Human activity recognition (HAR) is an evolving technique that offers innovative solutions across various domains, such as healthcare, sports training, and human-computer interactions. This paper addresses the novel challenge of video-based activity recognition, focusing on detecting and classifying athletes' actions to enable precision sports training. Conventional HAR methods based on direct video analysis incur excessive computational overhead and constrained applicability. In contrast, our novel transformer-based framework, namely RSFomer, converts videos into multivariate time series, and then detects and classifies the athletes' actions. However, sports videos often suffer from severe occlusion, which introduces significant noise to the converted time series and thus deteriorates recognition performance. To address this challenge, we implement several innovative strategies to improve the robustness of our framework. First, we propose a dual-scale filtering mechanism that leverages the unscented Kalman filter and kinematic constraints to reduce noise and outliers in the converted time series. Second, we incorporate the masking mechanism and temporal slicing mechanism to enhance the transformer's ability to handle anomalies and extract multi-scale features for accurate action recognition. We perform extensive evaluations on our Boxing dataset as well as the UEA and FineGym datasets. The results demonstrate that our RSFomer is effective, outperforming existing state-of-the-art methods with significant advantages.

Many human-computer interactions and automated processes are dependent on data. We need to correctly store data and retrieve data to support these processes. Large Language Models (LLMs) have made significant progress in comprehension and reasoning. However, we have found that their ability to handle professional data is weak, which significantly limits their applications in automated processes and professional applications. These scenarios often involve a large number of similar data entries, which are challenging for LLMs such as ChatGPT-4 Omni. In this paper, we propose a Rational Intelligence Model that comprehends human experts' knowledge of data structure and process requirements of data, automatically extracts data from conversations with end users, and effectively stores the data and retrieves it for supporting the interaction and process. Experiments show that ChatGPT-4o can achieve 0% error in simple queries, 2.6% errors when data entries are out of order, and 38% errors when the queries are reasonably complex. However, with our proposed Rational Intelligence Model (RIM), we can achieve 0% error rate in all tests. RIM fundamentally changes software engineering and expert system development approaches. Instead of having a software engineer understand expert knowledge of data processing, this is now achieved by RIM, which means it is much more flexible, lower in cost, and requires much less development time.

This research proposed a model called rational intelligence and studied its reasoning capability as compared to ChatGPT-4o in counterfactual reasoning tasks. Unlike traditional AI models that rely heavily on data-driven approaches, rational intelligence allows for reasoning over abstract principles and hypothetical scenarios, similar to human cognitive processes. The proposed Rational Intelligence Model (RIM) applies Large Language Models (LLMs) to enable human-like comprehension, knowledge application, and problem-solving capabilities. In complex counterfactual reasoning tasks with scenarios proven to be challenging to human adults, we demonstrate that RIM achieves a clearly higher accuracy rate (76%) compared to ChatGPT-4o (68%), showing its enhanced reasoning capabilities. Additionally, RIM incorporates a self-reflection mechanism to manage knowledge conflicts and gaps, which can further improve its performance and adaptability.

This paper investigates data communication and access control in satellite-assisted Internet of Things. In particular, given the characteristics of an open communication environment, a multi-layer heterogeneous network, and a time-varying topology in the Space-Air-Ground-Sea Integrated Network (SAGSIN), traditional data communication and security mechanisms built upon the TCP/IP architecture may not fully leverage their potential. Current networks are primarily responsible for end-to-end transmission of binary data, lacking the capability to semantically perceive and handle dynamic and decentralized content. This leads to significant performance gaps in the network. In other words, it is better to retrieve expected information directly from the network and perform content protection on it with low dependency. We propose DARS, a Data Access scheme with Robust and Secure content communication for Satellite-assisted Internet of Things (S-IoT), leveraging the architectural benefits of content centric networks and rich attributes of IoT. DARS enables automatic data retrieval and access control without additional mechanisms, such as online certificate distribution, homogeneous network, and other presuppositions, which facilitates DARS to be applied in various environments. Additionally, DARS integrates a combination of techniques, including semantic representation, cryptographic technologies, and content caching, from a network-centric perspective. Theoretical analysis and experimental simulations show that DARS simplifies system operations and offers a viable solution for satellite-assisted IoT-based applications.

