# US 2021/0241066 A1 · Systems and methods for intelligently configuring and deploying a control structure of a machine learning-based dialogue system

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

- Number: US 2021/0241066 A1
- Document ID: US20210241066A1
- Status: Published application
- Kind: A1
- Application number: 17/149,786
- Filed: 2021-01-15
- Published: 2021-08-05
- Earliest claimed: 2018-12-14
- Assignees: Clinc, Inc.
- Applicants: Clinc, Inc.
- Inventors: Parker Hill; Jason Mars; Lingjia Tang; Michael A. Laurenzano; Johann Hauswald; Yiping Kang; Yunqi Zhang
- Web page: https://patentgrep.com/patent/US20210241066A1

## CPC

- G06F 40/35: Discourse or dialogue representation
- G06F 16/3329: Natural language query formulation
- G06F 18/22: Matching criteria, e.g. proximity measures
- G06F 40/30: Semantic analysis
- G06N 3/006: based on simulated virtual individual or collective life forms, e.g. social simulations or particle swarm optimisation [PSO]
- G06N 3/044: Recurrent networks, e.g. Hopfield networks
- G06N 3/0442: characterised by memory or gating, e.g. long short-term memory [LSTM] or gated recurrent units [GRU]
- G06N 3/045: Combinations of networks
- G06N 3/084: Backpropagation, e.g. using gradient descent
- G06N 3/088: Non-supervised learning, e.g. competitive learning
- G06N 3/09: Supervised learning
- G06N 5/01: Dynamic search techniques; Heuristics; Dynamic trees; Branch-and-bound
- G06N 5/04: Inference or reasoning models
- G06N 20/00: Machine learning
- G06N 20/20: Ensemble learning
- G06N 7/01: Probabilistic graphical models, e.g. probabilistic networks
- G10L 15/1815: {Semantic context, e.g. disambiguation of the recognition hypotheses based on word meaning}
- G10L 15/1822: {Parsing for meaning understanding}
- G10L 15/22: Procedures used during a speech recognition process, e.g. man-machine dialogue
- G06F: ELECTRIC DIGITAL DATA PROCESSING (computer systems based on specific computational models G06N)
- G06N: COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G10L: SPEECH ANALYSIS TECHNIQUES OR SPEECH SYNTHESIS; SPEECH RECOGNITION; SPEECH OR VOICE PROCESSING TECHNIQUES; SPEECH OR AUDIO CODING OR DECODING

## Abstract

A system and method of configuring a graphical control structure for controlling a machine learning-based automated dialogue system includes configuring a root dialogue classification node that performs a dialogue intent classification task for utterance data input; configuring a plurality of distinct dialogue state classification nodes that are arranged downstream of the root dialogue classification node; configuring a graphical edge connection between the root dialogue classification node and the plurality of distinct state dialogue classification nodes that graphically connects each of the plurality of distinct state dialogue classification nodes to the root dialogue classification node, wherein (i) the root dialogue classification node, (ii) the plurality of distinct classification nodes, (iii) and the transition edge connections define a graphical dialogue system control structure that governs an active dialogue between a user and the machine learning-based automated dialogue system.

## Claims

### Claim 1

A computer-implemented method for designing a dialogue flow of a machine learning-based dialogue agent, the method comprising:
receiving, via a dialogue flow interface, one or more user inputs;
in response to the one or more user inputs, constructing a dialogue flow of a machine learning-based agent including:
(i) positioning within the dialogue flow interface a starting dialogue interface object,
(ii) positioning within the dialogue flow interface one or more dialogue state interface objects, and
(iii) connecting the starting dialogue interface object to the one or more dialogue state interface objects using one or more connection interface objects.

### Claim 2 (depends on 1)

The computer-implemented method according to claim 1, wherein
the dialogue flow of the machine learning-based dialogue agent, when executed by one or more computers, directs one or more operations of the machine learning-based dialogue agent in building one or more digital responses to utterance input from a user.

### Claim 3 (depends on 1)

The computer-implemented method according to claim 1, wherein
the dialogue flow of the machine learning-based dialogue agent, when executed by one or more computers, directs one or more operations of the machine learning-based dialogue agent starting from the starting dialogue interface object then along at least one of the one or more connection interface objects toward the one or more dialogue state interface objects.

