Imperial College London

DrKezhiLi

Faculty of EngineeringDepartment of Electrical and Electronic Engineering

Visiting Researcher
 
 
 
//

Contact

 

+44 (0)7859 995 590kezhi.li

 
 
//

Location

 

Royal School of MinesSouth Kensington Campus

//

Summary

 

Publications

Publication Type
Year
to

50 results found

Zhu T, Kuang L, Piao C, Zeng J, Li K, Georgiou Pet al., 2024, Population-Specific Glucose Prediction in Diabetes Care With Transformer-Based Deep Learning on the Edge., IEEE Trans Biomed Circuits Syst, Vol: PP

Leveraging continuous glucose monitoring (CGM) systems, real-time blood glucose (BG) forecasting is essential for proactive interventions, playing a crucial role in enhancing the management of type 1 diabetes (T1D) and type 2 diabetes (T2D). However, developing a model generalized to a population and subsequently embedding it within a microchip of a wearable device presents significant technical challenges. Furthermore, the domain of BG prediction in T2D remains under-explored in the literature. In light of this, we propose a population-specific BG prediction model, leveraging the capabilities of the temporal fusion Transformer (TFT) to adjust predictions based on personal demographic data. Then the trained model is embedded within a system-on-chip, integral to our low-power and low-cost customized wearable device. This device seamlessly communicates with CGM systems through Bluetooth and provides timely BG predictions using edge computing. When evaluated on two publicly available clinical datasets with a total of 124 participants with T1D or T2D, the embedded TFT model consistently demonstrated superior performance, achieving the lowest prediction errors when compared with a range of machine learning baseline methods. Executing the TFT model on our wearable device requires minimal memory and power consumption, enabling continuous decision support for more than 51 days on a single Li-Poly battery charge. These findings demonstrate the significant potential of the proposed TFT model and wearable device in enhancing the quality of life for people with diabetes and effectively addressing real-world challenges.

Journal article

Zhu T, Li K, Georgiou P, 2023, Offline Deep Reinforcement Learning and Off-Policy Evaluation for Personalized Basal Insulin Control in Type 1 Diabetes., IEEE J Biomed Health Inform, Vol: 27, Pages: 5087-5098

Recent advancements in hybrid closed-loop systems, also known as the artificial pancreas (AP), have been shown to optimize glucose control and reduce the self-management burdens for people living with type 1 diabetes (T1D). AP systems can adjust the basal infusion rates of insulin pumps, facilitated by real-time communication with continuous glucose monitoring. Deep reinforcement learning (DRL) has introduced new paradigms of basal insulin control algorithms. However, all the existing DRL-based AP controllers require extensive random online interactions between the agent and environment. While this can be validated in T1D simulators, it becomes impractical in real-world clinical settings. To this end, we propose an offline DRL framework that can develop and validate models for basal insulin control entirely offline. It comprises a DRL model based on the twin delayed deep deterministic policy gradient and behavior cloning, as well as off-policy evaluation (OPE) using fitted Q evaluation. We evaluated the proposed framework on an in silico dataset generated by the UVA/Padova T1D simulator, and the OhioT1DM dataset, a real clinical dataset. The performance on the in silico dataset shows that the offline DRL algorithm significantly increased time in range while reducing time below range and time above range for both adult and adolescent groups. Then, we used the OPE to estimate model performance on the clinical dataset, where a notable increase in policy values was observed for each subject. The results demonstrate that the proposed framework is a viable and safe method for improving personalized basal insulin control in T1D.

Journal article

Zhu T, Li K, Herrero P, Georgiou Pet al., 2023, GluGAN: Generating Personalized Glucose Time Series Using Generative Adversarial Networks., IEEE J Biomed Health Inform, Vol: 27, Pages: 5122-5133

