Text Processing

This expanded view of task automation includes new capabilities in the the traditional area of instrument automation and in the somewhat newer related field of robotics. In addition it includes a number of functions which are not new to the office and business environment but have only recently become readily available in the laboratory. These are tools such as data base management, scientific text processing, and electronic mail and document transfer. One way to improve technical productivity Is by giving the scientist more time to do science. This can be accomplished through improved efficiency In the office, communication, and information retrieval functions which must be performed as well as by allowing science to be done In new and more efficient ways through the use of computers. 

Meaningful codes may or may not be needed for modem dictionaries. For example, the new Adverse Drag Reactions On-Line Information Tracking (ADROIT) dictionaries do not use meaningful codes, but rely on linkage of related terms and effective text processing. Where codes are considered necessary, they should be as short as possible (Westland, 1991). Whenever a system is used for AEs from the literature, spontaneous reports, clinical trials or a combination of these, the needs of the users of the system will influence the selection of the dictionary. 

The authors are indebted to the Bell Laboratories Text Processing group, especially T. I. Howard, J. Alder, and M. McCann. Special thanks also go to J. H. Bruning for his assistance with Chapter 2. 

Figure 21.10 Modifications by which proto-oncogenes can be activated. Conversion to an oncogene or excess activation of proto-oncogene results in excess growth-promoting activity, and consequently increases the risk of tumour development (see text). Process (iv) can result in many cells in the host being infected with an oncogene that has been formed from a protooncogene within the virus.

The authors and editors are indebted to the AT T Bell Laboratories Text Processing Group, especially Susan Pope and Carmela Patuto. Special thanks are due to AT T Bell Laboratories Art Department and Roy T. 

The first chapter is dedicated to scientific word and text processing. Wipke describes the philosophy, purposes and abilities of word processors. The chapter is not written with the intent to promote a particular make of word processor, but to show the scientist what to look for when choosing a new product or evaluating an existing one. A careful reader will notice a subtle difference between the layout of the first and those of the rest of the chapters. In order to show the high quality manuscript that any chemist can produce in his or her own office without the assistance of artists, typists, editors, etc., Wipke has supplied his manuscript in camera-ready form made on a PC-based word processor, hence the minor differences. 

Throughout this document (prepared using RUNOFF, a text processing package) you will see examples of text formatting automatic generation of... 

The work which my associates and I have done in the use ot text as data by chemists and for chemists draws on two sources. The first ot these is computer text processing. It is even more clear today than it was 13 years ago when this work was started that computers may not only be used for number crunching but for a number ot other jobs in which letters and words as well as numbers can be used. Tasks such as extracting, sorting, reformatting, associating, and counting can be done not only with numeric information, but also with alphanumeric information. Whereas such operations may have been useful but esoteric 13 years ago, the tremendous use ot word... 

Preceding chapters have dealt largely with pure substances or with constant-composition mixtures. e.g., air. However, composition changes are the desired outcome, not only of chemical reactions, but of a number of industrially important mass-transfer operations. Thus composition becomes a primary variable in the remaining chapters of tliis text. Processes such as distillation, absorption, and extraction bring phases of different composition into contact, and when tlie phases are not in equilibriimi, mass transfer between the phases alters their compositions. Botli tlie extent of change and tlie rate of transfer depend on the departure of the system from equilibrium. Thus, for quantitative treatment of mass transfer the equilibrium T, P, and phase compositions must be known. 

Editorial support. Dirk De Weerdt text processing, correspondence, lay-out, figure design, index, cover design, cover illustration. Contact ddwdesign telenet.be... 

Tanner, L. (1991). Selecting a text-processing system as a qualitative multiple criteria problem, European Journal of Operational Research 50, pp. 179-187. 

Salton, G. 1989. Automatic text processing. Reading, MA Addison-Wesley. 

Text processing by P i cA Publishing Services from original typescripts and illustrations provided by the authors... 

The idea of using symbol frequencies to guide the selection of an indexing vocabulary for search in the two-dimensional domain of chemical structure representations then prompted us to look at the applicability of these ideas in the text processing context - an example of phenomenon - driven research. This led to a series of studies of the... 

Today s systems are massively more accurate with respect to text analysis, and so any text processing errors in systems such as MITalk have to be eliminated if a comparison of synthesis encoding with respect to intelligibility is to be performed. 

Many tasks in text processing involved assigning a label to a unit this is called classification. 

Text analysis A sensible starting point for any prediction would be to extrapolate current trends. To this extent, it is reasonable to assume that TTS will become entirely data driven. I think it is incontestable that the front end, or text processing component will become entirely statistical. In recent years the advances in statistical NLP have been enormous, fueled by the use of search engines and the need to translate documents. I think many of these techniques are directly applicable to TTS and will be adopted. 

In coimnercial TTS, text analysis is seen as vitally important since any error can cause an enourmous negative impression witii a hstener. Despite this, the problem has received only sporadic attention amongst academic researchers. Notable exceptions include various works by Richard Sproat [410], [411], [412], [413], who tackles nearly all the problems in text analysis (and many in prosody, explained in file next chapter). The Bell Labs system tiiat Sproat worked on is particularly well known for its very high accuracy with regard to text processing. In a separate line of work Sproat et al. investigated a number of machine-learning approaches to semiotic classification and verbahsation. 

Hron, A., Kurbjuhn, L, Mandl, H., and Schnotz, W. Structural inferences in reading and listening. In Inferences in text processing (G. Rickheit and H. Suohner, ed.). North-Holland Elsevier Science Publishers, 1985. 

Rouet, J-F. (1990c) Interactive text processing by inexperienced (hyper-) readers. ECHT 90. Paris, Novembie. 

---