perm filename DOYLE.XGP[S80,JMC] blob sn#519252 filedate 1980-06-30 generic text, type T, neo UTF8
/LMAR=0/XLINE=3/FONT#0=NONM/FONT#1=NONMI/FONT#2=NONMB/FONT#3=NGR40/FONT#4=NONH/FONT#5=NGR30/FONT#6=BDR40
␈↓ ↓H␈↓α␈↓ E1


␈↓ ↓H␈↓β␈↓ ∧jSupplementary Proposal to


␈↓ ↓H␈↓∧␈↓ ¬\Defense
␈↓ ↓H␈↓∧␈↓ αCAdvanced Research Projects Agency


␈↓ ↓H␈↓β␈↓ ¬Qfor research in


␈↓ ↓H␈↓∧␈↓ ∧NComputer Science

␈↓ ↓H␈↓β␈↓ β8John McCarthy, Professor of Computer Science
␈↓ ↓H␈↓β␈↓ ¬∨Principal Investigator









␈↓ ↓H␈↓β␈↓ ¬mJune 1980

␈↓ ↓H␈↓ε␈↓ ∧BComputer Science Department
␈↓ ↓H␈↓∧␈↓ ∧8Stanford University
␈↓ ↓H␈↓α␈↓ E1


␈↓ ↓H␈↓α␈↓ ↓h1.  Modelling Deliberation, Action and     ␈↓ εx␈↓Carrying this out accurately and
␈↓ ↓H␈↓ β_␈↓αIntrospection                              ␈↓ εx␈↓effectively involves self-observation akin
                                           ␈↓ εx␈↓to human introspection.  Self-observation
␈↓ ↓H␈↓        This is a proposal to extend the work␈↓ εx␈↓includes making and examining traces of
␈↓ ↓H␈↓of the Formal Reasoning Group of the       ␈↓ εx␈↓inferences made so that when a conclusion
␈↓ ↓H␈↓Stanford Artificial Intelligence Laboratory␈↓ εx␈↓has to be revised, the reasoning that led
␈↓ ↓H␈↓to study programs modelling deliberation,  ␈↓ εx␈↓to it can be identified and the assumption
␈↓ ↓H␈↓action and introspection.  The work will be␈↓ εx␈↓or conjectural conclusion that has to be
␈↓ ↓H␈↓done by Jon Doyle and by John McCarthy.    ␈↓ εx␈↓taken back can be located.
␈↓ ↓H␈↓Doyle's work requires additional DARPA
␈↓ ↓H␈↓support amounting to $62,055 for the       ␈↓ εx␈↓        Much human decision making involves
␈↓ ↓H␈↓period October 1, 1980 thru October 31,    ␈↓ εx␈↓processes that may be called ␈↓↓dialectical
␈↓ ↓H␈↓1981 as described in the budgetary         ␈↓ εx␈↓↓argumentation␈↓.  Reasons for and against a
␈↓ ↓H␈↓section of this proposal.  No additional   ␈↓ εx␈↓contemplated course of action are
␈↓ ↓H␈↓support is requested for McCarthy.         ␈↓ εx␈↓developed and played against one another.
                                           ␈↓ εx␈↓This process, which we believe is also
␈↓ ↓H␈↓        Advanced intelligent computer      ␈↓ εx␈↓required for advanced computer
␈↓ ↓H␈↓programs must reason about the effects of  ␈↓ εx␈↓intelligence, is quite different from the
␈↓ ↓H␈↓their potential future actions, and this   ␈↓ εx␈↓mathematical deductions heretofore
␈↓ ↓H␈↓includes reasoning about their own ability ␈↓ εx␈↓carried out by computer programs.
␈↓ ↓H␈↓to solve problems by reason and action.
␈↓ ↓H␈↓For example, a person familiar with air    ␈↓ εx␈↓        In particular, it involves non-
␈↓ ↓H␈↓travel knows that when he reaches an       ␈↓ εx␈↓monotonic reasoning of the kinds recently
␈↓ ↓H␈↓intermediate airport, he will be able to   ␈↓ εx␈↓studied by McCarthy (1980), McDermott
␈↓ ↓H␈↓determine the gate and time of his         ␈↓ εx␈↓and Doyle (1980), Doyle (1979) and Reiter
␈↓ ↓H␈↓outgoing airplane, and he knows that he    ␈↓ εx␈↓(1980).  The identification of non-
␈↓ ↓H␈↓doesn't need this information in advance.  ␈↓ εx␈↓monotonic reasoning as a process  distinct
␈↓ ↓H␈↓However, a non English speaking person     ␈↓ εx␈↓from logical deduction but just as formal
␈↓ ↓H␈↓needs to plan how he is going to learn this␈↓ εx␈↓was accomplished in various formalisms in
␈↓ ↓H␈↓information - by getting someone to write  ␈↓ εx␈↓the above cited papers.  It represents a
␈↓ ↓H␈↓appropriate names and requests on a card   ␈↓ εx␈↓major discovery by workers in AI.
␈↓ ↓H␈↓or by finding an interpreter.  A person    ␈↓ εx␈↓Logicians and philosophers have from time
␈↓ ↓H␈↓unfamiliar with air travel may not even    ␈↓ εx␈↓to time asserted the existence of non-
␈↓ ↓H␈↓know what he has to learn, but he probably ␈↓ εx␈↓deductive modes of reasoning but have
␈↓ ↓H␈↓does know that the air travel system will  ␈↓ εx␈↓generally supposed them to be non-
␈↓ ↓H␈↓take care of him if he has no unusual      ␈↓ εx␈↓formalizable.  Now that AI has established
␈↓ ↓H␈↓problems.  More generally, whenever an     ␈↓ εx␈↓an entry into this field, logicians and
␈↓ ↓H␈↓executive decides that getting a job done  ␈↓ εx␈↓philosophers are also beginning to work on
␈↓ ↓H␈↓right requires his own presence at the     ␈↓ εx␈↓it.
␈↓ ↓H␈↓scene of action, he is anticipating problems
␈↓ ↓H␈↓that require his own special knowledge,    ␈↓ εx␈↓        Consider what happens when a
␈↓ ↓H␈↓priorities for different goals, ability and␈↓ εx␈↓person proposes an argument and then
␈↓ ↓H␈↓authority.                                 ␈↓ εx␈↓thinks of a counter-argument.  The original
                                           ␈↓ εx␈↓argument leads to a certain conclusion.  If
␈↓ ↓H␈↓        A promising approach is to regard  ␈↓ εx␈↓this conclusion were a logical
␈↓ ↓H␈↓reasoning itself as a species of action that␈↓ εx␈↓consequence, in the sense of mathematical
␈↓ ↓H␈↓whose effects can be reasoned about.       ␈↓ εx␈↓logic, then no additional considerations
␈↓ ↓H␈↓Thus the program must reason about what    ␈↓ εx␈↓would change the conclusion unless some
␈↓ ↓H␈↓it would be able to do or would do in      ␈↓ εx␈↓of the premises were found to be
␈↓ ↓H␈↓hypothetical future circumstances.         ␈↓ εx␈↓incorrect.  Actually such arguments are
␈↓ ↓H␈↓αModelling Deliberation, Action and Introspection␈↓ E2


