New Summer School in AI Coming Up!
Due to take place in Oxford on 31 August - 4 September
The new Summer School in AI is due to take place at Christ Church, Oxford, on 31 August - 4 September
Our unique Summer School in AI is due to take place at one of Oxford's oldest and most beautiful colleges, Christ Church, on 31 August - 4 September, 2019. Revise the basics and proceed to state-of-the-art data science and machine learning (ML) techniques during the most intensive training available anywhere in the world in only three working days and one weekend!
It is unique, because it blends the academic rigour of academics from the world's top universities with decades of practical experience of top data scientists and algorithmic traders who have applied their knowledge to the most challenging datasets.
The flipped classroom approach focusses attention on practical applications during the Tutorials based on real datasets, where the students (many of whom are seasoned data scientists, quantitative analysts, engineers, and developers) test their existing knowledge and, in collaboration, acquire new skills. These are followed by Lectures, where these skills are further strengthened through added intellectual rigour and state-of-the-art techniques are presented.
The programme concludes with a Hackathon, during which the students can apply their newly acquired skills to some of the most challenging datasets representing some of the most pressing open problems.
Discounts are available for existing clients, students, academics, and groups. Please contact us on education@thalesians.com for more information.
Furthermore, if you are interested in sponsoring the programme and contributing a dataset to our Hackathon (bearing in mind the calibre of our students, you can expect useful and nontrivial results), please contact us ASAP.
Register online while places are still available: https://ai.thalesians.com/courses/thalesians-ai/
The class of Spring 2019
Prof. Matthew Dixon explaining neural networks
Dr. Paul Bilokon teaching the mathematical prerequisites
Places still remain on our unique big data, high-frequency data, and machine learning with kdb+/q course!
q is a programming language for array processing, developed by Arthur Whitney on the basis of Kenneth E. Iverson’s APL. The kdb+ database built on top of q is a de facto standard technology for dealing with rapidly arriving, high-frequency, big data.
kdb+/q has taken the world of electronic, including algorithmic, trading by storm. It is used by numerous sell-side and buy-side institutions, including some of the most successful hedge funds and electronic market makers.
kdb+/q experts are highly regarded and are actively sought in the City of London, on New York's Wall Street, and centres of high-frequency trading, such as Chicago.
Beyong the world of electronic trading, kdb+/q is used in retail, gaming, manufacturing, telco, IoT, life sciences, utilities, and aerospace industries.
Your course will take you through the foundations of kdb+/q and explain why it is a language of choice for Big Data, high-frequency data, and real-time event processing.
We shall explain how to work with tables and q-sql effectively, how to set up tickerplants, real-time, and historical instances, and how to apply kdb+/q to machine learning problems.
We shall consider advanced applications to tree-based regression and classification, random forests, deep learning, Google DeepMind and Monte Carlo search, producing demonstrations on real-life data examples.
A NEW COURSE: Bayesian methods, filtering, and Markov chain Monte Carlo
Suppose that there is a latent, unobserved process, X, and an observable process Y, which, in a certain specific sense, depends on X. Both X and Y evolve over time. The filtering problem consists in using our observations of Y to estimate, in some optimal sense, the state of X. This is, in a nutshell, the filtering problem. The information about the current state of X can also be used to forecast both X and Y.
We may further suppose that X and/or Y depend on certain parameters, θ. Our task may be complicated by the need to estimate these parameters. We may either use Frequentist methods of their estimation, relying on maximum likelihood, or Bayesian methods, such as Markov chain Monte Carlo.
We shall consider both approaches and introduce Markov chain Monte Carlo packages such as WinBUGS and PyMC3, which greatly simplify the task of parameter estimation.
The filtering problem was first introduced in the context of radar communications, ballistics, and rocket science. The first stochastic filters were used in the Apollo 11 programme to help land on the Moon. Particle filters are nowadays used in computer vision and self-driving cars, and modern econometrics. In our course we shall also consider applications of stochastic filtering to finance.
