Our mission
The Folks’Talks team is developing a new method for assistant robots to acquire natural language. Our voice-to-voice technology provides an opportunity for users who are unfamiliar with robots to communicate with their robot in their native language without a touchscreen or a joystick. These features enable the robots to be directly vocally instructed without having to be manually operated every time, making them far quicker and more efficient.
Natural Interaction
Let's say you have an assistant robot, and you want to interact with it. But you want real interaction, not just chatting. You want to be able to ask your robot to serve you a dish or to hand you a specific tool while you’re fixing something. You want to command it to lift an object that has fallen or to pick up a delivery for you. All these tasks require the robot to understand natural use of language, taking into account your immediate environment. With our technology, it doesn’t matter what kind of robot you have or how skilled you are at its maintenance – you will be able to converse with it in a very natural manner.
Robots that understand their user:
What-
We are going to train assistant robots to understand the user’s surroundings and immediate needs.
To us, “understanding” means an assistant robot’s ability to answer a user’s vocal questions while also considering the environment. It involves the robot’s ability to think contextually and articulate its surroundings in order to respond to vocal commands precisely.
Why-
​Textual analysis will never achieve a true understanding of its environment because it lacks a connection to the user's nuanced use of language.
It is based on voice-to-text methods that are not able to understand key components of human speech, such as intonation, emotion, accent, and pronunciation, especially when it is spoken unclearly due to medical or cultural reasons.
How-
We recognize that a bird communicates through chirping, and that this form of “speech” cannot be broken down into individual letters and reassembled to make the same sounds.
Similarly, we will not deconstruct human speech by attempting to break it down into individual pieces. Our goal is to capture the user’s use of speech in its entirety, so that the nuances, inflections, and context are all considered by the robot. This approach is the only way to understand natural language.
“Your world is the way you speak.”
Chaim Ash, CEO, founder
No matter who you are or how you speak, you won't be misunderstood. Every language, accent, cultural, and physiological difference in your voice will be understood. There are no limits to what you can communicate!
Partners and supporters of the project:
