The　volume　of　the　data　has　been　rocketed　since　the　new　information　era　arrives.　How　to　protect　information　privacy　and　detect　the　threat　whenever　the　intrusion　happens　has　become　a　hot　topic.　In　this　essay,　we　are　going　to　look　into　the　latest　machine　learning　techniques　(including　deep　learning)　which　are　applicable　in　intrusion　detection,　malware　detection,　and　vulnerability　detection.　And　the　comparison　between　the　traditional　methods　and　novel　methods　will　be　demonstrated　in　detail.　Specially,　we　would　examine　the　whole　experiment　process　of　representative　examples　from　recent　research　projects　to　give　a　better　insight　into　how　the　models　function　and　cooperate.　In　addition,　some　potential　problems　and　improvements　would　be　illustrated　at　the　end　of　each　section.　©　2021　Institute　of　Physics　Publishing.　All　rights　reserved.