␈↓ ↓H␈↓almost never logical deductions but non-   ␈↓ εx␈↓α␈↓ λLEarlier Work
␈↓ ↓H␈↓monotonic consequences of two kinds.
                                           ␈↓ εx␈↓        The proposed work will be based on
␈↓ ↓H␈↓        One kind of non-monotonic conclusion␈↓ εx␈↓ideas in Jon Doyle's (1980) PhD
␈↓ ↓H␈↓is the default.  A default is represented by␈↓ εx␈↓dissertation and on approaches to non-
␈↓ ↓H␈↓a sentence that is taken to be true        ␈↓ εx␈↓monotonic reasoning by Doyle (1979) and
␈↓ ↓H␈↓provided other sentences being considered  ␈↓ εx␈↓by McCarthy (1980).
␈↓ ↓H␈↓don't refute it.  An example from Doyle
␈↓ ↓H␈↓(1979) is ␈↓↓"The meeting is on Wednesday     ␈↓ εx␈↓        Doyle's thesis investigates the
␈↓ ↓H␈↓↓unless there is a reason why not"␈↓.  A      ␈↓ εx␈↓problem of controlling or directing the
␈↓ ↓H␈↓program will use this default to conclude  ␈↓ εx␈↓reasoning and actions of a computer
␈↓ ↓H␈↓that the meeting is on Wednesday unless it ␈↓ εx␈↓program.  The basic approach explored is
␈↓ ↓H␈↓has a sentence asserting something         ␈↓ εx␈↓to view reasoning as a species of action,
␈↓ ↓H␈↓incompatible like a conflicting meeting on ␈↓ εx␈↓so that a program might apply its reasoning
␈↓ ↓H␈↓Wednesday.                                 ␈↓ εx␈↓powers to the task of deciding what
                                           ␈↓ εx␈↓inferences to make as well as deciding
␈↓ ↓H␈↓        Another kind of non-monotonic      ␈↓ εx␈↓what other actions to take.  Doyle
␈↓ ↓H␈↓consequence occurs when the facts at a     ␈↓ εx␈↓proposed a design for the architecture of
␈↓ ↓H␈↓person's or program's disposal show the    ␈↓ εx␈↓reasoning programs.  This architecture
␈↓ ↓H␈↓existence of certain objects of a given    ␈↓ εx␈↓involves several of the features mentioned
␈↓ ↓H␈↓kind, and the person or program concludes  ␈↓ εx␈↓earlier, including self-consciousness,
␈↓ ↓H␈↓that these are all of the objects of the   ␈↓ εx␈↓intentional actions, deliberate adaptations,
␈↓ ↓H␈↓given kind.  Thus  we may know that a boat ␈↓ εx␈↓and a form of decision-making based on
␈↓ ↓H␈↓has a leak and lacks oars, and we may      ␈↓ εx␈↓dialectical argumentation.
␈↓ ↓H␈↓conjecture (non-monotonically) that these
␈↓ ↓H␈↓are the only "things" wrong with the boat. ␈↓ εx␈↓        A program based on this architecture
                                           ␈↓ εx␈↓would inspect itself, describe aspects of
␈↓ ↓H␈↓        Argument, whether with another     ␈↓ εx␈↓itself to itself, and use this self-reference
␈↓ ↓H␈↓person or with oneself, often involves     ␈↓ εx␈↓and these self-descriptions in making
␈↓ ↓H␈↓finding reasons permitting such non-       ␈↓ εx␈↓decisons and taking actions.  The
␈↓ ↓H␈↓monotonically obtained conclusions and     ␈↓ εx␈↓program's mental life would include
␈↓ ↓H␈↓then finding counter arguments.  In the    ␈↓ εx␈↓awareness of its own concepts, beliefs,
␈↓ ↓H␈↓above examples, a counter argument might   ␈↓ εx␈↓desires, intentions, inferences, actions,
␈↓ ↓H␈↓involve a non-monotonic deduction that     ␈↓ εx␈↓and skills.  All of these are represented by
␈↓ ↓H␈↓there is another group in Wednesday's      ␈↓ εx␈↓self-descriptions in a single sort of
␈↓ ↓H␈↓meeting room or that the boat also has a   ␈↓ εx␈↓language, so that the program has access
␈↓ ↓H␈↓broken rudder.  Counter-counter arguments  ␈↓ εx␈↓to all of these aspects of itself, and can
␈↓ ↓H␈↓may involve asserting the existence of     ␈↓ εx␈↓reason about them in the same terms.
␈↓ ↓H␈↓another room or a plan for fixing the
␈↓ ↓H␈↓rudder.                                    ␈↓ εx␈↓α␈↓ λYObjectives

