Leemann, S.C. Byrne, W.E. Cuneo, D.P. Ehrlichman, M.P. Hellert, T. Hexemer, A. Lu, Y. Marcus, M. Melton, C.N. Nishimura, H. Penn, G. Sannibale, F. Shapiro, D.A. Sun, C. Ushizima, D. Venturini, M. Wallén, E.J. JACoW Publishing Geneva, Switzerland In state-of-the-art synchrotron light sources the overall source stability is presently limited by the achievable level of electron beam size stability. This source size stability is presently on the few-percent level, which is still 1–2 orders of magnitude larger than already demonstrated stability of source position/angle (slow/fast orbit feedbacks) and current (top-off injection). Until now source size stabilization has been achieved through corrections based on a combination of static predetermined physics models and lengthy calibration measurements (feed-forward tables), periodically repeated to counteract drift in the accelerator and instrumentation. We now demonstrate for the first time* how application of machine learning allows for a physics- and model-independent stabilization of source size relying only on previously existing instrumentation in ALS. Such feed-forward correction based on neural networks that can be continuously online-retrained achieves source size stability as low as 0.2 microns rms (0.4%) which results in overall source stability approaching the sub-percent noise floor of the most sensitive experiments.