This paper introduces an adaptive background compensation technique to mitigate errors of the analog signal path of Frequency-Interleaved ADCs (FI-ADC). The key ingredients of the proposed technique are MIMO equalization and the backpropagation algorithm often used to train neural networks. The technique introduced here is applicable to digital receivers such as those used in coherent optical communications. Although compensation of FI-ADCs has been discussed in the technical literature, adaptive background techniques have not been reported so far. The latter are important to automatically compensate errors caused by process, voltage, and temperature variations in CMOS implementations of the data converters and therefore ensure high manufacturing yield. Simulations show that the impairments of the analog signal path are accurately compensated and their effect essentially eliminated, resulting in a high performance ADC system.