Can an instrument learn from experiments done by itself?

Abstract

The hardware and software requirements for intelligent analytical instruments are discussed. Existing technology in both fields allows the production of a new generation of instruments that can learn from their own operation and even transfer the acquired knowledge to machines of a later generation. The hardware needed is a processor with the power of at least the Intel 80386 processor, 4–5 Mbyte RAM, storage equivalent to a 1.6-Gbyte hard disk and a special encode/decode processor for compressing complex instrumental information. This hardware can support the transmission, transformation, storage and handling of about 100 000 complex measurements (up to 3 kbyte of numbers per measurement) and a knowledge base of about the same size. In this context, the space for software can be neglected or, in the case of neural networks, the “software” can be hardwired. At preent, there are two different approaches to handling the required amount of data, extracting vital information, and acquiring (in close cooperation with experts) new knowledge. The first is neural networks and the second is hierarchical clustering. It is argued that combination of both methods forms a very powerful basis for software development in intelligent instruments.