Workshop 40: Smart Internet of ThingsTitle: Lung Cancer Detection by the Analysis of A Biomarker Toluene Based on Multi-task Simpleformer Model
Workshop 40: Smart Internet of Things
Title: Lung Cancer Detection by the Analysis of A Biomarker Toluene Based on Multi-task Simpleformer Model
Keywords: Lung Cancer, Electronic Nose, Multi-task, Transformer, Toluene
Summary:
Lung cancer is one of the most threatening malignancies to human health and life. Early detection and treatment can improve the five-year survival rate. As a key biomarker for lung cancer diagnosis, toluene can be used for breath simulation analysis of lung cancer patients, which can complement the existing lung cancer diagnosis methods and improve the diagnosis rate of lung cancer. In this paper, a multi-task simple transformer model is proposed in the electronic nose system to conduct gas-sensing experiments on toluene in low-concentration environments. Encoder blocks are shared and two tasks are performed simultaneously, achieving fast and accurate target recognition and concentration prediction at low concentrations. The accuracy of the model on the target recognition task is 98.52%, which is better than the K-Nearest Neighbors (KNN) (89.06%) and the Support Vector Machine (SVM) (91.17%). The root mean square error on the concentration prediction task is 0.0369ppm, outperforming the Support Vector Regression (SVR) (0.0898ppm) and the Extreme Gradient Boosting (XGBoost) (0.0725ppm). The experimental results show that the MT-Simpleformer has good prediction performance for the identification and concentration of a lung cancer biomarker (toluene) under dynamic, low-concentration conditions. It provides an effective reference method for rapid, economical, and non-invasive lung cancer screening.
Chair: Prof. Yatao Yang
| Shenzhen Univeristy
Yatao Yang received his B.Sc. degree in optical instrumentation engineering from Zhejiang University, China, and Ph.D. degree in fiber optics from Glasgow Caledonia University, U.K.. After B.Sc. he joined Institute of Optics and Electronics, Chinese Academy of Sciences, where he was involved in semiconductor equipment and optoelectronic device development. In 1996, he was with the University of Leeds, U.K., as a research officer involved in areas of optical fiber laser materials. In 1997, he joined Resonance Ltd., in Canada, where he was developing spectral gas sensors. In 1998, he joined JDSU Corporation, Canada, where he was developing optical fiber devices. In 2000, he joined Chorum Technologies Inc., US, where he was developing optical fiber devices. In 2004, he joined JDSU Corporation, US, where he was developing optical fiber devices and fiber lasers. In 2009, he joined NeoPhotonics Corporation, as VP of R&D, developing optical fiber devices. In2014, he founded Shenzhen Dade Laser Technology Co. , Ltd. In 2017, he was appointed as a Distinguished Professor with the College of Information Engineering, Shenzhen University, China. In 2018, he became the Head of the Smart IoT Center, Shenzhen University. His research interests include optical networking, optical sensors, optical data transmission and data processing, optical nanomaterials, optical-wireless communications, and lasers.