Lenat, D. B. 1990. CYC: Toward programs with common sense. Communications of the ACM, 33(8), 30--49.  Home/Search   Document Not in Database   Summary   ACM   TOC   Related Articles   Check

....51 11 Conclusions and Future Work 53 1 Introduction Artificial intelligence research falls roughly into two categories: formal and implementational. This division is not completely firm: there are implementational studies based on (formal or informal) theories (e.g. CYC [ Lenat and Guha, 1990, Lenat et al. 1990, Guha and Lenat, 1990 ] SOAR [ Laird et al. 1987, Rosenbloom et al. 1991 ] OSCAR [ Pollock, 1989, Pollock, 1992 ] and there are theories framed with an eye toward implementability (e.g. predicate circumscription [ McCarthy, 1980 ] Nevertheless, formal theoretical work tends to focus ....

....Work 53 1 Introduction Artificial intelligence research falls roughly into two categories: formal and implementational. This division is not completely firm: there are implementational studies based on (formal or informal) theories (e.g. CYC [ Lenat and Guha, 1990, Lenat et al. 1990, Guha and Lenat, 1990 ] SOAR [ Laird et al. 1987, Rosenbloom et al. 1991 ] OSCAR [ Pollock, 1989, Pollock, 1992 ] and there are theories framed with an eye toward implementability (e.g. predicate circumscription [ McCarthy, 1980 ] Nevertheless, formal theoretical work tends to focus on very narrow ....

[Article contains additional citation context not shown here]

D. Lenat, R. V. Guha, K. Pittman, D. Pratt, and M. Shepherd. Cyc: Toward programs with common sense. Communications of the ACM, 33(8):30--49, August 1990.

....English are used to parse English texts. Semantic and discourse interpretation then produce fully scoped logical forms representing their possible linguistic meaning(s) Domain specific reasoning is then used both to convert the linguistically motivated logical forms to Cyc style representations [12], and also to disambiguate them by ruling out linguistically possible interpretations that are implausible given the domain. Within this setting, an adequate treatment of prevention statements is necessary even for low level information collating tasks, such as matching the objects referred to in ....

....is strikingly close to a knowledge representation that we believe supports tractable inference. It also validates the correct monotonicity entailments and existence commitments, as we will shortly see. 5. A Knowledge Representation with Reified Concepts In implemented reasoning systems like Cyc [12], concepts are first order entities distinct from their instantiations. Basic concepts are reified as part of a static ontology. While non basic concepts need to be constructed dynamically, the existence of a static ontology covering much of the domain substantially facilitates the reasoning ....

Lenat, D., R. V. Guha, K. Pittman, D. Pratt, and M. Shepherd. Cyc: Towards programs with common sense. Communications of the ACM 33(8), August 1990.

....Knowledge Bases: CYC CYC is an attempt to build a massive knowledge base so that one (person or machine) can apply it to some reasoning mechanism. The importance of building very large knowledge bases as one of the steps for developing programs with some sort of intelligence, is well motivated in [40]. In this paper the authors name the lack of very large knowledge bases, covering a wide area of (human) knowledge as one of the reasons why expert systems failed in their attempt to be regarded as intelligent programs [40] Suppose an expert system has the following four rules: IF frog(x) ....

.... developing programs with some sort of intelligence, is well motivated in [40] In this paper the authors name the lack of very large knowledge bases, covering a wide area of (human) knowledge as one of the reasons why expert systems failed in their attempt to be regarded as intelligent programs [40]. Suppose an expert system has the following four rules: IF frog(x) THEN amphibian(x) IF amphibian(x) THEN laysEggsInWater(x) IF laysEggsInWater(x) THEN livesNearLotsOf(x, Water) IF livesNearLotsOf(x, Water) THEN :livesInDesert(x) 33 Given the assertion frog(Freda) those rules could be ....

D. B. Lenat, R. V. Guha, K. Pittman, D. Pratt, and M. Shepherd. Cyc: toward programs with common sense. Communications of the ACM, 33(8):30-49, August 1990.

....are used to parse English sentences. Semantic and discourse interpretation then produces fully scoped logical forms representing their possible linguistic meaning(s) Domain speci c reasoning is then used both to convert the linguistically motivated logical forms to a CYC style representation [4], and also disambiguate by ruling out linguistically possible interpretations that are implausible given the domain. Within this setting, an adequate treatment of prevention statements must address both linguistic and knowledge representation concerns. 1 In the following we present a set of ....