␈↓ ↓H␈↓        We propose to extend the work of   ␈↓ εx␈↓        During the thirteen month period of
␈↓ ↓H␈↓our Formal Reasoning Group to develop      ␈↓ εx␈↓this addition to work, the studies will
␈↓ ↓H␈↓theories and write programs that will      ␈↓ εx␈↓mainly be conceptual.  This is partly
␈↓ ↓H␈↓decide what to do by reasoning that        ␈↓ εx␈↓because these ideas require additional
␈↓ ↓H␈↓includes introspection and dialectical     ␈↓ εx␈↓theoretical work before they can be
␈↓ ↓H␈↓argumentation.                             ␈↓ εx␈↓embodied in programs and partly because
                                           ␈↓ εx␈↓programs that keep a trace of their own
                                           ␈↓ εx␈↓reasoning processes may require bigger
                                           ␈↓ εx␈↓memories than are currently available for
␈↓ ↓H␈↓αModelling Deliberation, Action and Introspection␈↓ E3


␈↓ ↓H␈↓single processes at Stanford.  In any case,␈↓ εx␈↓α␈↓ λ≥CURRICULUM VITAE
␈↓ ↓H␈↓implementation will be a large enough      ␈↓ εx␈↓α␈↓ 	_of
␈↓ ↓H␈↓project to require very careful planning.  ␈↓ εx␈↓α␈↓ λaJon Doyle

