Data-driven synthesis

In all types of data driven synthesis, we not only require data, but require this data to be labelled in some way. As a minimum, this normally means we require the words and phones, but any feature that we require for model building or unit selection must be provided by some means. In this section, we concentrate on one such way of carrying out this labelling, based on the HMM principles introduced above. 

Chonghun Han, Ramachandran Lakshmanan, Bhavik Bakshi, and George Stephanopoulos, Nonmonotonic Reasoning The Synthesis of Operating Procedures in Chemical Plants Pedro M. Saraiva, Inductive and Analogical Learning Data-Driven Improvement of Process Operations... 

Despite the broad definition of chemometrics, the most important part of it is the application of multivariate data analysis to chemistry-relevant data. Chemistry deals with compounds, their properties, and their transformations into other compounds. Major tasks of chemists are the analysis of complex mixtures, the synthesis of compounds with desired properties, and the construction and operation of chemical technological plants. However, chemical/physical systems of practical interest are often very complicated and cannot be described sufficiently by theory. Actually, a typical chemometrics approach is not based on first principles—that means scientific laws and mles of nature—but is data driven. Multivariate statistical data analysis is a powerful tool for analyzing and structuring data sets that have been obtained from such systems, and for making empirical mathematical models that are for instance capable to predict the values of important properties not directly measurable (Figure 1.1). 

Microwave-assisted synthesis is attractive to researchers for many reasons, including speed, yields, and the potential for reduced solvent use. Raman monitoring offers a convenient way to elucidate the chemical mechanism while instantly, continuously monitoring reaction kinetics. This enables rapid, data-driven process optimizations without concerns about safely and accurately sampling out of a microwave vessel stopped mid-reaction. Pivonka and Empheld of AstraZeneca Pharmaceuticals describe the continuous acquisition of Raman spectra of an amine or Knoevenagel coupling reaction in a sealed microwave reaction vessel at elevated temperatures and pressures [134]. 

Prominence prediction by deterministic means is actually one of the most successful uses of non-statistical methods in speech synthesis. This can be attributed to a number of factors, for example the fact that the rules often don t interact or the fact that many of the rules are base on semantic features (such that even if we did use a data driven technique we would still have to come up with the semantic taxonomy by hand). Sproat notes [410] that statistical approaches have had only limited success as the issue (especially in compound noun phrases) is really one of breadth and not modelling regardless of how the prominence algorithm actually works, what it requires is a broad and exhaustive list of examples of compound nouns. Few complex generalisations are present (what machine learning algorithms are good at) and once presented with an example, the rules are not difficult to write by hand. 

The unit selection synthesis technique described in Chapter 16 uses an entirely data driven approach, whereby recorded speech waveforms are cut up, rearranged and concatenated to say new sentences. Given the success of this approach in normal synthesis, a number of researchers have applied these algorithms to FO synthesis [296], [310], [311]. 

Formant synthesis was the first genuine synthesis technique to be developed and was the dominant technique imtil the early 1980s. Formant synthesis is often called synthesis-by-rule a term invented to make clear at the time that this was synthesis from scratch (at the time the term s5mthesis was more commonly used for the process of reconstructing a waveform that had been parameterised for speech coding purposes). As we shall see, most formant synthesis techniques do in fact use rules of the traditional form, but data driven teehniques have also been used. 

Ease of data acquisition Whether the system is rule-driven or data-driven , some data has to be acquired, even if this is just to help the rule-writer determine appropriate values for the rules. Here linear prediction clearly wins, because its parameters can easily be determined from any real speech w aveform. When formant synthesisers were mainly being developed, no fully reliable formant trackers existed, so the formant values had to be determined either manually or semi-manually. While better formant traekers now exist, many other parameters required in formant S5mthesis (e.g. zero loeations or bandwidth values) are still somewhat difficult to determine. Articulatory synthesis is partieularly interesting in that in the past it was next to impossible to acquire data. Now, various techniques such as EMA and MRI have made this much easier, and so it should be possible to collect much bigger databases for this purpose. The inability to collect accurate articulatory data is certainly one of the main reasons why articulatory synthesis never really took off. 

Hogberg, J. Data driven formant synthesis. InProceedings of Eurospeech 1997 (1997). 

---