Papers by Richard Lippmann
Using missing feature theory to actively select features for robust speech recognition with interruptions, filtering and noise
5th European Conference on Speech Communication and Technology (Eurospeech 1997)
Speech recognizers trained with quiet wide-band speech degrade dramatically with high-pass, low-p... more Speech recognizers trained with quiet wide-band speech degrade dramatically with high-pass, low-pass, and notch filtering, with noise, and with interruptions of the speech input. A new and simple approach to compensate for these degradations is presented ...
Journal of Machine Learning Research, Dec 1, 2006
The prevalent use of computers and internet has enhanced the quality of life for many people, but... more The prevalent use of computers and internet has enhanced the quality of life for many people, but it has also attracted undesired attempts to undermine these systems. This special topic contains several research studies on how machine learning algorithms can help improve the security of computer systems.
In the area of research and development effort for cloud computing, Cloud security is considered ... more In the area of research and development effort for cloud computing, Cloud security is considered as one of challenging issues. Most commonly faced attacks are Distributed Denial-of-Service (DDoS) attacks. DDoS attacks are variation of DOS attack at distributed and large-scale level. Firstly attacker tries to discover the vulnerabilities or we can say loopholes of a cloud system and takes control over the virtual machines. And then gets success in deploying DDoS at large scale. Such attacks includes certain actions at initial stage such as exploitation in multiple steps, scanning for uncommon or less occurring vulnerabilities, identified vulnerabilities are utilized against virtual machines to use them as zombies and finally DDOS is achieved through these compromised zombies. To avoid vulnerable virtual machines from being compromised in the cloud system, proposed approach uses multiphase vulnerability detection at distributed level, measurement, countermeasure selection mechanism ca...
An Interactive Attack Graph Cascade and Reachability Display
VizSEC 2007
Attack graphs for large enterprise networks improve security by revealing critical paths used by ... more Attack graphs for large enterprise networks improve security by revealing critical paths used by adversaries to capture network assets. Even with simplification, current attack graph displays are complex and difficult to relate to the underlying physical networks. We have developed a new interactive tool intended to provide a simplified and more intuitive understanding of key weaknesses discovered by attack graph
Lecture Notes in Computer Science, 2002
Vulnerability scanning and installing software patches for known vulnerabilities greatly affects ... more Vulnerability scanning and installing software patches for known vulnerabilities greatly affects the utility of network-based intrusion detection systems that use signatures to detect system compromises. A detailed timeline analysis of important remote-to-local vulnerabilities demonstrates (1) Vulnerabilities in widely-used server software are discovered infrequently (at most 6 times a year) and (2) Software patches to prevent vulnerabilities from being exploited are available before or simultaneously with signatures. Signature-based intrusion detection systems will thus never detect successful system compromises on small secure sites when patches are installed as soon as they are available. Network intrusion detection systems may detect successful system compromises on large sites where it is impractical to eliminate all known vulnerabilities. On such sites, information from vulnerability scanning can be used to prioritize the large numbers of extraneous alerts caused by failed attacks and normal background traffic. On one class B network with roughly 10 web servers, this approach successfully filtered out 95% of all remote-to-local alerts.
Experience Using Active and Passive Mapping for Network Situational Awareness
Fifth IEEE International Symposium on Network Computing and Applications (NCA'06)
Passive network mapping has often been proposed as an approach to maintain up-to-date information... more Passive network mapping has often been proposed as an approach to maintain up-to-date information on networks between active scans. This paper presents a comparison of active and passive mapping on an operational network. On this network, active and passive tools found largely disjoint sets of services and the passive system took weeks to discover the last 15% of active services. Active and passive mapping tools provided different, not complimentary information. Deploying passive mapping on an enterprise network does not reduce the need for timely active scans due to non-overlapping coverage and potentially long discovery times
Proceedings 2002 IEEE Symposium on Security and Privacy
An integral part of modeling the global view of network security is constructing attack graphs. I... more An integral part of modeling the global view of network security is constructing attack graphs. In practice, attack graphs are produced manually by Red Teams. Construction by hand, however, is tedious, error-prone, and impractical for attack graphs larger than a hundred nodes. In this paper we present an automated technique for generating and analyzing attack graphs. We base our technique on symbolic model checking [4] algorithms, letting us construct attack graphs automatically and efficiently. We also describe two analyses to help decide which attacks would be most costeffective to guard against. We implemented our technique in a tool suite and tested it on a small network example, which includes models of a firewall and an intrusion detection system.
Proceedings of the 12th ACM SIGSOFT twelfth international symposium on Foundations of software engineering, 2004
Five modern static analysis tools (ARCHER, BOON, Poly-Space C Verifier, Splint, and UNO) were eva... more Five modern static analysis tools (ARCHER, BOON, Poly-Space C Verifier, Splint, and UNO) were evaluated using source code examples containing 14 exploitable buffer overflow vulnerabilities found in various versions of Sendmail, BIND, and WU-FTPD. Each code example included a "BAD" case with and a "OK" case without buffer overflows. Buffer overflows varied and included stack, heap, bss and data buffers; access above and below buffer bounds; access using pointers, indices, and functions; and scope differences between buffer creation and use. Detection rates for the "BAD" examples were low except for PolySpace and Splint which had average detection rates of 87% and 57%, respectively. However, average false alarm rates were high and roughly 50% for these two tools. On patched programs these two tools produce one warning for every 12 to 46 lines of source code and neither tool accurately distinguished between vulnerable and patched code.