### Claim 4

A computer-implemented method of creating a visual dialogue flow for implementing a digital dialogue agent using a dialogue flow, the computer-implemented method comprising:
positioning a starting node onto a dialogue flow interface;
positioning at least one dialogue node onto the dialogue flow interface that is distinct from the starting node; and
connecting the starting node with the at least one dialogue node within the dialogue flow interface.

### Claim 5 (depends on 4)

The computer-implemented method of claim 4, wherein
the connecting includes positioning a graphical connector that connects the starting node to the at least one dialogue node, and
the graphical connector directs a dialogue flow with a user to a computed dialogue intent state associated with the at least one dialogue node.

### Claim 6 (depends on 4)

The computer-implemented method of claim 4, wherein
upon connecting the starting node with the at least one dialogue node enables a dialogue path for a computer-implemented dialogue agent.

### Claim 7 (depends on 4)

The computer-implemented method of claim 4, wherein
the starting node comprises a parent node to the at least one dialogue node, the parent node passes conversational data to the at least one dialogue node based on (a) performing a classification of a conversational input and (b) an existence of a direct graphical connection between the starting node and the at least one dialogue node.

### Claim 8 (depends on 4)

The computer-implemented method of claim 4, wherein
the starting node comprises a parent node to the least one dialogue node, the parent node passes utterance data to the at least one dialogue node based on (a) performing a classification of an utterance input and (b) an existence of a direct graphical connection between the starting node and the at least one dialogue node.

### Claim 9 (depends on 4)

The computer-implemented method of claim 4, wherein
the dialogue flow interface comprises a web-based user interface that enables a user to visually build the dialogue flow based on positioning dialogue interface objects onto the dialogue flow interface.

### Claim 10 (depends on 4)

The computer-implemented method of claim 4, wherein
upon connecting the starting node with the at least one dialogue node, creates a dialogue path from the starting node to the at least one dialogue node that controls one or more operations of a dialogue agent upon execution of one or more functions of dialogue flow interface objects within the dialogue flow interface.

### Claim 11 (depends on 4)

The computer-implemented method of claim 4, wherein
the at least one dialogue node is positioned downstream of the starting node and configured to perform a respective dialogue intent operation based on input derived from a real-time conversational dialogue with a user.

### Claim 12 (depends on 4)

The computer-implemented method of claim 4, further comprising:
executing one or more dialogue flow interface objects within the dialogue flow interface, wherein:
upon executing the one or more dialogue flow interface objects activates a dialogue agent, and
the at least one dialogue node comprises a confirmational node that confirms aspect of utterance input from a user.

### Claim 13 (depends on 4)

The computer-implemented method of claim 4, wherein
the at least one dialogue node comprises a first dialogue node having a first dialogue intent state connected to the starting node and a second dialogue node having a second dialogue intent state positioned downstream of the first dialogue node and that is graphically connected to the first dialogue node.

### Claim 14 (depends on 13)

The computer-implemented method of claim 13, wherein
the first dialogue node and the second dialogue node are connected via an update transition to graphically connect the first dialogue node to the second dialogue node.

### Claim 15 (depends on 14)

The computer-implemented method of claim 14, wherein
the update transition reverts back to the first dialogue node indicating an action to be performed by the first dialogue node for obtaining additional information.

### Claim 16 (depends on 4)

The computer-implemented method of claim 4, wherein
the at least one dialogue node comprises a first dialogue node having a first dialogue intent state connected to the starting node and a second dialogue node having a second dialogue intent state graphically connected to the starting node.

### Claim 17 (depends on 4)

The computer-implemented method of claim 4, wherein
the positioning of the starting node and positioning of the at least one dialogue node is performed, via the dialogue flow interface, by a subscriber to an automated dialogue service that provides the dialogue flow interface over the web.

### Claim 18 (depends on 4)

The computer-implemented method of claim 4, further comprising:
setting, via the dialogue flow interface, a logic transition connector that extends from the least one dialogue node,
wherein the logic transition connector, when executed, causes a transition from the at least one dialogue node along to another dialogue node of the dialogue flow interface based on an input.

### Claim 19 (depends on 4)

The computer-implemented method of claim 4, further comprising:
identifying, via the dialogue flow interface, a response to utterance input of a user based on an execution of the visual dialogue flow.

### Claim 20 (depends on 4)

The computer-implemented method of claim 4, further comprising:
providing a dialogue response via the digital dialogue agent based on dialogue inputs handled by the visual dialogue flow.