nl-kr-request@CS.ROCHESTER.EDU (10/05/88)
NL-KR Digest (10/04/88 20:38:35) Volume 5 Number 16 Today's Topics: Systems Engineering Level in KBS state and change/continuous actions common sense knowledge of continuous action Model-based Reasoning - References? NL interfaces to Rule Based Expert Systems (Seek info. & address) Submissions: NL-KR@CS.ROCHESTER.EDU Requests, policy: NL-KR-REQUEST@CS.ROCHESTER.EDU ---------------------------------------------------------------------- Date: Fri, 16 Sep 88 15:44 EDT From: James P. Davis <jdavis@gollum.UUCP> Subject: Systems Engineering Level in KBS I am looking for any pointers to references regarding the "systems engineering level" of knowledge as defined for knowledge base management systems (KBMS). The only reference I have is in Brodie et al. *On Knowledge Based Management Systems*, where Brachman and Levesque discuss the various levels associated with knowledge representation and knowledge systems (knowledge level, symbol level, organization level). They mention this Systems Engineering level in passing, but do not fully define it. Does anyone have any references, or is anyone doing work in this area of further defining these "levels" (Ron and Hector, are you out there)? The nature of my research in this area involves the definition of a "level" which allows structure and organization to be imposed on the Universe of Discourse (which doesn't conform to Newell's Knowledge Level, which deals specifically with what can be stated or implied about the world on a functional basis independent of organization or implementation). However, I am looking at this imposition of organization independent of how the knowledge schema is implemented or manipulated to carry out rational behavior (which doesn't confrom to Newell's Symbol Level either, which deals with issues of how rational behavior is realized on a machine, addressing such issues as how to exploit the syntactic properties of a representation technique to effectively produce rational actions, e.g., inheritance in frame systems). The perspective that I am approaching this from is based on the ideas from Database and data modeling involving the construction of an "enterprise model" of a domain, which is primarily a structural description (in some formalism such as any number of deviations of the E-R model which have been researched) that captures domain objects, relationships, and constraints according to some set of model-dependent wff's. This description is a declarative representation of the UoD. What I am looking at is the correlation between this process in database/data modeling and constructing knowledge schemas for a domain in AI. The goal is to define an architecture for the tight coupling of database and knowledge based systems as KBMS'. It seems that some of the work that I am doing at this level between the KNowledge and Symbol Levels (which I call the "Enterprise Level" may be what has been termed the "Systems Engineering" Level. Is this Systems Engineering Level defined sufficiently? Is anyone working on it? Are there references? Anyone want to correspond regarding these levels? Any and all responses are appreciated. jdavis@Gollum.Columbia.NCR.COM Jim Davis Advanced Systems Development NCR Corporation ------------------------------ Date: Fri, 16 Sep 88 17:25 EDT From: Paul Fishwick <fishwick@uflorida.cis.ufl.EDU> Subject: state and change/continuous actions An inquiry into concepts of "state" and "change": In browsing through Genesereth's and Nilsson's recent book "Logical Foundations of Artificial Intelligence," I find it interesting to compare and contrast the concepts described in Chapter 11 - "State and Change" with state/change concepts defined within systems theory and simulation modeling. The authors make the following statement: "Insufficient attention has been paid to the problem of continuous actions." Now, a question that immediately comes to mind is "What problem?" Perhaps, they are referring to the problem of defining semantics for "how humans think about continuous actions." This leads to some interesting questions: 1) Clearly, the vast literature on math modeling is indicative of "how humans think about continuous actions." This knowledge is in a compiled form, and use of this knowledge has served science in an untold number of circumstances. 