You can register online for this course: https://ai.thalesians.com/courses/bayesian-methods-filtering-and-markov-chain-monte-carlo/
The Machine Learning and Big Data with kdb+/q book is available for pre-order
Upgrade your programming language to more effectively handle high-frequency data, ML, and Big Data with kdb+q. The book offers quants, programmers, and algorithmic traders a practical entry into the powerful but non-intuitive kdb+ database and q programming language. Ideally designed to handle the speed and volume of high-frequency financial data at sell- and buy-side institutions, these tools have become the de facto standard; this book provides the foundational knowledge that practitioners need to work effectively with this rapidly-evolving approach to analytical trading. The discussion follows the natural progression of working strategy development to allow hands-on learning in a familiar sphere, illustrating the contrast of efficiency and capability between the q language and other programming approaches. Rather than an all-encompassing "bible"-type reference, this book is designed with a focus on real-world practicality to help you quickly get up to speed and become productive with the language. Understand why kdb+/q is the ideal solution for high-frequency data. Delve into the "meat" of q programming to solve practical economic problems. Perform everyday operations, including basic regressions, cointegration, volatility estimation, modelling and more. Learn advanced techniques from market impact and microstructure analyses to machine learning techniques including neural networks. The kdb+/q database and its programming language q offer unprecedented speed and capability. As trading algorithms and financial models grow ever more complex against the markets they seek to predict, they encompass an even-larger swath of data - more variables, more metrics, more responsiveness, and altogether more "moving parts". Traditional programming languages are increasingly failing to accommodate the growing speed and volume of data, and lack the necessary flexibility that cutting-edge financial modelling demands. Machine Learning and Big Data with kdb+/q opens up the technology and flattens the learning curve to help you quickly adopt a more effective set of tools.
Our article "C++ or Java? Which is best for trading systems?" has just been published in eFinancialCareers
Should you just clean up your Python code from the Jupyter notebook (make it work, make it right, make it fast), package it as a collection of modules - as a library - and put it into production for real-time use by others?
Sometimes the answer is yes. More often than not, though, it is 'no'. Python is not an ideal language for production use. It is slower than many other languages. There's that global interpreter lock (GIL). Python's duck- and dynamic typing are a double-edged sword: those annoying checks that you were happy to forego while prototyping could come back to bite you when a corner case surfaces in real-life use. And they always do surface.
So, you begin to ponder another question: C++ or Java?
Read on on eFinancialCareers: https://news.efinancialcareers.com/uk-en/3001368/c-or-java
A delegation of the Thalesians attends a meeting of the All-Party Parliamentary Group meeting on Artificial Intelligence
The APPG AI is co-chaired by Stephen Metcalfe MP and Lord Clement-Jones CBE. The Group Officers are Chris Green MP, The Right Reverend Doctor Steven Croft, Baroness Kramer, Lord Janvrin, Lord Broers, Lord Holmes of Richmond, Lord Willetts, Baroness McGregor-Smith, Mark Hendrick MP and Carol Monaghan MP.
Big Innovation Centre is the APPG AI Secretariat.
The Thalesians present neocybernetics at UK China AI Innovation Forum
As a technology which can change human life in the future, artificial intelligence (AI) has attracted intensive attention in the global world. And the globalisation is an inevitable trend for AI development. This forum is focussed on the AI globalisation and business cooperation, especially the AI International cooperation between UK and China.
The delegates reviewed the current status of AI in both UK and China, and key areas of unmet needs in the next 3 years, cross-border technology collaboration between the UK and China and the key lessons, private capital/VC investment into AI from China, co-development and collaborations in AI with chinese partners, and commercialisation of AI in China.
Our special thanks go to Mona Ye Zhang for welcoming us to this event.
Dr. Paul A. Bilokon addresses the Cambridge University Algorithmic Trading Society
Most of the present advances in ML/AI have been on natural language and images. It is time to address the Holy Grail: noisy time series. Dr. Bilokon showed how how this quest is in line with Norbert Wiener's original work on cybernetics and explained how Thalesians Ltd are using our experience from building some of the world's best high- and medium-frequency trading systems to advance ML/AI in this area.
Learn linear algebra from Thalesians on YouTube!
In data science, machine learning (ML), and artificial intelligence (AI), we usually deal not with single numbers but with multivariate (i.e. containing multiple elements or entries) lists of numbers - mathematically speaking, vectors, - and multivariate tables of numbers - mathematically speaking, matrices. Therefore we solve multivariate equations, apply multivariate calculus to find optima of multivariate functions, etc.
