Perhaps because the ML community is dominated by CS people, not Statistics people, and the former do not care so much about statistical significance?
Perhaps because the ML community is dominated by CS people, not Statistics people, and the former do not care so much about statistical significance?