Parallel Speech Language Identification Paths
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing language translation applications require prior knowledge of the source language, limiting their ability to translate unknown languages and often suffering from latency issues when processing multiple languages.
Innovation Solution
A system and method that utilize a database of supported languages and a feed-forward deep neural network (DNN) to identify the language of an input speech signal without prior knowledge, employing x-vector embeddings, multilingual bottleneck features, data augmentation, and a discriminative Gaussian classifier.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If the application processes many different languages, then the language identification capability is improved, but the latency increases
Solution Approach 1:
The patent segments the language identification task into multiple parallel processing paths, each dedicated to a specific language. This allows the system to process multiple languages simultaneously without sequential overhead, resolving the contradiction between supporting many languages and maintaining low latency.
Solution Approach 2:
The system dynamically selects and activates only the processing paths corresponding to languages that are actually present in the input speech signal. This dynamic adaptation allows the system to maintain low latency for the specific language detected while having the capability to handle many different languages, thus resolving the contradiction between versatility and speed.
2Ease of operation
If the source language is not known, then the translation application cannot perform translation, but requiring users to specify the source language reduces ease of operation
Solution Approach 1:
The system implements self-service by automatically detecting and identifying the source language of the input speech signal without requiring user specification. The multi-language processing paths autonomously determine which language is being spoken, eliminating the need for user input while ensuring the translation capability remains reliable.
Solution Approach 2:
The patent performs preliminary language identification action before the translation process begins. By pre-processing the speech signal to determine the source language and activating the appropriate processing path in advance, the system ensures that translation can proceed reliably without requiring users to manually specify the source language.
Data Source
AI summary
A method, computer program product, and computing system for speech language identification. An input speech signal in a particular language of a plurality of languages is received and processed by a plurality of speech recognition processing paths, each speech recognition processing path being configured to recognize a subset of the plurality languages. Each of the plurality of speech recognition processing paths processes the input speech signal using machine learning to identify a language in the associated subset of languages which is a closest match to the particular language of the input speech signal. The processing of the input speech signal by the plurality of speech recognition processing paths results in a plurality of identified languages. The input speech signal and an indication of each of the plurality of identified languages are processed in a further speech recognition processing path to recognize one of the plurality of identified languages as a closest match to the particular language of the input speech signal.