DESC:Technical field
The present invention pertain to the chatbot/virtual assistants. More particularly, to the real time learning of the chatbot/virtual assistants during interaction with the correspondent.

Background of the invention
Computers can store, modify, and retrieve large amounts of data much more quickly than humans. They are also more accurate and precise in their computations and less prone to error than most conscientious human beings. However, computers cannot cope with many of the simple tasks that humans perform every day. In particular, they completely fail in generalizing and guessing. Also they have great difficulty working with either partial or noisy information. For this reason scientists have designed parallel distributed processors that consist of a vast network of neuron-like units. These systems are called Artificial Neural Networks. The arena of artificial intelligence mainly function through theses system of neural networks for smart data processing which enables the system to impart more human friendly tasks.
A neural network comprises of output vector, input vector and target vector. An output vector is generated form the input vector and then compares to the target vector. This process is called "learning". If the output and target vectors are identical, no learning takes place. Otherwise the weights are changed to reduce the error. A neural network learns a mapping function between the input and target vectors by repeatedly observing patterns from a training set and modifying the weights of its synaptic links. Each pass through the training set is called a "cycle". Typically, a learning process consists of many thousands of cycles, and takes from several minutes to several hours to execute on a digital computer.
With the new wave of Artificial Intelligence (AI), some research effort has been directed to conversational information systems. Intelligent assistant or so called intelligent bot has emerged in recent years. Examples include Siri of Apple, Facebook Messenger, Amazon Echo, and Google Assistant.
A chat bot (or robot) is a computer server or other computerized device that can carry on a conversation with a human. Conventional chat bot systems require many hand written rules and manually labelled training data for the systems to learn the communication rules for each specific domain. This led to expensive human-labeling efforts and, hence, high costs. In addition, developers of conventional chat bot systems are required to write and debug source codes themselves. This makes the system less user friendly.
There is the mention of use of neural networks in chatbot/virtual assistants or for artificial intelligence functions wherein the AI is provided with extensive data feed and the training to obtain the relevant output for each of the specific input. This system of training cycles is time consuming and requires more processing time for the output.
Therefor there is need for the neural network system which has less processing time for the out and provides the quick learning of the chatbot/virtual assistants with minimal training sets.

Object of the present invention
The present invention objects to provide the chatbot/virtual assistant/virtual assistant system which is capable of learning during real-time interaction with the correspondent or other chatbot/virtual assistant systems.
Another object of the present invention is to provide the chatbot/virtual assistant/virtual assistant system that is capable of storing perceived and learned information in the form of neural signature to form the memory.
Summary of the invention
The present invention pertain to the chatbot/virtual assistants. More particularly, to the real time learning of the chatbot/virtual assistants during interaction with the correspondent.
The present invention comprises of the chatbot/virtual assistants which are enabled for real time learning wherein the chatbot/virtual assistants are capable of memorizing the information perceived during the real interaction with the correspondent.
The present invention pertaining to the chatbot/virtual assistant system which comprises of neural signatures wherein every information or data perceived is stored as a memory in the form of neural signature. Unlike the neural network system with the input and output nodes that processes the information perceived, the neural network in the given invention makes use of neural signatures which are the codes assigned to the information or the data perceived stored as the memory, at the next encounter with the same information the memory with the neural signature for the same information is triggered to give the output. The system is capable of memorizing and assigning the neural signature to every information and data perceived even during the real or actual interaction with the correspondent which can be a human or the other neural network system. Thus this system is capable of addition and the change in the information stored in the form of neural signature during the real time interaction with the correspondent.

Detail Description of the invention
The present pertains to the chatbot/virtual assistant systems wherein the chatbot/virtual assistant system is capable of real time learning without the training cycle as in other systems. The present invention comprises of the neural network of neural signatures which are the unique codes assigned for the information or the data perceived. The output for a query or the required information is through these neural signatures forming the memory. The data stored in the form of neural signature can be easily availed without the prolonged processing as in the other neural networks. Also the chatbot/virtual assistants with such systems do not require the training cycles for giving the output as the information is memorized and stored for the next instant.
In the present invention, the chatbot/virtual assistant includes a texting interface wherein the conversation or the exchange of the data takes place. The input is in the form of the answer provided or feed to the chatbot/virtual assistant against the question or the query during the real time interaction between the chatbot/virtual assistant and the correspondent. The answer or the feed provided to the chatbot/virtual assistant in real time is stored as the information in the coded form, forming the neural signature. The neural signature is the unique code assigned to the data received as the feed. The coding for the neural signature can be hexadecimal coding, octal coding, binary coding, cryptographic hash or like. The data or information stored is thus coded to form the neural signature wherein these neural signature constitutes the neural network. At the next instant when the same question or query is encountered by the chatbot/virtual assistant, the output is obtained by triggering of that specific neural signature stored as the memory. In the present invention once the information is stored in the form of neural signature, there is no requirement of any training cycle or pattern formation to get the desired output. This system of the present invention thus mimics the human brain memory wherein the information perceived becomes the memory which is automatically triggered at the next instant of encounter with the same information or data. Thus each neural signature functions as the neurons.
The present invention is capable of using different coding systems like hexadecimal coding, octal coding, binary coding and cryptographic hash wherein each system has different coding pattern. Thus with different coding systems, the number and combination of the unique digital neurons created would also vary substantially. This provides large set of digital neurons which are used uniquely to code large amount of data or information, such as HEX coding system is capable of providing 4,228,250,625 unique digital neurons.
The application of the real time learning system in the present invention is not limited for the chatbot/virtual assistants alone wherein it can also be applied to the artificial intelligence (AI) systems and the AI robots as well.
Example
Suppose the information or the data feed to the chatbot/virtual assistant system of the present invention is “Table is a furniture”. This feed of information will be saved as the memory in the form of neural signature. Now if the query or the question posed to the chatbot/virtual assistant as the input is “What are different types of furniture?” the answer as the output will be “Table”.
In the second scenario of the learning, the chatbot/virtual assistant is feed with the information that “Chair is the furniture”. Now when the same query is posed to the chatbot/virtual assistant system as “What are the different types of furniture?” the answer retrieved now is “Table and Chair”. Thus there is real time learning of the chatbot/virtual assistant that “Chair” is also a furniture along with the “Table”.
,CLAIMS:We claim:
1. A method of real-time learning in a chatbot comprising the steps of:
a. input of the query/question on the chatbot interface;
b. input of the answer to the query/question to the said query/question on the chatbot interface;
c. coding of the said answer to form a unique neural signature;
d. said neural signature is stored to form the memory in the neural network,
wherein the said above method is the part of the chatbot system.
2. A method as claimed in claim 1 wherein, the interacting chatbot interface is a texting interface.
3. A method as claimed in claim 1 wherein, the query/question includes query/question pertaining to any topic/subject.
4. A method as claimed in claim 1 wherein, the real-time learning also includes the input of any feed/knowledge in the chatbot through chatbot interface.
5. A method as claimed in claim 1 wherein, the coding of the feed/answer includes any type of coding system such as HEX coding, binary coding and like.
6. A method as claimed in claim 1 wherein, the neural signature is a unique code assigned to every answer/feed received as the information/data by the chatbot system.
7. A method as claimed in claim 1 wherein, the linking of neural signatures constitutes the neural network.
8. A method as claimed in claim 1 wherein, the said neural signature functions as the memory in the neural network.
9. A method as claimed in claim 1 wherein, the said method includes the quick retrieving of the information at the next instant that is learned through real-time interaction.
10. A method as claimed in claim 1 wherein, the said method can also include the virtual assistant with real-time learning through verbal interaction.