Download Advances in Artificial Intelligence. PRICAI 2000 Workshop by Graham J. Williams, Dickson Lukose2 (auth.), Ryszard PDF

By Graham J. Williams, Dickson Lukose2 (auth.), Ryszard Kowalczyk, Seng Wai Loke, Nancy E. Reed, Graham J. Williams (eds.)

This e-book constitutes the completely refereed joint post-proceedings of 4 workshops held in the course of the Pacific Rim foreign convention on synthetic Intelligence, PRICAI 2000, held in Melbourne, Australia, in August/September 2000.
The 32 revised complete papers offered have been rigorously chosen in the course of rounds of reviewing and revision. based on the 4 workshops represented, the ebook is prepared in topical sections on purposes of synthetic intelligence in undefined, synthetic intelligence in digital trade, clever details brokers, and teamwork and adjustable autonomy in brokers.

Show description

Read Online or Download Advances in Artificial Intelligence. PRICAI 2000 Workshop Reader: FourWorkshops held at PRICAI 2000 Melbourne,Australia,August 28 - September 1, 2000 Revised Papers PDF

Best nonfiction_7 books

Adaptive Modelling, Estimation and Fusion from Data: A Neurofuzzy Approach

In an international of just about everlasting and speedily expanding digital info availability, concepts of filtering, compressing, and examining this knowledge to rework it into invaluable and simply understandable details is of maximum value. One key subject during this sector is the potential to infer destiny process habit from a given facts enter.

Extra resources for Advances in Artificial Intelligence. PRICAI 2000 Workshop Reader: FourWorkshops held at PRICAI 2000 Melbourne,Australia,August 28 - September 1, 2000 Revised Papers

Example text

We extract measures from BSM audit data and transform them into one dimensional data using SOM that is a neural network capable of clustering input patterns into fixed vectors automatically. The reduced data can be used to model normal behaviors for intrusion detection. To select the optimal measures of a BSM event, we evaluate the effect of the measures by investigating the performance in IDS. Second, modeling plan and a modeling tool are needed. To determine current behavior is an anomaly, a normal behavior model must be built against which to compare.

We can also see that probability of usage is above 50 % when the map size is smaller than 50. Therefore, useful map size should be selected between 25 and 50. 1 5 20 40 60 Map Size 80 100 0 20 40 60 Map Size 80 100 Fig. 6. Variation of quantization error and map usage according to the map size Useful Measures. In this paper, we extract 3 measures from the data that are related to system call ID, return value, return status), process (ID, IPC ID, IPC permission, exit value, exit status) and file access (access mode, path, argument length, file system).

Experimental results are shown in Section 4. Anomaly Detection of Computer Usage 2 33 Backgrounds The first phase of intrusion detection is to collect audit data from user and system activities. IDS extracts several measures from audit data using BSM, and reduces them in order to profile user’s normal behaviors. After normal behavior modeling, IDS can determine whether current behavior is normal or not. Typically measures that are related to the system bugs exploited by invaders for a long time or important system resources are selected.

Download PDF sample

Rated 4.59 of 5 – based on 23 votes