ArchIE uses logistic regression on population genetic statistics to infer archaic local ancestry without archaic reference genomes. ArchIE outperforms S* and S' in simulations at 2% admixture with AUROC 0.97 and AUPR 0.60. Minimum distance to reference population and skew of distance vector drive predictions. ArchIE identifies 2.04% of CEU European genomes as archaic at 20% FDR threshold. Archaic segments in Europeans show elevated Neanderthal match rates to Altai genome. Archaic ancestry elevated near BNC2 and OAS loci in Europeans. Archaic segments depleted in genomic regions under strong selective constraint. ArchIE robust to demographic misspecifications like population size and split times. Method trained on Neanderthal-human demography simulations with 50kb windows. Features include individual frequency spectrum, haplotype distances, private SNPs, and S*. Europeans used YRI Africans as reference to detect post-split admixture. ArchIE confirms prior Neanderthal introgression patterns without using Neanderthal data. Segments match reference-based Neanderthal calls with 91% matching rate at 1% detection. Method applicable to unknown archaic admixture like in Africans or Denisovans.