The role involves developing an understanding of the traditional and current adversarial machine learning (AML) landscape, concept drift, attacks and mitigation strategies in an attempt to improve robustness of ML models that may be used in intelligent transportation systems (ITS). Additional tasks:- Implementing convolutional neural networks on transport datasets (object detection, image / semantic / mask segmentation)- Recommending key AI technologies to National Highways - Collecting relevant, publicly available datasets- Collaboration with the research team to create an overview of AI / AML in the ITS landscape- Assist in writing an academic paper which aims to guide future works in the field

My role involved creating a data collection system, which I automated via Selenium (Python), and JavaScript to identify elements of web pages (like Twitter), pull them from these pages and post iteratively according to certain page characteristics. Additional tasks:- Literature review considering relevant machine learning models- Automated web-scraping in Python / JavaScript- Feature importance on publicly available datasets