␈↓ ↓H␈↓        By the end of 1981, we will have
␈↓ ↓H␈↓planned a system that will be able to      ␈↓ εx␈↓Upcoming Position:
␈↓ ↓H␈↓reason about its own actions and           ␈↓ εx␈↓        Research Associate
␈↓ ↓H␈↓"thoughts" and carry out internal          ␈↓ εx␈↓        Artificial Intelligence Laboratory
␈↓ ↓H␈↓arguments as well as simpler forms of non- ␈↓ εx␈↓        Stanford University
␈↓ ↓H␈↓monotonic reasoning.
                                           ␈↓ εx␈↓Present Position:
                                           ␈↓ εx␈↓        Scientist
                                           ␈↓ εx␈↓        Artificial Intelligence Laboratory
␈↓ ↓H␈↓α␈↓ β#References                                 ␈↓ εx␈↓        Massachusetts Institute of
                                           ␈↓ εx␈↓␈↓ πxTechnology
␈↓ ↓H␈↓␈↓αDoyle, J. (1979): "A truth maintainance
␈↓ ↓H␈↓αsystem", ␈↓↓Artificial Intelligence 12␈↓, 231-  ␈↓ εx␈↓Address:
␈↓ ↓H␈↓272                                        ␈↓ εx␈↓        305 Memorial Drive, #612-A
                                           ␈↓ εx␈↓        Cambridge, Massachusetts 02139
␈↓ ↓H␈↓␈↓αDoyle, J. (1980): "A model for
␈↓ ↓H␈↓αdeliberation, action, and introspection"   ␈↓ εx␈↓Telephone: (617) 494-9214
␈↓ ↓H␈↓α␈↓↓PhD Thesis, MIT AI Laboratory TR-581␈↓
                                           ␈↓ εx␈↓Citizenship: United States of America
␈↓ ↓H␈↓␈↓αMcCarthy, J. (1980): "Circumscription -
␈↓ ↓H␈↓αa form of non-monotonic reasoning"         ␈↓ εx␈↓Principal Fields of Professional Interest:
␈↓ ↓H␈↓α␈↓↓Stanford Artificial Intelligence Laboratory␈↓ εx␈↓        Intelligence
␈↓ ↓H␈↓↓Memo AIM-334␈↓                               ␈↓ εx␈↓        Theory of Computation
                                           ␈↓ εx␈↓        Philosophy
␈↓ ↓H␈↓␈↓αMcDermott, Drew and Jon Doyle (1980):      ␈↓ εx␈↓        Logic
␈↓ ↓H␈↓α"Non-monotonic logic I", ␈↓↓Artificial        ␈↓ εx␈↓        Mathematics
␈↓ ↓H␈↓↓Intelligence␈↓, to appear                    ␈↓ εx␈↓        Physics

␈↓ ↓H␈↓␈↓αReiter, Raymond (1980)␈↓: "A logic for       ␈↓ εx␈↓Education:
␈↓ ↓H␈↓default reasoning", ␈↓↓Artificial Intelligence␈↓,
␈↓ ↓H␈↓to appear                                  ␈↓ εx␈↓1971-72   South Texas Junior College,
                                           ␈↓ εx␈↓␈↓ π_Houston, Texas.

                                           ␈↓ εx␈↓1972-74   University of Houston, B.S. in
                                           ␈↓ εx␈↓␈↓ π_Mathematics, December 1974.  Senior
                                           ␈↓ εx␈↓␈↓ π_Honors Thesis under Prof. J. A. Schatz on
                                           ␈↓ εx␈↓␈↓ π_"Computational Investigations of Non-
                                           ␈↓ εx␈↓␈↓ π_Repetitive Sequences."

                                           ␈↓ εx␈↓1975-77   M.I.T., S.M. in Electrical
                                           ␈↓ εx␈↓␈↓ π_Engineering and Computer Science, June
                                           ␈↓ εx␈↓␈↓ π_1977.  Thesis under Prof. G. J. Sussman
                                           ␈↓ εx␈↓␈↓ π_on "Truth Maintenance Systems for
                                           ␈↓ εx␈↓␈↓ π_Problem Solving."

