190
9 Networked Minds
system is increasingly being replaced by online reviews and recommender systems
such as price comparison websites, which widens the circle of people contributing
information and improves the chances that each individual will make better decisions.
While the above approach collates existing knowledge well, some methods can
also create new knowledge, which is more than just the sum of its parts. Let us
assume that a particular problem needs to be solved. One would then invite everyone
to contribute to an online discussion of aspects that should be considered. With
suitable software, it is possible to structure these aspects in an argument map, and
to work out a number of different perspectives that matter.
23 Once, the problem has
been boiled down to anything between, say, 2 and 7 perspectives, it’s time to invite
the leading representatives of these perspectives to a round table. Political or other
decision makers could then moderate a deliberation process to develop an integrated
solution (or a few good alternatives to choose from). When many different ideas
“collide”, this often triggers innovative ideas, which may help to overcome conflicts
of interest in favor of a win-win situation. This leads to better solutions, which find
wider support. Therefore, rather than letting a majority impose partial interests on
others, the future success principle will be to get as many good ideas on board and
produce as many “winners” as possible.
The examples above illustrate how collective intelligence works. First, a number
of teams needs to tackle a problem independently using diverse methodologies.
Subsequently, these independent streams of knowledge need to be combined. If
there is too much communication at the beginning of this process, each team may be
tempted to copy promising approaches of others, which would reduce the diversity
of ideas. However, if there is too little communication at the end of the process, the
knowledge created by all these different approaches won’t be fully used.
9.15 What We Can Learn from IBM’s Watson Computer
It is also interesting to discuss how “cognitive computing” works in IBM’s Watson
computer.
24 The computer scans hundreds of thousands of sources of information,
including scientific publications, and extracts potentially relevant statements. It can
also formulate hypotheses and seek evidence to support or refute them. Afterwards, it
produces a list of possible answers and ranks them according to their likelihood. All
of this is achieved using algorithms based on the laws of probability, which calculate
how likely a hypothesis is based on the available data. Compared to humans, however,
Watson loses less information, which we would filter out due to limited memory and
attention spans, cognitive biases, and our preference for consistency. For example,
when applied in a medical context, Watson would come up with a ranked list of
diseases that are compatible with a number of symptoms. While a typical doctor
23 See https://en.wikipedia.org/wiki/Argument_map and http://sourceforge.net/projects/argumenta
tive/.
24 Kelly and Hamm [19].
9 Networked Minds
system is increasingly being replaced by online reviews and recommender systems
such as price comparison websites, which widens the circle of people contributing
information and improves the chances that each individual will make better decisions.
While the above approach collates existing knowledge well, some methods can
also create new knowledge, which is more than just the sum of its parts. Let us
assume that a particular problem needs to be solved. One would then invite everyone
to contribute to an online discussion of aspects that should be considered. With
suitable software, it is possible to structure these aspects in an argument map, and
to work out a number of different perspectives that matter.
23 Once, the problem has
been boiled down to anything between, say, 2 and 7 perspectives, it’s time to invite
the leading representatives of these perspectives to a round table. Political or other
decision makers could then moderate a deliberation process to develop an integrated
solution (or a few good alternatives to choose from). When many different ideas
“collide”, this often triggers innovative ideas, which may help to overcome conflicts
of interest in favor of a win-win situation. This leads to better solutions, which find
wider support. Therefore, rather than letting a majority impose partial interests on
others, the future success principle will be to get as many good ideas on board and
produce as many “winners” as possible.
The examples above illustrate how collective intelligence works. First, a number
of teams needs to tackle a problem independently using diverse methodologies.
Subsequently, these independent streams of knowledge need to be combined. If
there is too much communication at the beginning of this process, each team may be
tempted to copy promising approaches of others, which would reduce the diversity
of ideas. However, if there is too little communication at the end of the process, the
knowledge created by all these different approaches won’t be fully used.
9.15 What We Can Learn from IBM’s Watson Computer
It is also interesting to discuss how “cognitive computing” works in IBM’s Watson
computer.
24 The computer scans hundreds of thousands of sources of information,
including scientific publications, and extracts potentially relevant statements. It can
also formulate hypotheses and seek evidence to support or refute them. Afterwards, it
produces a list of possible answers and ranks them according to their likelihood. All
of this is achieved using algorithms based on the laws of probability, which calculate
how likely a hypothesis is based on the available data. Compared to humans, however,
Watson loses less information, which we would filter out due to limited memory and
attention spans, cognitive biases, and our preference for consistency. For example,
when applied in a medical context, Watson would come up with a ranked list of
diseases that are compatible with a number of symptoms. While a typical doctor
23 See https://en.wikipedia.org/wiki/Argument_map and http://sourceforge.net/projects/argumenta
tive/.
24 Kelly and Hamm [19].