Time series data generated by continuous glucose monitoring sensors offer unparalleled opportunities for developing data-driven approaches, especially deep learning-based models, in diabetes management. Although these approaches have achieved state-of-the-art performance in various fields such as glucose prediction in type 1 diabetes (T1D), challenges remain in the acquisition of large-scale individual data for personalized modeling due to the elevated cost of clinical trials and data privacy regulations. In this work, we introduce GluGAN, a framework specifically designed for generating personalized glucose time series based on generative adversarial networks (GANs). Employing recurrent neural network (RNN) modules, the proposed framework uses a combination of unsupervised and supervised training to learn temporal dynamics in latent spaces. Aiming to assess the quality of synthetic data, we apply clinical metrics, distance scores, and discriminative and predictive scores computed by post-hoc RNNs in evaluation. Across three clinical datasets with 47 T1D subjects (including one publicly available and two proprietary datasets), GluGAN achieved better performance for all the considered metrics when compared with four baseline GAN models. The performance of data augmentation is evaluated by three machine learning-based glucose predictors. Using the training sets augmented by GluGAN significantly reduced the root mean square error for the predictors over 30 and 60-minute horizons. The results suggest that GluGAN is an effective method in generating high-quality synthetic glucose time series and has the potential to be used for evaluating the effectiveness of automated insulin delivery algorithms and as a digital twin to substitute for pre-clinical trials.

Journal article

Zhu T, Kuang L, Daniels J, Herrero P, Li K, Georgiou Pet al., 2023, IoMT-Enabled Real-Time Blood Glucose Prediction With Deep Learning and Edge Computing, IEEE INTERNET OF THINGS JOURNAL, Vol: 10, Pages: 3706-3719, ISSN: 2327-4662

Journal article

Zhu T, Li K, Herrero P, Georgiou Pet al., 2023, Personalized blood glucose prediction for Type 1 diabetes using evidential deep learning and meta-learning., IEEE Transactions on Biomedical Engineering, Vol: 70, Pages: 193-204, ISSN: 0018-9294

The availability of large amounts of data from continuous glucose monitoring (CGM), together with the latest advances in deep learning techniques, have opened the door to a new paradigm of algorithm design for personalized blood glucose (BG) prediction in type 1 diabetes (T1D) with superior performance. However, there are several challenges that prevent the widespread implementation of deep learning algorithms in actual clinical settings, including unclear prediction confidence and limited training data for new T1D subjects. To this end, we propose a novel deep learning framework, Fast-adaptive and Confident Neural Network (FCNN), to meet these clinical challenges. In particular, an attention-based recurrent neural network is used to learn representations from CGM input and forward a weighted sum of hidden states to an evidential output layer, aiming to compute personalized BG predictions with theoretically supported model confidence. The model-agnostic meta-learning is employed to enable fast adaptation for a new T1D subject with limited training data. The proposed framework has been validated on three clinical datasets. In particular, for a dataset including 12 subjects with T1D, FCNN achieved a root mean square error of 18.64±2.60 mg/dL and 31.07±3.62 mg/dL for 30 and 60-minute prediction horizons, respectively, which outperformed all the considered baseline methods with significant improvements. These results indicate that FCNN is a viable and effective approach for predicting BG levels in T1D. The well-trained models can be implemented in smartphone apps to improve glycemic control by enabling proactive actions through real-time glucose alerts.

Journal article

Zhu T, Chen T, Kuang L, Zeng J, Li K, Georgiou Pet al., 2023, Edge-Based Temporal Fusion Transformer for Multi-Horizon Blood Glucose Prediction, 56th IEEE International Symposium on Circuits and Systems (ISCAS), Publisher: IEEE, ISSN: 0271-4302

Conference paper

King Z, Farrington J, Utley M, Kung E, Elkhodair S, Harris S, Sekula R, Gillham J, Li K, Crowe Set al., 2022, Machine learning for real-time aggregated prediction of hospital admission for emergency patients, NPJ DIGITAL MEDICINE, Vol: 5, ISSN: 2398-6352

Journal article

Thygesen JH, Tomlinson C, Hollings S, Mizani MA, Handy A, Akbari A, Banerjee A, Cooper J, Lai AG, Li K, Mateen BA, Sattar N, Sofat R, Torralbo A, Wu H, Wood A, Sterne JAC, Pagel C, Whiteley WN, Sudlow C, Hemingway H, Denaxas Set al., 2022, COVID-19 trajectories among 57 million adults in England: a cohort study using electronic health records, LANCET DIGITAL HEALTH, Vol: 4, Pages: E542-E557

