Title of the talk: Topological and Machine Learning Methods for Portfolio Choice  and Insider-Trading Detection

Date & Time: Nov 14,2025 at 5 PM (ONLINE)

Abstract: This research seminar surveys my research journey on data-driven methods for portfolio design and fraud detection across three strands. I first present tail-risk-aware index tracking and enhanced indexing using two-tail Mixed-CVaR to penalize both upside and downside deviations from the benchmark, and a Mixed-STARR objective for return-seeking with controlled risk. Then, I discuss applications of topological data analysis for asset filtering and sparse index tracking. Finally, I present my recent work on Hidden Cascades models for information diffusion on graphs to infer node-level transmission probabilities by matching feature distributions between real and simulated cascades. Applications include insider-trading detection, where results on an empirical insider network align with synthetic benchmarks and recover target feature distributions. I will conclude with a research plan for IIT Jodhpur.

About the speaker: Dr. Anubha Goel is a Marie Curie Postdoctoral Research Fellow in the Unit of Computing Sciences at Tampere University, Finland. Before this, she was a postdoctoral researcher at the Swiss Finance Institute, EPFL (Jan 2020–Mar 2022). She earned her PhD in 2019 and her Masters in Mathematics in 2014, both from IIT Delhi. Her research focuses on portfolio optimization; financial time-series forecasting; TDA in finance; insider trading; hidden cascade models.