Name: Sung Il Choi
Course: CIS 203
Subject: Project Paper
Instructor: Dr Wang
Active
Logic
In AI. many researchers tries to make a system which is flexible enough able to take in a lot of different kinds of incoming information such as contradiction, nonsense, change of topic, ambiguity, and so on. To make such system, we need a system that includes detailed information about its own history. A system with its own history can be more flexible under varied conditions. Active Logic provides good system for IA. Program.
Active logic is a family of inference engines, that
use a history of their own reasoning as they run. Thus at any time T, an active logic has a record of its
reasoning at all times prior to T,
and it also knows that the current time is T.
As it continues to reason from time T,
that reasoning is also recorded in the history, marked at time T+1 as having occurred at time T. Thus an active logic records the
passage of time in discrete steps, and the "current" time slides
forward as the system runs. It is convenient to regard its current inferences
as occurring in a working memory, then is transferred to the history (or
long-term memory) in the next time-step.
There is simple examples:
Basic rule:
t: now(t)
-------------------
t(t + 1) now(t + 1)
-
-
-
t(t + n)
now(t + n)
Now has the value t, the next step, value of t is
increased by 1(t +1) and so on. This feature give the Active Logic to keep
track of step numbers and time.
The
contradiction rule:
t: p,
not(p)
-----------------------
t +1
not(p, not(p))
The database has p and not(p), at next step, add
not(p, not(p)) to the database. There will be other rules that will cause the
consequences of p and not(p) from the first step. The rule will try to resolve
the not(p, not(p)) and reinstate either p or not(p) to the database at later
time.
Modus ponens:
t: p,
p -> q
---------------------
t+1 q
The database has p and p->q , at next step,
conclude q.
Application
of Active logic
Active logic can be a framework for many application
dealing with presence of uncertainties. Some of application are:
·
Deadline
-coupled Planning
The time is tracking and
observation mechanisms of active logic give it useful in such applications that deal with
uncertainties while trying to meet a deadline.
·
Common
Sense Reasoning
·
Simple
Default Reasoning
·
Default
Reasoning with Preferences
·
Reasoning
with contradictions
·
Dialog
·
Assumption
·
Repair
·
Change
in meaning
Future
Application of Active logic
Time Sensitive
Automated Theorem Prover
A computer runs a given program for hundreds of
years in a search for a proof because a computer doesn't have aware of the
time. A human mathematician might, after months of pursuing a particular
problem, try a different tracks. For this kind of application, Active logic has
great advantage. Its built-in time sensitivity provides a good mechanism to
become more like a human mathematician.
Active Logic
Operation System
There will be several advantages incorporating
Active Logic. Operation system with Active logic is able to repair itself
because it keeps his old histories in long term memory. OS use old histories to
compare with current situation, then makes correction of the problem. Another
advantage is OS can update itself for its need.
Current Implementation of Active Logic
There is
research group called “Active Logic Research Group” in the Computer Science
Department at the University of Maryland at College Park. This research group is working currently in
implementation of Active Logic called Alma/Carne. Alma applies the rule of
inference to the formulas in the database at that step to produce a set of new
formulas. These are added to the database, and the process repeats at each
subsequent step. Carne is a process that communicates with Alma but runs
independently. The main use of Carne is to run non-logical computations
asynchronous from Alma steps. One of its roles is serve as an input-output
interface to Alma.
I think Active Logic needs future step or future
time to make more efficient and reliable. Now (t) is the current time or step
and, at same time, create two future steps. These future step will be Now(tp)
and Now(tn). Now(tp) are the positive result or conclusion and Now(tn) is the
negative or contradiction result. For next step, Active logic use future positive and negative to obtain next
conclusion with current step or time.
I think that Active Logic has theoretical value. A
system with Active Logic can have information about its own past, it then might
be able to analyze what it has been doing, recognize trouble, and make
decisions as to whether it needs to alter any of its past behavior. Active Logic also can provide mechanism to
deal with various forms of uncertainties. It provides a framework for several
application programs.
Active Logic has great future in AI. With Active
Logic we can pursue a system that can work in misidentifications,
contradiction, conflicting advice, context shifts, new words, new meanings for
old words, belief revision, tight deadlines, and ambiguities. Maybe in future,
we can have operation system with Active logic that can repair and update
itself.