What’s New? / About Me / Publications / Music

---

Welcome

I’m a researcher in statistical signal processing and machine learning. I work at Nokia Bell Labs, Espoo, Finland, where my research focuses on physical layer processing in wireless communications. Read more about me here.

What’s New?

• (On Nov 29, 2025) I attended the 33rd European Signal Processing Conference 2025 (EUSIPCO) in Palermo, Italy, 8-12 September. I co-authored with Prof. Esa Ollila and Dr. Xavier Mestre on the paper Approach to beamforming minimizing the signal power estimation error, which studies beamformers in terms of their signal estimation error, signal power estimation error, and signal power estimation bias. A feature of Capon and MMSE beamformers is that they either overestimate or underestimates the signal power. To address this issue, we propose Capon+ beamformer that strikes a favorable balance between signal estimation error, signal power estimation error, and signal power bias. We have recently also extended the paper to a journal version, which can be found from [arXiv].

• (On Sep 20, 2024) It was a pleasure to attend the 32nd European Signal Processing Conference 2024 (EUSIPCO) in Lyon, France, August 26-30, where I presented the paper Oracle approximating shrinkage covariance matrix estimators for complex elliptical distribution. The paper generalizes the popular Oracle Approximating Shrinkage (OAS) method by Chen et al. (2010) to complex elliptical (including complex Gaussian) distributions and robust shrinkage M-estimators of scatter matrix.

• (On Mar 14, 2023) I feel honored and grateful to have received the Aalto University Doctoral Thesis Award 2023.

• (On Aug 8, 2022) I’m excited to start in my new job as a Radio Research Engineer/Scientist at Nokia Bell Labs, Espoo, Finland.

• (On May 27, 2022) I successfully defended my thesis Contributions to theory and estimation of high-dimensional covariance matrices. The thesis proposes theory and methods for covariance matrix estimation, when the sample size is smaller than or comparable to the dimension of the data.

• (On Oct. 6, 2021) I delivered a talk On high-dimensional covariance matrix estimation for multiple classes at Statistical Learning for Signal and Image Processing (SLSIP) Workshop, Rüdesheim am Rhein, Germany, October 6th – 8th, 2021.

• (On Dec. 18, 2020) Prof. Esa Ollila delivered a talk about our work Linear Pooling of Sample Covariance Matrices in a virtual seminar at DataIA institute of University of Paris-Saclay. Check the [video] and [slides].