How　to　increase　both　autonomy　and　versatility　of　a　knowledge　discovery　system　is　a　core　problem　and　a　crucial　aspect　of　KDD　(Knowledge　Discovery　and　Data　Mining).　We　have　been　developing　a　multi-agent　based　KDD　methodology/system　called　GLS　(Global　Learning　Scheme)　for　performing　multi-aspect　intelligent　data　analysis　as　well　as　multi-level　conceptual　abstraction　and　learning.　With　multi-level　and　multi-phase　process,　GLS　increases　versatility　and　autonomy.　This　paper　presents　our　recent　development　on　the　GLS　methodology/system.　©　Springer-Verlag　Berlin　Heidelberg　2002.