Tuning Intrusion Detection to Work with a Two Encryption Key Version of IPsec
MILCOM 2007 - IEEE Military Communications Conference, 2007
Proceedings of the Seventh International Symposium on Visualization for Cyber Security, 2010
Throughout this work L. Braida and N. Durlach have provided continual guidance, review, and advic... more Throughout this work L. Braida and N. Durlach have provided continual guidance, review, and advice and seen to it that I received technical and financial support. Their recent and past efforts in reviewing this thesis have been heroic and they have provided valuable detailed reviews of all published forms of this work. E. Villchur has also been involved in this work from the beginning and his research was, in fact, one of the sparks that initiated this work. He provided much encouragement and detailed advice on many aspects of this research and he taught me much during the time he was working in our lab. K. Stevens has also provided advice concerning aspects of this work related to speech and speech perception. The consultants on the hearing-aid grant (R. Bilger, A
A taxonomy that uses twenty-two attributes to characterize Cprogram overflows was used to constru... more A taxonomy that uses twenty-two attributes to characterize Cprogram overflows was used to construct 291 small C-program test cases that can be used to diagnostically determine the basic capabilities of static and dynamic analysis buffer overflow detection tools. Attributes in the taxonomy include the buffer location (e.g. stack, heap, data region, BSS, shared memory); scope difference between buffer allocation and access; index, pointer, and alias complexity when addressing buffer elements; complexity of the control flow and loop structure surrounding the overflow; type of container the buffer is within (e.g. structure, union, array); whether the overflow is caused by a signed/unsigned type error; the overflow magnitude and direction; and whether the overflow is discrete or continuous. As an example, the 291 test cases were used to measure the detection, false alarm, and confusion rates of five static analysis tools. They reveal specific strengths and limitations of tools and suggest directions for improvements.
Spotting tasks require detection of target patterns from a background of richly varied non-target... more Spotting tasks require detection of target patterns from a background of richly varied non-target inputs. The performance measure of interest for these tasks, called the figure of merit (FOM), is the detection rate for target patterns when the false alarm rate is in an acceptable range. A new approach to training spotters is presented which computes the FOM gradient for each input pattern and then directly maximizes the FOM using b ackpropagati on. This eliminates the need for thresholds during training. It also uses network resources to model Bayesian a posteriori probability functions accurately only for patterns which have a significant effect on the detection accuracy over the false alarm rate of interest. FOM training increased detection accuracy by 5 percentage points for a hybrid radial basis function (RBF)-hidden Markov model (HMM) wordspotter on the credit-card speech corpus.
A 1/3 Octave-Band Noise Generator for Sound-Field Audiometric Measurements
Journal of Speech and Hearing Disorders, 1982
The design of a low-cost 1/3 octave-band noise generator is presented. This device produces 1/3 o... more The design of a low-cost 1/3 octave-band noise generator is presented. This device produces 1/3 octave bands of noise with center frequencies from 100 to 10,000 Hz using a recently introduced switch-capacitor filter with a frequency response which is similar to that of the 1/2 octave filters in the GR 1925 multifilter. The spectrum level of the noise bands produce by this device falls at a rate of 60 dB/octave or more. The noise generator may be used with an audiometer for sound-field measurements in non-anechoic audiometric testing rooms or for earphone measurements. It may also be used as a 1/3 octave-band filter with a center frequency from 100 to 10,000 Hz or as an octave-band filter with a center frequency from 125 to 8000 Hz.
Lecture Notes in Computer Science
Attack graphs are valuable tools in the assessment of network security, revealing potential attac... more Attack graphs are valuable tools in the assessment of network security, revealing potential attack paths an adversary could use to gain control of network assets. Creating an effective visualization for attack graphs is essential to their utility, but many previous efforts produce complex displays that are difficult to relate to the underlying networks. This thesis presents GARNET (Graphical Attack graph and Reachability Network Evaluation Tool), an interactive visualization tool intended to facilitate the task of attack graph analysis. The tool provides a simplified view of critical steps that can be taken by an attacker and of host-to-host network reachability that enables these exploits. It allows users to perform "what-if" experiments including adding new zero-day attacks, following recommendations to patch software vulnerabilities, and changing the attacker starting location to analyze external and internal attackers. Users are able to view a set of attack graph metrics that summarize different aspects of overall network security for a specific set of attacker models. An initial user evaluation of GARNET identified problematic areas of the interface that assisted in the development of a more functional design.