                                           ␈↓ εx␈↓1977-80   M.I.T., Ph.D. in Artificial
␈↓ ↓H␈↓αModelling Deliberation, Action and Introspection␈↓ E4


␈↓ ↓H␈↓␈↓ ↓hIntelligence, June 1980.  Thesis under     ␈↓ εx␈↓        MIT Symphony Orchestra (1977-78)
␈↓ ↓H␈↓␈↓ ↓hProf. G. J. Sussman on "A Model for
␈↓ ↓H␈↓␈↓ ↓hDeliberation, Action, and Introspection."  ␈↓ εx␈↓Offices:
␈↓ ↓H␈↓␈↓ ↓hProfs. M. Minsky, P. Szolovits, and D.     ␈↓ εx␈↓        Director of Univ. of Houston chapter
␈↓ ↓H␈↓␈↓ ↓hMcDermott (Yale), readers.                 ␈↓ εx␈↓␈↓ πxof Pi Mu Epsilon (1974)
                                           ␈↓ εx␈↓        MIT Ashdown House Executive
␈↓ ↓H␈↓History of Employment:                     ␈↓ εx␈↓␈↓ πxCommittee (1978-79)
                                           ␈↓ εx␈↓        Associate Editor of ACM SIGART
␈↓ ↓H␈↓May 1974 - August 1974  Symbiotics         ␈↓ εx␈↓␈↓ πxNewsletter (1978-79)
␈↓ ↓H␈↓␈↓ ↓hInternational, Inc., Houston, Texas
␈↓ ↓H␈↓␈↓ ↓hDevelopment and maintenance of             ␈↓ εx␈↓␈↓αPublications:␈↓
␈↓ ↓H␈↓␈↓ ↓hbusiness accounting programs.
                                           ␈↓ εx␈↓Papers in Refereed Journals:
␈↓ ↓H␈↓January 1975 - July 1975  Shell Oil
␈↓ ↓H␈↓␈↓ ↓hCompany, Houston, Texas Development        ␈↓ εx␈↓1. Jon Doyle and Ronald L. Rivest, "Linear
␈↓ ↓H␈↓␈↓ ↓hof a user environment for geophysical      ␈↓ εx␈↓␈↓ π_Expected Time of a Simple Union-Find
␈↓ ↓H␈↓␈↓ ↓hprocessing in the Technical Computing      ␈↓ εx␈↓␈↓ π_Algorithm", ␈↓↓Information Processing
␈↓ ↓H␈↓␈↓ ↓hDivision.                                  ␈↓ εx␈↓↓␈↓ π_Letters␈↓, Volume 5, Number 5, (November
                                           ␈↓ εx␈↓␈↓ π_1976), pp. 146-148.
␈↓ ↓H␈↓Awards Received:
␈↓ ↓H␈↓        1975-1980 Fannie and John Hertz    ␈↓ εx␈↓2. Jon Doyle, "A Truth Maintenance
␈↓ ↓H␈↓␈↓ αHFoundation Graduate Fellowship             ␈↓ εx␈↓␈↓ π_System" ␈↓↓Artificial Intelligence 12␈↓
␈↓ ↓H␈↓        1975 NSF Honorable Mention         ␈↓ εx␈↓␈↓ π_(1979), 231-272.   Also MIT AI Lab
␈↓ ↓H␈↓        1974 Summa Cum Laude, University   ␈↓ εx␈↓␈↓ π_Memo 521, June 1979.
␈↓ ↓H␈↓␈↓ αHof Houston
␈↓ ↓H␈↓        1974 Honors in Mathematics,        ␈↓ εx␈↓3. Drew McDermott and Jon Doyle, "Non-
␈↓ ↓H␈↓␈↓ αHUniversity of Houston                      ␈↓ εx␈↓␈↓ π_Monotonic Logic I", to appear in
␈↓ ↓H␈↓        1974 Honors Program, University of ␈↓ εx␈↓␈↓ π_␈↓↓Artificial Intelligence␈↓, 1980.  Also MIT AI
␈↓ ↓H␈↓␈↓ αHHouston                                    ␈↓ εx␈↓␈↓ π_Lab Memo 468, August 1978.  Abstract
␈↓ ↓H␈↓        1974 First prize Mathematics       ␈↓ εx␈↓␈↓ π_in ␈↓↓Notices of the AMS␈↓, V. 26, No. 1 (Jan.
␈↓ ↓H␈↓␈↓ αHContest, University of Houston             ␈↓ εx␈↓␈↓ π_1979), #79T-E4, p. A-16.
␈↓ ↓H␈↓        1973 Third prize Mathematics
␈↓ ↓H␈↓␈↓ αHContest, University of Houston             ␈↓ εx␈↓Papers in Other Journals
␈↓ ↓H␈↓        1973 Outstanding First-year Student
␈↓ ↓H␈↓␈↓ αHof Russian, University of Houston          ␈↓ εx␈↓1. Jon Doyle and Philip London,  "A
                                           ␈↓ εx␈↓␈↓ π_Selected Descriptor-Indexed
␈↓ ↓H␈↓Current Organization Memberships:          ␈↓ εx␈↓␈↓ π_Bibliography to the Literature on Belief
␈↓ ↓H␈↓        American Association for the       ␈↓ εx␈↓␈↓ π_Revision",  ␈↓↓ACM SIGART Newsletter␈↓, No.
␈↓ ↓H␈↓␈↓ αHAdvancement of Science                     ␈↓ εx␈↓␈↓ π_71, 7-23, 1980.   Also MIT AI Lab Memo
␈↓ ↓H␈↓        American Mathematical Society      ␈↓ εx␈↓␈↓ π_568, February 1980.
␈↓ ↓H␈↓        Association for Computing Machinery
␈↓ ↓H␈↓        Boston Museum of Fine Arts         ␈↓ εx␈↓2. Jon Doyle  "Historical Annotations and
␈↓ ↓H␈↓        Mathematical Association of America␈↓ εx␈↓␈↓ π_Humble Databases",  to appear in ␈↓↓ACM
␈↓ ↓H␈↓        MIT Musical Theater Guild          ␈↓ εx␈↓↓␈↓ π_SIGMOD Record␈↓.
␈↓ ↓H␈↓        Sigma Xi
                                           ␈↓ εx␈↓Proceedings of Refereed Conferences:
␈↓ ↓H␈↓Sigart, Pi Mu Epsilon, Phi Kappa Phi, Omega
␈↓ ↓H␈↓␈↓ αHDelta Kappa                                ␈↓ εx␈↓1. Johan de Kleer, Jon Doyle, Guy L. Steele
                                           ␈↓ εx␈↓␈↓ π_Jr. and Gerald Jay Sussman,  "AMORD:
␈↓ ↓H␈↓Past Organization Memberships:             ␈↓ εx␈↓␈↓ π_Explicit Control of Reasoning",  ␈↓↓Proc.
␈↓ ↓H␈↓        MIT Choral Society (1979)
␈↓ ↓H␈↓αModelling Deliberation, Action and Introspection␈↓ E5


