Dynamic Chatbot

The main purpose of our project is to implement a conversational AI chatbot but that chatbot will be dynamic chatbot that can respond to user?s questions with best suitable answers in text form. The chatbot will be dynamic in a sense of training means it will be able to train on any excel file g

Project Title

Dynamic Chatbot

Project Area of Specialization

Artificial Intelligence

Project Summary

The main purpose of our project is to implement a conversational AI chatbot but that chatbot will be dynamic chatbot that can respond to user’s questions with best suitable answers in text form.
The chatbot will be dynamic in a sense of training means it will be able to train on any excel file given to it containing question/answers data related to any domain. This chatbot will fall under the category of deep learning Retrieval based chatbots. Use cases of chatbots can be seen in many
industries like travel, retail, ecommerce, education and media. We are designing a chatbot for NED admission cell to facilitate students. At the time of admissions students face many problems in admission procedure and have many queries. So our chatbot will be able to resolve their queries
and tries its best to reply them with best answers. Data can be collected by distributing questionnaires to the coming fresher’s, asking them what queries did they have when they were seeking admission in NED. The chatbot model will be a general one that will be able to trained on any excel file but we will work on admission facilitation domain

Project Objectives

The objectives of our project are:
• To collect dataset of questions related to admission procedure and their corresponding answers
• To develop a deep learning model which can be able to trained on any excel file containing questions/answers data
• To deploy a model on AWS
• To create user interface where user can conversate with our ‘Dynamic Deep Learning Chatbot’

Project Implementation Method

Majority of bots in market are rule based but technology is evolving and so are chatbots. By incorporating Ai in chatbots businesses can be able to combine human intelligence with machine intelligence. In deep learning two kinds of chatbots are introduced. This dynamic chatbot will be a retrieval based chatbot that provide only predefined responses and do not generate new output. The vast majority of production systems today are retrieval-based, or a combination of retrieval-based and generative. The other deep learning chatbots are generative based but the problem is that generative models don’t work well in practice. In the area of NLP(natural language processing) many machine and deep learning algorithms/models are presented for text processing in which Recurrent Neural network (RNN) model can be seen
similarly effective or even better at specific natural language tasks. Other variants, such as long shortterm memory (LSTM) networks, residual networks (ResNets), and gated-recurrent networks (GRU). We are building our model using Deep learning algorithms. After model creation and testing we need to launch it in a place where people can interact with it. We are creating an application where user can conversate with our ‘Dynamic Deep Learning Chatbot’. We have created static frontend using Angular. Our backend will be developed on Python and AWS Lambda functions will be used. in later stages we will deploy our frontend on some hosting or dummy URL.

Benefits of the Project

Our proposed chatbot is poised to ease these frustrations by providing the real-time, on-demand information that candidates are seeking out. While numbers are great, a business case for chatbots also needs to spell out the benefits to the business in terms of cost savings and improved processes for users. Our chatbot will be able to train on any excel file given to it. means now companies and businesses with small datasets, does not have to go to developers to build their chatbot. they can only upload their file of questions answers and in few minutes their chatbot will be ready.

Technical Details of Final Deliverable

Following are the steps that we are going to follow to create our dynamic chatbot: 

1. Prepare Data: Our first step will be data preparation. We will use process of ontology to gather as many interactions as possible.
2. Data Reshaping: Then we’ll reshape our data into single rows of observations. These observations can be called message-response pairs that will be added to the classifier.
3. Pre-Processing: The next step in building a deep learning chatbot is that of pre-processing. The processes involved in this machine learning step are tokenizing, stemming, and lemmatizing the chats. This makes the chats readable for the deep learning chatbot. 
4. Generate Word Vectors: Since we are going to create a retrieval-based chatbot we need to Generate Word Vectors. We might use Word2Vec model, and python script to train our model. Alternatively, we can also use TensorFlow Seq2Seq function for the same. 
5. Model Creation and Optimization: We might use TensorFlow to create our model by writing a python script. We will try to follow the model strategy for our deep learning chatbot. 
6. Track the Process: We are going to observe how our deep learning chatbot gets trained via machine translation techniques. 
7. Add it to an Application: We’ll create an application where user can conversate with our ‘Dynamic Deep Learning Chatbot’.

Final Deliverable of the Project

Software System

Core Industry

Education

Other Industries

Core Technology

Artificial Intelligence(AI)

Other Technologies

Sustainable Development Goals

Industry, Innovation and Infrastructure

Required Resources

If you need this project, please contact me on contact@adikhanofficial.com