perm filename DOYLE.DOC[S80,JMC] blob sn#517179 filedate 1980-06-19 generic text, type C, neo UTF8
COMMENT āŠ—   VALID 00012 PAGES
C REC  PAGE   DESCRIPTION
C00001 00001
C00003 00002	Supplementary proposal
C00004 00003	                                                                     1
C00009 00004	Modelling Deliberation, Action and Introspection                     2
C00014 00005	Modelling Deliberation, Action and Introspection                     3
C00017 00006	Modelling Deliberation, Action and Introspection                     4
C00020 00007	Modelling Deliberation, Action and Introspection                     5
C00024 00008	Modelling Deliberation, Action and Introspection                     6
C00028 00009	Modelling Deliberation, Action and Introspection                     7
C00031 00010	Modelling Deliberation, Action and Introspection                     8
C00033 00011	Modelling Deliberation, Action and Introspection                     9
C00035 00012	
C00036 ENDMK
CāŠ—;
Supplementary proposal
Bob,

Here is the text of a supplementary proposal covering Jon Doyle's work, as
discussed earlier.  We have started pushing it through the paper mill, so
that a formal submission will reach you quite soon.  If you spot any
problems, please let us know.

Regards,
Les
-----------
                                                                     1


         1.  Modelling Deliberation, Action and Introspection

        This is a proposal to extend the work of the  Formal Reasoning
Group  of the  Stanford  Artificial Intelligence  Laboratory  to study
programs modelling deliberation,  action and introspection.   The work
will be done by Jon Doyle and by John McCarthy.  Doyle's work requires
additional DARPA support amounting  to $62,055 for the  period October
1, 1980 thru October 31, 1981 as described in the budgetary section of
this proposal.  No additional support is requested for McCarthy.

        Advanced intelligent computer  programs must reason  about the
effects of their potential future actions, and this includes reasoning
about their  own ability to  solve problems by  reason and  action.  A
promising  approach is  to  regard reasoning  itself as  a  species of
action that  whose effects  can be reasoned  about.  Thus  the program
must  reason  about  what it  would  be  able to  do  or  would  do in
hypothetical future circumstances.   Carrying this out  accurately and
effectively  involves  self-observation akin  to  human introspection.
Self-observation includes  making and  examining traces  of inferences
made so that when a  conclusion has to be revised, the  reasoning that
led  to  it  can  be  identified  and  the  assumption  or conjectural
conclusion that has to be taken back can be located.

        Much  human decision  making  involves processes  that  may be
called  dialectical   argumentation.   Reasons   for  and   against  a
contemplated course  of action  are developed  and played  against one
another.  This process, which we believe is also required for advanced
computer  intelligence,  is  quite  different  from  the  mathematical
deductions heretofore carried out by computer programs.

        In  particular,  it involves  non-monotonic  reasoning  of the
kinds recently studied by McCarthy (1980), McDermott and Doyle (1980),
Doyle (1979) and  Reiter (1980).  The identification  of non-monotonic
reasoning as a  process  distinct from  logical deduction but  just as
formal  was  accomplished in  various  formalisms in  the  above cited
papers.  It represents a major discovery by workers in  AI.  Logicians
and philosophers have from time to time asserted the existence of non-
deductive modes of  reasoning but have  generally supposed them  to be
non-formalizable.   Now that  AI has  established an  entry  into this
field, logicians and philosophers are also beginning to work on it.

        Doyle's proposals  for self-argumentation  go a  step farther.
Consider  what happens  when a  person proposes  an argument  and then
thinks  of  a  counter-argument.  The  original  argument  leads  to a
certain conclusion.  If this conclusion were a logical consequence, in
the  sense of  mathematical logic,  then no  additional considerations
would change the conclusion unless some of the premises were  found to
be  incorrect.   Actually  such  arguments  are  almost  never logical
deductions but non-monotonic consequences of two kinds.
Modelling Deliberation, Action and Introspection                     2


        One  kind  of  non-monotonic  conclusion  is  the  default.  A
default is represented by a sentence that is taken to be true provided
other sentences  being considered  don't refute  it.  An  example from
Doyle (1979) is "The meeting is on Wednesday unless there is  a reason
why  not".  A  program  will use  this  default to  conclude  that the
meeting is on Wednesday  unless it has a sentence  asserting something
incompatible like a conflicting meeting on Wednesday.