␈↓ ↓H␈↓↓␈↓ ↓hACM Conference on AI and Programming       ␈↓ εx␈↓1. Johan de Kleer, Jon Doyle, Guy L. Steele
␈↓ ↓H␈↓↓␈↓ ↓hLanguages␈↓, Rochester, New York, August     ␈↓ εx␈↓␈↓ π_Jr. and Gerald Jay Sussman,  "Explicit
␈↓ ↓H␈↓␈↓ ↓h1977.   Also MIT AI Lab Memo 427           ␈↓ εx␈↓␈↓ π_Control of Reasoning",  ␈↓↓Artificial
␈↓ ↓H␈↓␈↓ ↓h("Explicit Control of Reasoning"), June    ␈↓ εx␈↓↓␈↓ π_Intelligence: An MIT Perspective␈↓,  P. H.
␈↓ ↓H␈↓␈↓ ↓h1977.                                      ␈↓ εx␈↓␈↓ π_Winston and R. H. Brown, editors,  MIT
                                           ␈↓ εx␈↓␈↓ π_Press series in Artificial Intelligence,
␈↓ ↓H␈↓2. Jon Doyle,  "Truth Maintenance Systems  ␈↓ εx␈↓␈↓ π_Cambridge, 1979.
␈↓ ↓H␈↓␈↓ ↓hfor Problem Solving",  ␈↓↓Proc. Fifth
␈↓ ↓H␈↓↓␈↓ ↓hInternational Joint Conference on          ␈↓ εx␈↓2. Jon Doyle,  "A Glimpse of Truth
␈↓ ↓H␈↓↓␈↓ ↓hArtificial Intelligence␈↓, Cambridge,        ␈↓ εx␈↓␈↓ π_Maintenance",  ␈↓↓Artificial Intelligence: An
␈↓ ↓H␈↓␈↓ ↓hMassachusetts, August 1977.   Also MIT     ␈↓ εx␈↓↓␈↓ π_MIT Perspective␈↓,  P. H. Winston and R. H.
␈↓ ↓H␈↓␈↓ ↓hAI Lab TR-419, January 1978.               ␈↓ εx␈↓␈↓ π_Brown, editors,  MIT Press series in
                                           ␈↓ εx␈↓␈↓ π_Artificial Intelligence, Cambridge, 1979.
␈↓ ↓H␈↓3. Jon Doyle, "A Glimpse of Truth
␈↓ ↓H␈↓␈↓ ↓hMaintenance",  ␈↓↓Proc. Fourth Workshop on    ␈↓ εx␈↓Internal Reports:
␈↓ ↓H␈↓↓␈↓ ↓hAutomated Deduction␈↓, Austin, Texas,
␈↓ ↓H␈↓␈↓ ↓hFebruary 1979.   Also ␈↓↓Proc. Sixth          ␈↓ εx␈↓1. Jon Doyle,  "Analysis by Propagation of
␈↓ ↓H␈↓↓␈↓ ↓hInternational Joint Conference on          ␈↓ εx␈↓␈↓ π_Constraints in Elementary Geometry
␈↓ ↓H␈↓↓␈↓ ↓hArtificial Intelligence␈↓, Tokyo, Japan,     ␈↓ εx␈↓␈↓ π_Problem Solving",  MIT AI Lab WP-108,
␈↓ ↓H␈↓␈↓ ↓hAugust 1979.   Also MIT AI Lab Memo        ␈↓ εx␈↓␈↓ π_June 1976.
␈↓ ↓H␈↓␈↓ ↓h461, February 1978.
                                           ␈↓ εx␈↓2. Jon Doyle,  "The Use of Dependency
␈↓ ↓H␈↓4. Drew McDermott and Jon Doyle,  "Non-    ␈↓ εx␈↓␈↓ π_Relationships in the Control of
␈↓ ↓H␈↓␈↓ ↓hMonotonic Logic I (extended abstract)",    ␈↓ εx␈↓␈↓ π_Reasoning",  MIT AI Lab WP-133,
␈↓ ↓H␈↓␈↓ ↓h␈↓↓Proc. Fourth Workshop on Automated         ␈↓ εx␈↓␈↓ π_November 1976.
␈↓ ↓H␈↓↓␈↓ ↓hDeduction␈↓, Austin, Texas, February
␈↓ ↓H␈↓␈↓ ↓h1979.                                      ␈↓ εx␈↓3. Jon Doyle,  "Hierarchy in Knowledge
                                           ␈↓ εx␈↓␈↓ π_Representations",  MIT AI Lab WP-159,
␈↓ ↓H␈↓5. Drew McDermott and Jon Doyle,  "An      ␈↓ εx␈↓␈↓ π_November 1977.
␈↓ ↓H␈↓␈↓ ↓hIntroduction to Non-Monotonic Logic",
␈↓ ↓H␈↓␈↓ ↓h␈↓↓Proc. Sixth International Joint            ␈↓ εx␈↓4. Johan de Kleer, Jon Doyle, Charles Rich,
␈↓ ↓H␈↓↓␈↓ ↓hConference on Artificial Intelligence␈↓,     ␈↓ εx␈↓␈↓ π_Guy L. Steele Jr., and Gerald Jay
␈↓ ↓H␈↓␈↓ ↓hTokyo, Japan, August 1979.                 ␈↓ εx␈↓␈↓ π_Sussman,  "AMORD: A Deductive
                                           ␈↓ εx␈↓␈↓ π_Procedure System",  MIT AI Lab Memo
␈↓ ↓H␈↓Unrefereed Conferences:                    ␈↓ εx␈↓␈↓ π_435, January 1978.

