Kirtivardhan Singh

ACE Ampethatic Bot

Combining multiple AI models to implement an AI most suitable for the creation of a personalized conversational Bot that can help users in dealing with emotional stress and anxiety.

Many types of research have been done on Empathetic Bots, on how to make them more humanized, and creating an effective bot to deal with Human emotions and situations which require someone to talk to. Still, their effectiveness in dealing has not been remarkable. So to add on or improve on already models, we are observing how the models interact differently or respond with different persons and how effective they are in certain situations. To enhance technology with digital health.

Combining multiple conversational AI models to achieve the most effective results for the empathetic bot. Personalizing the bot - to know and memorize the nature of the person the bot is chatting with. Check for chat patterns to decide whether a user is under mental stress and hence effectively communicate. To check the effectiveness of bot in real chat scenarios. To replicate the human feel to talk with someone who is in a distressed situation or dealing with anxiety.

Implemented and tested multiple models for the AI(OpenAI GPT, seq to seq, transfer learning, persona chat) All AI implementation was done using Python. Developed a mobile and Web App for chatting with the bot. Web Backend is implemented with Flask. For Frontend, Flutter is used to create a minimalist UI which helps in providing a pleasing experience and to minimize the number of click ratios to provide a seamless interface.

More detailed project description coming soon!
In the meantime, check out some of my other projects.

Close Project

Inventory Management System

The main objective is to increase the productivity of any small business.

We have implemented the application on 13 products for the initial stage and processed revenue calculation on the basis of previous inflow and outflow of inventory. We also have reached two companies for the real-world implementation of the project with more than 1k data.

This project is about a full-fledged Inventory Management System Desktop application that can help different firms and businesses to overcome the problem of amorphous stock trackings.

More detailed project description coming soon!
In the meantime, check out some of my other projects.

Close Project

Sanskrit-Devanagari Text Generator

A small approach to convert huge amount of text from different schemes

Here I Am using Bhagwad Gita for transliteration test. We are using Roman as the base scheme to add Devanagari scheme. I am using indic-transliteration library to extract Devanagari Text.

I started working on an other project where this project came sideways.

More detailed project description coming soon!
In the meantime, check out some of my other projects.

Close Project

cat non-cat classifier

Neural Network Implementation of Image classifier with multiple hidden layers.

In this project I have used Neural Network Implementation of Image classifier with multiple hidden layers. I've used two datsets: "test_catvnoncat.h5" as test dataset and "train_catvnoncat.h5" as train dataset and used logistic regression to classify whether an image is of a cat or non cat.

More detailed project description coming soon!
In the meantime, check out some of my other projects.

Close Project

Planar-Data-Classification-with-One-Hidden-Layer

Visualising the dataset using matplotlib. The data looks like a "flower" with some red (label y=0) and some blue (y=1) points. Number of training examples is 400. The goal is to build a model to fit this data. The Dataset contains a numpy-array (matrix) X that contains the features (x1, x2) and a numpy-array (vector) Y that contains your labels (red:0, blue:1). The size of hidden layer is 4.

The Neural Network model with one hidden layer has learnt the leaf patterns of the flower! 90% Accurate

More detailed project description coming soon!
In the meantime, check out some of my other projects.

Close Project