Accurately Estimating Total COVID-19 Infections using Information Theory DOI Creative Commons
Jiaming Cui, Bijaya Adhikari,

Arash Haddadan

et al.

medRxiv (Cold Spring Harbor Laboratory), Journal Year: 2021, Volume and Issue: unknown

Published: Sept. 22, 2021

Abstract One of the most significant challenges in combating against spread infectious diseases was difficulty estimating true magnitude infections. Unreported infections could drive up disease spread, making it very hard to accurately estimate infectivity pathogen, therewith hampering our ability react effectively. Despite use surveillance-based methods such as serological studies, identifying is still challenging. This paper proposes an information theoretic approach for number total Our built on top Ordinary Differential Equations (ODE) based models, which are commonly used epidemiology and We show how we can help models better compute identify parametrization by need fewest bits describe observed dynamics reported experiments COVID-19 that leads not only substantially estimates but also forecasts than standard model calibration methods. additionally learned helps modeling more accurate what-if scenarios with non-pharmaceutical interventions. provides a general method improving epidemic applicable broadly.

Language: Английский

Integrating artificial intelligence with mechanistic epidemiological modeling: a scoping review of opportunities and challenges DOI Creative Commons

Yang Ye,

Abhishek Pandey,

Carolyn E. Bawden

et al.

Nature Communications, Journal Year: 2025, Volume and Issue: 16(1)

Published: Jan. 10, 2025

Integrating prior epidemiological knowledge embedded within mechanistic models with the data-mining capabilities of artificial intelligence (AI) offers transformative potential for modeling. While fusion AI and traditional approaches is rapidly advancing, efforts remain fragmented. This scoping review provides a comprehensive overview emerging integrated applied across spectrum infectious diseases. Through systematic search strategies, we identified 245 eligible studies from 15,460 records. Our highlights practical value models, including advances in disease forecasting, model parameterization, calibration. However, key research gaps remain. These include need better incorporation realistic decision-making considerations, expanded exploration diverse datasets, further investigation into biological socio-behavioral mechanisms. Addressing these will unlock synergistic modeling to enhance understanding dynamics support more effective public health planning response. Artificial has improve diseases by incorporating data sources complex interactions. Here, authors conduct use summarise methodological advancements identify gaps.

Language: Английский

Citations

4
Physics-informed radial basis network (PIRBN): A local approximating neural network for solving nonlinear partial differential equations DOI Creative Commons
Jinshuai Bai, Guirong Liu, Ashish Gupta

et al.

Computer Methods in Applied Mechanics and Engineering, Journal Year: 2023, Volume and Issue: 415, P. 116290 - 116290

Published: Aug. 3, 2023

Our recent intensive study has found that physics-informed neural networks (PINN) tend to be local approximators after training. This observation leads this novel radial basis network (PIRBN), which can maintain the property throughout entire training process. Compared deep networks, a PIRBN comprises of only one hidden layer and "activation" function. Under appropriate conditions, we demonstrated PIRBNs using gradient descendent methods converge Gaussian processes. Besides, studied dynamics via tangent kernel (NTK) theory. In addition, comprehensive investigations regarding initialisation strategies were conducted. Based on numerical examples, been more effective efficient than PINN in solving PDEs with high-frequency features ill-posed computational domains. Moreover, existing techniques, such as adaptive learning, decomposition different types loss functions, are applicable PIRBN. The programs regenerate all results at https://github.com/JinshuaiBai/PIRBN.

Language: Английский

Citations

34
A stochastic particle extended SEIRS model with repeated vaccination: Application to real data of COVID‐19 in Italy DOI Creative Commons
Vasileios E. Papageorgiou, George Tsaklidis

Mathematical Methods in the Applied Sciences, Journal Year: 2024, Volume and Issue: 47(7), P. 6504 - 6538

Published: Feb. 22, 2024

The prediction of the evolution epidemics plays an important role in limiting transmissibility and burdensome consequences infectious diseases, which leads to employment mathematical modeling. In this paper, we propose a stochastic particle filtering extended SEIRS model with repeated vaccination time‐dependent parameters, aiming efficiently describe demanding dynamics time‐varying epidemics. validity our is examined using daily records COVID‐19 Italy for period 525 days, revealing notable capacity uncover hidden pandemic. main findings include estimation asymptomatic cases, well‐known feature current Unlike other proposed models that employ extra compartments force proportion significantly increase model's complexity, approach evaluation without additional computational burden. Other confirm appropriateness robustness are its parameter more ICU‐admitted cases compared official during most prevalent infection wave January 2022, attributed intensified admissions may have led full occupancy ICUs. As vast majority datasets contain time series total recovered vaccinated statistical algorithm estimate currently protected through cases. This necessity arises from attenuation antibodies after vaccination/infection necessary long‐time interval predictions. Finally, not only present novel epidemiological test efficiency but also investigate properties, such as existence stability epidemic equilibria, giving new insights literature. latter provides details concerning system's long‐term behavior, while conclusions drawn index provide perspectives on severity future

Language: Английский

Citations

7
Machine learning for data-centric epidemic forecasting DOI
Alexander Rodríguez, Harshavardhan Kamarthi,

Pulak Agarwal

et al.

Nature Machine Intelligence, Journal Year: 2024, Volume and Issue: unknown

Published: Sept. 27, 2024

Language: Английский

Citations

4
Machine Learning for Infectious Disease Risk Prediction: A Survey DOI Creative Commons
Mutong Liu, Yang Liu, Jiming Liu

et al.

ACM Computing Surveys, Journal Year: 2025, Volume and Issue: unknown

Published: Feb. 24, 2025

Infectious diseases place a heavy burden on public health worldwide. In this paper, we systematically investigate how machine learning (ML) can play an essential role in quantitatively characterizing disease transmission patterns and accurately predicting infectious risks. First, introduce the background motivation for using ML risk prediction. Next, describe development application of various models prediction, categorizing them according to models’ alignment with vital concerns specific two distinct phases propagation: (1) pandemic epidemic (the P-E phaseS) (2) endemic elimination E-E phaseS), each presenting its own set critical questions. Subsequently, discuss challenges encountered when dealing model inputs, designing task-oriented objectives, conducting performance evaluations. We conclude discussion open questions future directions.

Language: Английский

Citations

0
Epidemiology-informed Network for Robust Rumor Detection DOI
Wei Jiang, Tong Chen, Xinyi Gao

et al.

Published: April 22, 2025

Language: Английский

Citations

0
Optimizing ODE-derived Synthetic Data for Transfer Learning in Dynamical Biological Systems DOI Creative Commons
Julian Zabbarov, Simon Witzke,

Maximilian Kleissl

et al.

bioRxiv (Cold Spring Harbor Laboratory), Journal Year: 2024, Volume and Issue: unknown

Published: March 29, 2024

Abstract Motivation Successfully predicting the development of biological systems can lead to advances in various research fields, such as cellular biology and epidemiology. While machine learning has proven its capabilities generalizing underlying non-linear dynamics systems, unlocking predictive power is often restrained by limited availability large, curated datasets. To supplement real-world data, informing transfer with data simulated from ordinary differential equations emerged a promising solution. However, success this approach highly depends on designed characteristics synthetic data. Results We optimize dataset size, diversity, noise equation-based time series datasets three relevant representative systems. achieve this, we here, for first time, present framework systematically evaluate influence design choices performance one place. improvement up 92% mean absolute error our optimized simulation-based compared non-informed deep learning. find strong interdependency between size diversity effects. The optimal setting heavily relies well coherence data’s dynamics, emphasizing relevance framework. Availability Implementation code available at https://github.com/DILiS-lab/opt-synthdata-4tl .

Language: Английский

Citations

2
Neural networks for endemic measles dynamics: comparative analysis and integration with mechanistic models DOI Open Access
Wyatt Madden, Wei Jin, Benjamin A. Lopman

et al.

medRxiv (Cold Spring Harbor Laboratory), Journal Year: 2024, Volume and Issue: unknown

Published: May 31, 2024

Abstract Measles is an important infectious disease system both for its burden on public health and as opportunity studying nonlinear spatio-temporal dynamics. Traditional mechanistic models often struggle to fully capture the complex dynamics inherent in measles outbreaks. In this paper, we first develop a high-dimensional feed-forward neural network model with spatial features (SFNN) forecast endemic outbreaks systematically compare predictive power that of classical (TSIR). We illustrate utility our using England Wales data from 1944-1965. These present multiple modeling challenges due interplay between metapopulations, seasonal trends, related demographic changes. Our results show that, while TSIR yields more accurate very short-term (1 2 biweeks ahead) forecasts highly populous cities, overall, outperforms other forecasting windows. Furthermore, spatial-feature model, without imposing assumptions priori , can uncover gravity-model-like hierarchy spread which major cities play role driving regional then turn attention integrative approaches combine machine learning models. Specifically, investigate how be utilized improve state-of-the-art approach known Physics-Informed-Neural-Networks (PINN) explicitly combines compartmental networks. facilitate reconstruction latent susceptible dynamics, improving parameter inference within PINN. summary, appropriately designed network-based outperform traditional short long-term forecasts, simultaneously providing interpretability. work also provides valuable insights into effectively integrating enhance responses similar systems. Author summary Mechanistic have been foundational developing understanding transmission diseases including measles. contrast their counterparts, techniques networks primarily focused accuracy inferring Effectively these two remains central challenge. spatiotemporal detailed dataset describing 1944-1965, one best-documented most-studied all mechanism hierarchical where drive models, inference. offers effective utilization integration enhancing outbreak

Language: Английский

Citations

1
Physics-guided Active Sample Reweighting for Urban Flow Prediction DOI
Wei Jiang, Tong Chen, Guanhua Ye

et al.

Published: Oct. 20, 2024

Language: Английский

Citations

1
Physics-guided deep learning framework for computational mechanics DOI Creative Commons
Jinshuai Bai

Published: Jan. 1, 2024

This thesis developed a novel, effective and robust numerical framework based on the physics-guided deep learning technique for wide range of mechanics modelling. In thesis, thorough investigations regarding proposed have been conducted from both theoretical aspects. It has demonstrated that great advantages over traditional methods when facing challenges, such as nonlinearity free-surface tracking problems. The possibilities integrating state-of-the-art techniques into computational opened new avenue

Language: Английский

Citations

0