**Angelos Filos**
****************
* .---. *
* | α +---+ *
* '--^' | *
* ^ | v *
* | | .-----. *
* | +-+ | *
* | | φ | *
* +---+ | *
* '-----' *
****************
Education
===================================================================
2018–Present
: *DPhil Candidate at University of Oxford*, Oxford UK
:
_Department of Computer Science_
: - *Supervisor*: Yarin Gal at Oxford Theoretical and Applied Machine Learning (OATML) group.
: - *Thesis*: Learning Sequential Strategies Under Non-Stationary Dynamics.
: - *Research Interests*: Meta-Reinforcement Learning, Multi-Agent Systems, Online Learning, Imitation Learning.
2014–2018
: *MEng at Imperial College London*, London UK
:
_Department of Electrical and Electronic Engineering, top student of the final class list_
: - *Dissertation*: Reinforcement Learning for Portfolio Management (arXiv, GitHub), supervised by Danilo P. Mandic.
: - *Courses*: Optimal Control Theory, Signal Processing, System Identification, Mathematics of Machine, Reinforcement Learning.
: - *Teaching Assistant*: Multivariate Calculus (EE1-10), Complex Calculus (EE2-08A), Probability Theory and Elements of Statistics (EE2-08B), Semiconductor Devices (EE1-03), Digital Design with Verilog (EE2-LABE), High-Level Programming (EE3-22).
2013–2014
: *Panhellenic Exams*, Athens Greece
:
_3rd General Lyceum of Neo Iraklio_
: - Science Stream 19.603/20.000.
: - *Courses*: Science Mathematics (20), Science Physics (19.8), Chemistry (19.6), Science Biology (19.8), Elements of Statistics and Probability (19.8), Modern Greek (17.4).
Experience
===================================================
2018–Present
: *Quantitative Researcher at JPMorgan Chase & Co.*, London UK
:
_Part-time employed during DPhil_
: - *Mechanism Design for Fair Markets*: $n$-players differentiable market games with convergence guarantees and fairness.
: - *Explicit Opponent Modeling in Auction Markets*: Analysis of the impact of opponent modeling on eponymous auction markets.
: - *Learning from Expert Demonstrations to Market Make*: Inverse reinforcement learning for limit order book derivatives trading.
: - *Systematic Derivatives Trading*: Portfolio optimization for equity derivative markets.
2017
: *Data Scientist at JPMorgan Chase & Co.*, London UK
:
_Six months internship_
: - *Anomaly Detection*: State-space time-series modeling for automated regime-switching and anomaly detection.
: - *Implied Volatility Prediction for Illiquid Assets*: Predictive modeling with conditional Generative Adversarial Networks (GANs) for forecasting implied volatility for illiquid assets, using financial indicators (e.g. fundamental factors).
: - *Analytics Platform*: Data visualization & monitoring platform build on web technologies and GitHub Electron.
2017-2018
: *Head of Talent Development at Imperial College Data Science Society*
:
_Volunteering work_
: - *Workshops*: Demystifying Artificial Intelligent concepts and teaching practical skills.
: - *Mentoring*: Leading and training the Imperial Advanced Data Science Team, participating in Hackathons and partnering with industrial sponsors to provide intelligence in their problems.
2016
: *Mixed Signal Engineer at Dialog Semiconductor*, Swindon UK
:
_Summer internship_
: - *Asynchronous Design*: Functionally described the operation of a legacy PMIC Buck Converter digital controller. Implemented a new digital controller using Petrify and MPSAT asynchronous logic design. Functionally verify the operation of the controller using UVM.
Publications
===============================================================
+ *Can Autonomous Vehicles Identify, Recover From, and Adapt to Distribution Shifts?*
*Angelos Filos*♮, Panagiotis Tigas♮, Rowan McAllister, Nicholas Rhinehart, Sergey Levine, Yarin Gal.
_Under review_.
Paper
+ *Invariant Causal Prediction for Block MDPs*
Amy Zhang♮, Clare Lyle♮, Shagun Sodhani, *Angelos Filos*, Marta Kwiatkowska, Joelle Pineau, Yarin Gal, Doina Precup.
_arXiv preprint arXiv:2003.06016_.
Paper
+ *Robust Imitative Planning: Planning from Demonstrations Under Uncertainty*
Panagiotis Tigas♮, *Angelos Filos*♮, Rowan McAllister, Nicholas Rhinehart, Sergey Levine, Yarin Gal.
_Machine Learning for Autonomous Driving NeurIPS 2019 Workshop_.
Paper
+ *A Systematic Comparison of Bayesian Deep Learning Robustness in Diabetic Retinopathy Tasks*
*Angelos Filos*, Sebastian Farquhar, Aidan N. Gomez, Tim G. J. Rudner, Zachary Kenton, Lewis Smith, Milad Alizadeh, Arnoud de Kroon, Yarin Gal.
_Spotlight at Bayesian Deep Learning NeurIPS 2019 Workshop_.
Paper Code
+ *On the Importance of Opponent Modelling in Auction Markets*
Mahmoud Mahfouz♮, *Angelos Filos*♮, Cyrine Chtourou, Joshua Lockhart, Samuel Assefa, Danilo Mandic, Tucker Balch, Manuela Veloso.
_Robust AI in Financial Services NeurIPS 2019 Workshop_.
Paper
+ *Generalizing from a few environments in safety-critical reinforcement learning*
Zachary Kenton, *Angelos Filos*, Owain Evans, Yarin Gal.
_SafeML ICLR 2019 Workshop_.
Paper Blog
+ *Towards Inverse Reinforcement Learning for Limit Order Book Dynamics*
Jacobo Roa-Vicens, Cyrine Chtourou, *Angelos Filos*, Francisco Rullan, Yarin Gal, Ricardo Silva.
_Contributed Talk at AI in Finance: Applications and Infrastructure for Multi-Agent Learning Workshop at ICML 2019_.
Paper
Open Source
===================================================================
bdl-benchmarks
ML Benchmark
: Bayesian deep learning benchmarks with a transparent, modular and
: consistent interface for the evaluation of deep probabilistic models.
qtrader
Algo Trading
: Reinforcement learning for portfolio management.
: MEng dissertation at Imperial College London, supervised by Danilo P. Mandic.
Service
==================================================================
Conference Workshop Organizing
------------------------------------------------------------------
- Multimodal Brain Tumor Segmentation (BraTS) Challenge @ MICCAI 2019
- Emergent Communication: Towards Natural Language @ NeurIPS 2019
Program Committees and Reviewing
-----------------------------------------------------------------
- Bayesian Deep Learning @ NeurIPS 2018
- International Conference on Machine Learning (ICML) 2019
- BrainLes @ MICCAI 2019
- Conference on Neural Information Processing Systems (NeurIPS) 2019
- Bayesian Deep Learning @ NeurIPS 2019
- International Conference on Machine Learning (ICML) 2020