2) If commonsense knowledge representation is the issue then we might want to ask a fundamental question "Why do we care about representing commonsense knowledge about continuous actions?" I can see 2 possible goals: One goal is to validate some given theory of commonsense "continuous action" knowledge against actual psychological data. Then we could say, for instance, that Theory XYZ reflects human thought and is therefore useful. I don't think it would be useful to increase our knowledge of mechanics or fluidics, for instance, but perhaps a psycho-therapist might find this knowledge useful. A second goal is to obtain a better model of the continuous action (this reflects the "AI is an approach to problem solving" method where one can study "how Johnny reasons when balls are bounced" and obtain a scientifically superior model regardless of its actual psychological validity). Has anyone seen a commonsense model of continuous action that is an improvement over systems of differential equations, graph based queueing models (and other assorted formal languages for systems and simulation)? Obviously, I'm trying to spark some inter-group discussion and so I hope that any responses will post to both the AI group (comp.ai) AND the SIMULATION group (comp.simulation). In addition (sci.math) and (comp.theory.dynamic-sys) may be appropriate. I believe that Genesereth and Nilsson are quite correct that "reasoning about time and continous actions" is an important issue. However, an even more important issue revolves around people discussing concepts about "state," "time," and "change" by crossing disciplines. Any thoughts? -paul +------------------------------------------------------------------------+ | Prof. Paul A. Fishwick.... INTERNET: fishwick@bikini.cis.ufl.edu | | Dept. of Computer Science. UUCP: gatech!uflorida!fishwick | | Univ. of Florida.......... PHONE: (904)-335-8036 | | Bldg. CSE, Room 301....... FAX is available | | Gainesville, FL 32611..... | +------------------------------------------------------------------------+ ------------------------------ Date: Sat, 17 Sep 88 12:14 EDT From: Greg Lee <lee@uhccux.uhcc.hawaii.edu> Subject: Re: state and change/continuous actions From article <18249@uflorida.cis.ufl.EDU>, by fishwick@uflorida.cis.ufl.EDU (Paul Fishwick): " " 2) If commonsense knowledge representation is the issue then we " might want to ask a fundamental question "Why do we care about " representing commonsense knowledge about continuous actions?" " I can see 2 possible goals: One goal is to validate some given " ... To reason about continuous actions where the physics hasn't been worked out or is computationally infeasible. How about that as a third goal? " Obviously, I'm trying to spark some inter-group discussion and so I hope " that any responses will post to both the AI group (comp.ai) AND " the SIMULATION group (comp.simulation). In addition (sci.math) and " (comp.theory.dynamic-sys) may be appropriate. Tsk, tsk. Left out sci.lang. The way people think about these things is reflected in the tense/aspect systems of natural languages. " I believe that Genesereth and Nilsson are quite correct that "reasoning " about time and continous actions" is an important issue. However, an " even more important issue revolves around people discussing " concepts about "state," "time," and "change" by crossing disciplines. " Any thoughts? In English, predicates which can occur with Agent subjects, those capable of deliberate action, can also occur in the progressive aspect, expressing continuous action. This suggests some connection between intent and continuity whose nature is not obvious, to me anyway. Greg, lee@uhccux.uhcc.hawaii.edu ------------------------------ Date: Sun, 18 Sep 88 21:18 EDT From: Steven Ryan <smryan@garth.UUCP> Subject: Re: state and change/continuous actions >Foundations of Artificial Intelligence," I find it interesting to >compare and contrast the concepts described in Chapter 11 - "State >and Change" with state/change concepts defined within systems >theory and simulation modeling. The authors make the following statement: >"Insufficient attention has been paid to the problem of continuous >actions." Now, a question that immediately comes to mind is "What problem?" Presumably, they are referring to that formal systems are strictly discrete and finite. This has to do to with `effective computation.' Discrete systems can be explained in such simple terms that is always clear exactly what is being done. Continuous systems are computably using calculus, but is this `effective computation?' Calculus uses a number of existent theorems which prove some point or set exists, but provide no method to effectively compute the value. Or is knowing the value exists sufficient because, after all, we can map the real line into a bounded interval which can be traversed in finite time? It is not clear that all natural phenomon can be modelled on the discrete and finite digital computer. If not, what computer could we use? >Any thoughts? ------------------------------ Date: Mon, 19 Sep 88 11:18 EDT From: John McCarthy <JMC@SAIL.Stanford.EDU> Subject: common sense knowledge of continuous action If Genesereth and Nilsson didn't give an example to illustrate why differential equations aren't enough, they should have. The example I like to give when I lecture is that of spilling the water glass on the lectern. If the front row is very close, it might get wet, but usually not even that. The Navier-Stokes equations govern the flow of the spilled water but are entirely useless in this common sense situation. No-one can acquire the initial conditions or integrate the equations sufficiently rapidly. Moreover, absorbtion of water by the materials it flows over is probably a strong enough effect, so that more than the Navier-Stokes equations would be necessary. Thus