The National Aeronautic and Space Administration has a large investment in Bayesian research. NASA's Ames Research Center is interested in deep-space exploration and knowledge acquisition. In gathering data from deep-space observatories and planetary probes, an apriori imposition of structure or pattern expectations is inappropriate. Researchers do not always know what to expect or even have hypotheses for which to test when gathering such data. Bayesian inference is useful because it allows the inference system to construct its own potential systems of meaning upon the data. Once any implicit network is discovered within the data, the juxtaposition of this network against other data sets allows for quick and efficient testing of new theories and hypotheses.
The AutoClass project is an attempt to create Bayesian applications that can automatically interpolate raw data from interplanetary probes, and deep space explorations.[15] [1] A graphical example of AutoClass's capabilities is displayed in Figure 3. Incidentally, the source code for AutoClass is available in both LISP and C on an Open Source basis.
Figure 3: An AutoClass interpolation of raw data with no predefined categories. Sorted data is grouped by colour and shape. The top area is sorted into green-blue shapes, the middle into blues, and the bottom into red-orange-yellow shapes.
An applied example of AutoClass's capabilities was the input of infrared spectra. Although no differences among this spectra were initially suspected, AutoClass successfully distinguished two subgroups of stars.[16] [2]
Searching for a solution to a problem is usually an NP-hard problem resulting in a combinatorial explosion of possible solutions to investigate. This problem is often ameliorated through the use of heuristics, or sub-routines to make "intelligent" choices along the decision tree. An appropriately defined heuristic can quicken the search by eliminating obviously unsuccessful paths from the search tree. An inappropriately defined heuristic might eliminate the successful solutions and result in no evident solution.
Bayesian networks can replace heuristic methods by introducing a method where the probabilities are updated continually during search.
One class of search algorithms called Stochastic searching utilizes what are known as "Monte-Carlo" procedures. These procedures are non-deterministic and do not guarantee a solution to a problem. As such they are very fast, and repeated use of these algorithms will add evidence that a solution does not exist even though they never prove that such a solution is non-existent.
"Coupling such procedures with knowledge of properties of the distribution from which problem instances are drawn may be an effective way of extending the utility of these algorithms"[17] [3] by helping to focus in on areas of the search tree not previously studied.
Microsoft began work in 1993 on Lumiere, its project to create software that could automatically and intelligently interact with software users by anticipating the goals and needs of these users.
This research was in turn based on earlier research on pilot-aircraft interaction.[18] [5] The concern of this investigation was the ability of a system to supply a pilot with information congruent with the pilot's current focus of attention. Extraneous information or information not related to the pilot's current task list was suppressed.
"This ability to identify a pilot's focus of attention at any moment during a flight can provide an essential link to the provision of effective decision support. In particular, understanding the current goals of a pilot decision maker can be applied to select the presentation of alternative systems and displays."[19] [6]
The Lumiere project at Microsoft eventually resulted in the "Office Assistant" with the introduction of the Office 95 suite of desktop products.[20] [7]
Links
[1] https://niedermayer.ca/node/31#fn14
[2] https://niedermayer.ca/node/31#fn15
[3] https://niedermayer.ca/node/31#fn16
[4] https://niedermayer.ca/user/login?destination=node/37%23comment-form
[5] https://niedermayer.ca/node/31#fn17
[6] https://niedermayer.ca/node/31#fn18
[7] https://niedermayer.ca/node/31#fn19