Search

The first process in medical diagnosis is gathering useful patient data from physical examination. How to capture proper and correct morbidity during a consultation is a very difficult task. It usually takes many years of clinical training and practice for a physician to successfully do this. This paper proposes an intelligent interface agent to alleviate the problem. It includes an hypermedia human–machine interface, a medicine-related common sense and past QA verification of completeness, consistency, and rationality of the input data; and generation of complete patient data. Knowing the proficiency level of a user facilitates the system in soliciting relevant patient's data. Verification of the input data against the common sense allows the system to guarantee their validity. Generation of complete patient data is carried out by hypothesizing a major disease using the common sense, followed by producing more diseases through the interaction of the involved symptoms, and completed by adding those symptoms that are decided to be significant in the discovered diseases into the patient record. The intelligent system thus can correctly gather high-quality patient data from all user levels.

A methodology ontology is proposed to help requirement analysis. It helps the software engineer to generate methodical knowledge for requirement analysis (RA). It includes a method library and a set of modeling support entities. The method library contains method templates and components, e.g., primitive methods, inference functions, inference structures, domain models, and a domain ontology. The modeling support entities use the library to construct the desired methodical knowledge. An example is given showing how this approach constructs an object-oriented RA method. Thus generated methodical knowledge can be further coupled with domain knowledge to form a domain-specific RA tool, which relieves the software engineer of selecting and applying a distinct RA tool for a domain. This approach allows ready enhancement, since it is designed as an open architecture, which helps us to evolve it to assimilate new software methodology.

The Firing Sequence Graph (FSG) technique is proposed as a new technique to do behavior analysis and temporal verification of a system modeled by the real-time Petrinet model. It employs simple mechanical procedures to do analysis to reduce the involvement of human beings during firing sequence derivation. It derives firing sequences from local configurations by local sequence accumulation to alleviate the state-space explosion problem. It contains more transition sequencing information in a firing sequence. It allows the efficient derivation of temporal propagation expressions to do various temporal verification, including triggering, deadline, response, and period time constraints check against what the user specifies. All these features make the technique work better in system evaluation than traditional Petri-net languages, reachability tree analysis, incidence matrix technique, marking transition, or state graph technique.

A knowledge-based framework to support task-level programming and operational control of robots is described. Our bask intention is to enhance the intelligence of a robot control system so that it may carefully coordinate the interactions among discrete, asynchronous and concurrent events under the constraints of action precedence and resource allocation. We do this by integrating both off-line and on-line planning capabilities in a single framework. The off-line phase is equipped with proper languages for describing workbenches, specifying tasks, and soliciting knowledge from the user to support the execution of robot tasks. A static planner is included in the phase to conduct static planning, which develops local plans for various specific tasks. The on-line phase is designed as a dynamic control loop for the robot system. It employs a dynamic planner to tackle any contingent situations during the robot operations. It is responsible for developing proper working paths and motion plans to achieve the task goals within designated temporal and resource constraints. It is implemented in a distributed and cooperative blackboard system, which facilitates the integration of various types of knowledge. Finally, any failures from the on-line phase are fed back to the off-line phase. This forms the interaction between the off-line and on-line phases and introduces an extra closed loop opportunistically to tune the dynamic planner to adapt to the variation of the working environment in a long-term manner.

A real time Petri net (RTPN) model is proposed to model the task specification of a real time system in terms of events, actions, states, and temporal constraints. The basic operations of a RTPN include executable transition identification, marking updating, temporal constraint check, and shared input/output calculation. These are time‐consuming processes on whatever computers they are executed. A hardware‐implementable execution model is proposed to improve their performance. Detailed hardware designs of the event and state matching modules in the execution model are described. They employ an associative technique to provide massive parallelism needed in the modules. Moreover, the time complexities of the event and state matching modules are constant, which allows us to better control or predict the behavior of a real time task — a very important feature in a real‐time system.

mwKAT is an interactive knowledge acquisition tool for acquiring domain knowledge about multimedia components. It constructs knowledge bases for a consulting system that produces the design specification for a multimedia workstation according to the user requirements.mwKAT is generated from and executed inGAS, a primitives-based generic knowledge acquisition meta-tool. It contains three acquisition primitives, namely, parameter proposing, constraint proposing, and fix proposing to construct an intermediate knowledge base represented by a dependency model. These primitives identify necessary domain knowledge and guide users to propose significant components, constraints, and fix methods into the dependency model.mwKAT also invokes knowledge verification and validation primitives to verify the completeness, consistency, compilability, and correctness of the intermediate knowledge base.