␈↓ ↓H␈↓1. Jon Doyle,  "Non-Repetitive Binary      ␈↓ εx␈↓Invited Lectures:
␈↓ ↓H␈↓␈↓ ↓hSequences",  727th Meeting of the
␈↓ ↓H␈↓␈↓ ↓hAmerican Mathematical Society,             ␈↓ εx␈↓1. "What if the Mayans had Escalators",
␈↓ ↓H␈↓␈↓ ↓hCambridge, Massachusetts, October,         ␈↓ εx␈↓␈↓ π_Yale University, Computer Science
␈↓ ↓H␈↓␈↓ ↓h1975.   Abstract in ␈↓↓Notices of the AMS␈↓,    ␈↓ εx␈↓␈↓ π_Department, June 23, 1978.
␈↓ ↓H␈↓␈↓ ↓hOct. 1975, p. A-660, #727-A5.
␈↓ ↓H␈↓␈↓ ↓hReferee recommended but scooped for        ␈↓ εx␈↓2. "Reflexive Interpreters: A Model for
␈↓ ↓H␈↓␈↓ ↓h␈↓↓J. Combinatorial Theory A␈↓.  My theorems    ␈↓ εx␈↓␈↓ π_Deliberation, Action, and Introspection",
␈↓ ↓H␈↓␈↓ ↓h1 and 2 paraphrase theorems 1 and 2 of     ␈↓ εx␈↓␈↓ π_University of Pennsylvania, Computer
␈↓ ↓H␈↓␈↓ ↓hF. M. Dekking, "On Repetitions of Blocks   ␈↓ εx␈↓␈↓ π_Science Department, October 30, 1979.
␈↓ ↓H␈↓␈↓ ↓hin Binary Sequences," ␈↓↓J. C. T. A 20␈↓ #3
␈↓ ↓H␈↓␈↓ ↓h(May 1976) 292-299.                        ␈↓ εx␈↓3. "Reasoned Deliberation and Decision-
                                           ␈↓ εx␈↓␈↓ π_Making"  SRI International, January 14,
␈↓ ↓H␈↓Articles in Books:                         ␈↓ εx␈↓␈↓ π_1980.
␈↓ ↓H␈↓αModelling Deliberation, Action and Introspection␈↓ E6