        Another  kind  of non-monotonic  consequence  occurs  when the
facts  at  a person's  or  program's disposal  show  the  existence of
certain objects of a given  kind, and the person or  program concludes
that these are  all of the  objects of the  given kind.  Thus   we may
know that  a boat has  a leak  and lacks oars,  and we  may conjecture
(non-monotonically) that  these are the  only "things" wrong  with the
boat.

        Argument, whether with  another person or with  oneself, often
involves  finding reasons  permitting such  non-monotonically obtained
conclusions  and  then  finding  counter  arguments.   In   the  above
examples, a counter  argument might involve a  non-monotonic deduction
that there is  another group in Wednesday's  meeting room or  that the
boat also has a broken rudder.  Counter-counter arguments  may involve
asserting  the existence  of another  room or  a plan  for  fixing the
rudder.

        We propose to extend the work of our Formal Reasoning Group to
develop theories  and write programs  that will decide  what to  do by
reasoning that  includes introspection and  dialectical argumentation.
This  work  will  be  based  on  ideas  in  Jon  Doyle's   (1980)  PhD
dissertation and on approaches to non-monotonic reasoning by Doyle and
by McCarthy.

        Doyle's  thesis  investigates the  problem  of  controlling or
directing the reasoning and actions of a computer program.   The basic
approach explored is to view reasoning as a species of action, so that
a program  might apply its  reasoning powers to  the task  of deciding
what inferences  to make  as well  as deciding  what other  actions to
take.   Doyle  proposed a  design  for the  architecture  of reasoning
programs.   This  architecture   involves  several  of   the  features
mentioned earlier, including self-consciousness,  intentional actions,
deliberate  adaptations,  and  a  form  of  decision-making  based  on
dialectical argumentation.

        A  program based  on this  architecture would  inspect itself,
describe aspects of itself to itself, and use this  self-reference and
these self-descriptions  in making decisons  and taking  actions.  The
program's mental  life would  include awareness  of its  own concepts,
beliefs, desires, intentions, inferences, actions, and skills.  All of
Modelling Deliberation, Action and Introspection                     3


these  are  represented  by  self-descriptions  in  a  single  sort of
language, so that  the program has access  to all of these  aspects of
itself, and can reason about them in the same terms.

        During the thirteen month period of this addition to work, the
studies will mainly be conceptual.  This is partly because these ideas
require additional  theoretical work  before they  can be  embodied in
programs and partly  because programs that keep  a trace of  their own
reasoning  processes may  require bigger  memories than  are currently
available   for  single   processes   at  Stanford.    In   any  case,
implementation will be a large enough project to require  very careful
planning.

        By the end of 1981, we will have planned a system that will be
able to  reason about  its own  actions and  "thoughts" and  carry out
internal  arguments  as   well  as  simpler  forms   of  non-monotonic
reasoning.



                              References

Doyle,   J.  (1979):   "A  truth   maintainance   system",  Artificial
Intelligence 12, 231-272

Doyle,   J.   (1980):   "A  model   for   deliberation,   action,  and
introspection" PhD Thesis, MIT AI Laboratory TR-581

McCarthy,  J.  (1980):  "Circumscription -   a  form  of non-monotonic
reasoning" Stanford Artificial Intelligence Laboratory Memo AIM-334

McDermott,  Drew  and  Jon  Doyle  (1980):  "Non-monotonic  logic  I",
Artificial Intelligence, to appear

Reiter, Raymond  (1980): "A logic  for default  reasoning", Artificial
Intelligence, to appear
Modelling Deliberation, Action and Introspection                     4


                           CURRICULUM VITAE
                                  of
                              Jon Doyle


Upcoming Position:
        Research Associate
        Artificial Intelligence Laboratory
        Stanford University

Present Position:
        Scientist
        Artificial Intelligence Laboratory
        Massachusetts Institute of Technology

Address:
        305 Memorial Drive, #612-A
        Cambridge, Massachusetts 02139

Telephone: (617) 494-9214

Citizenship: United States of America

Principal Fields of Professional Interest:
        Intelligence
        Theory of Computation
        Philosophy
        Logic
        Mathematics
        Physics

Education:

1971-72   South Texas Junior College, Houston, Texas.

1972-74   University of Houston, B.S. in Mathematics, December 1974.
  Senior Honors Thesis under Prof. J. A. Schatz on "Computational
  Investigations of Non-Repetitive Sequences."

