Fig. 3: DL-based prediction of R-tipping for paradigmatic systems. | Nature Machine Intelligence

Fig. 3: DL-based prediction of R-tipping for paradigmatic systems.

From: Deep learning for predicting rate-induced tipping

Fig. 3

a–c, Saddle-node system (a), Bautin system (b), compost-bomb system (c) and comparison to classical early warning indicators. First row: statistics of simulated time series for indicating the time-varying system states. Second row: estimated autocorrelation evolution. Third row: estimated variance. Fourth row: DL-derived R-tipping probabilities as functions of time. The red solid lines represent the composite mean values of the time series within the class exhibiting R-tipping, and light-red- and red-shaded areas depict the 99% and 75% confidence intervals, respectively. The corresponding non-tipping scenarios are indicated by blue. The unit on the time axis adopts the time step used in numerical integration (Methods).

Back to article page