Logically speaking, MindWizard isn't for everyone


MindWizard is an artificial intelligence program that's supposed to simulate your own
reasoning powers. It could give a hand to engineers, accountants, analysts or anyone else
who makes forecasts.


AI programs often fall short in emulating the human thinking processes. MindWizard's
developers recognize this. They promise future application-specific templates and macros
that would make the package take off.


I had a hard time finding a way to test it--the fault of the documentation as much as
anything. The package proclaimed, "MindWizard helps you create better results with
your spreadsheets and databases." But the index had no generic spreadsheet or
database lists.


I did find Microsoft Excel, Visual C++ and Visual Basic mentioned in the index, which
led me to a few pages on feeding data from other apps into MindWizard. It turned out to be
easy to import ASCII and Excel files.


MindWizard is a reasoning engine that links to applications. Spreadsheets are great for
manipulating data, but then you have to analyze the numerical output. If you can express
that in logical steps, MindWizard can automate the analysis stage just as the spreadsheet
does the calculating for you.


In MindWizard's workspace, you first set the input data types. In the Reasoning cells,
you link and configure Boolean logic, computational and timer cells that process the
input. There are statistical, table and macro cells to speed up programming. At the right
side of the screen are output cells.


A demo illustrates the complex reasoning followed by commodity traders. The sample
inputs include prices for soybeans, soy oil and soy meal. MindWizard recommends which to
buy or sell based on the price relationships.


When you run the simulation in batch mode, you see changing input prices and changing
highlights to illustrate the reasoning chains. Two graphs at the left of the screen show
prices and profits.


The demo has little relevance for government workers, but the same reasoning chain
would apply to many programs.


Caution: MindWizard is easy to program as long as the logical process you record is
well thought out. Never did the phrase "garbage in, garbage out" seem more
relevant.


Some AI programs deal with fuzzy logic--processes that can't be precisely defined.
MindWizard, however, is firmly grounded in Boolean logic, although you can assign
different weights to certain inputs.


If your project is complex but the logic is clear, you should have little difficulty
setting up MindWizard to reflect that thought process.


I was surprised to find that MindWizard has no wizards, which are common in Microsoft
Windows applications to walk users through far simpler tasks.


If you routinely analyze data, and your thought process lends itself to Boolean
expressions, then MindWizard will be a powerful tool. If you need help deciding how to
analyze data, it won't do you much good.


Given better documentation and real-world apps, MindWizard might become popular. At the
moment, it's suitable only for highly trained professionals who apply formal logic rules
to daily tasks.


The lingering problem with MindWizard and other AI software intended for everyday use
is that few workers can transfer their own thought processes to software--no matter how
easy it is to use.


 


John McCormick, a free-lance writer and computer consultant, has been working with
computers since the early 1960s.


inside gcn

  • machine learning

    Mitigating the risks of military AI

Reader Comments

Please post your comments here. Comments are moderated, so they may not appear immediately after submitting. We will not post comments that we consider abusive or off-topic.

Please type the letters/numbers you see above

More from 1105 Public Sector Media Group