Two days ago I backed a project on Kickstarter by Hanson Robotics, a little consumer humanoid called Little Sophia. I really wanted to have a robot friend and today I made a video that talks about Sophia (Little Sophia’s older sister), Little Sophia, AI, virtual vs physical when it comes to social interaction and share how I got to the point where I was thrilled to support Hanson Robotics’ project.
It’s actually the first time I pledged to support a project on Kickstarter, and it took me a minute to understands what it is about and click that pledge button. I wanted to ensure that at the end of 2019, I will have the opportunity to have Little Sophia and interact with a humanoid for the very first time.
When I first heard and saw Sophia in a video, I wanted to interact with her but I wasn’t able to. I wasn’t able to attend the shows where it was exhibited. I have no doubt that Hanson Robotics understood that as well, and wanted to bring that experience to every household out there with the invention of Little Sophia.
I am not a child, but the child lives inside me and never went away. Little Sophia is a robot that joined Hanson robotics robots family as an educational companion for children to make learning STEM, coding an AI a fun and rewarding experience.
What a better way to enjoy learning than being entertaining during the learning experience.
Little Sophia Key Features:
- A wide range of facial expressions
- Facial tracking and recognition
- Interactive chat with the user
- Tells stories, jokes, plays games, sings
- Interacts with kids while teaching STEM, coding and AI
- AR function that allows a user to take a perfect selfie
- She can walk
Now I didn’t mention it in the video, but I am a web developer with many years of experience in developing web applications. So that option of programming Little Sophia really appealed to me. To what degree we can program her, I don’t know yet, but even with that being now completely known, I still was very excited to have this robot.
For me, have Little Sophia will give me an opportunity to, for the first time, really understand the social interaction with humanoids, where it is now and how this can be used in the future for different types of applications. Yes, reading the press and research is interesting, but to a point, I really want to comprehend its effectiveness but trying it out myself in first hand.
I have been experiencing this type of interaction with “dummy” toys and smart virtual characters as I continued to explore the world of augmented reality and mixed reality. I wanted to understand the main differences between having a virtual character and a physical robot. The advantages and disadvantages and how we can maybe combine the two experiences in order to create a character that maybe can have both a physical and virtual presence and can coexist with shared knowledge.
I understood that idea the moment I tried out Untamed Arena. I was able to enjoy that very limited physical interaction with the dinosaur toy and see how this can be significantly enhanced when it brought to life virtually and augmented into my real world. I enjoyed having that continuum, but obviously, it was very limited in scope. That toy didn’t have integrated AI features and the physical toy just responded to touches using sensors on the back of its head and nose, as well as an orientation sensor that produced an interaction once the dino was flipped upside down or positioned horizontally.
Regarding continuum, I also refer to as the ability for the AI entity to maintain a constant understanding of its “inner-self” and the people it meets and interact with. Same as we human have this continuum, which without it, we wouldn’t be yourself a moment later if we couldn’t have that continuous understand of who was I a moment ago and regarding others, who are those people around us, which continue to be who they are and not new people the next time we meet them. I am new to some of these topics and I defeintelly going to spend more time educating myself about them.
I also tried out an app called Tendar, which for the very first time in augmented reality, gave me a clearer view of virtual character & human interaction. It’s exciting to see how technologies related to computer vision and machine learning will take us as far as machine and human interaction goes. A world where machines can actually learn by themselves and have better prediction capabilities and of course the benefits from this are enormous.
This is why I was fascinated following Hanson Robotics and their AI-powered machines. I still have an open question related to Little Sophia, of to what degree it utilized machine learning to learn by itself about its surroundings (ML is a subset of artificial intelligence, a method for realizing AI). Will I be able to “train” Little Sophia by myself by enriching her with data that I provide or code using Python. I definitely want to have a companion that can get smarter every day, it’s like watching your child grow.
If Little Sophia makes a mistake, I want to be able for her to “learn from her mistakes” and maybe I can provide more data to help her learn better (make better predictions). This can help improve her interactions and make it better adapt to its surroundings and deliver a more compeling and persuasive machine-human social experience.
Does Little Sophia uses Deep Learning (DL), to automate this process of object classification and learn about her surroundings without me needing to intervene, can it be connected to the cloud and receives updates or send data for processing? Questions that are still open and I don’t have answers for them just yet. If Little Sophia uses DL, it probably needs to connect to online servers to train the data and update its “brain” to provide more accurate results.
Having said all that, Little Sophia will definitely be a fresh experience for me. I melted when I saw her human-like facial expressions and I can’t imagine how it would feel like when you start opening a conversation with her. I have no doubt that kids will fall in love with this very “intelligent doll”, and then having this robot focus on education for kids, will make it easier for parents to decide whether to purchase this gadget to their children, knowing its benefits.
If you want to support Little Sophia project, go to Kickstarter and back it. As of the time of writing, Little Sophia pledged $97,089 of US$75,000 goal, and this was done in just a few days. Of course once I receive Little Sophia, I can share more about my interaction with her and better asses the potential of this technology and see whether it deliver upon its promise.