February 29th through March 1st, 2020 Computing Ph.D. student Amrina Ferdous attended and presented her current research at the “Data Science and Image Analysis Conference of the Pacific Northwest” at Washington State University in Pullman, WA. This year’s conference explored “efforts to address various data analytic methodologies with their scientific and social impacts of large-scale datasets and images”. Amrina presented her current work “A Comparison of Inverse Methods and Neural Networks in Geophysics”. In this presentation, she compared two different approaches: Inverse Methods and Neural Networks and explained how data collection practitioners could design useful methods that could substantially advance knowledge in a variety of applications, such as industry, academic, government institutions etc.
Amrina’s presentation emphasized parameter estimation to recover images in a nonlinear system from an incomplete data set. Amrina summarized her talk stating, “Inverse Methods and Neural Networks are used to estimate causal factors from a set of observations. Inversion sees the physics through the mathematical model while the Neural Networks learn the physics from the training data.”
Amrina is a current Computing Ph.D. student with Data Science emphasis at Boise State who is currently involved with working on tensor, inversion, tensor Neural Network, and deep Neural Network.