This paper introduces a novel threat prediction framework for network security. It starts by creating a network-specific attack plan, represented by colored Petri Nets (CPN), from reusable network security knowledge of attack profiles. The CPN-represented plan then allows us to generate a coverability tree for security situation assessment and attack scenario prediction. Exploit certainty degrees are finally introduced to indicate the likelihood of exploits being executed and are used to support the evaluation of threat scores for potential goals in the predicted attack scenarios. Our experiment shows the framework can make correct threat prediction about a multistage attack. With this, the network defender can have a better chance to take mitigation actions before the attacker fulfills his malicious intentions.

We propose two neural network architectures involving feedforward and probabilistic neural network models to simulate the blackboard demon subsystem, which is responsible for triggering knowledge sources in accordance with the complex monitoring conditions on the hypotheses in the blackboard. Both architectures involve a preprocessor module to perform the transformation between symbolic hypotheses and numerical ones. They learn and record the triggering relationships between hypotheses and knowledge sources in the network links and perform better blackboard monitoring function than the traditional symbolic demon subsystems. In comparison, the probabilistic neural network-like architecture performs better when there is a possibility of using a centralized knowledge representation and when it only involves one-to-one or one-to-many mapping between hypotheses and triggering patterns. The feedforward architecture may be useful when a distributed knowledge representation is possible and when time requirements for training the architecture to learn the complex mappings are not too strict.

This paper presents an intelligent system allowing handicapped aphasiacs to perform basic communication tasks. It has the following three key features: (1) A 6-sensor data glove measures the finger gestures of a patient in terms of the bending degrees of his fingers. (2) A finger language recognition subsystem recognizes language components from the finger gestures. It employs multiple regression analysis to automatically extract proper finger features so that the recognition model can be fast and correctly constructed by a radial basis function neural network. (3) A coordinate-indexed virtual keyboard allows the users to directly access the letters on the keyboard at a practical speed. The system serves as a viable tool for natural and affordable communication for handicapped aphasiacs through continuous finger language input.

In general, AAC systems are equipped with special access mechanisms, since the users lack the fine motor-control required to operate such systems with standard peripheral devices. The major problem with such AAC systems is low text typing speed. This paper presents a fast text-based communication system, enabled by finger language and supported by a virtual keyboard, to aim at efficient text entry. A handicapped aphasiac can enter text via the virtual keyboard by wearing a data glove to make proper finger gestures and use them to directly select the letters on the keyboard. The technique centers on the concept of finger language components. We have developed a finger language recognition system that could recognize static finger language components from the finger gestures of a handicapped aphasiac. We then extended the technique to continuous input of finger language components and developed a text-based communication environment for handicapped aphasiacs.

This paper describes a finger language recognition system for handicapped aphasiacs, who are able to express their intents only by using 'finger language'. Finger language is different from sign language in the sense that it is composed of simple hand gestures, each representing a predefined meaning. The system consists of a low-cost fiber data glove, a small-scale finger language recognition subsystem, and a commercial text-to-speech (TTS) subsystem. The data glove receives the optical fiber signals associated with a given hand gesture, and the recognition subsystem translates the intended meaning of the given hand gesture. The recognized meaning is fed to the TTS subsystem, which converts it to vocal signals. Our experiment shows the system can achieve 99.97% accuracy rate.

There are various security sensors deployed in the network to protect the information assets from destruction. These sensors produce a huge amount of alerts in different event granularities and semantics. Such a huge amount of alerts is hard to be comprehended; as a result, a timely response to the attacks is difficult. Thus, a better technique for alert analysis and management is imperative for promoting the network security. We have developed a two layered PA-based (primitive attack-based) correlation approach to tackle this problem. The first layer does PA construction by integrating related alerts into proper PAs. The second layer is the attack subplan-based correlation layer, which does attack scenario correlation from recognized PAs by employing attack subplan templates to guide the correlation process. This paper focuses on the second layer, by discussing how to automatically construct attack subplan templates, how to eliminate invalid attack subplan templates, how the attack subplan templates can direct the process of scenario correlation, and how a proper fusion process can reduce superfluous attack information. Our experiments show this correlation approach can effectively unearth the attack strategies of the attackers, reduce the cognitive loading of the network manager in comprehending alert reports, and also relieve him from the time consuming and skillful analysis task required in manually analyzing alerts to construct correlation knowledge.