JP7253760B2 - Method, computer program and apparatus for implementing negation in logic programming - Google Patents

Description

Application of Patent Law Article 30, Paragraph 2 ▲１▼ MITOU Advanced Business Project Proposal (Copy) ▲２▼ 71st Awagasaki Business Salon Supplementary materials (distributed on the day) ▲３▼ 71st Awagasaki Business Salon Slides for use (Explanation on the day, distributed to Kanazawa City via Professor Yamada) (4) Kanazawa City CVCK 2017/3/8 presentation slides & Q&A materials uploaded to slide share on 2017/3/9 and posted on Facebook public page linked)

The present invention relates to a method, computer program and apparatus for implementing negation in logic programming.

Traditional PROLOG and its functional subsets are extended to correspond to natural languages and the Internet (for example, our company's "Patent Document 1" below), etc. are known as logical (or declarative) programming languages. and its development/execution environment (application software) is well maintained.
These languages and environments generally adopt an interpretation policy called the "closed-world hypothesis," which regards "propositions that do not exist in memory and cannot be automatically proved by rules as false."
The reason is that it is better to divide the types of logic into two types (binary) of things that can be (automatically) proven and things that cannot be (automatically) proven by combining evidence (facts, facts) and logic (inference rules, rules). Expression of knowledge and contextual state is simple, and solution search and inference processing including pattern matching (unification) and syllogism (automatic proof, recursive) for answering queries can be easily realized (high speed). This is thought to be because it was easy to verify the legitimacy (a certain degree of completeness and soundness).
The so-called "NOT" when this policy is adopted is often called "NOT as failure of proof", and many PROLOG textbooks include "!" ) is shown (about two lines).
However, when it is necessary to exclude ideas such as "science, technology, business, public policy, and ethical improbability" from the solution candidates, it is possible to more accurately and efficiently express real-world situations and problem settings. In many situations where it is necessary, ``NOT as a failure to prove'' is insufficient in its descriptive ability, and an interpretation policy that considers ``a proposition that does not exist in memory and cannot be automatically proved by rules to be false'' is easy to match with this. has the harmful effect of unreasonably restricting the exploration of new ideas (especially ontology (ontology) and pragmatics (especially the synthesis of new concepts and ideas by combining words) that are free for users of natural language). .
In other words, the expressive ability and semantics (meaning by humans and the interpretation method by computers) that are more advanced than the conventional "NOT" can make use of free ontology and pragmatics by natural language especially from users ( It was required from the systems of Patent Documents 1, 2, and 3).

Patent Application No. 2013-087008 Multi-length matching of literals with variables embedded in natural language and management of infinite sets, etc. Patent Application No. 2016-009324 Reverse question from computer to user and addition of dynamic facts (temporary context) Patent Application No. 2017-092047 Addition of Dynamic Facts (Temporary Context) by Forward Inference

In view of such problems, the present invention provides a method of efficiently expressing knowledge and interpreting it by a computer even when it is necessary to exclude "science/technology/business/public order/ethical improbable" ideas from solution candidates. , an apparatus and a computer program.
It also aims to make it easier for users to take advantage of the freedom of natural language through its advanced expressiveness and semantics.

In order to solve such problems, the present invention
Claim 1 of the present invention is
In a question, fact, rule condition or rule effect entered by a user, or in a logic program fact, rule condition or rule effect,
In addition to the affirmation (true) as the existence of evidence facts or the success of automatic proof and its denial (absence of evidence facts and automatic proof failure/impossibility),
Permitting the use of the word "impossible" to imply the existence of evidence to the contrary or the success of automatic proof to the contrary;
When using an affirmative expression that does not include the variable from the beginning or does not include it due to variable binding,
Automatically confirming that the expression "impossible" corresponding to the affirmative expression does not exist and cannot be automatically proved,
To provide a computer program for restricting the use of the corresponding affirmative expression when the "impossible" expression exists or can be automatically proved.
: Here, ``corresponding to an affirmative expression'' means, for example, that the expression ``impossible'' corresponding to the affirmative sentence ``Taro is the eldest son'' is ``!Taro is the eldest son.'' When Taro coexists in a logic program or temporary context, by prioritizing ``! judgment becomes possible.
This allows certain "economically and ethically implausible" ranges to be succinctly expressed and eliminated from the idea, for example the list below.

