Oxford-Man Institute of Quantitative Finance, University of Oxford

Congratulations to Fernando Moreno-Pino, PhD, Alvaro Arroyo, Harrison Waldon, Xiaowen Dong, and Álvaro Cartea! Their paper “Rough Transformers: Lightweight Continuous-Time Sequence Modelling with Path Signatures” has been accepted at NeurIPS 2024. They present Rough Transformers, a more efficient Transformer that operates on continuous-time representations of time-series data and is robust to irregular sampling: • They achieve significant compression without compromising performance by modifying the input representation rather than sparsifying the attention mechanism. • They utilize the Signature Transform to capture local and global temporal patterns, compactly describing the properties of the time-series. • Their method enhances sample efficiency, accuracy, and is computationally efficient, reducing training costs by up to 25x compared to Neural ODE models while remaining competitive with SSMs and other baselines. • Rough Transformers also demonstrate robustness to irregular sampling, showing strong performance across different drop percentages. Link to paper: https://lnkd.in/dZbgxKuS

Our director Álvaro Cartea will be giving a talk at The Chinese University of Hong Kong Distinguished Lectures in Quantitative Finance on the 26th of November. The talk will also be online, details given below.

The OMI is delighted to share its November 2024 Newsletter. Link to the PDF: https://lnkd.in/gmBMj75i #machinelearning #quantitativefinance 

The OMI together with the Man Group are excited to be hosting a Workshop on New Challenges and Risks in Finance on the 14th of November 2024. The workshop will provide an opportunity for junior and senior leading academics to discuss new challenges and risks that investors face in an ever-evolving financial landscape. We have a great lineup of speakers: Ricardo De la O, Javier Gil-Bazo, Qi Jin, Robert L. Kosowski, and Mungo Wilson. Register on the conference website: https://cvent.me/00GYRD

Thank you to everyone who joined our Financial Economics and Microstructure Workshop. We had a number of great talks from Álvaro Cartea, Patrick Chang, Carole Comerton-Forde, Winston Dou, Giuliano Graziani, Albert J. Menkveld, Joshua Mollner, Chaojun Wang, and Basil Williams. Link to photos from the event: https://lnkd.in/gTNc88Bk

The OMI together with the Man Group are excited to be hosting a Workshop on New Challenges and Risks in Finance on the 14th of November 2024. The workshop will provide an opportunity for junior and senior leading academics to discuss new challenges and risks that investors face in an ever-evolving financial landscape. We have a great lineup of speakers: Ricardo De la O, Javier Gil-Bazo, Qi Jin, Robert L. Kosowski, and Mungo Wilson. Register on the conference website: https://cvent.me/00GYRD

The OMI is delighted to share its October 2024 Newsletter. Link to the PDF: https://lnkd.in/ghqzkz7e #machinelearning #quantitativefinance 

Congratulations to John J. Hopfield and Geoffrey E. Hinton. They have been awarded the Nobel Prize in Physics for their foundational discoveries and inventions that enable machine learning with artificial neural networks. We are excited for their recognition. Machine Learning and Data Science have quickly become fundamental tools in modern finance.

The OMI in conjunction with Harvard University is excited to be hosting the Oxford—Harvard Conference on Decentralised Finance and Market Microstructure on the 20th and 21st of January 2025. Blockchain technology, asset digitisation, and automated market makers are reshaping the future of the financial landscape and challenging traditional stakeholders. The conference will provide an opportunity for junior and senior leading academics in Finance, Economics, Computer Science, and industry to discuss the emerging challenges in this field with a focus on the microstructure and the design of decentralised marketplaces. Registration is open on the conference website: https://lnkd.in/giiPTdpF Organisers and scientific committee: Agostino Capponi, Álvaro Cartea, Fayçal Drissi, Matheus Venturyne Xavier Ferreira, PhD, David Parkes. 

Congratulations to Álvaro Cartea, Sam Cohen, Rob Graumans, Saad Labyad, Leandro Sánchez-Betancourt, and Leon van Veldhuijzen. Their paper “Statistical Predictions of Trading Strategies in Electronic Markets” has been accepted in the Journal of Financial Econometrics. Using proprietary data (with algorithm identification and member identification) from Euronext exchange: • They obtain high out-of-sample accuracies when predicting the behaviour of trading algorithms. • They analyse key features that trading algorithms use to make decisions. • They find three clusters of stylised trading behaviour: directional traders, opportunistic traders, and market markers. • Only one third of firms registered as liquidity providers fall within the cluster of market makers. Link to paper: https://lnkd.in/edtih3Hs