HMVI (pronounced like Humvee) is an experimental system that enables vocal interaction between humans and machines, similar to Apple Siri or Amazon Alexa.
The following figure shows the main components of HMVI:
- Speech To Text: When a user says something, this module will translate that utterance into text. This is also called Speech Recognition.
- Intent Recognizer: Receive text from A, parse the text, and apply different algorithms to determine the intent of the text. Based on the intent of the text, it will query one or more of the Knowledge Providers.
- Knowledge Providers: Receive queries from B, forward the query to appropriate provider, and return the results back. The data returned usually will be formatted in a programmer friendly fashion, like JSON or XML.
- Data To Text: Translate the formatted data from C into human readable format like English.
- Text To Speech: Receive text from D and convert it to speech (sound waves). This is also called Speech Synthesis.
- HMVI Service Example: What is IBM stock price
The HMVI Advantage
|HMVI||Amazon Alexa||Apple Siri||Google Assistant||Microsoft Cortana|
|Voice to Text||✓||✓||✓||✓||✓|
|Text to Voice||✓||✓||✓||✓||✓|
|TDA owns framework||✓|
|Quick expandable knowledge||✓||⚑||⚑||⚑|
|Service Oriented API||✓||⚑|
|SMS text messages||✓|
- ✓ Well supported
- ⚑ Partial support (as of March 2017)
You can reach us at the Advanced Technology Group, TD Ameritrade, Ann Arbor, Michigan.