# US 2018/0024989 A1 · Automated building and sequencing of a storyline and scenes, or sections, included therein

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

- Number: US 2018/0024989 A1
- Document ID: US20180024989A1
- Status: Published application
- Kind: A1
- Application number: 15/213,867
- Filed: 2016-07-19
- Published: 2018-01-25
- Earliest claimed: 2016-07-19
- Assignees: International Business Machines Corporation
- Applicants: International Business Machines Corporation
- Inventors: Harish Bharti; Sanjib Choudhury; Kshitij K. Raval; Rajesh K. Saxena
- Web page: https://patentgrep.com/patent/US20180024989A1

## CPC

- G06F 40/289: Phrasal analysis, e.g. finite state techniques or chunking
- G06F 40/35: Discourse or dialogue representation
- H04N 21/8405: represented by keywords
- H04N 21/8541: involving branching, e.g. to different story endings
- G06F 16/9024: {Graphs; Linked lists (G06F16/9027 takes precedence)}
- G06F: ELECTRIC DIGITAL DATA PROCESSING (computer systems based on specific computational models G06N)
- H04N: PICTORIAL COMMUNICATION, e.g. TELEVISION

## Abstract

A method uses random walk traversal(s) (for example, Lévy's flight algorithm) of a story data graph to help select and/or order nodes for automated story writing purposes. Dialogue metadata (for example, inverted quotation marks) is added, to indicate natural language dialogue in nodes of the story data graph. A cyclic, non-directed story data graph is partitioned into non-cyclic tree(s) (also called tree logical data structures) using nodes and connections from the story data graph.

## Claims

### Claim 1

A computer-implemented method for composing an automated story, the computer-implemented method comprising:
receiving a story data graph comprising a plurality of nodes including content and a plurality of connections among and between the nodes;
selecting a starting node from the plurality of nodes;
performing, using a content curator algorithm, a first random walk according to a first random walk algorithm, with the first random walk: (i) starting with the starting node, (ii) traversing a plurality of first-random-walk-traversed nodes of the plurality of nodes, and (iii) following connections of the plurality of connections of the story data graph on the first random walk; and
outputting a first-random-walk-traversed data set that indicates an identity and order of traversal of the first-random-walk-traversed nodes to further stages of an automated story writing program.

### Claim 2 (depends on 1)

The computer-implemented method of claim 1 further comprising:
receiving raw content and associated metadata; and
constructing the story data by forming the plurality of nodes and the plurality of connections as data structures, with the nodes and connections being based, at least in part, upon the raw content and associated metadata.

### Claim 3 (depends on 1)

The computer-implemented method of claim 1 further comprising:
writing, by machine logic and based, at least in part, upon the first-random-walk-traversed data set, an automated story using content from at least some of the first-random-walk-traversed nodes at least partially in the order in which the first-random-walk-nodes were traversed on the first random walk.

### Claim 4 (depends on 1)

The computer-implemented method of claim 1 wherein the first random walk algorithm is isotropic.

### Claim 5 (depends on 1)

The computer-implemented method of claim 1 wherein the first random walk algorithm is characterized by a probability distribution that is heavy tailed.

### Claim 6 (depends on 1)

The computer-implemented method of claim 1 wherein the first random walk algorithm is a Lévy's algorithm type of random walk algorithm.

### Claim 7 (depends on 1)

The computer-implemented method of claim 1 further comprising:
repeatedly performing, using the content curator algorithm, a plurality of random walks to respectively generate a plurality of traversed node sets;
ranking the traversed node sets to obtain a preferred-traversed-node data set indicative of an identity and order of traversal of a highest ranked traversed node data set of the plurality of traversed node data sets; and
outputting the preferred-traversed-node data set to further stages of an automated story writing program.

### Claim 8 (depends on 7)

The computer-implemented method of claim 7 further comprising:
writing, by machine logic and based, at least in part, upon the preferred-traversed-node data set, an automated story using content from at least some nodes of the preferred-traversed-node data set at least partially in the order indicated by the first preferred-node data set.

### Claim at position 9

9-20. (canceled)