Natural Language Interaction Outcome Prediction

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Conversation-based collaboration tools lack the ability to automatically identify and analyze workflow patterns in natural language interactions, making it difficult for users to predict outcomes and take corrective actions, as existing solutions require manual effort and are cumbersome.

Innovation Solution

A method using natural language analysis and a state prediction model to classify messages, forecast interaction outcomes, and suggest responses, which automates the identification of workflow patterns and improves interaction outcomes by ranking next message classes based on predicted success probabilities.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If manual analysis of workflow patterns is used, then users can identify interaction outcomes, but the process is cumbersome and requires significant manual effort

Engineering Contradiction:

Improveoutcome prediction accuracyVSAvoidmanual effort required

Core Design Contradiction:

Measurement precisionVSEase of operation

Solution Approach 1:

The system automatically analyzes workflow patterns and predicts interaction outcomes without requiring manual user intervention. The state prediction model self-services by continuously monitoring conversation states and generating outcome predictions, eliminating the need for users to manually track and analyze each interaction step.

Inventive Principle:

Principle #25Self-service

Solution Approach 2:

The patent replaces manual mechanical analysis with an automated computational system. The state prediction model uses natural language processing and machine learning algorithms to substitute human cognitive effort in analyzing workflow patterns and predicting outcomes, transforming a manual intellectual task into an automated computational process.

Inventive Principle:

Principle #28Mechanics substitution (Replace mechanical system)

2Reliability

If automated state prediction is implemented, then interaction outcomes can be forecasted accurately, but the system complexity increases

Engineering Contradiction:

Improveoutcome prediction reliabilityVSAvoidsystem complexity

Core Design Contradiction:

ReliabilityVSDevice complexity

Solution Approach 1:

The system segments the complex prediction task into distinct message classes (e.g., task creation, task completion, meeting scheduling). Each message class is analyzed independently through classification, and the state prediction model transitions between these segmented states, making the overall complex system manageable through modular classification categories.

Inventive Principle:

Principle #1Segmentation

Solution Approach 2:

The patent introduces message classes as an intermediary layer between raw conversation data and outcome predictions. The state prediction model uses these intermediate message class states to bridge the gap between unstructured natural language input and structured outcome forecasts, simplifying the overall system architecture through this mediating classification layer.

Inventive Principle:

Principle #24Intermediary (Mediator)

3Loss of time

If real-time analysis of natural language interactions is performed, then timely interventions can be made, but the processing time and computational resources increase

Engineering Contradiction:

Improveintervention response timeVSAvoidcomputational resource consumption

Core Design Contradiction:

Loss of timeVSUse of energy by moving object

Solution Approach 1:

The system performs preliminary classification of incoming messages into predefined message classes before full analysis. This preliminary action of categorizing messages upfront allows the state prediction model to efficiently process only relevant features, reducing subsequent computational requirements while maintaining real-time prediction capabilities.

Inventive Principle:

Principle #10Preliminary action

Solution Approach 2:

The patent applies partial analysis by focusing on key message classes and critical workflow states rather than analyzing every aspect of each conversation. The state prediction model selectively processes relevant message features and transitions, performing sufficient analysis to generate accurate predictions without the excessive computational overhead of complete exhaustive analysis.

Inventive Principle:

Principle #16Partial or excessive action

Data Source

PatentUS11443112B2Outcome of a natural language interaction

Publication Date: 2022.09.13 INTERNATIONAL BUSINESS MACHINE CORPORATION

AI summary

Using a natural language analysis, a current message is classified into a current message class, the current message being a portion of an interaction in narrative text form. For the interaction using a state prediction model, an interaction outcome corresponding to the current message class is forecasted, the forecasting comprising computing a probability that the current message class will result in a successful message class. Using the state prediction model, a set of next message classes and a set of predicted interaction outcomes are determined, each message in the set of next message classes corresponding to the current message class, each predicted interaction outcome in the set of predicted interaction outcomes corresponding to a next message class in the set of next message classes. According to the corresponding predicted interaction outcome, the set of next message classes is ranked.