Research map / Time series / Foundation models for time series
HEARTS: Benchmarking LLM Reasoning on Health Time Series
Sirui Li, Shuhan Xiao, Mihir Joshi, Ahmed Metwally, Daniel McDuff, Wei Wang, Yuzhe Yang
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 changed | Where | Measured | Effect |
|---|---|---|---|
| use visual (image) input instead of text or file access (ablation) | Grok 4.1 Fast · HeaRTS 10 short-sequence tasks | score: drops to 0.54 | Hurt |
| give direct text input instead of file access (ablation) | Llama 4 Maverick and Grok 4.1 Fast · HeaRTS 10 short-sequence tasks | score | No clear effect |
| add auxiliary signals to the single-channel input (ablation) | CodeAct LLM agents · HeaRTS 6 paired tasks (13 tasks) | score | Mixed |
| use the Biomni domain-specialist agent instead of CodeAct (ablation) | GLM 4.7 and Llama 4 Maverick · HeaRTS | score | No clear effect |
| increase input sequence length and sampling frequency | LLM agents · HeaRTS | score | Hurt |
| use a higher general intelligence index model | LLM agents · HeaRTS | score | No clear effect |
| use OpenTSLM instead of CodeAct-enabled LLMs | OpenTSLM · HeaRTS 3 tasks | score | Hurt |
| scale up or use newer model within a family | GPT, Qwen, Gemini series · HeaRTS | score | No clear effect |
| use reasoning model Gemini 3.1 Pro instead of non-reasoning DeepSeek V3.1 | LLM agents · HeaRTS overall | overall score: 0.64 -> 0.69 | Improved |
Closest papers by meaning
- Medmarks: A Comprehensive Open-Source LLM Benchmark Suite for Medical Tasks (NeurIPS 2026 · similarity 0.86)
- Evaluating LLMs Across Multi-Cognitive Levels: From Medical Knowledge Mastery to Scenario-Based Problem Solving (ICML 2025 · similarity 0.86)
- MedMemoryBench: Benchmarking Agent Memory in Personalized Healthcare (NeurIPS 2026 · similarity 0.86)
- Fisher-R1: Training LLM Agents for Reliable Hypothesis Testing (NeurIPS 2026 · similarity 0.85)
- Optimizing Inference-Time Compute for Medical Reasoning via Uncertainty Quantification (ICML 2026 · similarity 0.85)
- MDAgents: An Adaptive Collaboration of LLMs for Medical Decision-Making (NeurIPS 2024 · similarity 0.85)
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.