Software vulnerabilities, also known as flaws, bugs or weaknesses, are common in modern information systems, putting critical data of organizations and individuals at cyber risk. Due to the scarcity of resources, initial risk assessment is becoming a necessary step to prioritize vulnerabilities and make better decisions on remediation, mitigation, and patching. Datasets containing historical vulnerability information are crucial digital assets to enable AI-based risk assessments. However, existing datasets focus on collecting information on individual vulnerabilities while simply storing them in relational databases, disregarding their structural connections. This article constructs a compact vulnerability knowledge graph, VulKG, containing over 276 K nodes and 1 M relationships to represent the connections between vulnerabilities, exploits, affected products, vendors, referred domain names, and more. We provide a detailed analysis of VulKG modeling and construction, demonstrating VulKG-based query and reasoning, and providing a use case of applying VulKG to a vulnerability risk assessment task, i.e., co-exploitation behavior discovery. Experimental results demonstrate the value of graph connections in vulnerability risk assessment tasks. VulKG offers exciting opportunities for more novel and significant research in areas related to vulnerability risk assessment. The data and codes of this article are available at https://github.com/happyResearcher/VulKG.git .

The prosperity of software applications brings fierce market competition to developers. Employing third-party libraries (TPLs) to add new features to projects under development and to reduce the time to market has become a popular way in the community. However, given the tremendous TPLs ready for use, it is challenging for developers to effectively and efficiently identify the most suitable TPLs. To tackle this obstacle, we propose an innovative approach named PyRec to recommend potentially useful TPLs to developers for their projects. Taking Python project development as a use case, PyRec embeds Python projects, TPLs, contextual information, and relations between those entities into a knowledge graph. Then, it employs a graph neural network to capture useful information from the graph to make TPL recommendations. Different from existing approaches, PyRec can make full use of not only project-library interaction information but also contextual information to make more accurate TPL recommendations. Comprehensive evaluations are conducted based on 12,421 Python projects involving 963 TPLs, 9,675 extra entities, 121,474 library usage records, and 73,277 contextual records. Compared with five representative approaches, PyRec improves the recommendation performance significantly in all cases.

Abstract For modern information systems, robust access control mechanisms are vital in safeguarding data integrity and ensuring the entire system’s security. This paper proposes a novel semi-supervised learning framework that leverages heterogeneous graph neural network-based embedding to encapsulate both the intricate relationships within the organizational structure and interactions between users and resources. Unlike existing methods focusing solely on individual user and resource attributes, our approach embeds organizational and operational interrelationships into the hidden layer node embeddings. These embeddings are learned from a self-supervised link prediction task based on a constructed access control heterogeneous graph via a heterogeneous graph neural network. Subsequently, the learned node embeddings, along with the original node features, serve as inputs for a supervised access control decision-making task, facilitating the construction of a machine-learning access control model. Experimental results on the open-sourced Amazon access control dataset demonstrate that our proposed framework outperforms models using original or manually extracted graph-based features from previous works. The prepossessed data and codes are available on GitHub,facilitating reproducibility and further research endeavors.

Explainable recommendation systems provide explanations for recommendation results to improve their transparency and persuasiveness. The existing explainable recommendation methods generate textual explanations without explicitly considering the user's preferences on different aspects of the item. In this paper, we propose a novel explanation generation framework, namelyAspect-guidedExplanation generation withSyntaxGraph (AESG), for explainable recommendation. Specifically, AESG employs a review-based syntax graph to provide a unified view of the user/item details. An aspect-guided graph pooling operator is proposed to extract the aspect-relevant information from the review-based syntax graphs to model the user's preferences on an item at the aspect level. Then, an aspect-guided explanation decoder is developed to generate aspects and aspect-relevant explanations based on the attention mechanism. The experimental results on three real datasets indicate that AESG outperforms state-of-the-art explanation generation methods in both single-aspect and multi-aspect explanation generation tasks, and also achieves comparable or even better preference prediction accuracy than strong baseline methods.

Abstract Knowledge graph, as an extension of graph data structure, is being used in a wide range of areas as it can store interrelated data and reveal interlinked relationships between different objects within a large system. This paper proposes an algorithm to construct an access control knowledge graph from user and resource attributes. Furthermore, an online learning framework for access control decision-making is proposed based on the constructed knowledge graph. Within the framework, we extract topological features to represent high cardinality categorical user and resource attributes. Experimental results show that topological features extracted from knowledge graph can improve the access control performance in both offline learning and online learning scenarios with different degrees of class imbalance status.

