Trace-based synthesis

The complexity and power of MIS compare favorably with the trace-based synthesis mechanisms of Biermann and Summers (see Section 3.3.2). Since it is parameterized on a specialization operator, it is much more adaptable than these other mechanisms. However, MIS is much less data-efficient than the latter. 

The work of [Jorge and Brazdil 94] is another attempt at trace-based synthesis of logic programs. Their specifications by examples are augmented with partial execution traces, called algorithm sketches. 

A few researchers have tackled this lack of discipline in the synthesis of recursive logic programs from examples for instance, [Tinkham 90] and [Sterling and Kirschenbaum 93] investigate the use of schemas to guide synthesis. Curiously, the now virtually defunct research on trace-based synthesis of functional programs from examples [Summers 77] [Biermann 78] did not suffer from such a marked lack of discipline , even though this research preceded ILP research. 

In Chapter 3, we survey the use of inductive inference in automatic programming. Specifications by examples are concise and natural, but are usually also incomplete and ambiguous, due to the insufficient expressive power of examples. As inductive inference is much less known than deductive inference, we first survey this field. Inductive synthesis from specifications by examples can be classified into trace-based synthesis and model-based synthesis. We survey the achievements of inductive synthesis of LISP functions and Prolog predicates. 

Organization of the Value Trace based synthesis system, parsing a Value Trace file to build a set of data structures, ASAP control step scheduling, a graphical Value Trace display, and Value Trace metrics. 

We have no empirical data (synthesis times, example consumption,. ..) yet about a comparison between SYNAPSE and its related systems, such as MIS [Shapiro 82], CLINT [De Raedt and Bruynooghe 92b], FOIL [Quinlan 90], or the other model-based systems of the ELP community (see Section 3.4), or the trace-based systems of the LISP community (see Section 3.3). But we may make some predictions by comparing the underlying synthesis mechanisms instead. This requires some analysis first. 

To separate the oil added an equal volume of fresh cool water (note waited until solution cooled before adding the water). The oil started to drop out perfectly, used DCM to extract all traces of the oil. This woik up is by far the cleanest, easiest and simplest to date... (This dreamer was tried all method of ketone synthesis)... Once the oil was extracted, the extracts were pooled washed with sodium bicarbonate lx, saturated solution of NaCI 1x, and two washes with fresh dHzO... Some time was required for the work up as there was a little emulsion from the use of the base wash and then with the first water wash. The JOC ref suggested using an alumina column to remove the catalyst (could be a better way to go). 

The methanation reaction is currently used to remove the last traces (<1%) of carbon monoxide and carbon dioxide from hydrogen to prevent poisoning of catalysts employed for subsequent hydrogenation reactions. Processes for conversion of synthesis gas containing large quantities of carbon monoxide (up to 25%) into synthetic natural gas have been investigated to serve plants based on coal-suppHed synthesis gas. 

In the first of these sequences, often called the Torgov-Smith synthesis, the initial step consists in condensation of a 2-alkyl-cyclopentane-l,3-dione with the allyl alcohol obtained from 6-methoxy-l-tetralone and vinylmagnesium chloride. Although this reaction at first sight resembles a classic SN displacement, the reaction is actually carried out with only a trace of base. 

---