1. PROBLEMATICS Maintenance operations as well as production launches are essentially based on procedures which describe how to install and use a product and how to maintain it. Due to the complexity of to-day's equipements, and to the complexity of their interactions it is difficult to maintain up-to-date documentations. These procedural documents become more and more complex, even if simplified language constraints and revision scenarios are imposed. According to several analysis, out of 377 technicians working in different domains, 45% of them indicate that they have identified major errors in maintenance documents. About 75% indicate that there are major gaps (missing instructions) or obscure or imcomplete instructions, and 78% admit that often need help because they feel they are not operating the right way. We are all confronted to situations where we wish to follow instructions (DIY, software installation, etc.) with pictures, diagrams, etc. and that these are not understandable, have obvious gaps or do not correspond to the situation at stake. In some industrial areas, such difficulties are common and lead to accidents (aeronautics, nuclear energy, health, etc.). Risk analysis and prevention are therefore a major concern.
The most frequently encountered parameters are, informally:
- presence of ’complex’ manners (e. g. very slowly), by
complex we mean either a manner which is inherently difficult
2. A DOMAIN-INDEPENDENT ANALYSIS to realize or a manner reinforced by an adverb of intensity,
- technical complexity of the verb or the verb compound
OF RISKY SITUATIONS FROM TEXT ANAL-used: if most instructions include a verb which is quite
simYSIS ple, some exhibit quite technical verbs, metaphorical uses,
Procedural texts consist of a sequence of instructions, de- or verbs applied to unexpected situations, for which an
elabsigned with some accuracy in order to reach a goal (e.g. oration is needed.
assemble a computer) [
For some of these criteria, some application-dependent knowledge linguistic resources are needed: some lexical data, basic ontological data, and a few business rules. These observations allow us to introduce a very preliminary measure of complexity. To be able to have an indicative evaluation, each of the points above counts for 1, independently of its importance or strength in the text. Complexity c therefore [procedure [purpose Writing a paper: [elaboration Read light sources, then thorough ]] [assumption/circumstance Assuming you’ve been given a topic,] [circumstance When you conduct research], move from light to thorough resources [purpose to make sure you’re moving in the right direction].
Begin by doing searches on the Internet about your topic [purpose to familiarize yourself with the basic issues;]
[temporal−sequence then ] move to more thorough research on the Academic Databases;
[temporal−sequence finally ], probe the depths of the issue by burying yourself in the library.
[warning Make sure that despite beginning on the Internet, you don’t simply end there.
[elaboration A research paper using only Internet sources is a weak paper, [consequence which puts you at a disadvantage... ]]]
While the Internet should never be your only source of information, [contrast it would be ridiculous not to utilize its vast sources
of information. [advice You should use the Internet to acquaint yourself with the topic more before you dig into more academic
texts. ]]]
ranges from 0 to 6. The complexity rate di of instruction i
is c/6 to keep it in [
Expliciteness characterizes the degree of accuracy of an instruction. Several marks, independently of the domain, contribute to making more explicit an instruction: - when appropriate: existence of means or instruments, - pronominal references as minimal as possible, and predicate argument constructions as comprehensive as possible, - length of action explicit when appropriate (stir for 10 minutes), - list of items to consider as explicit and low level as possible (mix the flour with the sugar, eggs and oil), - presence of an argument, advice or warning, - presence of some help elements like images, diagrams, etc. - presence of elaborations, illustrations or goal specification, - presence of a frame or a condition to limit the scope of the action.
Those criteria may be dependent on the domain, for
example length of an action is very relevant in cooking, somewhat
in do-it-yourself, and much less in the society domain.
Similarly as for d, each item counts for 1 at the moment,
expliciteness e therefore ranges from 0 to 8. The expliciteness
rate is ti = e/8 to keep it in [
Now, if we consider the product di × (1 − ti), the more it tends towards 1, the higher the risk is for the action to fail. Therefore, when di is high, it is also necessary that ti is high to compensate the difficulty. Given that di remains unchanged (if the instruction cannot be simplified), the strategy is then to increase ti as much as possible.
A number of factors of risk are clearly domain-dependent. The difficulty is to be able to identify and evaluate risks without any access to a deep semantic analysis of the different actions of the domain at stake since this is seldom available.
In a first stage, as an exploration, our strategy is to extract from a large corpus of documents of the domain, for each action, the set of warnings associated with it. An action is characterized by a verb and its object argument(s), whatever their position in the instruction. Following argumentation theory, instructions with warnings have the following form: instruction because warning, as in Carefully plug-in the mother card vertically, otherwise you will damage the connectors, where the otherwise section is the support: it indicates the risks of not doing the action correctly. In this work, if the action is ’plug-in the mother card’ the risks are the list of those warnings associated with it over the whole corpus.
In this short paper, we presented the main lines of a preliminary approach to risk identification in procedures. This is a huge problem in the industry, to prevent accidents (humans and ecological). We proposed a simple solution to capture domain dependent knowledge acquired from procedure warnings. Obviously, this is just one useful facet of the problem, since a lot of knowledge is implicit and almost never expressed. Our users estimates is that we cover about 40% of the risks using this approach.