The goal of sequential recommendation is to predict the next item that a user would like to interact with, by capturing her dynamic historical behaviors. However, most existing sequential recommendation methods do not focus on solving the long-tail item recommendation problem that is caused by the imbalanced distribution of item data. To solve this problem, we propose a novel sequential recommendation framework, named MASR (ie Memory Bank Augmented Long-tail Sequential Recommendation). MASR is an "Open-book'' model that combines novel types of memory banks and a retriever-copy network to alleviate the long-tail problem. During inference, the designed retriever-copy network retrieves related sequences from the training samples and copies the useful information as a cue to improve the recommendation performance on tail items. Two designed memory banks provide reference samples to the retriever-copy network by memorizing the historical samples appearing in the training phase. Extensive experiments have been performed on five real-world datasets to demonstrate the effectiveness of the proposed MASR model. The experimental results indicate that MASR consistently outperforms baseline methods in terms of recommendation performance on tail items.

Knowledge graph embedding, which aims at learning the low-dimensional representations of entities and relationships, has attracted considerable research efforts recently. However, most knowledge graph embedding methods focus on the structural relationships in fixed triples while ignoring the temporal information. Currently, existing time-aware graph embedding methods only focus on the factual plausibility, while ignoring the temporal smoothness, which models the interactions between a fact and its contexts, and thus can capture fine-granularity temporal relationships. This leads to the limited performance of embedding related applications. To solve this problem, this article presents a Robustly Time-aware Graph Embedding (RTGE) method by incorporating temporal smoothness. Two major innovations of our article are presented here. At first, RTGE integrates a measure of temporal smoothness in the learning process of the time-aware graph embedding. Via the proposed additional smoothing factor, RTGE can preserve both structural information and evolutionary patterns of a given graph. Secondly, RTGE provides a general task-oriented negative sampling strategy associated with temporally aware information, which further improves the adaptive ability of the proposed algorithm and plays an essential role in obtaining superior performance in various tasks. Extensive experiments conducted on multiple benchmark tasks show that RTGE can increase performance in entity/relationship/temporal scoping prediction tasks.

Commonsense knowledge is critical in human reading comprehension. While machine comprehension has made significant progress in recent years, the ability in handling commonsense knowledge remains limited. Synonyms are one of the most widely used commonsense knowledge. Constructing adversarial dataset is an important approach to find weak points of machine comprehension models and support the design of solutions. To investigate machine comprehension models' ability in handling the commonsense knowledge, we created a Question and Answer Dataset with common knowledge of Synonyms (QADS). QADS are questions generated based on SQuAD 2.0 by applying commonsense knowledge of synonyms. The synonyms are extracted from WordNet. Words often have multiple meanings and synonyms. We used an enhanced lesk algorithm to perform word sense disambiguation to identify synonyms for the context. ELECTRA achieves the state-of-art result on the SQuAD 2.0 dataset in 2019. With about 1/10 scale, ELECTRA can achieve similar performance as BERT does. However, QADS shows that ELECTRA has little ability to handle commonsense knowledge of synonyms. In our experiment, ELECTRA-small can achieve 70% accuracy on SQuAD 2.0, but only 20% on QADS. ELECTRA-large did not perform much better. Its accuracy on SQuAD 2.0 is 88% but dropped significantly to 26% on QADS. In our earlier experiments, BERT, although also failed badly on QADS, was not as bad as ELECTRA. The result shows that even top-performing NLP models have little ability to handle commonsense knowledge which is essential in reading comprehension.

In real-world situations, several threat alerts are being investigated by the specialised staff. In order to prompt response to serve incidents or ignore false alarms, alerts are prioritised and analysed. Security professionals rely on information provided in the alert message. Insufficient information in alert messages raises challenges for security analysts that require them to keep track of all internal and external sources to identify the relevant information. In this paper, a Narrative Analytics-Assisted System (NAAS) is proposed, and a knowledge graph is used in the proposed system to present the relationships. The knowledge graph is proposed to capture the complex relationships between the alert and relevant information from the Internal and External knowledge bases to reduce the cognitive effort in information digestion and to understand a wealth of security data. To enable cooperation in the cyber risk management process, it is an inevitable necessity to generate the knowledge graph and interpret it in a human-friendly format. The current machine-friendly formats for reporting incidents from alerts are complex and of an extensive nature. These characteristics hamper the readability and contribution, therefore preventing humans from understanding and being up to date about the incident. NAAS contains four life cycles to assist an analyst to have a better perception of the elements of the environment by involving more staff in the risk management: (1) Analyses the alert, (2) designs the knowledge graph with the natural language sentences, (3) automatically implements the incident report in natural language by applying novel storytelling techniques from the knowledge graph, and (4) maintains it with the contribution of different levels of expertise. The performance of various NAAS's cycles is demonstrated in a case study with an example scenario from the Security Operations Centre (SOC) at an educational institution, highlighting its useability.