....like prevent in a way more suitable for computational processing required for knowledge representation and reasoning. We propose a rst order representation in which cause and prevent predicates have a concept denoting terms as their second argument. In implemented reasoning systems (e.g. CYC [4]) frequently occurring concepts are rei ed as rst order entities. Even though we still have to allow for constructed concepts to be generated dynamically, the existence of a static ontology that covers much of the domain facilitates the reasoning process substantially. Description logics ( 1] ....

D. Lenat, R. V. Guha, K. Pittman, D. Pratt, and M. Shepherd. Cyc: Towards programs with common sense. Communications of the ACM, 33(8), August 1990.

....a normal (i.e. healthy) liver is typically neither big nor firm. The issues of common sense and its necessary use if we are to build machines that perform human tasks effectively have long been raised and hotly debated. Attempts have been made at compiling large databases of common sense facts [Lenat et al. 90] and formalizing common sense [Lifschitz 91, Hobbs and Moore 85, Davis 90] Most formalizations are variants of first order predicate logic. These schemes are theoretically attractive and expressively powerful, but the machinery is heavy and difficult for the novice to deal with, and few ....

Lenat, D.B., Guha, R.V., Pittman, K., Pratt, D., and Shepherd, M. Cyc: Toward Programs With Common Sense. Communications of the ACM, 33, 8 (Aug. 1990), 30-49.

....defined as the ability that any functional human being has to cope with the world, and to quickly make useful inferences in the presence of incomplete and uncertain information. The lack of such simple mechanisms in AI models has become known as the brittleness bottleneck in the AI community [LEN90, MCC84, SUN92]. Much work has been expended in characterizing common sense [FOR89, LER92, LIF88, SUN92, WOL94] formalizing its representation [DAV90, HOB85, LIF90, ZLA90] constructing logics (or implementations thereof) powerful enough to handle it [BRE91, GOL93, 13 LUK90, REI80] and developing models ....

....[SUN93] However, FLARE is able to handle its protocols either extensionally or intensionally. Most reasoning systems, because they do not have any inductive mechanisms, are confined to the intensional approach. The extensional approach endows the system with more selfadaptivity. Work on Cyc [LEN90, LEN94] is also relevant to the context of the present research as it seeks to capture most of human commonsense knowledge into a large database that can be used by a variety of models such as FLARE. Relevant contexts of Cyc s knowledge base might be instantiated to guide FLARE s learning. 3. FLARE A ....

[Article contains additional citation context not shown here]

Lenat, D.B., et al. (1990). Cyc: Toward Programs With Common Sense. Communications of the ACM, Vol. 33, No. 8, 30-49.

....is about learning meaningful representations. The other conditions say that a reliable indicator relationship must exist and be exploited by an agent for some purpose. Thus, the relationship between mean daily tem # Projects such as Cyc emphasize the denotational meanings of representations [14, 13]. Terms in Cyc are associated with axioms that say what the terms mean. It took a collosal e ort to get enough terms and axioms into Cyc to support the easy acquisition of new terms and axioms. perature (the indicator) and ice cream sales (the indicated) is apt to be meaningful to ice cream ....

D. B. Lenat. Cyc: Towards programs with common sense. ############## ## ### ###, 33(8), 1990.

.... likely interpretation of incomplete or vague requirements, but rather to force the expression of requirements as explicit as possible (to minimize the probability of misunderstandings) an application geared to requirement engineering could dispense with huge knowledge bases like the Cyc ontology [13, 14] or ACAPULCO [11] Interactivity. Requirement writing is an interactive process. The user can repeatedly modify a requirement until it can be parsed by a tool; in our experience, this e ort always makes the requirement clearer and often exposes real problems in the original version. Given the ....

D. B. Lenat, R. V. Guha, K. Pittman, D. Pratt, and M. Shepherd. Cyc: Toward programs with common sense. Communications of the ACM, 33(8), Aug. 1990.