Journal article

Zhu T, Uduku C, Li K, Herrero Vinas P, Oliver N, Georgiou Pet al., 2022, Enhancing self-management in type 1 diabetes with wearables and deep learning, npj Digital Medicine, Vol: 5, ISSN: 2398-6352

People living with type 1 diabetes (T1D) require lifelong selfmanagement to maintain glucose levels in a safe range. Failure to do socan lead to adverse glycemic events with short and long-term complications. Continuous glucose monitoring (CGM) is widely used in T1Dself-management for real-time glucose measurements, while smartphoneapps are adopted as basic electronic diaries, data visualization tools, andsimple decision support tools for insulin dosing. Applying a mixed effectslogistic regression analysis to the outcomes of a six-week longitudinalstudy in 12 T1D adults using CGM and a clinically validated wearablesensor wristband (NCT ID: NCT03643692), we identified several significant associations between physiological measurements and hypo- andhyperglycemic events measured an hour later. We proceeded to developa new smartphone-based platform, ARISES (Adaptive, Real-time, and Intelligent System to Enhance Self-care), with an embedded deep learning algorithm utilizing multi-modal data from CGM, daily entries of mealand bolus insulin, and the sensor wristband to predict glucose levels andhypo- and hyperglycemia. For a 60-minute prediction horizon, the proposed algorithm achieved the average root mean square error (RMSE)of 35.28±5.77 mg/dL with the Matthews correlation coefficients fordetecting hypoglycemia and hyperglycemia of 0.56±0.07 and 0.70±0.05,respectively. The use of wristband data significantly reduced the RMSEby 2.25 mg/dL (p < 0.01). The well-trained model is implemented onthe ARISES app to provide real-time decision support. These resultsindicate that the ARISES has great potential to mitigate the risk ofsevere complications and enhance self-management for people with T1D.

Journal article

King Z, Farrington J, Utley M, Kung E, Elkhodair S, Harris S, Sekula R, Gillham J, Li K, Crowe Set al., 2022, Machine Learning for Real-Time Aggregated Prediction of Hospital Admission for Emergency Patients

<jats:title>Abstract</jats:title><jats:p>Machine learning for hospital operations is under-studied. We present a prediction pipeline that uses live electronic health-records for patients in a UK teaching hospital’s emergency department (ED) to generate short-term, probabilistic forecasts of emergency admissions. A set of XGBoost classifiers applied to 109,465 ED visits yielded AUROCs from 0.82 to 0.90 depending on elapsed visit-time at the point of prediction. Patient-level probabilities of admission were aggregated to forecast the number of admissions among current ED patients and, incorporating patients yet to arrive, total emergency admissions within specified time-windows. The pipeline gave a mean absolute error (MAE) of 4.0 admissions (mean percentage error of 17%) versus 6.5 (32%) for a benchmark metric. Models developed with 104,504 later visits during the Covid-19 pandemic gave AUROCs of 0.68-0.90 and MAE of 4.2 (30%) versus a 4.9 (33%) benchmark. We discuss how we surmounted challenges of designing and implementing models for real-time use, including temporal framing, data preparation, and changing operational conditions.</jats:p>

Journal article

Shen F, Wang Z, Ding G, Li K, Wu Qet al., 2022, 3D Compressed Spectrum Mapping With Sampling Locations Optimization in Spectrum-Heterogeneous Environment, IEEE TRANSACTIONS ON WIRELESS COMMUNICATIONS, Vol: 21, Pages: 326-338, ISSN: 1536-1276

Journal article

Thygesen JH, Tomlinson C, Hollings S, Mizani M, Handy A, Akbari A, Banerjee A, Cooper J, Lai A, Li K, Mateen B, Sattar N, Sofat R, Torralbo A, Wu H, Wood A, Sterne JAC, Pagel C, Whiteley W, Sudlow C, Hemingway H, Denaxas Set al., 2021, Understanding COVID-19 trajectories from a nationwide linked electronic health record cohort of 56 million people: phenotypes, severity, waves &amp; vaccination