Learning Management Systems (LMS) is a combination of different information technology tools that is the core of electronic Learning systems and is widely adopted in many academic sections and institutions. Several research have investigated the driving factors impacting positive learning outcomes in higher education. This research investigates the impact of the learners' use of LMS, attitude towards e-Learning, and perceived admin presence on positive learning outcomes in terms of learners' satisfaction, engagement, and learning outcomes. To achieve this, data from undergraduate and postgraduate students studying in one of the departments of an Australian university is collected through an online questionnaire, and statistical analysis using Partial Least Square is undertake to examine the hypotheses developed. Overall, the results confirmed the positive impact of the learners' use of LMS, attitude towards eLearning, and perceived admin presence on positive learning outcomes.

'Synonym' is an imperative instrument of commonsense knowledge that we apply to make a good sense and sound judgement of our reading. To investigate the ability of machine comprehension models in handling the synonym commonsense knowledge, we developed an innovative approach to automatically generate a dataset based on the Stanford Question Answering Dataset (SQuAD 2.0). The brand-new dataset consists of additional distracting sentences or questions spawned using synonym commonsense knowledge. We formulated new questions by replacing noun entities of the original ones in SQuAD 2.0 with their synonyms. This approach followed the two fundamental principles of SQuAD 2.0 dataset: relevancy and plausibility (incorrect answers are more challenging if they are relevant and plausible). It improves the robustness/abstraction of the question set. To improve the synonym selection strategy in Word Sense Disambiguation (WSD) problem, we designed a new algorithm Multiple Source Adapted Lesk Algorithm (MSALA). Rather than only using WordNet as the source of gloss for adapted Lesk algorithm, we used both lexical database WordNet and commonsense database ConceptNet. This fusion provides a rich hierarchy of semantic relations for the MSALA algorithm. Using this method, we devised 11,000 questions and evaluated the performance of the state-of-the-art question answering system-BERT. Our result shows that the accuracy of the contemporary BERT-Base model dropped from 74.98% to 63.24%. This 10+% accuracy drop revealed the limitations of BERT in handling synonym commonsense knowledge.

Social Media platforms have become an imperative source of information related to urban functions and social behaviour. Analysis of social media data can provide meaningful insights into the personality and behaviour traits of its users. Behaviour studies have been typically conducted using questionnaire and survey data. Since most of our activities have moved online, our behaviour on social media has become a proxy for our real-world behaviour. In this paper, we collected data from Swarm and Twitter, two widely used social media platforms, to detect user's antisocial behaviour with a high degree of accuracy. The behaviour was then categorized into different classes of antisocial behaviour using cutting edge deep learning technology. Four different deep learning models and two different word embeddings were experimented with to achieve a final model accuracy of 99%. Visually enhanced interpretation of the classification process and the word clouds for some of the most commonly used antisocial terms are presented, along with model training and error analysis.

The World Health Organization announced the novel coronavirus (COVID-19) outbreak as a global pandemic on March 11, 2020. To date, this virus has infected over 13 million people worldwide, causing a significant impact on human societies. More than 570,000 people have lost their lives. This study tries to understand the concern and experiences of the general public in this pandemic via social media data analysis. Twitter is a very popular social media platform where people share their ideas and express their views in a timely manner. It is also one of the most comprehensive sources of public conversation. In this work, we collected COVID-19 related Tweets during this pandemic, analyzed the volume, location, frequent words, main topics and people’s emotion. We further visualized the evolution of the main topics along with the development of this pandemic. The analysis and visualization can help us to construct a picture of people’s concerns and to understand their behaviors. It would also be helpful to the public health systems to devise proper policies to better guide and support the pubic.

Detecting Melanoma, the deadliest skin cancer, at the early stage can exceptionally escalate the possibility of the cure up to 99.2% 5-year survival rate. Manual examination by the dermatologist continues to be used as the core and most trusted method till today, albeit a decade of effort in using the technology. Therefore, given the low supply of dermatologists, it is impossible to proactively run the surveillance on people at highest risk for early finding. Deep Convolutional Neural Networks (DCNNs) have demonstrated a dramatic breakthrough in automatic skin lesion classification, which is imperative to improve the diagnostic performance over the mass population with limited access to specialists. Web-application based dermoscopic imaging with integrated artificial intelligence (AI) is a plausible accessible method for future skin lesion analysis. The artificially intelligent plugin, consequently, can serve as enablers for the dermatology community. In this paper, we report the findings of our investigation of using DCNNs for automated Melanoma region segmentation in dermoscopy images. For training and evaluation, we use the HAM10000 public dataset. We aim to realize our model by creating a web tool that can tell general practitioners (GP) and lab technologists the probability diagnoses for a given skin lesion. This automation will assist in fast segregation of the high-risk patients and speed up the follow-up diagnosis and treatment workflow.

