aiqbench analysis
This research demonstrates how to optimize Transformer inference on FPGAs specifically for real-time outlier detection in financial time series, addressing latency and accuracy challenges caused by noisy data. By leveraging the reconfigurability of hardware like PYNQ-Z2 boards, the study provides a practical pathway to deploy robust anomaly detection systems that can handle increasing volumes of market data without downstream instability.
Optimizing Transformer Neural Network for Real-Time Outlier Detection on FPGAs
arXiv:2607.22786v1 Announce Type: cross Abstract: In this work, we explore how the inference time of a Transformer Neural Network can be efficiently optimized with applications to real-time anomaly detection in financial time series. The financial time series are price series such as asset prices. Unfortunately, the data is often with errors or outliers that make the downstream data processing tasks useless, unstable or even harmful. Moreover, the amount of financial time-series data has been significantly increasing. Hence, there is a need for better data-cleaning methods in terms of accuracy and in terms of processing speed. Transformers as a neural network architecture have achieved superior performances in many tasks such as Natural Language Processing and Computer Vision. Time series modelling and especially anomaly detection tasks can benefit from the features of transformers architecture in multiple ways, including the capacity to capture long-range dependencies and interactions. Increasingly powerful hardware, such as field-programmable gate arrays (FPGAs), have seen increasing usage in recent years due to their reconfigurability and high performance. They can be efficiently utilized to speed up the computations of the Transformer architecture. We explore different Transformer architectures for time series modelling and how they can be efficiently implemented on an FPGA board (PYNQ-Z2). In particular, we examine the application of Transformers to detect anomalies in time series and we show how they can be efficiently implemented on an FPGA board to minimize latency. The code is available at https://github.com/thxi/icl_thesis
Abstract:In this work, we explore how the inference time of a Transformer Neural Network can be efficiently optimized with applications to real-time anomaly detection in financial time series. The financial time series are price series such as asset prices. Unfortunately, the data is often with errors or outliers that make the downstream data processing tasks useless, unstable or even harmful. Moreover, the amount of financial time-series data has been significantly increasing. Hence, there is a need for better data-cleaning methods in terms of accuracy and in terms of processing speed. Transformers as a neural network architecture have achieved superior performances in many tasks such as Natural Language Processing and Computer Vision. Time series modelling and especially anomaly detection tasks can benefit from the features of transformers architecture in multiple ways, including the capacity to capture long-range dependencies and interactions. Increasingly powerful hardware, such as field-programmable gate arrays (FPGAs), have seen increasing usage in recent years due to their reconfigurability and high performance. They can be efficiently utilized to speed up the computations of the Transformer architecture. We explore different Transformer architectures for time series modelling and how they can be efficiently implemented on an FPGA board (PYNQ-Z2). In particular, we examine the application of Transformers to detect anomalies in time series and we show how they can be efficiently implemented on an FPGA board to minimize latency. The code is available at this https URL
Machine Learning (cs.LG); Artificial Intelligence (cs.AI); Hardware Architecture (cs.AR); Distributed, Parallel, and Cluster Computing (cs.DC); Performance (cs.PF); Computation (stat.CO)
I.2.6; C.3; I.5.4
arXiv:2607.22786 [cs.LG]
(or arXiv:2607.22786v1 [cs.LG] for this version)
https://doi.org/10.48550/arXiv.2607.22786 arXiv-issued DOI via DataCite (pending registration)
The Journal of FinTech, Vol. 5, No. 1, 2550001 (2025)
https://doi.org/10.1142/S2705109925500014 DOI(s) linking to related resources
Full text under the publisher's open licence · 288 words
