Fig. 6: Composition-dependent distribution of Δq extracted from experimental and generated XRD spectra.
From: Artificial intelligence can recognize metallic glasses in vast compositional space with sparse data
a Δq for 20 randomly selected alloys from the Zr-Cu-Al alloy system for model training. b Δq extracted from XRD spectra generated by the model trained with spectra from the 20 alloys shown in (a). c Δq extracted from XRD spectra by high-throughput experiments covering 639 Zr-Cu-Al alloys. d Δq for 100 randomly selected alloys from the Zr-Cu-Ni-Al alloy system. e Δq extracted from XRD spectra generated by the model trained with spectra from the 100 alloys shown in (d). The black dots in the plot mark the previously reported bulk glass-forming alloys in the system. f Δq extracted from XRD spectra by high-throughput experiments covering 1704 Zr-Cu-Ni-Al alloys.
