18244063. METHOD AND SYSTEM FOR AUTO-CORRECTION OF AN ONGOING SPEECH COMMAND simplified abstract (Samsung Electronics Co., Ltd.)

METHOD AND SYSTEM FOR AUTO-CORRECTION OF AN ONGOING SPEECH COMMAND

Organization Name

Samsung Electronics Co., Ltd.

Inventor(s)

Prashant Inbavaluthi of Noida (IN)

Vikas Kapur of Noida (IN)

Ramakant Singh of Noida (IN)

METHOD AND SYSTEM FOR AUTO-CORRECTION OF AN ONGOING SPEECH COMMAND - A simplified explanation of the abstract

This abstract first appeared for US patent application 18244063 titled 'METHOD AND SYSTEM FOR AUTO-CORRECTION OF AN ONGOING SPEECH COMMAND

Simplified Explanation

The system described in the patent application is a voice assistant that can receive voice commands from users, convert them into text, extract features from the audio and text inputs, determine connections between them using a multi-modal unified attention sequence tagger, tag specific words in the text, decode revised text on-the-fly, and generate a response using natural language processing.

• Voice assistant receiving voice commands
• Speech to text convertor
• Feature extractor for acoustic and textual features
• Multi-modal unified attention sequence tagger
• On-the-fly decoder
• Natural language processing for response generation

Potential Applications

This technology can be applied in:

• Voice-controlled devices
• Virtual assistants
• Speech recognition systems

Problems Solved

This technology solves the following problems:

• Improving accuracy of voice command interpretation
• Enhancing user experience with voice assistants
• Streamlining communication between users and devices

Benefits

The benefits of this technology include:

• Faster and more accurate voice command processing
• Seamless integration of audio and text inputs
• Real-time feedback and response generation

Potential Commercial Applications

This technology can be commercially benefit in:

• Smart home devices
• Customer service chatbots
• Language translation services

Possible Prior Art

One possible prior art for this technology could be existing speech recognition systems that use similar techniques for processing voice commands and generating responses.

How does this technology handle different accents and speech patterns?

The patent application does not provide specific details on how the system adapts to various accents and speech patterns. It would be interesting to know if the technology has been tested with a diverse range of users to ensure accuracy across different linguistic backgrounds.

What measures are in place to protect user privacy and data security?

The patent application does not mention any specific protocols or features related to user privacy and data security. It would be important to understand how the system handles sensitive information and ensures the confidentiality of user data.

Original Abstract Submitted

The system includes a voice assistant receiving a voice command as input from user. The speech to text convertor converts the voice command into a text. A feature extractor extracts acoustic features from raw waveform of voice command and textual features from converted text for determining nearby context tokens. A multi modal unified attention sequence tagger determines a connection between the audio and the text based on an individual contextual embedding and a fused contextual embedding at context tokens level. It further tags replacement, cue and correction words sequentially based on determined connection between the audio and the text. An on-the-fly decoder decodes revised text on-the-fly based on tagged replacement, cue and correction words, to display the decoded revised text on user interface and sends the decoded revised text to NLP to generate a response corresponding to the input speech.