AI Paper Atlas

Research map / Time series / Causal and generative time series

How (Not) to Hybridize Neural and Mechanistic Models for Epidemiological Forecasting

Yiqi Su, Ray Lee, Jiaming Cui, Naren Ramakrishnan

ICML 2026 · Time series › Causal and generative time series · approach: Robust forecasting and evaluation · 0 citations (OpenAlex)

Read the paperCheck your idea against 137,217 papers

AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
add decomposition-controlled latent neural ODEs with SIRS dynamics (EpiNode) versus next-best LatentODEEpiNode · SIRS (Fixed)RMSE: 0.0046 -> 0.0022Improved
EpiNode versus EINN-SIRS under physics mismatchEpiNode · SIR (Fixed)RMSE: 0.0305 -> 0.0223Improved
EpiNode versus LatentODE under physics mismatchEpiNode · SEIRS (Fixed)RMSE: 0.0111 -> 0.0041Improved
EpiNode versus KAN-ODEs on real ILI dataEpiNode · ILIRMSE: 0.0209 -> 0.0093Improved
EpiNode versus EINN-SIRS on beta sign agreementEpiNode · SIRS (Varying) parameter inferencesign agreement of beta (%): 59.8 -> 66.7Improved
EpiNode versus EINN-SIRS on gamma sign agreementEpiNode · SIRS (Varying) parameter inferencesign agreement of gamma (%): 52.0 -> 50.6Hurt
EpiNode versus ARIMA on peak value error in SIR (Fixed)EpiNode · SIR (Fixed)peak value error: -0.0063 -> -0.0594Hurt
add three collaborative latent ODEs (trend/seasonal/residual) instead of a single latent ODE (ablation)EpiNode · five synthetic and ILI datasetsRMSEImproved
increase decomposition from 1 to 3 components (VMD) (ablation)EpiNode · four of five datasetsRMSEImproved
add a third decomposition component (ablation)EpiNode · SIRS (Varying)RMSEHurt
add time-delay embedding of control signals (ablation)EpiNode · datasets other than SIRS (Fixed)RMSEImproved
add time-delay embedding of control signals (ablation)EpiNode · SIRS (Fixed)RMSENo clear effect

Earlier methods this paper compares with

Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 1 on most shared results.

Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.

Closest papers by meaning

Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy