About Us

JUSense is a team of enthusiastic members on a mission to utilise state-of-the-art artificial intelligence for societal benefits. We focus on basic and advanced research in a multitude of areas including IoT-Cloud, Social Network Analysis and Geoinformatics. Our research focuses on using computer algorithms, machine learning, deep learning and data science to understand the complex multi-modal (sensor, image, audio, text, etc.) data collected by us. Our members include Ph.D. scholar, masters and undergaduate students of the Computer Science and Engineering department of Jadavpur University and other institutes and organizations.

In the Spotlight

An attention enhanced hybrid model for particulate matter forecasting

Our article entitled “Attention enhanced hybrid model for spatiotemporal short-term forecasting of particulate matter concentrations” has been accepted in the journal of Sustainable Cities and Society (Impact factor=10.696). In this paper, a hybrid deep learning framework - AGCTCN (Attention based Graph Convolution and Temporal Convolution Network), based on spatial attention, graph convolution and temporal convolution is presented for short-term forecasting of particulate matter (PM) levels.

A study on Covid-19 is published

Our article entitled “Spatio-temporal variation of Covid-19 health outcomes in India using deep learning-based models” has been accepted in the journal of Technological Forecasting and Social Change (Impact factor=10.884). The article attempts to perform a case study that investigates the spatio-temporal variation in the performance of deep-learning-based methods for predicting COVID-19 health outcomes in India. Various widely applied deep learning models namely CNN (convolutional neural network), RNN (recurrent neural network), Vanilla LSTM (long short-term memory), LSTM Autoencoder, and Bidirectional LSTM are considered to investigate their spatio-temporal performance variation. The effectiveness of the models is assessed using various metrics based on COVID-19 mortality time-series from 36 states and union territories of India.

Classification of Mars imagery captured by the Curiosity rover

Our article entitled “Mars-TRP: Classification of Mars imagery using dynamic polling between transferred features” has been accepted in the journal of Engineering Applications of Artificial Intelligence (Impact factor=7.802). The article attempts to analyze and augment the collected dataset of MSL (Mars Science Laboratory) Mars surface imagery and formulate a supervised multi-class image classification problem for further study. It demonstrates the design and working principle of the Mars-Transfer Routing Perceptrons, implements the same, and applies it to the dataset.