TEMPO: Transformers for Temporal Disease Progression from Cross-Sectional Data

Hongtao Hao, Joseph L Austerweil,  Alzheimer’s Disease Neuroimaging Initiative
Proceedings of the 7th Conference on Health, Inference, and Learning, PMLR 333:614-643, 2026.

Abstract

Event-Based Models (EBMs) infer biomarker progression from cross-sectional data but typically only as ordinal sequences and rely on rigid model assumptions. We propose Tempo, a Transformer architecture that learns both ordinal and continuous event sequences through simulation-based supervised learning. Tempo uses two Transformer modules: one treats biomarkers as tokens to infer event sequencing; the other treats patients as tokens, representing each by their per-biomarker abnormality profile, to infer patients’ disease stages. On synthetic benchmarks, Tempo reduces normalized Kendall’s Tau distance by 52.89% and staging MAE by 25.33% compared to state-of-the-art SA-EBM, with larger reductions in high-dimensional settings (58.88% and 61.10%). Applied to ADNI, Tempo recovers a biologically plausible Alzheimer’s progression: early medial temporal atrophy, followed by amyloid accumulation and cognitive decline, and late-stage tau pathology with terminal acceleration of global neurodegeneration—broadly consistent with established disease models. Tempo also eliminates the need to derive custom inference algorithms and enables rapid empirical comparison of generative hypotheses.

Cite this Paper

BibTeX
@InProceedings{pmlr-v333-hao26a, title = {TEMPO: Transformers for Temporal Disease Progression from Cross-Sectional Data}, author = {Hao, Hongtao and Austerweil, Joseph L and {{Alzheimer's Disease Neuroimaging Initiative}}}, booktitle = {Proceedings of the 7th Conference on Health, Inference, and Learning}, pages = {614--643}, year = {2026}, editor = {Healey, Elizabeth and Fries, Jason and Pollard, Tom and Tang, Shengpu and Zink, Anna and Hartvigsen, Tom and Agrawal, Monica and Finlayson, Sam and Glicksberg, Benjamin and Beaulieu-Jones, Brett and Wang, Kai and Fontalvo, Daseyra and Sarker, Tasmie and Chen, Irene and Alsentzer, Emily}, volume = {333}, series = {Proceedings of Machine Learning Research}, month = {29--30 Jun}, publisher = {PMLR}, pdf = {https://raw.githubusercontent.com/mlresearch/v333/main/assets/hao26a/hao26a.pdf}, url = {https://proceedings.mlr.press/v333/hao26a.html}, abstract = {Event-Based Models (EBMs) infer biomarker progression from cross-sectional data but typically only as ordinal sequences and rely on rigid model assumptions. We propose Tempo, a Transformer architecture that learns both ordinal and continuous event sequences through simulation-based supervised learning. Tempo uses two Transformer modules: one treats biomarkers as tokens to infer event sequencing; the other treats patients as tokens, representing each by their per-biomarker abnormality profile, to infer patients’ disease stages. On synthetic benchmarks, Tempo reduces normalized Kendall’s Tau distance by 52.89% and staging MAE by 25.33% compared to state-of-the-art SA-EBM, with larger reductions in high-dimensional settings (58.88% and 61.10%). Applied to ADNI, Tempo recovers a biologically plausible Alzheimer’s progression: early medial temporal atrophy, followed by amyloid accumulation and cognitive decline, and late-stage tau pathology with terminal acceleration of global neurodegeneration—broadly consistent with established disease models. Tempo also eliminates the need to derive custom inference algorithms and enables rapid empirical comparison of generative hypotheses.} }
Endnote
%0 Conference Paper %T TEMPO: Transformers for Temporal Disease Progression from Cross-Sectional Data %A Hongtao Hao %A Joseph L Austerweil %A Alzheimer’s Disease Neuroimaging Initiative %B Proceedings of the 7th Conference on Health, Inference, and Learning %C Proceedings of Machine Learning Research %D 2026 %E Elizabeth Healey %E Jason Fries %E Tom Pollard %E Shengpu Tang %E Anna Zink %E Tom Hartvigsen %E Monica Agrawal %E Sam Finlayson %E Benjamin Glicksberg %E Brett Beaulieu-Jones %E Kai Wang %E Daseyra Fontalvo %E Tasmie Sarker %E Irene Chen %E Emily Alsentzer %F pmlr-v333-hao26a %I PMLR %P 614--643 %U https://proceedings.mlr.press/v333/hao26a.html %V 333 %X Event-Based Models (EBMs) infer biomarker progression from cross-sectional data but typically only as ordinal sequences and rely on rigid model assumptions. We propose Tempo, a Transformer architecture that learns both ordinal and continuous event sequences through simulation-based supervised learning. Tempo uses two Transformer modules: one treats biomarkers as tokens to infer event sequencing; the other treats patients as tokens, representing each by their per-biomarker abnormality profile, to infer patients’ disease stages. On synthetic benchmarks, Tempo reduces normalized Kendall’s Tau distance by 52.89% and staging MAE by 25.33% compared to state-of-the-art SA-EBM, with larger reductions in high-dimensional settings (58.88% and 61.10%). Applied to ADNI, Tempo recovers a biologically plausible Alzheimer’s progression: early medial temporal atrophy, followed by amyloid accumulation and cognitive decline, and late-stage tau pathology with terminal acceleration of global neurodegeneration—broadly consistent with established disease models. Tempo also eliminates the need to derive custom inference algorithms and enables rapid empirical comparison of generative hypotheses.
APA
Hao, H., Austerweil, J.L. & Alzheimer’s Disease Neuroimaging Initiative, . (2026). TEMPO: Transformers for Temporal Disease Progression from Cross-Sectional Data. Proceedings of the 7th Conference on Health, Inference, and Learning, in Proceedings of Machine Learning Research 333:614-643 Available from https://proceedings.mlr.press/v333/hao26a.html.

Related Material