Network security is a big issue at present for large organizations. There are many types of Intrusion Detection Systems like Host Based IDS, Network Based IDS, Anomaly Based IDS and Misuse Based IDS etc. There are various methods to implement IDS that are based on some artificial intelligence concepts which improve performance of IDS. Some of those techniques are Neural Network, Data Mining, Genetic algorithm, Fuzzy Logic etc. The paper reviews these techniques and their comparison in brief.