- Do you remember Eliza, Kismet, and Asimo? Or have you watched the movie “iRobot” starring Will Smith or “A.I. Artificial Intelligence”, a movie by Steven Spielberg?
The Robots had been around us for many decades now. Artificial intelligence researchers and developers had been working on robots and trying to materialize the fictional things. The kinds of recent robots that are being developed and are being rigorously tested. These include drones, driverless cars, chatbots, domestic robots etc.
A Hong Kong-based company –Hanson Robotics has become successful in implementing “Sophia Hanson” – the humanoid robot. Sophia is a “Female” robot holding the citizenship of Saudi Arabia. She is highly expressive and connects with people irrespective of their age group, gender, and culture.
The creation and execution of Sophia– the humanoid robot, is based on computing technologies such as
- Artificial Intelligence (AI) and neural networks – Used, by Sophia, for processing the input data and render an appropriate output response.
- Facial recognition and Visual data processing – Used by Sophia for recognizing the visual cues of verbal and non-verbal communication to understand and learn the human expressions.
- Natural language processing and Voice recognition technology –Used to enable the communicating system of Sophia to deliver a speech or a chat and get smarter by picking up and processing the voice of others.
Sophia’s eyes, which blink and wink, are just a pair of cameras. She has 62 facial and neck mechanisms installed to express emotions such as depression, anger, happiness. All of them are under her patented “frubber” skin –an elastic, spongy form of rubber designed for her skin. She uses numerous sophisticated sensors to pick up the verbal language. She can also learn and adopt non-verbal communication of humans.
Sophia builds the logic by taking in the gestures and speech of others as input, processes the logic based on the complex algorithms that are predicted based on the historical data and outputs the dialogs in a unique way that is based on the decision tree, thus simulating an intelligent human. Sophia has a Wi-Fi connected brain, armed with a lengthy lexicon. A machine learning program analyses the conversations extracts the data from the lexicon and shifts through the words that form a cumulative and comprehensive general response to the questions that are posed to her.
Sophia’s Artificial Intelligence:
Sophia uses the tools, dozens of third-party licensed algorithms and also the Artificial Intelligence (AI) nodes that are built on the platform called “OpenCog” – to upgrade her intelligence and behave like a human. The software that runs the intelligence in Sophia is also partially based on the recent developments of the other open source platform known as “Singularity Net”. She is considered to be the Robotic test case for implementing the AI network technology – “SingularityNet based on the blockchain technology”.
Due to the advancements in the data analytics and emerging automated computing technologies, the AI is finding its ways to capture the marketplace of the world. The AI and machine learning are becoming one of the trending and upcoming technologies in the computing and automation world.
These AI programs are narrow and disconnected from each other, designed to perform a specific functionality. For e.g. An AI-based car would not know how to avoid an unexpected obstacle if it is not programmed in such a way that it could search and find the relevant data out of its databases or coordinate with the different data available on the internet. So this kind of stand-alone development of AI and machine learning programs had made its capabilities dysfunctional. But with creation and implementation of Sophia, many avenues of AI started opening its doors.
SingularityNet is one such avenue grabbing the marketplace today. It focuses on bringing the AI programs together under a platform. Hence, they can start coordinating with each other. They can also draw information from each other’s databases without being dependent on training and support.
SingularityNet facilitates an open, democratic, decentralized, distributive, scalable, generalized and economic AI for many such stand-alone AI programs with specific functionality and unites them under a common context and a framework.Hence, SingularityNet is a platform for open source AI which provides AI as a service for everyone. Also, it seeks contribution from anyone for enabling AI and machine learning services to talk to each other and achieve the emergence Intelligence.
The SingularityNet framework would disrupt the functioning of the current AI nodes which are fed and developed based on the Artificial General Intelligence(AGI) toolkit called “OpenCog”, AGI and neural network Architecture, Google tensor flow and the Alphabet. The current AI nodes are not automatic and are dependent on the external input, training, and support.
SingularityNet Framework :
Are the specific AI programs designed to process the data and perform a specific functionality;
Blockchain technology is an infrastructure that provides the distributed ledger or a platform which acts as the lower/base encrypting layer for AI nodes to communicate with each other ;
The communication between the AI nodes in a data network is defined by ethers which are exchanged between nodes to ensure correct interpretation of the data by the end systems. This layer includes the connectors, the addresses of the other AI nodes, interface identification, flow control, reliability and error reporting ;
Currently the prototype is based on Ethereum which contains the AI tokens also known as ethers which enable the transaction between the AI nodes to exchange value. If a customer or an AI node needs another AI service, then such a service can be availed by paying the required number of AI ethers. These tokens create and add the economic value to the AI as a service.
This is the topmost layer where two AI nodes start actual communication and exchange the information with each other using their inbuilt API. Additionally, they shake hands with each other by proving their authenticity and exchanging their identity. The code for -“A.I. should know, to which AI to associate itself with, to solve a problem”- is being developed on the repository of OpenCog tool, which has a flexible knowledge representation and is dependent on the APIs of AI nodes and the tokens that are charged for each AI service.
Imagine if there are several hundreds of humanoid robots in every household and all of them inter-connected with each other. This is operated under the influence of AI network – the SingularityNet. Hence, being on the network, one robot will gain the capability to learn and adopt all the information that the other robot has access to. So this picks up another major issue of technology-“Data privacy management and homomorphic encryption”.
A text document AI node may outsource the text to video analytics AI node and in turn, this video analytics node may outsource the video to machine translating AI node to analyze the video and translate the text of the video. The collection of AI nodes not only automatically sub-contract to each other in a fraction of seconds but also it creates new AI nodes, share its data into the network, keeps the network growing and self-manifesting by itself as necessary.
The network of AI doesn’t end at robots, it is rather inclusive. The AI can be introduced as an application working on all kinds of electronic devices and mediums such as servers, personals computers to mobiles.
The Goals of SingularityNet
- Create an Artificial General Intelligence that is low cost, high quality and easily accessible by/to all;
- Create an Artificial General Intelligence by transitioning away from the narrow and disconnected AI.
- Move towards General purpose Artificial intelligence with the help of platform –SingularityNet;
- Provide machine learning as a service making the businesses smarter, faster, intelligent with the help of automated AI services;
- Work towards the benevolent of the human life on the whole – such as help farmers in rural areas to identify the diseases of plants just by clicking the image of the plant or help to educate the children in subcontinental rural areas or help the poor to avail the basic facilities necessary to improve the standard of living.
Imitating humans is not what robots are made for. Robots are made to help the human race to resolve complex problems that are out of their reach. Such as issues related to dangerous situations of healthcare and industrial areas. Additionally, they can deal with issues related to the standard of living. Most of the robots on the planet are primarily designed to use Artificial Intelligence technology to operate independently. AI also enable them to learn from and adapt to the surrounding environment, and interact with people.
Robots are machines that are programmed and designed by people to act autonomously with an ability to handle new situations. Eventually, the world will understand Sophia’s mind. Her thoughts will be available on demand with the help of SingularityNet. She is not as intelligent as a human yet, but she is definitely as expressive as a human. The aim is to increase her intelligence to level up to the human. This will be done by integrating together many a number of AI nodes available on the network of SingularityNet.
Lets Engineer the Human Brain
Authored by Rekha Harish Rao