no code implementations • ICML 2018 • Matthew Mirman, Dimitar Dimitrov, Pavle Djordjevic, Timon Gehr, Martin Vechev
We investigate the effectiveness of trace-based supervision methods for training existing neural abstract machines.