15: !feed $X $Y and make money :- $X is livestock; $Y is poison;
16: How to dispose of $Y's waste using $X :- Make $X eat $Y ;
17: Make $X eat $Y and make money :- $X eats $Y ;
18: Goats are Livestock
19: Paper with arsenic is poisonous
20: Goats eat paper with rapeseed oil
21: Goats eat arsenic paper. The solution candidate "waste disposal method" is not generated because the ground (sufficient condition) for the solution "make money by feeding goats with arsenic paper" is restricted by line 15. Become.
:Suppression (restriction) of such generation is obvious at the time when the idea becomes a fixed character string by the literal binding process in which the variable is embedded in the natural language in the above-mentioned Patent Document 1, the calculation in a realistic time is clear. Therefore, it is advantageous to search for all logical programs at such timings and suppress/exclude them by, for example, line 15.
: In addition, when ``!'' is uniformly agreed to mean ``impossible'', the relationship between the fact ``Taro is not the eldest son'' and the fact ``!Taro is not the eldest son'' becomes the fact ``Taro is the eldest son'' In fact, it is the same as "! Taro is the eldest son." If it is agreed that the end of the sentence "is not possible" uniformly means "impossible", the above "!" It is possible to mechanically interpret "impossible" (automatic restriction of solution candidates, etc.).
> Constraints, meaning "cannot be" or "not", are extremely important knowledge representations for solving constraint satisfaction problems that express real problems as constraints, and the ability of logic programs to express real problems. will be greatly improved.

As explained above, claim 2 of the present invention is
To provide a computer program characterized by uniformly prioritizing (or uniformly subordinating) a knowledge expression meaning "impossible" to a corresponding affirmative expression.
In the "preferred" case, if you want to prove to the "impossible" side, the search cost for the computer program to check the absence of the affirmative expression is unnecessary. Even if the opposite policy of "subordination" (affirmative expression priority) is adopted, the same cost reduction effect can be obtained.
In addition, in the conditional expression of the rule, conventional NOT (absence of evidence and failure of automatic proof) can be written together with knowledge expression meaning "impossible". is clearly necessary and useful, whether or not exclusion as "is necessary/unnecessary.
In this case, for example, the knowledge expression meaning "impossible" can be (!...), and the conventional knowledge expression meaning NOT can be (...).
This conventional NOT (above, …) in a method that allows literals that embed variables in natural language is the inversion of logic (exists/automatic proof False if it can, and true if it does not exist and cannot be automatically proved), and if the variable exists in the literal, it means that the variable binding that satisfies the condition that can be removed by NOT is excluded from the solution candidates of the rule ( similar to conventional PROLOG).
This exclusion (~) means that the knowledge expression (!), which means "impossible", identifies and detects contradictions in the global scope through the fixed character string that is the basis, and eliminates the path of binding and automatic proof ( Contrary to what we are trying to base on (the fact that there is no variable), we do not detect whether it is inconsistent with the concrete fixed string (the fact that there is no variable), etc. (In terms of implementation, by managing a set of value pairs that must not be bound to the variable pairs on the rules, even if the condition "~$X is a man" becomes "$ This function can be realized even if the set of solution candidates for $X, such as "X is Japanese," appears before the first appearance (usually on the left side of the body).
It should be noted that it is difficult or expensive to express the same knowledge content as the above list only by using this conventional NOT (...) as the condition of the body part. This is because the combination of livestock and poison through the N-argument is viewed as an economic and ethical problem. This is because we must be able to express the nuance that we do not regard the ideas that are meant after the N-stage reasoning as problematic.

In addition, claim 3 of the present invention is
3. A computer program according to claim 1 or 2, characterized in that logic translation is performed with priority given to the most recent user input.
Here, the inversion of logic is classified into the following three types.
1) Conversion from a positive expression to an expression meaning "impossible" and vice versa 2) Conversion adding or deleting an affirmative expression 3) Adding or deleting an expression meaning "impossible" Intention The reason for giving priority to the "latest user input" is to reflect the more specific or the latest recognition facts of the user who knows the site and the situation in the knowledge inference.
Specifically, in line with the latest "user input", in the case of 1), delete or comment out (with reason and timestamp) the expression with the opposite meaning from the memory that stores the logic program or temporary context. be able to. It would be advantageous to have an optional logical program protection mode, in which only the logical program can be suppressed from being deleted or commented out at runtime (temporary context can be changed).
If ``impossible'' is given priority, even if affirmative sentences are left as they are when translating to ``impossible'', the overall logic of the logic program as a whole does not change, but from the point of view of computational speed and readability. It is advantageous to delete or comment out, and if you comment out and add the reason and time stamp as the comment automatically as in the example below, the version management of the logic program and temporary context at the time of inference execution is also become easier.