1975-77   M.I.T., S.M. in Electrical Engineering and Computer Science,
  June 1977.    Thesis under Prof. G. J. Sussman on "Truth Maintenance
  Systems for Problem Solving."

1977-80   M.I.T., Ph.D. in Artificial Intelligence, June 1980.
  Thesis under Prof. G. J. Sussman on "A Model for Deliberation,
  Action, and Introspection."  Profs. M. Minsky, P. Szolovits, and
  D. McDermott (Yale), readers.

Modelling Deliberation, Action and Introspection                     5


History of Employment:

May 1974 - August 1974  Symbiotics International, Inc., Houston, Texas
  Development and maintenance of business accounting programs.

January 1975 - July 1975  Shell Oil Company, Houston, Texas
  Development of a user environment for geophysical processing in the
  Technical Computing   Division.
nofill;

Awards Received:
        1975-1980 Fannie and John Hertz Foundation Graduate Fellowship
        1975 NSF Honorable Mention
        1974 Summa Cum Laude, University of Houston
        1974 Honors in Mathematics, University of Houston
        1974 Honors Program, University of Houston
        1974 First prize Mathematics Contest, University of Houston
        1973 Third prize Mathematics Contest, University of Houston
        1973 Outstanding First-year Student of Russian,
              University of Houston

Current Organization Memberships:
        American Association for the Advancement of Science
        American Mathematical Society
        Association for Computing Machinery
        Boston Museum of Fine Arts
        Mathematical Association of America
        MIT Musical Theater Guild
        Sigma Xi

Sigart, Pi Mu Epsilon, Phi Kappa Phi, Omega Delta Kappa

Past Organization Memberships:
        MIT Choral Society (1979)
        MIT Symphony Orchestra (1977-78)

Offices:
        Director of Univ. of Houston chapter of Pi Mu Epsilon (1974)
        MIT Ashdown House Executive Committee (1978-79)
        Associate Editor of ACM SIGART Newsletter (1978-79)

Publications:

Papers in Refereed Journals:

1. Jon Doyle and Ronald L. Rivest, "Linear Expected Time of a Simple
  Union-Find Algorithm", Information Processing Letters, Volume 5,
  Number 5, (November 1976), pp. 146-148.

2. Jon Doyle, "A Truth Maintenance System" Artificial Intelligence 12
  (1979), 231-272.   Also MIT AI Lab Memo 521, June 1979.
Modelling Deliberation, Action and Introspection                     6


3. Drew McDermott and Jon Doyle, "Non-Monotonic Logic I", to appear in
  Artificial Intelligence, 1980.  Also MIT AI Lab Memo 468, August
  1978.  Abstract in Notices of the AMS, V. 26, No. 1 (Jan. 1979),
  #79T-E4, p. A-16.

Papers in Other Journals

1. Jon Doyle and Philip London,  "A Selected Descriptor-Indexed
  Bibliography to the Literature on Belief Revision",  ACM SIGART
  Newsletter, No. 71, 7-23, 1980.   Also MIT AI Lab Memo 568, February
  1980.

2. Jon Doyle  "Historical Annotations and Humble Databases",  to
  appear in ACM SIGMOD Record.

Proceedings of Refereed Conferences:

1. Johan de Kleer, Jon Doyle, Guy L. Steele Jr. and Gerald Jay
  Sussman,  "AMORD: Explicit Control of Reasoning",  Proc. ACM
  Conference on AI and Programming Languages, Rochester, New York,
  August 1977.   Also MIT AI Lab Memo 427 ("Explicit Control of
  Reasoning"), June 1977.

2. Jon Doyle,  "Truth Maintenance Systems for Problem Solving",  Proc.
  Fifth International Joint Conference on Artificial Intelligence,
  Cambridge, Massachusetts, August 1977.   Also MIT AI Lab TR-419,
  January 1978.

3. Jon Doyle, "A Glimpse of Truth Maintenance",  Proc. Fourth Workshop
  on Automated Deduction, Austin, Texas, February 1979.   Also Proc.
  Sixth International Joint Conference on Artificial Intelligence,
  Tokyo, Japan, August 1979.   Also MIT AI Lab Memo 461, February
  1978.

