|
I am a first year Masters student in the Department of Computer Science at Georgia Tech . I completed my undergraduate studies in Electronics & Communication Engineering at Manipal Institute of Technology in Manipal, Karnataka. My research interests lie in developing safe, robust, and autonomous solutions to handle issues involving climate change, conservation, and sustainability using Computer Vision and Deep learning techniques. Prior to this, I worked for two years as a Machine Learning Engineer at Hyperverge, a Fintech and Geospatial startup based out of India. At Hyperverge, I developed various Computer Vision applications in the geospatial and facial spoof detection sector. Parallelly, I have also pursued my interests in environmental conservation by contributing to Prof. Ryan Kastner's group UCSD Engineers for Exploration to develop the pipeline for animal behavior analysis. In my free time, I love going trekking, camping, and pursuing adventure sports. I am also an avid reader and enjoy discussions related to music spanning various genres, languages, and cultures. If you have any questions / want to collaborate / discuss research, feel free to send me an email at apal72@gatech.edu CV / Google Scholar / LinkedIn / Github |
|
|
|
|
Generated a synthetic aerial imagery dataset with different weather and layout variations Conducted ablation studies on leveraging diverse simulated aerial data to improve visual recognition accuracy and robustness under differing real world conditions on both synthetic and real aerial imagery. |
|
|
Analyzing environmental risk due to warehouse construction & freight transportation in Henry County, Georgia using geospatial analytics and visualization Designing the decision support tool to quantify the impact of various development scenarios while channeling growth towards social, environmental and economic resilience |
|
Developed custom passive video liveness solution for Face PAD(Presentation Attack Detection) robust across different ethnicities using vision transformer based approach and achieved accuarcy of >99% on spoof and live classes across different benchmarks. Led the AI efforts of the Geospatial team towards designing the change detection and classification system using a combination of segmentation models, automatic imagery alignment adjustment techniques, vectorization and shape enhancement methods on satellite/aerial imagery to simplify and improve the efficiency of the property tax collection system. |
|
Enabled FY20 (Financial Year 2020) GTM (Go to Market) strategy for Microsoft Public Sector Department by conducting a thorough analysis of the smart city landscape in India and suggesting a roadmap for efficient integration of cloud and AI services. Facilitated easy adoption of smart city solutions by understanding the niche technical requirements of each smart city element and designing prototypes using Microsoft products and services. |
|
Designed a computationally efficient deep learning-based semantic segmentation algorithm for self- driving cars that achieved 10x faster speed, 5x less memory consumption, and 224x fewer parameters with similar accuracy compared to the state-of-the-art segmentation algorithms by introducing the concept of transposed depthwise separable convolutions and spark module. |
|
|
|
Developing an automated system to monitor shifting shoreline trends in high-impact regions of the world using satellite imagery (Landsat, SPOT) and deep learning based segmentation techniques. Achieved an accuracy of 97% in delineating shorelines in sandy beaches worldwide through semi-supervised learning techniques on a custom dataset using a 3 class approach (sea, land, and runup) to counter tidal noise. |
|
At University of California, San Diego Engineers for Exploration worked in collaboration with San Diego Zoo to develop a real-time machine learning system to monitor and analyze behavioral cues in Aye-Aye for early detection of health issues. |
|
Conducted extensive survey and experimentation on the different Hyperparameter Optimization techniques such as Turbo, Hyperopt, PySOT, and Nevergrad as part of the BBO challenge. Identified the advantage of trust-region based exploration-exploitation provided by Turbo and the superior candidate selection method (TPE) used by Hyperopt to design a hybrid approach. |
|
Trained neural networks that are robust to post-hoc pruning using Targeted Dropout. The idea was to cut off the snow-ball effect of error buildup in the network’s activations by removing the impact of most weights. Tested the feasibility of this technique by experimenting with FSGM and PGD attacks on CIFAR 10/100 datasets implemented on ResNet models augmented with Targeted Dropout. |
|
|
|
Georgia Institute of Technology, Atlanta | Aug 2022 - Present Specialization in Machine Learning |
|
Manipal Institute of Technology, Manipal | July 2016 - June 2020 Minor Specialization in Data Science |