Neural network software

The first contact that many scientific users have with neural networks is through commercially available software. The manuals and Help systems in off-the-shelf neural network software may not offer much on the principles of the method, so, on a first encounter, ANNs might appear to be black boxes whose inner workings are both mysterious and baffling (Figure 2.3). 

The artificial neural network (ANN) is a relatively new technique and possibly the preferred one for current and future (Q)SAR development. Basically, ANNs can be regarded as multinonlinear regression methods. Thus, the neural network software simply multiplies the input by a set of weights that in a nonlinear way transforms the... 

Chen, Y. Jiao, T. McCall, T.W. Baichwal, A.R. Meyer, M.C. Comparison of four artificial neural network software programs used to predict the in vitro dissolution of controlled release tablets. Pharm. Dev. Technol. 2002, 7, 373-379. 

Extensions to the earlier chemometric approaches include a toxicological assessment approach based on neural network software to ascertain whether the methods provide a robust approach, which could lead to automatic toxin classification. The neural network approach to sample classification, based on NMR spectra of urine, was in general predictive of the sample class. It appears to be reasonably robust and once the network is trained, the prediction of new samples is rapid and automatic. However, the principal disadvantage is common to all neural network studies in that it is difficult to ascertain from the network which of the original sample descriptors are responsible for the classification. Although recently it has been suggested that probabilistic... 

SAIC s neural network software products are more expensive ANSim and Shells are environments for implementing frequently-used neural net models ANSpec is a language for specifying and developing new models. Each package runs on a PC/AT or compatible, or on an AT enhanced with SAIC s Delta accelerator card. 

Figure 14.8 Sensor array responses for seven aviation fuels. A sensor array of three TGS 813 sensors produced a collection of outputs that could be successfully classified by neural network software. 

Some studies have been undertaken in the area of cancer diagnosis using perchloric acid extracts of various types of human brain tumor tissue [47], and the spectra could be classified using neural network software giving -85% correct classification. Tissues can be studied by metabonomics through the MAS NMR technique, and published examples include prostate cancer [16,48], renal cell carcinoma [18], breast cancer [19, 49], and various brain tumors [50]. Other recent studies include an NMR-based urinary metabonomic study of multiple sclerosis in humans and non-human primates [51]. 

Neural Network Software BrainMaker (California Scientific). 

To provide a user s guide for the application of neural network software, be it commercial or in-house. This guide should help you decide which type of network to use, what to do with your data before submitting it to a neural network, how to assess your results, and what to do when things go wrong i.e., when it appears that a neural network cannot solve your problem. 

Natural muscles are controlled by neurons and network of neurons. We can imagine artificial neurons and network of artificial neurons as well. Artificial muscles with motor proteins are studied and attract attention[79]. One direction is to develop deformable machine with real motor proteins, actins and myosins, and neurons. Another direction is to develop neural network software to control distributed artificial muscles. The author has been developing open brain simulator which can emulate the activities of human nervous system for estimating internal state of human through external observation [231]. Such software is also applicable to control artificial muscle systems, which is implemented on the personal robots and humanoid robots in the future. 

Brainmaker neural network software is available from California Scientific Software, Sierra Madre, CA 91024. NeuralWare, Inc., Pittsburgh, PA 15276 is supplier of NeuralWorks Explorer and NeuralWorks ll/Plus software. ALN software is supplied by General Research Corp., Advanced Technology Div., Arlington, VA 22209. 

One of us (P.R.G.) has incorporated neural network software into a low-resolution OP/FT-IR spectrometer. Once the components had been identified, an online PLS calibration and prediction was carried out Fheliminary results are promising. As an example, a mixture of acetone, methyl ethyl ketone (MEK), and diisobutyl ketone was allowed to evaporate close to the beam of an OP/FT-IR instrument. The software recognized the presence of these three compounds and allowed their quantification. The data obtained in this way are shown in Figure 22.8. Acetone and MEK are seen at relatively high concentrations at the start... 

---