4. Drew McDermott and Jon Doyle,  "Non-Monotonic Logic I (extended
  abstract)",  Proc. Fourth Workshop on Automated Deduction, Austin,
  Texas, February 1979.

5. Drew McDermott and Jon Doyle,  "An Introduction to Non-Monotonic
  Logic",  Proc. Sixth International Joint Conference on Artificial
  Intelligence, Tokyo, Japan, August 1979.

Unrefereed Conferences:

1. Jon Doyle,  "Non-Repetitive Binary Sequences",  727th Meeting of
  the American Mathematical Society, Cambridge, Massachusetts,
  October, 1975.   Abstract in Notices of the AMS, Oct. 1975, p. A-
  660, #727-A5.   Referee recommended but scooped for J. Combinatorial
Modelling Deliberation, Action and Introspection                     7


  Theory A.  My theorems 1 and 2 paraphrase theorems 1 and 2 of F. M.
  Dekking, "On Repetitions of Blocks in Binary Sequences," J. C. T. A
  20 #3 (May 1976) 292-299.

Articles in Books:

1. Johan de Kleer, Jon Doyle, Guy L. Steele Jr. and Gerald Jay
  Sussman,  "Explicit Control of Reasoning",  Artificial Intelligence:
  An MIT Perspective,  P. H. Winston and R. H. Brown, editors,  MIT
  Press series in Artificial Intelligence, Cambridge, 1979.

2. Jon Doyle,  "A Glimpse of Truth Maintenance",  Artificial
  Intelligence: An MIT Perspective,  P. H. Winston and R. H. Brown,
  editors,  MIT Press series in Artificial Intelligence, Cambridge,
  1979.

Internal Reports:

1. Jon Doyle,  "Analysis by Propagation of Constraints in Elementary
  Geometry Problem Solving",  MIT AI Lab WP-108, June 1976.

2. Jon Doyle,  "The Use of Dependency Relationships in the Control of
  Reasoning",  MIT AI Lab WP-133, November 1976.

3. Jon Doyle,  "Hierarchy in Knowledge Representations",  MIT AI Lab
  WP-159, November 1977.

4. Johan de Kleer, Jon Doyle, Charles Rich, Guy L. Steele Jr., and
  Gerald Jay Sussman,  "AMORD: A Deductive Procedure System",  MIT AI
  Lab Memo 435, January 1978.

Invited Lectures:

1. "What if the Mayans had Escalators",  Yale University, Computer
  Science Department, June 23, 1978.

2. "Reflexive Interpreters: A Model for Deliberation, Action, and
  Introspection",  University of Pennsylvania, Computer Science
  Department, October 30, 1979.

3. "Reasoned Deliberation and Decision-Making"  SRI International,
  January 14, 1980.

4. "Reasoned Deliberation and Decision-Making"  Hewlett-Packard, Inc.,
  January 15, 1980.

5. "Reasoned Deliberation and Decision-Making"  Stanford University,
  Computer Science Department, January 16, 1980.
Modelling Deliberation, Action and Introspection                     8


6. "Reasoned Deliberation and Decision-Making"  Xerox Palo Alto
  Research Center, January 17, 1980.

7. "Reasoned Deliberation and Decision-Making"  USC Information
  Sciences Institute, January 24, 1980.

8. "Reasoned Deliberation and Decision-Making"  Bell Laboratories,
  February 14, 1980.

Personal Background and Interests:

Born and raised in Houston, Texas and, in summers, Dundee, Wisconsin,
  by Leo M. Doyle and Marilyn C. Doyle.  Unmarried.  Interests in
  people, writing, literature, history, economics, musical and
  poetical composition, conducting, viola, recorders, painting,
  sculpture, technology, swimming and other sports.
Modelling Deliberation, Action and Introspection                     9


                     Basic AI & Formal Reasoning
Proposed budget addition to Contract MDA 903-80-C-0102 (DARPA Order
No. 2494) for the period 1 October 1980 through 31 October 1981.

                             Professional
Doyle, Jon                                                      27,952
                                                            ----------
   Salary total                                                 27,952
Staff benefits (21.15% of salaries)                              5,912
Travel                                                           1,173
Computer cost for 1 person(s)                                    2,002
Other direct expenses                                            2,236
                                                            ----------
   Total direct costs                                           39,275
Indirect costs (58% of above)                                   22,780
                                                            ----------
   Project total                                                62,055