<jats:title>Abstract</jats:title><jats:sec><jats:title>Background</jats:title><jats:p>Updatable understanding of the onset and progression of individuals COVID-19 trajectories underpins pandemic mitigation efforts. In order to identify and characterize individual trajectories, we defined and validated ten COVID-19 phenotypes from linked electronic health records (EHR) on a nationwide scale using an extensible framework.</jats:p></jats:sec><jats:sec><jats:title>Methods</jats:title><jats:p>Cohort study of 56.6 million people in England alive on 23/01/2020, followed until 31/05/2021, using eight linked national datasets spanning COVID-19 testing, vaccination, primary &amp; secondary care and death registrations data. We defined ten COVID-19 phenotypes reflecting clinically relevant stages of disease severity using a combination of international clinical terminologies (e.g. SNOMED-CT, ICD-10) and bespoke data fields; positive test, primary care diagnosis, hospitalisation, critical care (four phenotypes), and death (three phenotypes). Using these phenotypes, we constructed patient trajectories illustrating the transition frequency and duration between phenotypes. Analyses were stratified by pandemic waves and vaccination status.</jats:p></jats:sec><jats:sec><jats:title>Findings</jats:title><jats:p>We identified 3,469,528 infected individuals (6.1%) with 8,825,738 recorded COVID-19 phenotypes. Of these, 364,260 (11%) were hospitalised and 140,908 (4%) died. Of those hospitalised, 38,072 (10%) were admitted to intensive care (ICU), 54,026 (15%) received non-invasive ventilation and 21,404 (6%) invasive ventilation. Amongst hospitalised patients, first wave mortality (30%) was higher than the second (23%) in non-ICU settings, but remained unchanged for ICU patients. The highest mortality was for patients receiving critical care outside of ICU in wave 1 (51%). 13,0

Journal article

Zhu T, Li K, Herrero P, Georgiou Pet al., 2021, Deep Learning for Diabetes: A Systematic Review, IEEE JOURNAL OF BIOMEDICAL AND HEALTH INFORMATICS, Vol: 25, Pages: 2744-2757, ISSN: 2168-2194

Journal article

Zhu T, Li K, Herrero P, Georgiou Pet al., 2021, Basal Glucose Control in Type 1 Diabetes Using Deep Reinforcement Learning: An <i>In Silico</i> Validation, IEEE JOURNAL OF BIOMEDICAL AND HEALTH INFORMATICS, Vol: 25, Pages: 1223-1232, ISSN: 2168-2194

Journal article

Wang Z, Xu X, Qiang X, Li Ket al., 2021, Learning Vector Quantization-Aided Detection for MIMO Systems, IEEE COMMUNICATIONS LETTERS, Vol: 25, Pages: 874-878, ISSN: 1089-7798

Journal article

Kuang L, Zhu T, Li K, Daniels J, Herrero P, Georgiou Pet al., 2021, Live Demonstration: An IoT Wearable Device for Real-time Blood Glucose Prediction with Edge AI, IEEE Biomedical Circuits and Systems Conference (IEEE BioCAS), Publisher: IEEE

Conference paper

Zhu T, Kuang L, Li K, Zeng J, Herrero P, Georgiou Pet al., 2021, Blood Glucose Prediction in Type 1 Diabetes Using Deep Learning on the Edge, IEEE International Symposium on Circuits and Systems (IEEE ISCAS), Publisher: IEEE, ISSN: 0271-4302

Conference paper

Zhu T, Li K, Chen J, Herrero P, Georgiou Pet al., 2020, Dilated Recurrent Neural Networks for Glucose Forecasting in Type 1 Diabetes, JOURNAL OF HEALTHCARE INFORMATICS RESEARCH, Vol: 4, Pages: 308-324, ISSN: 2509-4971

Journal article

Zhu T, Li K, Kuang L, Herrero P, Georgiou Pet al., 2020, An Insulin Bolus Advisor for Type 1 Diabetes Using Deep Reinforcement Learning, SENSORS, Vol: 20

Journal article

Zhang J, Cong S, Ling Q, Li K, Rabitz Het al., 2020, Quantum State Filter With Disturbance and Noise, IEEE TRANSACTIONS ON AUTOMATIC CONTROL, Vol: 65, Pages: 2856-2866, ISSN: 0018-9286

