A&C_IADI Logo
HOME ABOUT VOLUMES SEARCH

Volumes / Supplements / 6 - 2014 / Buscema P.

The general philosophy of Artificial Adaptive Systems (AAS)

Click on any image to open slideshow

Fig. 1 – The diagram shows how the analysis of Natural and/or Cultural Processes, that need to be understood, starts from a theory which, adequately formalized (Formal Algebra), is able to generate Automatic Artificial Models of those Natural and/or Cultural Processes. Lastly, the generated Automatic Artificial Models must be compared with the Natural and/or Cultural Processes of which they profess to be the model and the explanation.

Fig. 1 – The diagram shows how the analysis of Natural and/or Cultural Processes...

Fig. 2 – Taxonomic tree of the disciplines that make up the Artificial Sciences system.

Fig. 2 – Taxonomic tree of the disciplines that make up the Artificial Sciences ...

Fig. 3 – The diagram shows in more detail the formalization, automation and comparison between Natural and/or Cultural Processes and Automatic Artificial Models seen from two points of view (Classical Computation and NC). Each point of view can be seen as a cycle that can repeat itself several times. This allows one to deduce that the human scientific process characterizing both the cycles resembles more the NC than the Classical Computation one.

Fig. 3 – The diagram shows in more detail the formalization, automation and comp...

Fig. 4 – Artificial Adaptive Systems: general diagram.

Fig. 4 – Artificial Adaptive Systems: general diagram.

Fig. 5 – Types of possible connections.

Fig. 5 – Types of possible connections.

Fig. 6 – Example of Multi-Layer Perceptron.

Fig. 6 – Example of Multi-Layer Perceptron.

Fig. 7a – Topology of Conic Net (in blue the Hidden layer).

Fig. 7a – Topology of Conic Net (in blue the Hidden layer).

Fig. 7b – Close up of one macro hidden node of a Conic Net.

Fig. 7b – Close up of one macro hidden node of a Conic Net.

Fig. 8 – AutoContractive Map.

Fig. 8 – AutoContractive Map.

Fig. 9 – Example of Unsupervised ANN for natural clustering – Self-Organizing Map.

Fig. 9 – Example of Unsupervised ANN for natural clustering – Self-Organizing Ma...

Fig. 10 – SOM with n-nodes of Input, with (mrmc) units of Kohonen’s layer. This architecture allows the Inputs to be classified into m2 classes, each being a subclass represented by a codebook.

Fig. 10 – SOM with n-nodes of Input, with (mrmc) units of Kohonen’s layer. This ...

Fig. 11 – Topology of the neighborhood Space of a Winner Unit in a square and in a rhomb; in the illustration v is the degree of proximity of the K units to the WU.

Fig. 11 – Topology of the neighborhood Space of a Winner Unit in a square and in...

Fig. 12 – Example of the topology of the neighborhood space with matrix K (8r × 8c), where the WU is the K55 unit. The first matrix shows a neighborhood in a square while the second a neighborhood in a rhomb. We can notice from the illustration that, for example, while in the matrix to the left the v distance of the K66 unit to the WU is 1, in the matrix to the right the v distance of the K66 unit to the WU is 2.

Fig. 12 – Example of the topology of the neighborhood space with matrix K (8r × ...

Fig. 13 – The illustration shows how, when the parameter ??(1, 2, 3) changes – parameter that determines the correction curve of the neighborhood function – the number of units that are part of the neighborhood and the extent of the correction (v1, v2, v3) made on the weights also change.

Fig. 13 – The illustration shows how, when the parameter ??(1, 2, 3) changes – p...

Buscema P. 2014, The general philosophy of Artificial Adaptive Systems (AAS), Archeologia e Calcolatori 6