"Basically, Cyc is based on getting humans to tell computers what words
mean. Webmind can be taught in this way – we have our own knowledge
entry language called KNOW, which we believe is simpler than Cyc and
closer to everyday human psychology. But KNOW expressions inspire rather
than control Webmind’s internal definitions, which in general are much
bigger and messier than KNOW Cyc definitions. In dealing with most
practical situations, humans, or Webminds, don’t refer to this kind of
abstract definition at all, but we use more detailed, context-specific
patterns of knowledge.
Cyc tried to divorce information from learning, but it can't be done. A
mind can only make intelligent use of information that it has figured
out for itself. Despite sixteen years of programming, Cyc never
succeeded in emulating an eight year old child. Nor has anyone yet found
much use for a CD-ROM full of formal, logical definitions of common
sense information. ...
... Based partically from the “Global Workspace” concept<
http://www.phil.vt.edu/ASSC/esem2.html >, Webmind 1.0 is scheduled to be
completed in 2001. Webmind 2.0 will incorporate the ability to reason
about and modify its own source code, and also one additional sensory
modality: sound processing (voice and music). This is expected about 18
months after Webmind 1.0. ...
... Webmind is not the only AI system to seek a middle way between
symbolic and connectionist AI. Gerald Edelman, a Nobel Prize winning
biologist, proposed a theory of "neuronal group selection" or Neural
Darwinism, which describes how the brain constructs larger scale
networks called "maps" out of neural modules, and selects between these
maps in an evolutionary manner, in order to find maps of optimum
performance. And Marvin Minsky, the champion of rule-based AI, had moved
in an oddly similar direction, proposing a "Society of Mind" theory in
which is mind is viewed as a kind of society of actors or processes that
send messages to each other and form alliances into temporary working
groups. < http://www.intelligenesis.net/webmindandotherai.html >...
< http://www.intelligenesis.net/webmindphilosophy.html >
< http://www.intelligenesis.net/communicatorframework.html >
3. The Webmind AI Architecture
“Activation,” the Webmind form of energy, spreads through the network
between nodes and links according to neural-net-like dynamics (embodying
Peirce’s Law of Mind). Activation-spreading, in itself is goal-less and
spontaneous, driven purely by the complex nonlinear dynamics of Peircean
generalized association. But this isn’t the whole story: a substantial
percentage of system activity is goal-directed. A goal activates things
that help achieve it, also recording as lasting knowledge information
about which things helped achieve it. Thus activation spreading, in
large part, works in the service of activating things that can help
fulfill basic system goals.
< http://www.intelligenesis.net/webmindaiengine.html >
KNOW (Knowledge Norm Of Webmind; likely be made public - open source?)
is a language used for knowledge representation in Webmind. It is
simpler and more cognitively natural than the predicate-logic knowledge
representation languages used in many other AI systems, such as CYC. The
main goals of KNOW are to enable
* Humans to enter knowledge into Webmind unambiguously
* Webmind to report its knowledge unambiguously
* Simple import of knowledge from external resources (e.g. databases,
other AI systems) into Webmind
To serve these purposes, KNOW was designed to keep a subtle balance. In
its relation to Webmind, the structure of the sentences in KNOW is
closely related to how knowledge is represented within Webmind (i.e.,
the nodes and links).
A text in KNOW is composed of sentences, with each sentence having
strength and confidence values. A sentence is a relation that has
arguments. A relation can be either built-in or user-defined. The
following example shows a simple text in KNOW:
{ [give John Mary book1 (1.0 0.9)]
[Inheritance book1 book (1.0 0.9)]
[author book1 John (1.0 0.9)] }
This text in composed of three sentences. Give, inheritance and author
are relations, and John, Mary, book and books are the arguments. A text
in KNOW can also be represented in XML format.
The complete KNOW knowledge representation grammar will likely be made
public at some future point, to enable maximally fluid knowledge
exchange between Webmind and other AI systems and computer programs.
This archive was generated by hypermail 2b29 : Thu Jul 05 2001 - 23:58:52 PDT