....developed in this text analysis area have been interesting since 1960. Initially, in the 60s and in the 70s, the Author for correspondence. Information Retrieval Systems (IRS) Salton, 1968; Salton, 1989; Cohen Kjeldsen, 1987; Doyle, 1975) emerged. Later, the CYC project (Lenat Guha, 1989; Lenat, Guha, Pittman, Pratt Shepherd, 1990) which was developed during one decade, was proposed. Several works were recently proposed in the Text Mining and knowledge discovering areas (Feldman Hirsh, 1996; Feldman Dagan, 1997; Feldman, Aumann, Fresko, Liphstat, Rosenfeld Schler, 1999; Aumann, Feldman, Yehuda, Landau, Liphstat ....

Lenat D.B., Guha R.V., Pittman K., Pratt D. & Shepherd M. (1990). CYC: Toward programs with common sense. Communications Of The ACM, 33(8), 31-49.

.... likely interpretation of incomplete or vague requirements, but rather to force the expression of requirements as explicit as possible (to minimize the probability of misunderstandings) an application geared to requirement engineering could dispense with huge knowledge bases like the Cyc ontology [22, 23] or ACAPULCO [19] In this case, ignorance is a blessing [6] ffl Interactivity. Requirement drafting is an interactive process. The user can repeatedly modify a requirement until it can be understood by a tool; in our experience, this effort always makes the requirement clearer and often exposes ....

D. B. Lenat, R. V. Guha, K. Pittman, D. Pratt, and M. Shepherd. Cyc: Toward programs with common sense. Communications of the ACM, 33(8), Aug. 1990.

....There will be terabytes of spatial and geographic data, including information on vegetative mapping, land use, demography, seismology, cadastral boundaries, water quality, etc. Although it has been argued that a huge knowledge base is necessary to perform large scale inferential reasoning (Lenat et al. 1990), previous natural language processing systems have not utilized such a volume of information. We believe the large scale of this raw data will make possible the kind of assigned inferential indexing GIPSY is trying to provide. The second advantage provided by Project Sequoia 2000 will be the ....

Lenat, D. B., R. V. Guha, K. Pittman, D. Pratt, & M. Shepard (1990). Cyc: Towards programs with common sense. Communications of the ACM, 33(8):30--49.

....system is unavailable, we will resort to a subject thesaurus with hierarchical relationships. Artificial intelligence representations such as semantic networks, expert systems, or ontologies represent another approach to capturing knowledge, e.g. Lenat s CYC common sense knowledge base [24] [25] [13] Their representations are often richer and more fine grained and the goal of capturing human intelligence is ambitious and difficult. Due to the granularity required of such representations, knowledge creation is slow and painstaking. Only experimental prototypes in small, limited domains ....

D. B. Lenat, R. Guha, K. Pittman, D. Pratt, and M. Shepherd. CYC: Toward programs with common sense. Communications of the ACM, 33(8):30-49, August 1990.

....Assertions and denials are used to add statements to viewpoints, where assertion means asserting some statement is true, while denial means asserting some statement is false. Similarly, retraction can be used to remove any commitment. The scheme can also be compared to that used in the Cyc project [Lenat et al. 1990], although Cyc uses a different definition of deny , as it does not distinguish between axioms and inferences in the knowledge base. In summary, the commitment reasoning scheme attaches one of four values to any statement within a viewpoint, and these values are altered by assertions and ....

Lenat, D. B., Guha, R. V., Pittman, K., Pratt, D., and Shepherd, M., 1990, "Cyc: Towards Programs with Common Sense", Communications of the ACM, Vol 33, No 8, Aug 1990, p30-49.

....training operations, and the use of programmed knowledge as contained, for example, in dictionaries. We believe that learning will need to play a large part of the overall task if we are to achieve the various kinds of robustness discussed in this paper. Purely programmed systems such as CYC [18] would appear to have an inherent limitation in this regard. It is quite likely that if all this is to be achieved, natural language inputs will play a major role. Corpora suitably annotated with the appropriate levels of knowledge will have to be created. Substantial infrastructure will need to ....

....a much smaller role. We are suggesting that the way to evaluate this architecture is by experimentation. In that connection we note here one aspect. Most general approaches to AI attempt to describe a uniform method for building a knowledge base starting from a blank slate. For example in CYC [18] facts and reasoning about the most universal concepts such as time and space are formalized in the same framework as the more specialized knowledge. In the neuroidal framework there is room reserved for treating 24 the universal concepts differently from the others. In particular, the features ....

D.B. Lenat, et al. CYC: Toward programs with common sense. In CACM, volume 33:8, pages 30--49, 1990.