Journal article

Spence R, Li K, Uduku C, Zhu T, Redmond L, Herrero P, Oliver N, Georgiou Pet al., 2020, A NOVEL HAND-HELD INTERFACE SUPPORTING THE SELF-MANAGEMENT OF TYPE 1 DIABETES, Publisher: MARY ANN LIEBERT, INC, Pages: A58-A58, ISSN: 1520-9156

Conference paper

Li K, Daniels J, Liu C, Herrero-Vinas P, Georgiou Pet al., 2020, Convolutional recurrent neural networks for glucose prediction, IEEE Journal of Biomedical and Health Informatics, Vol: 24, Pages: 603-613, ISSN: 2168-2194

Control of blood glucose is essential for diabetes management. Current digital therapeutic approaches for subjects with Type 1 diabetes mellitus (T1DM) such as the artificial pancreas and insulin bolus calculators leverage machine learning techniques for predicting subcutaneous glucose for improved control. Deep learning has recently been applied in healthcare and medical research to achieve state-of-the-art results in a range of tasks including disease diagnosis, and patient state prediction among others. In this work, we present a deep learning model that is capable of forecasting glucose levels with leading accuracy for simulated patient cases (RMSE = 9.38±0.71 [mg/dL] over a 30-minute horizon, RMSE = 18.87±2.25 [mg/dL] over a 60-minute horizon) and real patient cases (RMSE = 21.07±2.35 [mg/dL] for 30-minute, RMSE = 33.27±4.79\% for 60-minute). In addition, the model provides competitive performance in providing effective prediction horizon ( PHeff) with minimal time lag both in a simulated patient dataset ( PHeff = 29.0±0.7 for 30-min and PHeff = 49.8±2.9 for 60-min) and in a real patient dataset ( PHeff = 19.3±3.1 for 30-min and PHeff = 29.3±9.4 for 60-min). This approach is evaluated on a dataset of 10 simulated cases generated from the UVa/Padova simulator and a clinical dataset of 10 real cases each containing glucose readings, insulin bolus, and meal (carbohydrate) data. Performance of the recurrent convolutional neural network is benchmarked against four algorithms. The proposed algorithm is implemented on an Android mobile phone, with an execution time of 6ms on a phone compared to an execution time of 780ms on a laptop.

Journal article

Zhu T, Li K, Georgiou P, 2019, A Dual-Hormone Closed-Loop Delivery System for Type 1 Diabetes Using Deep Reinforcement Learning

We propose a dual-hormone delivery strategy by exploiting deep reinforcementlearning (RL) for people with Type 1 Diabetes (T1D). Specifically, doubledilated recurrent neural networks (RNN) are used to learn the hormone deliverystrategy, trained by a variant of Q-learning, whose inputs are raw data ofglucose \& meal carbohydrate and outputs are dual-hormone (insulin andglucagon) delivery. Without prior knowledge of the glucose-insulin metabolism,we run the method on the UVA/Padova simulator. Hundreds days of self-play areperformed to obtain a generalized model, then importance sampling is adopted tocustomize the model for personal use. \emph{In-silico} the proposed strategyachieves glucose time in target range (TIR) $93\%$ for adults and $83\%$ foradolescents given standard bolus, outperforming previous approachessignificantly. The results indicate that deep RL is effective in buildingpersonalized hormone delivery strategy for people with T1D.

Journal article

Li K, Liu C, Zhu T, Herrero P, Georgiou Pet al., 2019, GluNet: A deep learning framework for accurate glucose forecasting., IEEE Journal of Biomedical and Health Informatics, Vol: 24, Pages: 414-423, ISSN: 2168-2194

