24 Oct 2024
Two Works at ECAI 2024
Xinpeng Hong celebrates two accepted in-network ML papers
Congratulations of DPhil student Xinpeng Hong, who presented two papers at ECAI 2024.
Xinpeng's first work, MIND, detects transaction fraud in real time using in-network machine learning, processing 4.8 billion transactions per second with micro-scale latency.
This joint work with Changgang Zheng and Noa Zilberman.
Xinpeng's second work, accelerates machine learning inference for trading using programmable switches, lowering average latency by 45% with almost no loss of accuracy. If shows that not only stock mid-price can be predicted within the switch, but also other aspects such as volatility.
This joint work with Changgang Zheng, Stefan Zohren and Noa Zilberman.
Xinpeng received EurAI Travel Grant and AISB Student Grant for these works.
Congratulations Xinpeng!