MILCOM 2006, 2006
Defense in depth is a common strategy that uses layers of firewalls to protect Supervisory Contro... more Defense in depth is a common strategy that uses layers of firewalls to protect Supervisory Control and Data Acquisition (SCADA) subnets and other critical resources on enterprise networks. A tool named NetSPA is presented that analyzes firewall rules and vulnerabilities to construct attack graphs. These show how inside and outside attackers can progress by successively compromising exposed vulnerable hosts with the goal of reaching critical internal targets. NetSPA generates attack graphs and automatically analyzes them to produce a small set of prioritized recommendations to restore defense in depth. Field trials on networks with up to 3,400 hosts demonstrate that firewalls often do not provide defense in depth due to misconfigurations and critical unpatched vulnerabilities on hosts. In all cases, a small number of recommendations was provided to restore defense in depth. Simulations on networks with up to 50,000 hosts demonstrate that this approach scales well to enterprise-size networks.
Experiments demonstrated that sigmoid multilayer perceptron (MLP) networks provide slightly bette... more Experiments demonstrated that sigmoid multilayer perceptron (MLP) networks provide slightly better risk prediction than conventional logistic regression when used to predict the risk of death, stroke, and renal failure on 1257 patients who underwent coronary artery bypass operations at the Lahey Clinic. MLP networks with no hidden layer and networks with one hidden layer were trained using stochastic gradient descent with early stopping. MLP networks and logistic regression used the same input features and were evaluated using bootstrap sampling with 50 replications. ROC areas for predicting mortality using preoperative input features were 70.5% for logistic regression and 76.0% for MLP networks. Regularization provided by early stopping was an important component of improved perfonnance. A simplified approach to generating confidence intervals for MLP risk predictions using an auxiliary "confidence MLP" was developed. The confidence MLP is trained to reproduce confidence intervals that were generated during training using the outputs of 50 MLP networks trained with different bootstrap samples.
Machine Learning, 2010
Whenever machine learning is used to prevent illegal or unsanctioned activity and there is an eco... more Whenever machine learning is used to prevent illegal or unsanctioned activity and there is an economic incentive, adversaries will attempt to circumvent the protection provided. Constraints on how adversaries can manipulate training and test data for classifiers used to detect suspicious behavior make problems in this area tractable and interesting. This special issue highlights papers that span many disciplines including email spam detection, computer intrusion detection, and detection of web pages deliberately designed to manipulate the priorities of pages returned by modern search engines. The four papers in this special issue provide a standard taxonomy of the types of attacks that can be expected in an adversarial framework, demonstrate how to design classifiers that are robust to deleted or or corrupted features, demonstrate the ability of modern polymorphic engines to rewrite malware so it evades detection by current intrusion detection and antivirus systems, and provide approaches to detect web pages designed to manipulate web page scores returned by search engines. We hope that these papers and this special issue encouragesthe multidisciplinary cooperation required to address many interesting problems in this relatively new area including predicting the future of the arms races created by adversarial learning, developing effective long-term defensive strategies, and creating algorithms that can process the massive amounts of training and test data available for internet-scale problems.
Computer Networks, 2000
Eight sites participated in the second DARPA off-line intrusion detection evaluation in 1999. A t... more Eight sites participated in the second DARPA off-line intrusion detection evaluation in 1999. A test bed generated live background traffic similar to that on a government site containing hundreds of users on thousands of hosts. More than 200 instances of 58 attack types were launched against victim UNIX and Windows NT hosts in three weeks of training data and two weeks of test data. False alarm rates were low (less than 10 per day). Best detection was provided by network-based systems for old probe and old denialof-service (DoS) attacks and by host-based systems for Solaris user-to-root (U2R) attacks. Best overall performance would have been provided by a combined system that used both host-and network-based intrusion detection. Detection accuracy was poor for previously unseen new, stealthy, and Windows NT attacks. Ten of the 58 attack types were completely missed by all systems. Systems missed attacks because protocols and TCP services were not analyzed at all or to the depth required, because signatures for old attacks did not generalize to new attacks, and because auditing was not available on all hosts. Promising capabilities were demonstrated by host-based systems, by anomaly detection systems, and by a system that performs forensic analysis on file system data.
Advances in Neural Information Processing …, 1992
A new boundary hunting radial basis function (BH-RBF) classifier which allocates RBF centers cons... more A new boundary hunting radial basis function (BH-RBF) classifier which allocates RBF centers constructively near class boundaries is described. This classifier creates complex decision boundaries only in regions where confusions occur and corresponding RBF outputs are similar. A predicted square error measure is used to determine how many centers to add and to determine when to stop adding centers. Two experiments are presented which demonstrate the advantages of the BH-RBF classifier. One uses artificial data with two classes and two input features where each class contains four clusters but only one cluster is near a decision region boundary. The other uses a large seismic database with seven classes and 14 input features. In both experiments the BH-RBF classifier provides a lower error rate with fewer centers than are required by more conventional RBF, Gaussian mixture, or MLP classifiers.
Uploads
Papers by Richard Lippmann