Abstract: The present disclosure provides a self-similar regenerative amplification method and an apparatus. The apparatus includes a broadband seed source, a spectrum shaping broader, a self-similar regenerative amplifier and a pulse compressor disposed in order of a light path. The spectrum shaping broader includes a time domain broader and a spectrum shaper. The time domain broader is configured to broaden the seed pulses, and fine-tune a width of the seed pulse. The spectrum shaper is configured to perform spectrum shaping on the broadened pulses to obtain saddle chirped pulses. The pulse regenerative amplification component includes a gain crystal and a nonlinear crystal. The self-similar regenerative amplifier receives the saddle chirped pulses, performs multiple stepwise amplifications and multiple nonlinear spectrum broadenings back and forth on the saddle chirped pulses, and output high-energy chirped pulses to the pulse compressor.

Abstract: The disclosure describes a method of monitoring the dynamic power consumption of ReRAM crossbars and determines the occurrence of faults when a changepoint is detected in the monitored power-consumption time series. Statistical features are computed before and after the changepoint and train a predictive model using machine-learning techniques. In this way, the computationally expensive fault localization and error-recovery steps are carried out only when a high fault rate is estimated. With the proposed fault-detection method and the predictive model, the test time is significantly reduced while high classification accuracy for well-known AI/ML datasets using a ReRAM-based computing system (RCS) can still be ensured.

Abstract: The disclosure describes a method of monitoring the dynamic power consumption of ReRAM crossbars and determines the occurrence of faults when a changepoint is detected in the monitored power-consumption time series. Statistical features are computed before and after the changepoint and train a predictive model using machine-learning techniques. In this way, the computationally expensive fault localization and error-recovery steps are carried out only when a high fault rate is estimated. With the proposed fault-detection method and the predictive model, the test time is significantly reduced while high classification accuracy for well-known AI/ML datasets using a ReRAM-based computing system (RCS) can still be ensured.