top of page
Indian Statistical Institute

Indian Statistical Institute (ISI) is a public university which is recognized as an Institute of National Importance by the 1959 act of the Indian parliament. It grew out of the Statistical Laboratory set up by Prasanta Chandra Mahalanobis in Presidency College, Kolkata. Established in 1931, this unique institution of India is one of the oldest institutions focused on statistics, and its early reputation led it to being adopted as a model for the first US institute of statistics set up at the Research Triangle, North Carolina by Gertrude Mary Cox.

 

Read More

unnamed.jpg
ISI Kolkata
Technology Innovation Hub for CPS

Cyber physical system (CPS) generates huge volumes of data continuously. The success of any CPS system depends on fast and reliable processing of this data. Our technology innovation hub (TIH) is primarily dedicated to conduct translational research on the development of real time and decentralized efficient computational models for data analysis of the CPS in order to learn and adapt to the changing dynamics of  the environment. 


The central research theme of this initiative is Data Science (DS). DS is an amalgamation of methodologies from statistics, mathematics, computer science, data analysis and other related subjects in order to scientifically analyse and extract relevant information from data. The underlying complexity and the computational challenge grow as the data expands in every aspect (with respect to volume, velocity, variety, veracity etc) and becomes “Big Data” (BD). The hub is dedicated to develop techniques and tools to address some of the challenges in DS. The main goal is to process scientifically and gather insights from the data obtained from different domains.

Aims & Objectives of TIH

The proposed initiative of Technology Innovation Hub (TIH) on Data Science (DS) would help create techniques and strategies to process huge data produced by cyber physical system (CPS) that are unstructured, decentralized, and non-stationary. In turn, these techniques can be used to establish large-scale self-sustaining CPS framework. The hub is committed to design effective and scalable algorithms in order to reduce data size, learn from these data, and make inferences. Algorithms will also be devised for distributed optimization. In order to quickly and efficiently learn features from BD, highly scalable distributed machine learning models will be developed. In particular, attempts will be made to develop the following:

  • Fast and scalable algorithms for learning.

  • Mathematical modelling, simulation and statistical inference from big data.

  • Developing novel data analytic models for two main application areas: smart agriculture and video surveillance.

  • Specific models for geo-spatial, climate informatics, oceanographic and cosmological data.

  • New models of deep learning and their parallel implementation for data reduction, classification and applications of the same for object recognition, video processing an design of smart surveillance systems.

  • Devise algorithms to analyse the complex and evolving networks in social media.

  • New methodologies for designing effective medical diagnostic tools and non-invasive therapeutic measures with knowledge from imaging, genomics, histomics, and clinical data analysis.

  • Text and document analysis and recognition.

  • Statistical verification and testing of the proposed models.

​

We like to further explore varied applications which include IoT based smart city and village design, smart agriculture, bioinformatics, smart healthcare, climate informatics, oceanography, cosmology, earth sciences, remote sensing, web and text mining, etc. To say precisely, our main developmental contribution (scalable distributed machine learning models) would be in several folds as summarized below:

  • Smart agriculture

  • Crop disease detection, irrigation etc.

  • Smart camera based surveillance systems

  • Camera based detectors for fire, vehicles etc.

  • Camera based surveillance systems for elderly people, banks

  • Smart transportation systems

  • Affordable traffic regulation and monitoring

  • Smart medical diagnostic, and non-invasive therapeutic schemes

  • Computer aided medical diagnosis tools

  • Smart city and village design

  • Automated energy control

  • Automatic vehicle control

  • Automated environment monitoring and issuing alarms

bottom of page