The Hudson & Thames / Thalesians Evening Conference on Algorithmic Trading & ML

PLEASE NOTE THAT THIS EVENING CONFERENCE WILL START ON TUESDAY, 19 JULY, 2022, AT 6:30 PM BST (LONDON TIME) [1.30 PM EDT (NEW YORK TIME)]

Proceeds from this event will be donated to withukraine.org - the official fundraising effort by the Embassy of Ukraine to the United Kingdom.

FULL TITLE: The Hudson & Thames / Thalesians Evening Conference on Algorithmic Trading & Machine Learning

18:30 - 19:15: Hudson & Thames – Jacques Joubert (Co-Founder)

Title: Meta-labeling

Abstract

Meta-labeling is a machine learning (ML) layer that sits on top of a base primary strategy, to help size positions, filter out false-positive signals, and improve metrics such as the Sharpe ratio and maximum drawdown. This article consolidates the knowledge of from several publications into a single work, providing practitioners with a clear framework to support the application of meta-labeling to investment strategies. The relationships between binary classification metrics and strategy performance are explained, alongside answers to many frequently asked questions regarding the technique. The author also deconstructs meta-labeling into three components, using a controlled experiment to show how each component helps to improve strategy metrics and what types of features should be considered in the model specification phase.

19:15 - 20:00: Hudson & Thames – Oleksandr Proskurin (Co-Founder)

Title: Using Absolute Rules of Disqualification to Detect Mean-reverting Spreads for Pairs Trading and Statistical Arbitrage

Abstract

There are various ways to filter out pairs and spreads for long/short, statistical arbitrage strategies. However, cointegration is one of the most widespread, well-researched approaches many practitioners apply. ARODS framework allows us to apply all critical cointegration filters (ADF, half-life of mean-reversion, Hurst exponent, etc.) in one framework to choose stationary spreads and maximise the probability of preserving their mean-reverting nature on out-of-sample. In this lecture, Alex will describe the framework in detail and apply it to the universe of US large-cap stocks to find mean-reverting spreads.

20:00 - 20:45 Thalesians – Saeed Amen (Co-Founder)

Title: Speeding up Python

Abstract

Python isn't the fastest language. However, there are many ways we can speed up Python code! In this talk, we shall discuss various tips and tricks to speed up financial calculations, showing how techniques like Numba can yield a very large speed up and also using parallelization with tools like Dask. We shall present several use cases from finance including from option pricing.

20:45 - 21:30: Thalesians – Adithya Sireesh

Title: The power of NISQ algorithms

Abstract

Back in the 1980s, the notable physicist Richard Feynman was thinking about a new kind of computer that could simulate the dynamics of very small systems running on the principles of quantum mechanics. What stemmed from studying that question was the field of quantum computing. In the four decades since then, quantum researchers and companies have made immense strides both on the quantum hardware and quantum algorithms, with one of the most notable algorithms being the Shor’s factoring algorithm. However, since these algorithms require very accurate and error corrected computations known as “Fault Tolerance”, we might still be a few decades away from running these theoretically proposed algorithms such as Shor's factoring. However, these constraints lead to the study of a new class of algorithms known as NISQ (Noisy Intermediate Scale Quantum) algorithms that could be run even on the noisy and error prone quantum computers (Near term) we have right now, and have found applications in areas such as quantum chemistry, combinatorial optimisation and finance. What other important problems can we solve on near term quantum computers, and what lies ahead for the future of quantum computation?