Li et al.
Scaling law for physiological waveforms favors more training data over larger models
A 1.2x model size increase paired with an 8.2x increase in pretraining hours is compute-optimal for clinical prediction tasks.
Academic
Pingzhi Li · Jie Peng · Shuqing Luo · Zachary Plotkin · Tianlong Chen
University of North Carolina at Chapel Hill
Research Digest··3 min read
3M hours of data and derive a scaling law that predicts downstream clinical performance.
Why this paper
From University of North Carolina at Chapel Hill
In one line
Larger physiological waveform models and more pretraining hours reduce the need for labeled clinical data.
What we could check
- ·No code link found
- ·No weights link found
- ·No dataset link found
- ✓Compute or model size stated (params 2.1B)
- ·No stated limitations found
- ✓Reports numbers on named benchmarks (2 benchmarks)
Observed from the paper text and links we have. Absence here means we did not find it, not that it does not exist.
§