For people with Type 1 diabetes (T1D), forecasting of \red{blood glucose (BG)} can be used to effectively avoid hyperglycemia, hypoglycemia and associated complications. The latest continuous glucose monitoring (CGM) technology allows people to observe glucose in real-time. However, an accurate glucose forecast remains a challenge. In this work, we introduce GluNet, a framework that leverages on a personalized deep neural network to predict the probabilistic distribution of short-term (30-60 minutes) future CGM measurements for subjects with T1D based on their historical data including glucose measurements, meal information, insulin doses, and other factors. It adopts the latest deep learning techniques consisting of four components: data pre-processing, label transform/recover, multi-layers of dilated convolution neural network (CNN), and post-processing. The method is evaluated in-silico for both adult and adolescent subjects. The results show significant improvements over existing methods in the literature through a comprehensive comparison in terms of root mean square error (RMSE) (8.88 ± 0.77 mg/dL) with short time lag (0.83 ± 0.40 minutes) for prediction horizons (PH) = 30 mins (minutes), and RMSE (19.90 ± 3.17 mg/dL) with time lag (16.43 ± 4.07 mins) for PH = 60 mins for virtual adult subjects. In addition, GluNet is also tested on two clinical data sets. Results show that it achieves an RMSE (19.28 ± 2.76 mg/dL) with time lag (8.03 ± 4.07 mins) for PH = 30 mins and an RMSE (31.83 ± 3.49 mg/dL) with time lag (17.78 ± 8.00 mins) for PH = 60 mins. These are the best reported results for glucose forecasting when compared with other methods including the neural network for predicting glucose (NNPG), the support vector regression (SVR), the latent variable with exogenous input (LVX), and the auto regression with exogenous input (ARX) algorithm.

Journal article

Zhang J, Cong S, Ling Q, Li Ket al., 2019, An Efficient and Fast Quantum State Estimator With Sparse Disturbance, IEEE TRANSACTIONS ON CYBERNETICS, Vol: 49, Pages: 2546-2555, ISSN: 2168-2267

Journal article

Zhao Z, Li K, Toumazou C, Kalofonou Met al., 2019, A computational model for anti-cancer drug sensitivity prediction, IEEE Biomedical Circuits and Systems Conference (BioCAS), Publisher: IEEE, ISSN: 2163-4025

Conference paper

Zhang K, Cong S, Ding J, Zhang J, Li Ket al., 2019, Efficient and Fast Optimization Algorithms for Quantum State Filtering and Estimation, 10th International Conference on Intelligent Control and Information Processing (ICICIP), Publisher: IEEE, Pages: 7-13

Conference paper

Li K, Chandrasekera TC, Li Y, Holland DJet al., 2018, A Non-Linear Reweighted Total Variation Image Reconstruction Algorithm for Electrical Capacitance Tomography, IEEE SENSORS JOURNAL, Vol: 18, Pages: 5049-5057, ISSN: 1530-437X

Journal article

Song N, Li K, Chen W, 2018, ROBUST VISUAL TRACKING VIA ADAPTIVE STRUCTURE-ENHANCED PARTICLE FILTER, IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP), Publisher: IEEE, Pages: 1578-1582

Conference paper

Yang J, Cong S, Liu X, Li Z, Li Ket al., 2017, Effective quantum state reconstruction using compressed sensing in NMR quantum computing, Physical Review A, Vol: 96, ISSN: 1050-2947

Compressed sensing (CS) has been verified as an effective technique in the reconstruction of quantum state; however, it is still unknown if CS can reconstruct quantum states given the incomplete data measured by nuclear magnetic resonance (NMR). In this paper, we propose an effective NMR quantum state reconstruction method based on CS. Different from the conventional CS-based quantum state reconstruction, our method uses the actual observation data from NMR experiments rather than the data measured by the Pauli operators. We implement measurements on quantum states in practical NMR computing experiments and reconstruct states of two, three, and four qubits using fewer number of measurement settings, respectively. The proposed method is easy to implement and performs more efficiently with the increase of the system dimension size. The performance reveals both efficiency and accuracy, which provides an alternative for the quantum state reconstruction in practical NMR.

Journal article

This data is extracted from the Web of Science and reproduced under a licence from Thomson Reuters. You may not copy or re-distribute this data in whole or in part without the written consent of the Science business of Thomson Reuters.

Request URL: http://wlsprd.imperial.ac.uk:80/respub/WEB-INF/jsp/search-html.jsp Request URI: /respub/WEB-INF/jsp/search-html.jsp Query String: respub-action=search.html&id=00558981&limit=30&person=true