# US 2024/0202547 A1 · Abductive reasoning apparatus, abductive reasoning method, and storage medium

> Agents: this data is available as JSON from the patentgrep API and MCP server — see https://patentgrep.com/llms.txt

- Number: US 2024/0202547 A1
- Document ID: US20240202547A1
- Status: Published application
- Kind: A1
- Application number: 18/286,386
- Filed: 2021-04-16
- Published: 2024-06-20
- Earliest claimed: 2021-04-16
- Assignees: NEC Corporation
- Applicants: NEC Corporation
- Inventors: Kazeto YAMAMOTO
- Web page: https://patentgrep.com/patent/US20240202547A1

## CPC

- G06N 5/022: Knowledge engineering; Knowledge acquisition
- G06N 5/04: Inference or reasoning models
- G06N: COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS

## Abstract

An abductive reasoning apparatus (10) includes: a generation section (12) that generates a plurality of hypotheses with reference to background knowledge information and query information; a construction section (13) that constructs a constraint set by listing some of a plurality of logical constraints to be satisfied between a plurality of elements constituting each of the plurality of hypotheses; a retrieval section (14) that retrieves any one of the plurality of hypotheses as a candidate solution hypothesis with reference to the constraint set; a determination section (15) that determines whether the candidate solution hypothesis satisfies a logical constraint which is among logical constraints to be satisfied between elements constituting the candidate solution hypothesis and which is not included in the constraint set; and an output section (16) that outputs the candidate solution hypothesis as a solution hypothesis in a case where the candidate solution hypothesis satisfies the logical constraint.

## Claims

### Claim 1

An abductive reasoning apparatus comprising at least one processor, the at least one processor carrying out:
an acquisition process for acquiring background knowledge information which represents background knowledge by at least one logical formula and query information which represents an observed fact by at least one logical formula;
a generation means that generates process for generating a plurality of hypotheses with reference to the background knowledge information and the query information;
a construction process for constructing a constraint set by listing some of a plurality of logical constraints to be satisfied between a plurality of elements constituting each of the plurality of hypotheses;
a retrieval process for retrieving any one of the plurality of hypotheses as a candidate solution hypothesis with reference to the constraint set;
a determination process for determining whether the candidate solution hypothesis satisfies a logical constraint which is among logical constraints to be satisfied between elements constituting the candidate solution hypothesis and which is not included in the constraint set; and
an output process for outputting the candidate solution hypothesis as a solution hypothesis in a case where the candidate solution hypothesis satisfies the logical constraint which is not included in the constraint set.

### Claim 2 (depends on 1)

The abductive reasoning apparatus according to claim 1, wherein
in a case where the candidate solution hypothesis does not satisfy the logical constraint which is not included in the constraint set, in the determination process, the at least one processor adds the logical constraint to the constraint set and carries out the retrieval process.

### Claim 3 (depends on 1)

The abductive reasoning apparatus according to claim 1, wherein
in the retrieval process, the at least one processor excludes the candidate solution hypothesis that does not satisfy the logical constraint which is not included in the constraint set, and retrieves a new candidate solution hypothesis with reference to the constraint set.

### Claim 4 (depends on 1)

The abductive reasoning apparatus according to claim 1, wherein
assuming that a logical constraint that occurs for each combination of the plurality of elements constituting each of the plurality of hypotheses is to be subjected to determination, in the construction process, the at least one processor constructs the constraint set by listing logical constraints which are among the plurality of logical constraints and which are not to be subjected to determination.

### Claim 5 (depends on 4)

The abductive reasoning apparatus according to claim 4, wherein
the logical constraint that is to be subjected to determination includes a logical constraint on an equivalence relation between the plurality of elements.

### Claim 6 (depends on 4)

The abductive reasoning apparatus according to claim 4, wherein
the logical constraint that is to be subjected to determination includes a logical constraint on an order relation between the plurality of elements.

### Claim 7

An abductive reasoning method comprising:
(a) acquiring background knowledge information which represents background knowledge by at least one logical formula and query information which represents an observed fact by at least one logical formula;
(b) generating a plurality of hypotheses with reference to the background knowledge information and the query information;
(c) constructing a constraint set by listing some of a plurality of logical constraints to be satisfied between a plurality of elements constituting each of the plurality of hypotheses;
(d) retrieving any one of the plurality of hypotheses as a candidate solution hypothesis with reference to the constraint set;
(e) determining whether the candidate solution hypothesis satisfies a logical constraint which is among logical constraints to be satisfied between elements constituting the candidate solution hypothesis and which is not included in the constraint set; and
(f) outputting the candidate solution hypothesis as a solution hypothesis in a case where the candidate solution hypothesis satisfies the logical constraint which is not included in the constraint set,
(a) to (f) being carried out by an abductive reasoning apparatus.

### Claim 8

A non-transitory computer-readable storage medium storing therein a program for causing a computer to function as an abductive reasoning apparatus,
the program causing the computer to carry out:
an acquisition process for acquiring background knowledge information which represents background knowledge by at least one logical formula and query information which represents an observed fact by at least one logical formula;
a generation process for generating a plurality of hypotheses with reference to the background knowledge information and the query information;
a construction process for constructing a constraint set by listing some of a plurality of logical constraints to be satisfied between a plurality of elements constituting each of the plurality of hypotheses;
a retrieval process for retrieving any one of the plurality of hypotheses as a candidate solution hypothesis with reference to the constraint set;
a determination process for determining whether the candidate solution hypothesis satisfies a logical constraint which is among logical constraints to be satisfied between elements constituting the candidate solution hypothesis and which is not included in the constraint set; and
an output process for outputting the candidate solution hypothesis as a solution hypothesis in a case where the candidate solution hypothesis satisfies the logical constraint which is not included in the constraint set.
