The authors built QiYao-I around a sampling-conditioned temporal manifold attention mechanism.
QiYao-I models irregular timing and asynchronous variables in one forecaster
The model combines timestamp-aware attention with frequency-guided information exchange between variables observed at different times.
Chinese Tech
Linfeng Wang · Ruitong Zhang · Kai Zhao · Yang Shu · Zhongwen Rao · Meng Wang · +3 more
East China Normal University · Huawei Technologies Ltd.
Research Digest··2 min read
Wang and colleagues introduce QiYao-I, a foundation model for forecasting multivariate time series whose measurements are unevenly spaced and not synchronized across variables.
Why this paper
From Huawei Technologies Ltd. and East China Normal University
In one line
QiYao-I forecasts irregular multivariate time series better than prior foundation and end-to-end models, with strong zero-shot and few-shot generalization.
What we could check
- ·No code link found
- ·No weights link found
- ·No dataset link found
- ·No compute details found
- ·No stated limitations found
- ·No benchmark numbers found
Observed from the paper text and links we have. Absence here means we did not find it, not that it does not exist.
§