TODO :o) and even more todo ;^)

We want to be more precise about:

• when things started (since, since ever)

Ever since computers were invented, we have wondered whether they might be made to learn.

• things happening at the same time (as, while, whilst, when)

Alice inputs X into the first box and X' into the second, while Bob inputs Y into both boxes.

• one thing following another (after, before, as soon as, when, once, now that)

In a quantum experiment like this one, it is generally the case that the outcome of the measurement is obtained as soon as the measurement is performed.

Before we start to introduce the framework of the monogenic phase, we motivate why we want to use phase-based methods at all.

• one thing following another very quickly (no sooner...than, the moment...)

• every time (when, whenever, every time)

Briefly, whenever a node A provides state information to a node B, it attaches a timestamp to the message.

The property therefore makes sure that, when B_FAULTY is alive, its requests are never consistently ignored, which is what the user wishes to check.

• when things finish (until, by the time, at which point)

The procedure is iteratively processed until a convergence criterion is fulfilled.

• -ing clauses (after, before, when, while, on + -ing)

When watching an animated view of Figure 2 it is relatively easy to determine when indices are being updated correctly and when they are out of phase.

„When“ is quite overloaded.

Such a method is indirect because it relies on external mechanisms. Sensor networks place higher demands on scalability because every node is by design a potential router.

Because the scheme is optimized for noisy data, we perform edge detection on the cameraman image corrupted with additive white noise. Because there are so many possible paths, the learner is not confident that they have covered the entire information space.

Rules have the advantage of being very intelligible for users since they model information explicitly. Illumination is not a serious problem , since large variations in lighting are not expected over the rather short duration of a shot.

Since filtering is an essential step in obtaining a useful potential surface, we briefly consider the subject of intensity image filtering in Section III-B. Since the above constraints are all linear, the set of boxes with a given number of inputs and outputs is a polytope

We construct and maintain a small number of alternative paths that can be used in case the primary path fails. This step is included to recover from overspecialization in case the imperfect domain theory includes irrelevant literals.

(different use!!!)

In case there are multiple hypotheses consistent with the training examples, FIND-S will find the most specific.

In case a query can not be routed through the content provider nor through the recommender, it must be routed using a default routing strategy, e.g. flooding.

Having detected outliers, we can now safely use a fast, but non-robust method, such as LS.

Eliminating Route 2, we select Route 4 as our power efficient route when we use maximum PA scheme.