there is no "scientific theory" involving differential equations, queuing theory, etc. that can be used by a robot to determine what can be expected when a glass of water is spilled, given what information is actually available to an observer. To use the terminology of my 1969 paper with Pat Hayes, the differential equations don't form an epistemologically adequate model of the phenomenon, i.e. a model that uses the information actually available. While some people are interested in modelling human performance as an aspect of psychology, my interest is artificial intelligence. There is no conflict with science. What we need is a scientific theory that can use the information available to a robot with human opportunities to observe and do as well as a human in predicting what will happen. Thus our goal is a scientific common sense. The Navier-Stokes equations are important in (1) the design of airplane wings, (2) in the derivation of general inequalities, some of which might even be translatable into terms common sense can use. For example, the Bernouilli effect, once a person has (usually with difficulty) integrated it into his common sense knowledge can be useful for qualitatively predicting the effects of winds flowing over a house. Finally, the Navier Stokes equations are imbedded in a framework of common sense knowledge and reasoning that determine the conditions under which they are applied to the design of airplane wings, etc. ------------------------------ Date: Mon, 19 Sep 88 11:38 EDT From: Paul Fishwick <fishwick@uflorida.cis.ufl.EDU> Subject: Re: common sense knowledge of continuous actions I very much appreciate Prof. McCarthy's response and would like to comment. The "water glass on the lectern" example is a good one for commonsense reasoning; however, let's further examine this scenario. First, if we wanted a highly accurate model of water flow then we would probably use flow equations (such as the NS equations) possibly combined with projectile modeling. Note also that a lumped model of the detailed math model may reduce complexity and provide an answer for us. We have not seen specific work in this area since spilt water in a room is of little scientific value to most researchers. Please note that I am not trying to be facetious -- I am just trying to point out that *if* the goal is "to solve the problem of predicting the result of continuous actions" then math models (and not commonsense models) are the method of choice. Note that the math model need not be limited to a single set of PDE's. Also, the math model can be an abstract "lumped model" with less complexity. The general method of simulation incorporates combined continuous and discrete methods to solve all kinds of physical problems. For instance, one needs to use notions of probability (that a water will make it to the front row), simplified flow equations, and projectile motion. Also, solving of the "problem of what happens to the water" need not involve flow equations. Witness, for instance, the work of Toffoli and Wolfram where cellular automata may be used "as an alternative to" differential equations. Also, the problem may be solved using visual pattern matching - it is quite likely that humans "reason" about "what will happen" to spilt liquids using associative database methods (the neural netlanders might like this approach) based on a huge library of partial images from previous experience (note Kosslyn's work). I still haven't mentioned anything about artificial intelligence yet - just methods of problem solving. I agree that differential equations by themselves do not comprise an epistemologically adequate model. But note that no complex problem is solved using only one model language (such as DE's). The use of simulation is a nice example since, in simulating a complex system, one might use many "languages" to solve the problem. Therefore, I'm not sure that epistemological adequacy is the issue. The issue is, instead, to solve the problem by whatever methods available. Now, back to AI. I agree that "there is no theory involving DE's (etc.) that can be used by a robot to determine what can be expected when a glass of water is spilled." I would like to take the stronger position that searching for such a singular theory seems futile. Certainly, robots of the future will need to reason about the world and about moving liquids; however, we can program robots to use pattern matching and whatever else is necesssary to "solve the problem." I supposed that I am predisposed to an engineering philosophy that would suggest research into a method to allow robots to perform pattern recognition and equation solving to answer questions about the real world. I see no evidence of a specific theory that will represent the "intelligence" of the robot. I see only a plethora of problem solving tools that can be used to make future robots more and more adaptive to their environments. If commonsense theories are to be useful then they must be validated. Against what? Well, these theories could be used to build programs that can be placed inside