Collected sources and patterns will appear here. Add from search, explore, or the patterns library.
TimeSeries -> SymmetricForecast
Run inference on both the original and sign-flipped input time-series, then reverse and average the outputs to guarantee symmetric predictions.
Problem it solves
Deep models often predict asymmetric trajectories for identical upwards vs downwards historical patterns.
Consumes
Emits
The real projects this mechanism was found in. Attribution is the point — this is how the best teams actually do it.