Skip to main content

Staff Profile

Dr Sneha Verma

Postdoctoral Research Associate in Artificial Intelligence

  • Email: sneha.verma@ncl.ac.uk
  • Telephone: 07404858116
  • Address: Stephenson Building, Newcastle University, Newcastle upon Tyne NE1 7RU
Background

As a passionate and driven postdoctoral research associate, I specialise in uncertainty quantification methodologies, including bootstrapping ensembling and Monte Carlo dropout methods, utilising deep learning neural networks. My academic journey, crowned with a PhD from the City, University of London, highlights my pioneering work in nanophotonics for sensing applications. With a diverse skill set in C++, NanoPhotonics, Matlab, and Comsol, and a deep understanding of physics, science, and engineering, I offer extensive expertise in advancing research and development.

My academic pursuits are bolstered by over a year of hands-on industrial experience in robotics and automation, where I applied my knowledge in electronic devices and motion/gesture recognition. This valuable experience was gained at CSIR-Central Mechanical Engineering Research Institute, Durgapur. Additionally, my M.Tech in Optical Fibre Sensors from CSIR—Central Glass and Ceramic Research Institute, Kolkata—provided a strong foundation in cutting-edge technologies, shaping my career trajectory.

Committed to innovation, I continuously seek to bridge the gap between theoretical research and real-world applications. With a proven history of collaboration with leading research institutions, I am dedicated to pushing the boundaries of science and engineering, driving transformative advancements, and making a lasting impact on the global research landscape.





Research

With a solid foundation in artificial intelligence, machine learning, and neural networks, my research spans multiple domains, including digital twin technology, multimodal artificial intelligence, and the prediction of optical and structural properties using machine learning. Across different roles, I have developed AI solutions for complex systems while ensuring robustness and interpretability through uncertainty quantification methodologies.

In my current project, I am working in AI for Digital Twins. I focus on developing machine learning algorithms that enhance the real-time simulation, predictive analytics, and autonomous decision-making capabilities of digital twins. I specialise in integrating AI with IoT data streams to ensure continuous updates and reliable operations across large-scale systems. Through predictive models based on sensor data, I help anticipate equipment failures and optimise operational efficiency. My expertise in cloud platforms like AWS and Azure supports the deployment and scalability of digital twins in various industries.

My work also involves fusing infrared spectroscopy with digital pathology data to improve diagnostics using multimodal AI. Through the application of Convolutional Neural Networks (CNNs) and spectral data analysis techniques like Support Vector Machines (SVMs), I designed hybrid models that achieve superior predictive accuracy. These models, combined with SHAP and LIME for explainability, offer insights into the critical features that drive AI predictions. By utilizing Grad-CAM to visualize regions of interest in biomedical images, I further enhanced the interpretability of these multimodal systems, making the results more accessible to medical professionals.

I developed a deep neural network, XANESNET, for predicting the X-ray absorption spectra of first-row transition metals. This work utilised deep ensembles, Monte Carlo dropout and bootstrap resampling to improve the model's uncertainty predictions. The application of autoencoders allowed for forward and reverse predictions between spectra and nanostructures, enhancing the model’s capacity to link structural properties with their corresponding spectral data. My focus on uncertainty quantification has been crucial in ensuring model reliability, especially in high-stakes scientific simulations. I have focused on generating synthetic datasets to tackle data scarcity in specialised fields with the help of XANESNET. By utilising probabilistic models and curriculum learning strategies, I developed algorithms that simulate real-world conditions, enabling better generalisation in AI models. I have also leveraged cloud computing resources and high-performance computing clusters to accelerate large-scale computations and simulations, significantly enhancing the speed of model training and testing.

I also developed inverse machine learning algorithms using decision trees and deep learning frameworks like PyTorch and TensorFlow. This work was aimed at predicting the structural dimensions of nanomaterials based on their spectral properties. By integrating advanced data visualisation techniques using Pandas, NumPy, and Matplotlib, I created models capable of accurate predictions that informed the design of optoelectronic devices. The role also involved optimising topological structures and applying constrained optimisation algorithms to improve efficiency in sensor design and data processing.

I have also worked on video stitching, a complex task in computer vision. This involved seamlessly combining multiple video frames or streams to create a single, cohesive video. I developed algorithms that align overlapping frames based on features such as colour, motion, and geometry, ensuring smooth transitions and minimal artifacts. This work required precision in image processing, feature extraction, and optimisation to produce high-quality stitched videos.

In parallel, I applied machine learning to predict the optical properties of nanosensors. By leveraging datasets on material properties and sensor configurations, I trained machine learning models to predict behaviour under various physical conditions. This work significantly improved the design and optimisation of nanosensors by providing rapid, accurate predictions compared to traditional computational methods. Through these projects, I have demonstrated my ability to integrate machine learning with domain-specific knowledge, pushing the boundaries of sensor technology and data-driven science.



Awards and Recognitions:

  • Associate Fellow (AFHEA), UK Professional Standards Framework
  • London, UK (2022)
  • Recognized for contributions to higher education and teaching practices.
  • Travel Fund for Excellent Research Work, Worshipful Company of Tin Plate Workers
  • London, UK (2022)
  • Awarded travel funds in recognition of outstanding research contributions.
  • WCSIM Post-Graduate Award for Excellent Research Work
  • City of London, UK (2021)
  • Honored with the prestigious award for exemplary research efforts.
  • City University Scholarship for M.Phil and Doctoral Degree
  • London, UK (2019-2023)
  • Awarded a competitive scholarship to pursue M.Phil and PhD degrees at City, University of London.
  • CSIR Senior Research Fellowship at CMERI
  • India (2018-2019)
  • Received the highly esteemed fellowship for research excellence at the Central Mechanical Engineering Research Institute (CMERI).
  • International Conference Attendance Fund
  • India (2018)
  • Granted financial support to attend international conferences, promoting research dissemination.
  • India-Level Scholarship for Bachelor’s, Higher Education Commission (HEC)
  • India (2011-2015)
  • Secured a merit-based national scholarship for undergraduate studies through the HEC.
  • Master’s Project Funding, CGCRI
  • India (2016-2017)
  • Awarded project funds at the university level for research at the Central Glass and Ceramic Research Institute (CGCRI).



Professional and Academic Memberships:

  • Member of ML Commons
  • City of London (From 2023)
  • An active member of a prestigious machine learning community, contributing to advancements in AI research and applications.
  • Member of IGOAI (International Group of Artificial Intelligence)
  • London (From 2023)
  • Engaged in global AI initiatives and collaborative research within this prominent international group.
  • Member of Worshipful Company of Scientific Instrument Makers
  • City of London (From 2021)
  • Recognized for contributions to the field of scientific instrumentation, fostering collaboration and innovation.
  • Member of City AI Club, City University of London
  • City of London (From 2023)
  • Actively participating in the university’s AI research community, promoting knowledge sharing and research excellence.



Publications