working robots. Those robots that performed better (according to some statistical criterion) would validate respective theories used to program them. One must either 1) validate against real world data [the cornerstone to the method of computer simulation] , or 2) improved performance. Do commonsense theories have anything to say about these two "yardsticks?" Note that there are many AI research efforts that have addressed validation - expert systems such as MYCIN correctly answered "more and more" diagnoses as the program was improved. The yardstick for MYCIN is therefore a statistical measure of validity. My hat is off to the MYCIN team for proving the efficacy of their methods. Expert systems are indeed a success. Chess programs have a simple yardstick - their USCF or FIDE rating. This concentration of yardsticks and method of validation is not only helpful, it is essential to demonstrate the an AI method is useful. -paul +------------------------------------------------------------------------+ | Prof. Paul A. Fishwick.... INTERNET: fishwick@bikini.cis.ufl.edu | | Dept. of Computer Science. UUCP: gatech!uflorida!fishwick | | Univ. of Florida.......... PHONE: (904)-335-8036 | | Bldg. CSE, Room 301....... FAX is available | | Gainesville, FL 32611..... | +------------------------------------------------------------------------+ ------------------------------ Date: Sun, 18 Sep 88 20:55 EDT From: James P. Davis <jdavis@gollum.UUCP> Subject: Model-based Reasoning - References? I am looking for some good references on the subject of Model-based reasoning (MBR). I am also interested in finding out who is doing work/research in this area, and what domains are being investigated. Nobody seems to have put any special compendiums (like Morgan Kaufmann) in this area yet. Any of you out there? Specifically, I am looking at the area of using a modeling framework, which allows the structure and behavior for certain classes of domains to be expressed in some declarative form, to drive the reasoning process. My understanding of MBR is that it is an approach at exploiting the inherent structure and constraints of a system or enterprise to guide the process of reasoning about problems in the given domain. I am developing an "analogical" representation which allows the expression of domain semantics in terms of structure and constraint declaration constructs based on the syntactic construction of wff's in the modeling technique. The domain is information systems design. In theory, by developing a self-describing modeling formalism, in which the information systems design activity can take place, the nature of the solution space can be constrained such that only those solutions which adhere to the semantics of the formalism itself (in which are expressed the semantics of the domain application) are relevant. What's happening in MBR? How does it relate to "reasoning from first principles"? Any and all responses are appreciated. I can summarize to the net if requested. Jim Davis Advanced Systems Development NCR Corporation jdavis@Gollum.Columbia.NCR.COM ------------------------------ Date: Mon, 19 Sep 88 16:31 EDT From: Fabrizio Sebastiani <FABRIZIO%ICNUCEVM.BITNET@ICNUCEVM.CNUCE.CNR.IT> I am looking for papers on hybrid knowledge representation (MRS, KLONE, KRYPTON and the like); I am pretty familiar with the "KLONE world" literature (at least, with what has gone on up to 1985), but don't know much about: 1) what has been written past that date; 2) what has been written AGAINST this approach. Can anyone provide references to relevant papers on the subject? Is anyone interested to discuss the issue? Thanks Fabrizio Sebastiani ------------------------------ Date: Mon, 19 Sep 88 20:24 EDT From: ERIC Y.H. TSUI <munnari!aragorn.oz.au!eric@uunet.UU.NET> Subject: NL interfaces to Rule Based Expert Systems (Seek info. & address) I recently broadcasted and seek information on NL interfaces to rule based expert systems. There is no reply and I came across the following article: DATSKOVSKY-MOERDLER, G., McKEOWN, K.R. and ENSOR, J.R. (1987); Building Natural Language Interfaces for Rule-based Systems, IJCAI-87, p682-687. The first two authors are from Columbia University (NY) and the third author is from AT&T Bell Lab. (Holmdel, N.J.). Would anyone have their e-mail address ? (I am still interested to learn about pointers to other work.) Eric Tsui eric@aragorn.oz Division of Computing and Mathematics Deakin University Geelong, Victoria 3217 Australia ------------------------------ Date: Fri, 23 Sep 88 19:35 EDT From: Fabrizio Sebastiani <FABRIZIO%ICNUCEVM.BITNET@ICNUCEVM.CNUCE.CNR.IT> Does anybody know whether further studies have been carried out on Fagin and Halpern's notion of "awareness" in epistemic logics, as from their 1985 IJCAI paper? whether the notion had been previously discussed in the philosophy of language or the philosophy of mind? Anyone wishing to discuss the topic, provide references, send papers, etc., is invited to contact me. Fabrizio Sebastiani ------------------------------ End of NL-KR Digest *******************