Example: ///It's nice weather today///2017.5.1_12:21:21 The user entered "!It's nice weather today" and the same fact was added to the temporary context. As a result of giving priority to input, it is confirmed and proven by the reverse question (dynamic question from the computer to the user) in Patent Document 2 and the forward inference (automatic inference by computer) in Patent Document 3 due to user input. It is advantageous to ensure overall consistency, logic, and predictability by treating the fact (temporary context) as similar to the latest user input and prioritizing it over the opposite knowledge representation. is.

Therefore, claim 4 of the present invention is
characterized by deleting the expression of the opposite meaning from the memory storing the logic program or the temporary context when translating into logic,
A computer program according to claim 3 is provided.
This improves the readability of the logic program and temporary context after deletion, and facilitates the human and mechanical organization of comprehensive logic.
Here, the "logical program" means contents such as 15 to 21 in the list above, and the "temporary context" means, in addition to this, the fact that the user expresses a specific problem case. It is what was input and added to the memory (and the fact that the memory was added dynamically (=at the time of execution) in the reverse question and forward inference in Patent Documents 2 and 3) (same throughout the specification).

In addition, claim 5 of the present invention is
Characterized by commenting out the expression of the opposite meaning from the memory that stores the logic program or the temporary context when translating into logic,
A computer program according to claim 3 is provided.
This makes it possible to visualize the process of organizing the logic of the logic program and the temporary context after commenting out, facilitating mechanical arrangement of the general logic and version management (of the logic program and the temporary context) by humans or machines.
In addition, claim 6 of the present invention is
6. The computer program according to claim 5, wherein a time stamp or causal event identification information is automatically added to the comment.

In the above example, the head (effect) of the rule is added with "!" to exclude the solution of poisoning livestock in the global scope. case, the fact that the attached condition is "impossible" exists in the logical program or the temporary context, or the success of the automatic proof that the attached condition is "impossible" (the logic program in memory or the temporary context is a conditional statement that requires
A specific interpretation of "! ) is a simple string match, and its implementation is self-explanatory. On the other hand, with regard to automatic certification, the rule that omits the "!" in the rule with "!" An implementation method such as satisfying the condition if possible may be used. For a method of realizing this proof, please refer to Patent Document 1, which has already been published.
In general, such semantics (computer interpretation) has not been possible with conventional PROLOG, etc., and greatly facilitates the accurate and concise expression and computational solution of real-world problems. It is.

In addition, claim 7 of the present invention is
An apparatus using a computer program according to any one of claims 1 to 6 is provided.

Moreover, claim 8 of the present invention is
A method using a computer program according to any one of claims 1 to 6 or according to figure 2 is provided.

A logic program (fixed facts and rules), a temporary context already added by the user, a fact newly added by the user (temporary context) and a new addition by the user according to an embodiment of the present invention It is a figure explaining the relationship of the question sentence to be made. 2 is a diagram for explaining the most basic processing flowchart (steps as a method) of a computer program for implementing FIG. 1 and the like; FIG. It is an explanatory view showing a general configuration as a computer to implement.

The present invention can be implemented as an automatic idea synthesizing system using the logical AI (subset of PROLOG + α) and knowledge expression (which can be described in almost natural language) of Patent Document 1. Since the detailed configuration and algorithm contents are not the essence of the present invention, they will not be described in detail here.

FIG. 1 shows a logic program (fixed facts and rules), a temporary context already added by the user, a fact newly added by the user (temporary context) and the user It is a diagram for explaining the relationship between question sentences newly added by . The 15th to 21st lines of the in-memory array are shown, and the rest are omitted. Temporary contexts are stored from the 22nd line onwards.
On the 15th line, there is an expression as an "impossible" rule, "!Make $X eat $Y :- $X is livestock; $Y is poison;" A combination (value pair) of a value of $X known to be poisonous (e.g., "goat") and a value of $Y known to be poisonous (e.g., "paper with arsenic") is It is a knowledge expression that excludes (restricts) simultaneous binding of the variable set ($X, $Y) of "feed $Y and make money".

