AI Paper Atlas

Research map / Time series / Foundation models for time series

Forecasting Clinical Risk from Textual Time Series: Structuring Narratives for Temporal AI in Healthcare

Shahriar Noroozizadeh, Sayantan Kumar, Jeremy C. Weiss

AAAI 2026 · Time series › Foundation models for time series · approach: Time-series language reasoning

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
use encoder models instead of decoder LLMs (ablation)event occurrence forecasting · T2S2 F1 at 24h (DeepSeek-R1)F1: 0.482 -> 0.576Improved
fine-tune an MLP head instead of prompting an LLM (ablation)Llama-3.3-70B forecasting · T2S2 (Llama annotations) F1 at 24hF1: 0.460 -> 0.653Improved
MLP-head fine-tuning instead of fine-tuned masking (ablation)ModernBERT-base · T2S2 (DeepSeek-R1) F1 at 1wF1: 0.697 -> 0.877Improved
fine-tuned masking instead of MLP head (ablation)ModernBERT-base · T2S2 (Llama annotations) event orderingc-index: 0.594 -> 0.632Improved
add biomedical pretraining (BioClinical-ModernBERT) (ablation)ModernBERT-base encoder · T2S2 (DeepSeek-R1) event orderingc-index: 0.646 -> 0.677Improved
add biomedical pretraining (BioClinical-ModernBERT) (ablation)ModernBERT-base encoder · T2S2 (DeepSeek-R1) F1 at 24hF1: 0.576 -> 0.559No clear effect
add biomedical pretraining (BioClinical-ModernBERT) (ablation)ModernBERT-base encoder · sepsis-100 F1 at 24hF1: 0.607 -> 0.635Improved
use biomedical-pretrained decoder (MediPhi-PubMed) vs general decoder (ablation)LLM-MLP head · T2S2 (DeepSeek-R1) F1 at 24hF1: 0.482 -> 0.268Hurt
exclude PubMed from pretraining (RedPajama-INCITE-7B) (ablation)decoder LLM-MLP head · T2S2 event orderingc-index: 0.618 vs 0.618-0.632No clear effect
use decoder LLMs instead of encoders (ablation)survival analysis · T2S2 / sepsis-100 survivaltime-dependent c-index: 0.64 vs 0.60 at 0h; RedPajama 0.76 at 168h (sepsis-100)Improved
increase observation window to 168h for survival (ablation)survival models · T2S2 survivaltime-dependent c-index: DeepSeek-R1-Llama-70B 0.64 -> 0.59No clear effect
zero-shot masking with biomedical encoder (ablation)BioClinical-ModernBERT-base · T2S2 F1 at 1hF1: 0.246 F1 vs near-zero for BERT/RoBERTaImproved

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