System and methods for semiautomatic generation and tuning of natural language interaction applications

• US 8,903,711 B2
• Filed: 10/08/2013
• Issued: 12/02/2014
• Est. Priority Date: 04/10/2012
• Status: Active Grant

• Current Assignee: Artificial Solutions Iberia, S. L.
• Original Assignee: Artificial Solutions Iberia SL
• Inventors: Lundberg, Sonja Petrovic, Aili, Eric, Wieweg, Andreas, Jonsson, Rebecca, Hjelm, David
• Primary Examiner(s): Vo, Huyen X.
• Application Number:
  US14/049,168
• Publication Number:
  US 20140019116A1
• Time in Patent Office: 420 Days
• Field of Search: 704 1- 10, 704/251, 704/255, 704/257, 704/231, 704/275, 704/235, 704/270, 704/252, 706/62, 714/48
• US Class Current: 704/8
• CPC Class Codes: G06F 11/3476 Data logging G06F11/14, G06F11/2205 take precedence
• CPC Class Codes: G06F 11/3664 Environments for testing or debugging software
• CPC Class Codes: G06F 16/90332 Natural language query formulation or dialogue systems
• CPC Class Codes: G06F 16/972 Access to data in other repository systems, e.g. legacy data or dynamic Web page generation
• CPC Class Codes: G06F 40/247 Thesauruses; Synonyms
• CPC Class Codes: G06F 40/30 Semantic analysis
• CPC Class Codes: G06F 40/40 Processing or translation of natural language natural language analysis G06F40/20; semantic analysis G06F40/30
• CPC Class Codes: G06F 40/58 Use of machine translation, e.g. for multi-lingual retrieval, for server-side translation for client devices or for real-time translation
• CPC Class Codes: G06F 8/30 Creation or generation of source code
• CPC Class Codes: G06F 8/447 Target code generation
• CPC Class Codes: G10L 15/183 using context dependencies, e.g. language models

Abstract

A system for supervised automatic code generation and tuning for natural language interaction applications, comprising a build environment comprising a developer user interface, automated coding tools, automated testing tools, and automated optimization tools, and an analytics framework software module. Text samples are imported into the build environment and automated clustering is performed to assign them to a plurality of input groups, each input group comprising a plurality of semantically related inputs. Language recognition rules are generated by automated coding tools. Automated testing tools carry out automated testing of language recognition rules and generate recommendations for tuning language recognition rules. The analytics framework performs analysis of interaction log files to identify problems in a candidate natural language interaction application. Optimizations to the candidate natural language interaction application are carried out and an optimized natural language interaction application is deployed into production and stored in the solution data repository.

Claims

1. A system for supervised automatic code generation and tuning for natural language interaction applications, the system comprising:

a solution data repository component operating on a server computer and comprising a plurality of hierarchical ontological domain models and a plurality of hierarchical language models comprising a preconfigured, hierarchical arrangement of language objects, a plurality of language recognition rules expressed in terms of elements of the plurality of hierarchical language models;

a build environment comprising a plurality of software modules stored or operating on a computer, the plurality of software modules comprising a developer user interface, a plurality of automatic text analysis tools, a plurality of automated coding tools, a plurality of automated testing and tuning tools, and a plurality of automated optimization tools; and

an analytics framework software module operating on a server computer and adapted to receive and analyze a plurality of log files of natural language interaction sessions and a plurality of text corpora;

wherein a plurality of text samples is assigned to a plurality of input groups based on semantic content, each input group comprising a plurality of semantically related inputs, where the assignment of the plurality of text samples into input groups is performed either with automatic clustering or manually;

wherein for each of the plurality of input groups a plurality of language recognition rules, a plurality of language objects and a plurality of flow elements are automatically generated by the plurality of the automatic coding tools;

wherein the plurality of automated testing tools carries out automated testing of the plurality of language recognition rules and generates a plurality of recommendations for tuning or correcting language recognition rules;

wherein the analytics framework software module performs analysis of a plurality of log files of natural language interaction sessions between a plurality of users and a candidate natural language interaction application in order to identify problems in the candidate natural language interaction application.

2. A method for supervised automatic code generation and tuning for natural language interaction applications, the method comprising the steps of:

(a) importing, using an analytics framework software module stored and operating on a server computer, a plurality of text samples into a build environment comprising a plurality of software modules stored or operating on a computer, the plurality of software modules comprising a developer user interface, a plurality of automated coding tools, a plurality of automated testing tools, and a plurality of automated optimization tools;

(b) automatically assigning, using the analytics framework software module, the plurality of text samples to a plurality of input groups based on semantic content of the plurality of text samples, wherein the assignment is performed using automated clustering;

(c) automatically generating, using at least one of the plurality of automated coding tools, a plurality of language recognition rules, a plurality of language objects and a plurality of flow elements for each of the plurality of input groups;

(d) performing, using at least one of the plurality of automated testing tools, automated testing of language recognition rules generated in step (c);

(e) automatically generating, using at least one of the plurality of automated optimization tools, based on results obtained in step (d), a plurality of recommendations for tuning or correcting language recognition rules;

(f) providing the plurality of recommendations to a developer via the developer user interface;

(g) deploying, using the plurality of optimization tools, a candidate natural language interaction application to a plurality of users;

(h) performing analysis, using the analytics framework software module, of the log files to identify problems in the candidate natural language interaction application;

(i) carrying out, using the plurality of optimization tools, a plurality of optimizations to the candidate natural language interaction application; and

(j) deploying, using the plurality of optimization tools, an optimized natural language interaction application into production.