Shloak's Repository

Reinforcement Learning for Optimized Pipeline Maintenance Scheduling

The B. John Garrick Institute for the Risk Sciences - UCLA & California Energy Commission

Research starting in 2025

• Developing a novel multi-objective reinforcement learning approach for condition-based pipeline maintenance scheduling
• Implementing Bayesian network models to predict three types of corrosion: internal, external, and stress corrosion cracking
• Building upon previous research that achieved 58% maintenance cost reduction through Q-learning
• Enhancing pipeline longevity through comprehensive corrosion modeling and adaptive maintenance scheduling
• Collaboration with UCLA Risk Sciences Institute for computational resources and risk management expertise

Exploring Author Profiling for Fake News Detection

Published in the 2022 IEEE 46th Annual Computers, Software, and Applications Conference (COMPSAC)

The conference was due to take place in Turin, Italy, where I would have presented my paper. However, due to the pandemic the conference was online and live presentations were called off.

• Built author profiles to detect fake news authors based on 4 feature groups: stylometry, sentiment & emotion, named entities, and co-authorship patterns.
• Achieved an 83% True Positive Rate, a 5% False Positive Rate, and a 93% area under the curve score on a Repeater Operating Characteristics Curve.
• Virtually presented my research to 1000+ researchers from Amazon. inc, Princeton University, etc.
• Mentored by Dr. Maria Konte, research scientist at the School of Computer Science at Georgia Tech.
• Citation: S. Rathod, "Exploring Author Profiling for Fake News Detection," 2022 IEEE 46th Annual Computers, Software, and Applications Conference (COMPSAC), 2022, pp. 1614-1619, doi: 10.1109/COMPSAC54236.2022.00256.

Challenges in Equitable COVID-19 Vaccine Distribution: A Roadmap for Digital Technology Solutions

Published on arXiv under General Economics

• Worked under Prof. Ramesh Raskar, MIT Media Lab and co-authored the research.
• Identified challenges in logistics, health outcomes, user-centric matters, and communication associated with disease-related, individual, societal, economic, and privacy.
• Our primary aim is intended to succinctly compile challenges in Equitable COVID-19 Vaccine Distribution for governments, organizations, and individuals.
• Collaborated with researchers from Harvard Medical School, National Human Genome Research Institute (NIH), MIT, etc.
• Citation: Bae, Joseph, et al. "Challenges of equitable vaccine distribution in the covid-19 pandemic." arXiv preprint arXiv:2012.12263v3 (2022).