FIG. 2 is a diagram illustrating a processing flowchart (steps as a method) of the most basic forward inference of a computer program for implementing FIG. 1 and the like.
Step S1 is activated in a scene (in the execution environment program) where OO (some fixed character string) is to be used as a basis for forward/backward reasoning.
In step S2, the memory is searched to see if "!000" prefixed with "!" actually exists in the logic program or temporary context.
In step S3, if it exists, it means "impossible", so it is determined that it cannot be used as a basis (S7), but if it does not exist, further automatic confirmation is performed in step S4.
In step S4, for a rule of type "!□□□:-..." in either the logic program stored in the memory or in the temporary context, "□□□:-..." Automatically check whether 〇〇〇 can be automatically proved from the knowledge in the memory by the rule.
In step S5, if it can be confirmed that it can be automatically proved, it is determined that it is not used as a basis (S7), and if not, it is determined that it is used as a basis (S6).
In fact, in step 4, special unification (unification) such as multi-length matching detailed in Patent Document 1 (application published) and backward inference including backtracking are also performed in various aspects. Although it is advantageous, it is not the subject of this specification (and since ordinary PROLOG searches can be substituted to some extent), it will not be described in detail here (see the published application).

FIG. 3 is an explanatory diagram showing a general configuration as a computer that implements. Since there are no unique features of the present invention here, it will not be described in detail.

4 General computer used in the present invention (user-dedicated personal computer, shared file server, web server)
4a-4k general components of a general computer 4 (user-dedicated personal computer, etc.) 11 logic program (fixed facts and rules)
12 data (temporary context already added to array memory adjacent to the logic program)
13 1 new fact (temporary context) and 1 question 14 A rule that means that if 3 or more conditions on the right side are satisfied, the establishment of the proposition on the left side is automatically confirmed (automatically added & notified) 15 Added new facts 2 1 answer (but 1 answer candidate has been deleted by the present invention)
S1 In one embodiment, the step S2 of detecting situations/situations that are going to be based on the fact 〇〇 or the condition 〇〇 that does not contain/misses variables. In the same embodiment, the fact! Fact! If 〇〇〇 exists in the memory, the process proceeds to step S7; otherwise, the process proceeds to step S4. ..." type rule, in step S5 S4 for automatically confirming whether or not 〇〇〇 can be automatically proved from the knowledge in the memory by the "□□□:-..." rule without the "!" If automatic proof is possible with any of the "□□□:-..." rules that take "", the process moves to step S7, and if none of the rules can automatically prove it, the process moves to step S6. Step S7 to decide to hire based on 〇 Step to decide not to hire based on 〇〇〇

Claims (8)

A computer program,
In a question, fact, rule condition or rule effect entered by a user, or in the fact, rule condition or rule effect of a given logic program different from said computer program,
In addition to the affirmation (true) as the existence of evidence facts or the success of automatic proof and its denial (absence of evidence facts and automatic proof failure/impossibility),
allowing a predetermined symbol signifying the existence of a contradicting fact of evidence or the success of a contradicting automatic proof,
CPU
When using an affirmative expression that does not include the variable from the beginning or does not include it due to variable binding,
CPU
automatically confirming that an expression obtained by adding the predetermined symbol to the affirmative expression does not exist and cannot be automatically proved;
CPU
When an expression obtained by adding the predetermined symbol to the affirmative expression exists or can be automatically proved
A computer program adapted to limit the use of said affirmative expressions . CPU
A negative expression with a given sign added ,
It is characterized by uniformly logically prioritizing an affirmative expression that does not add the corresponding predetermined symbol.
2. A computer program as claimed in claim 1 . CPU
3. A computer program as claimed in claim 1 or 2, for translating logic regarding affirmative and negative facts, prioritizing the most recent user input. CPU
When rephrasing the above logic
Affirmative expression in case of conversion to negative or negative expression in case of conversion to positive
characterized by deleting from stored memory,
4. A computer program as claimed in claim 3. CPU
When rephrasing the above logic
Affirmative expression in case of conversion to negative or negative expression in case of conversion to positive
characterized by commenting out from stored memory,
4. A computer program as claimed in claim 3. CPU
6. The computer program according to claim 5, wherein the comment is automatically given a time stamp or that the reason is user input . Apparatus comprising a CPU, for causing the CPU to execute a computer program according to any one of claims 1 to 6. In the CPU provided in the device,
A method for executing a computer program according to any one of claims 1-6 .

Images