....kinds of inferences are often made while envisioning the design and in explanations about the results of simulations. For example, a designer of the manufacturing equipment in study needed to know that people do not like to be made wet by the machine they are operating all day. The Cyc Project [Lenat Guha, 1990; Lenat, Guha, Pittman, Pratt, Shepherd, 1990] is producing formalisms for representing this kind of knowledge, but it would be hard to imagine that the Cyc KB would be comprehensive enough to cover the space of background knowledge used by designers. The problem of background knowledge, then, ....

....are often made while envisioning the design and in explanations about the results of simulations. For example, a designer of the manufacturing equipment in study needed to know that people do not like to be made wet by the machine they are operating all day. The Cyc Project [Lenat Guha, 1990; Lenat, Guha, Pittman, Pratt, Shepherd, 1990] is producing formalisms for representing this kind of knowledge, but it would be hard to imagine that the Cyc KB would be comprehensive enough to cover the space of background knowledge used by designers. The problem of background knowledge, then, is an inherent limit to the capabilities of ....

Lenat, D. B., Guha, R. V., Pittman, K., Pratt, D., & Shepherd, M. (1990). Cyc: Toward Programs with Common Sense. Communications of the ACM , 33(8):30-49.

.... foundation has been laid, and we are drawing from representation work in design automation [34] decision support [1] design rationale [23] dependency management [27, 31] applied mathematics [13, 35] behavior modelling [6, 10, 30, 36] and the representation of everyday human experience [25]. We plan to develop a basic ontological foundation and then extend it in the directions indicated by how it is used in practice by engineers. The initial shade ontology (based on [17] and an informal study of group ontology development) will include classes for behavioral descriptions, ....

Douglas B. Lenat, Ramanathan V. Guha, Karen Pittman, Dexter Pratt, and Mary Shepherd. Cyc: Toward programs with common sense. Communications of the ACM, 33(8):30--49, 1990.

....in the language. 2 For example, the implementors of the Cyc knowledge base have identified a few tens of special purpose inference mechanisms, such as generalized inheritance. For each inference mechanism, Cyc has a special internal format and Lisp procedures for implementing the reasoning [17]. 3 ffl A mechanism that allows operational use of the ontologies within in a variety of implemented representation systems. ffl A syntax that facilitates the modular definition of terms in an ontology, and the modular packaging of ontologies. ffl A means for capturing conventions in knowledge ....

Douglas B. Lenat, Ramanathan V. Guha, Karen Pittman, Dexter Pratt, and Mary Shepherd. Cyc: Toward programs with common sense. Communications of the ACM, 33(8):30--49, 1990.

No context found.

Lenat, D.B., Guha, R. V., Pittman, K., Pratt, D., and Shepherd, M. CYC: Toward programs with common sense. Commun. ACM (Aug., 1990).

No context found.

Lenat, D. B. 1990. CYC: Toward programs with common sense. Communications of the ACM, 33(8), 30--49.

No context found.

Lenat, Douglas, Ramanathan Guha, Karen Pittman, Dexter Pratt, and Mary Shepherd (1990). "CYC: Toward Programs with Common Sense", in Communications of the ACM (CACM). vol. 33:8.

No context found.

D. B. Lenat et al., "Cyc: Toward Programs with Common Sense," Communications of the ACM, vol. 33, no. 8, pp. 30--49, August 1990.

No context found.

D. B. Lenat et al., "Cyc: Toward Programs with Common Sense," Communications of the ACM, 33(8): 30--49, August 1990.

No context found.

Lenat, D., R. V. Guha, K. Pittman, D. Pratt, and M. Shepherd. Cyc: Towards programs with common sense. Communications of the ACM 33(8), August 1990.

No context found.

Lenat, D. B., R. V. Guha, K. Pittman, D. Pratt and M. Shepherd, Cyc: toward programs with common sense, Comm. ACM, vol. 33(8), 1990.

No context found.

: 32-59. Lenat, D.B, Guha, R.V., Pittman, K., Pratt, D., & Shepherd, M. 1990. Cyc: Toward Programs with Common Sense. Communications of the ACM.

First 50 documents  Next 50

Online articles have much greater impact   More about CiteSeer.IST   Add search form to your site   Submit documents   Feedback  

CiteSeer.IST - Copyright Penn State and NEC