines, data encoders and recorders, clocks, etc. Von Neumann was unfortunately unable to finish the proof that his machine was a UTM before his death, but the proof was later completed and published by Arthur Burks [vonN66]. 

Suppose a state s is encoded as a binary string of 10 O s and 10 I s , s = 10101010..., While the number of raw bits defining s is huge, its algorithmic complexity is actually very small because it can be reproduced exactly by the following short program . On the other hand, a completely... 

We have defined above a way of quantifying the structure of water based on the profile of fx values that encode the number of each possible joined state of a molecule. It is now possible to use this profile as a measure of the structure of water at different temperatures. As an application of this metric it is possible to relate this to physical properties. We have shown the results of our earlier work in Table 3.3. The reader is encouraged to repeat these and to explore other structure-property relationships using the fx as single or multiple variables. A unified parameter derived from the five fx values expressed as a fraction of 1.0, might be the Shannon information content. This could be calculated from all the data created in the above studies and used as a single variable in the analysis of water and other liquid properties. 

In a cellular automata model of a solution, there are three different types of cells with their states encoded. The first is the empty space or voids among the molecules. These are designated to have a state of zero hence, they perform no further action. The second type of cell is the water molecule. We have described the rules governing its action in the previous chapter. The third type of cell in the solution is the cell modeling a solute molecule. It must be identified with a state value separate from that of water. 

Table 6.1. Rules encoding the hydropathic states of wall cells"...

The E-state indices are atomic descriptors composed of an intrinsic state value I and a perturbation AI that measures the interactions with all other atoms in a molecule. The Kier-Hall electronegativity is the starting point in the definition of the intrinsic state of an atom, which encodes its potential for electronic interactions and its connectivity with adjacent atoms. The intrinsic state of an atom i is [19, 21] ... 

---