Hybrid Voice Graphical Interface for Portable Information Retrieval
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Traditional information retrieval systems face challenges with user interfaces, particularly in portable devices, due to impracticality of full-sized keyboards, inadequate handling of equally relevant documents, and limitations of speech recognition, which lead to inefficiencies and user frustration.
Innovation Solution
A hybrid spoken and graphical user interface is developed, featuring multiple views with lists of potential targets, speech recognition for queries, and a language model-based ranking system that updates the interface to improve information retrieval efficiency without requiring a full keyboard, suitable for digital media libraries and portable applications.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of operation
If speech recognition is used for information retrieval in portable devices, then user interface convenience is improved, but processing requirements and memory usage increase
Solution Approach 1:
The system segments the information retrieval process into distinct phases: speech input collection, query translation using language models, and result presentation. This allows speech recognition to be used only when needed rather than continuously, reducing overall processing requirements while maintaining ease of operation.
Solution Approach 2:
Language models serve as intermediaries between speech input and the information retrieval system. The language model translates spoken queries into structured search queries, reducing the direct processing burden on the portable device while enabling natural language input.
2Measurement precision
If traditional probability ranking systems are used, then document retrieval follows relevance probability, but documents with equal relevance scores lack proper ordering
Solution Approach 1:
The system changes the ranking parameter from a single relevance score to a multi-dimensional ranking system that includes relevance score, document type, and recency. This resolves the tie-breaking problem for equally relevant documents by introducing additional sorting criteria beyond basic probability matching.
3Measurement precision
If speech recognition requires significant processing power, then accuracy may be improved, but portability and energy efficiency are reduced
Solution Approach 1:
The system applies partial speech recognition processing by using language models to handle the most critical aspects of query translation while relying on simpler matching algorithms for the remainder of the information retrieval process. This achieves sufficient accuracy without requiring full-scale speech recognition processing.
Data Source
AI summary
A method of information retrieval using a hybrid interface is provided. The method includes providing a graphical user interface having a plurality of views with a list of potential targets within each of the views, determining a selected view from a user selection of one of the views, translating a voice input into a query, and updating the selected view based on the query to assist in information retrieval.