I went down this rabbit hole earlier in the year. You generally don’t need error correction. Instead, you do a retraining of sorts where you basically fine tun

I went down this rabbit hole earlier in the year. You generally don’t need error correction. Instead, you do a retraining of sorts where you basically fine tune the model on the target chip architecture. Because these systems are so complex, errors tend to localize without affecting overall output. My understanding is the hardest part is actually noise in the analog system.