The branch of mathematics that studies vectors, matrices, and related mathematical objects is called linear algebra. It is one of the most practically useful areas of mathematics in applied work and a prerequisite for data science, ML, and AI.
In this video, we consider numbers as examples of mathematical objects; introduce a different kind of mathematical object - vector - first in two dimensions; demonstrate the importance of two-dimensional vectors in data science; introduce vector arithmetics: vector addition and multiplication by scalars; show how vectors and vector arithmetics can be implemented in Python; introduce the vector norm and relate it to the length of a vector; introduce the inner product and relate it to the angle between two vectors; consider vectors in three dimensions; show how vectors can be generalised to four- and higher-dimensional spaces; demonstrate the importance of higher-dimensional vectors in data science; consider vector spaces in general (i.e. not just the Euclidean vector spaces); show that functions also form a vector space; introduce linear combinations and examine the notions of linear (in)dependence, span, and basis; introduce subspaces; explain how one can obtain the equation of a (hyper)plane.
Rebranding
And yet, being rendered in a distinctly modern way, this particular column is in line with our use of the technology of the future.
The column is also reminiscent of the mortarboard hat, worn by scholars, and underlining our dedication to education.
In an alternative rendering, the lines of the column become connections in a dependency graph or a neural net, some of our objects of study, that we apply to time series data.
Finally, the notion of time is hinted at in the timelessness, the eternity of the Doric column.
Special thanks to Serhiy of design_13 for helping us with this rebranding: https://99designs.co.uk/profiles/1663775/about
Support for nanosecond precision timestamps in Java!
The advent of high-frequency trading had led to systems that dealt with messages with ever increasing frequency. First, intraday prices were processed. Eventually messages were considered at sub-millisecond frequency: first, several microseconds, and now nanoseconds. The ability to handle such messages is useful in fields outside high-frequency algorithmic/electronic trading, such as robotics, and electronic medicine.
Up until recently there was no support for timestamps with nanosecond precision in Java. We decided to change this by developing and open-sourcing a new Java library.
The key data type is NanoDateTime (but see also NanoDate and NanoTime). The data types were designed for immutability in order to be usable as value times in messages in neocybernetic systems.
If you are interested in reviewing this library and/or contributing to it, please contact us on info@thalesians.com
Using PyMC3 to calibrate time series models
Thalesians Ltd
The Level39 member fintech Thalesians Ltd are an artificial intelligence (AI) and machine learning (ML) company, founded on ancient philosophy, modern science, and the technology of the future.
We believe that we can create a healthier, more sustainable human environment by applying the lessons we are learning from building some of the world's most successful high- and medium-frequency trading systems to a wider range of data sets.
We specialise in dealing with time series data – by far the most challenging application of ML/AI.
Time series are sequences of timestamped updates, arriving in chronological order, on the state of a particular process evolving over time.
Time series data may arrive asynchronously from different sources requiring high-throughput, low-latency processing, often on the scale of nanoseconds. Individual time series may have low signal-to-noise ratios, but when taken together and processed intelligently, time series yield valuable insights.
Time series arise naturally in finance (stock prices, corporate bond prices/yields, interest rates, currency exchange rates, cryptocurrency exchange rates), economics (micro- and macroeconomic data), engineering (states of a particular machine or mechanism, e.g. systems operating on a ship, airplane, or on a space station), medicine (electrocardiogram (ECG) and electroencephalogram (EEG) test results, any medical test results evolving over time, fitband readings, at the microscopic level, metabolic chain states).
We create the new science – built on the foundation of quantitative finance, algorithmic trading, high-frequency trading and markets microstructure, ML, deep learning (DL), deep reinforcement learning (DRL), AI, reactive programming, and big data – neocybernetics, and use it to revolutionise finance, economics, insurance, tranportation, shipping, engineering, and medicine in the United Kingdom and worldwide.
We do this by implementing real-time ML/AI software for neocybernetic systems, providing education, and consulting services.
Email: info@thalesians.com
Website: https://ai.thalesians.com/
Location: Level39, London, UK
Phone: +44 (0)20 796 57587
Facebook: facebook.com/TheThalesians/
Twitter: @thalesians