Research

Research is the key focus at Solarillion Foundation. Students pursue research in their areas of interest, and regularly publish their work at reputed international conferences.

Research Areas

Edge Machine Learning

The ML Team focuses on solving data-driven problems in the real-world, industrial and research scenarios. They often handle large amounts of data and have also worked on optimised intelligence algorithms.

Current research teams are focused on deploying sophisticated models on wearables and other resource-constrained devices (edge devices). Research groups implement models with low-memory footprint to analyse sensor data for healthcare applications.

Robust Learning

The Adversarial Attacks team works on making Deep Learning models robust to adversarial attacks. They previously worked on defending against physically plausible adversarial patch attacks.

The Federated Learning team trains models across different data sites holding local data samples, and makes their models gain experience from various datasets without sharing the training data.

Deep Learning

The Computer Vision research subgroup focuses on the detection of objects and events from image and video data. Latest research by the team include real-time surveillance-based crime detection.

The Natural Language Processing team works on tasks like classification, sentiment analysis and machine translation. The most recent work in this domain introduced a Multi Context Transformer architecture for Sign Language Translation.