␈↓ ↓H␈↓4. "Reasoned Deliberation and Decision-
␈↓ ↓H␈↓␈↓ ↓hMaking"  Hewlett-Packard, Inc., January
␈↓ ↓H␈↓␈↓ ↓h15, 1980.

␈↓ ↓H␈↓5. "Reasoned Deliberation and Decision-
␈↓ ↓H␈↓␈↓ ↓hMaking"  Stanford University, Computer
␈↓ ↓H␈↓␈↓ ↓hScience Department, January 16, 1980.

␈↓ ↓H␈↓6. "Reasoned Deliberation and Decision-
␈↓ ↓H␈↓␈↓ ↓hMaking"  Xerox Palo Alto Research
␈↓ ↓H␈↓␈↓ ↓hCenter, January 17, 1980.

␈↓ ↓H␈↓7. "Reasoned Deliberation and Decision-
␈↓ ↓H␈↓␈↓ ↓hMaking"  USC Information Sciences
␈↓ ↓H␈↓␈↓ ↓hInstitute, January 24, 1980.

␈↓ ↓H␈↓8. "Reasoned Deliberation and Decision-
␈↓ ↓H␈↓␈↓ ↓hMaking"  Bell Laboratories, February 14,
␈↓ ↓H␈↓␈↓ ↓h1980.

␈↓ ↓H␈↓Personal Background and Interests:

␈↓ ↓H␈↓Born and raised in Houston, Texas and, in
␈↓ ↓H␈↓summers, Dundee, Wisconsin, by Leo M.
␈↓ ↓H␈↓Doyle and Marilyn C. Doyle.  Unmarried.
␈↓ ↓H␈↓Interests in people, writing, literature,
␈↓ ↓H␈↓history, economics, musical and poetical
␈↓ ↓H␈↓composition, conducting, viola, recorders,
␈↓ ↓H␈↓painting, sculpture, technology, swimming
␈↓ ↓H␈↓and other sports.
␈↓ ↓H␈↓αModelling Deliberation, Action and Introspection␈↓ E7


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

␈↓ ↓H␈↓α␈↓ β≡Professional
␈↓ ↓H␈↓Doyle, Jon ␈↓ ¬7 27,952
␈↓ ↓H␈↓ ␈↓ ¬≤----------
␈↓ ↓H␈↓   Salary total ␈↓ ¬7 27,952
␈↓ ↓H␈↓Staff benefits (21.15% of salaries) ␈↓ ε(
␈↓ ↓H␈↓ ∧t␈↓␈↓ α(5,912
␈↓ ↓H␈↓Travel ␈↓ ¬@  1,173
␈↓ ↓H␈↓Computer cost for 1 person(s) ␈↓ ¬@  2,002
␈↓ ↓H␈↓Other direct expenses ␈↓ ¬@  2,236
␈↓ ↓H␈↓ ␈↓ ¬≤----------
␈↓ ↓H␈↓   Total direct costs ␈↓ ¬7 39,275
␈↓ ↓H␈↓Indirect costs (58% of above) ␈↓ ¬7 22,780
␈↓ ↓H␈↓ ␈↓ ¬≤----------
␈↓ ↓H␈↓   Project total ␈↓ ¬7 62,055