The explosive popularity of e-commerce sites has reshaped users' shopping habits and an increasing number of users prefer to spend more time shopping online. This evolution allows e-commerce sites to collect rich data about users. The majority of traditional recommender systems have focused on the macro interactions between users and items, particularly the purchase history of customers. However, decision support only achieved limited performance due to the mismatch between the causality and the interaction sequence. It is especially challenging for products with low purchase frequency, such as refrigerators, or new users with little history data. To address the problem, we investigated how to leverage the heterogenous information, including decision making information to improve recommender systems, helping users approach their items easier and more accurately. Specifically, we propose to model users' purchasing reason information and knowledge graph of items to provide personalized recommendation. The new recommend model, called Decision-Behavior Knowledge Graph (DBKG), captures the decision-making knowledge during users' online purchasing, and update the decision making knowledge in the process of supporting users' purchase decision.

Most applications of Fuzzy Cognitive Maps (FCM) uses static causal links to connect different concepts. However, a causal impact may take effect immediately, or accumulate over a period of time. Consider two cognitive models with a same causal structure, state sets, decision functions and causal linkage strengths, if their causal links have different dynamics, they can have significantly different or even totally different hidden patterns. This paper proposes an easy to use model to represent the dynamics of causal relationships in fuzzy cognitive maps.

Online health information services and self diagnosis systems become popular recent years. We propose a computing model for collaborative medical diagnosis through multi agent argumentation. In this model, the agents are able to communicate with each other to share information, critique and verify each other's knowledge, and collaboratively make diagnosis based on multiple agents' knowledge through an argumentation process. Fuzzy Petri Net (FPN) is adopted as the agents' knowledge model. Different from the commonly used FPNs that assign tokens in places, we assign tokens on arcs and also give places capability in controlling the inference of FPN. The FPN based argumentation is automated with algorithms. The proposed model can be employed to achieve collaborative healthcare diagnosis systems, where agents with different expertise collaboratively argue with each other to come up with a mutually agreed diagnosis.

House fires can cause severe damages to property and even cause loss of life. Although complex fire detection systems are commercially available but generally standalone smoke detectors are utilized in domestic environments to indicate fire. Standalone smoke detectors ring alarm when they detect smoke and give an early warning to anyone present in the premises. One major issue with such smoke detectors is that they cannot be connected with other systems and hence can cause delay in taking appropriate action. In this paper we present a novel application of fire notification to remote parties. We propose that existing hardware like microphone and speakers from a desktop system can be linked with ordinary standalone smoke detectors to make a computer agent who can call and send notifications to someone in case of fire. The prototype of the proposed system has been developed by utilizing existing smoke detectors and well-known software like Skype, Media Player and Excel. The paper also presents the evaluation of prototype for its response time. The proposed solution takes very less effort to configure as well as it requires no change to the existing smoke detectors.

Fuzzy Cognitive Map (FCM) is a tool for modeling human beings' causal knowledge. As compared to many other knowledge models, it is much easier for domain experts to understand. However, domain experts with no computer science expertise still have difficulties in express their knowledge into FCM directly. Majority of the FCM applications in recent years are reported by computing experts in collaboration with domain experts. This paper provides a new FCM model to enable domain experts express their knowledge with FCM directly by removing the "specialized" tasks from the modeling process, including the definition of membership functions, the definition of decision functions and the normalization. This is an effort to reshape FCM as a tool for experts in a wide range of domains, aiming at a similar popularity or availability like that of concept maps.

Recently, the amount of personal medical information online is increasing exponentially, opening up new avenues for hackers to expose personal data that, unlike financial information, can result in a permanent violation of privacy. To protect the privacy of patient data, such as electronic health records (EHRs), access control was used before and attributed-based encryption is used recently. These techniques can effectively prevent from the outside attacks, but are hard to withstand the inside attacks, where the database administrator or the key manager is an attacker. In this paper, we provide a solution to protect the privacy of patient data (EHRs) under the multi-party framework where all EHRs are encrypted with the common public key and an encrypted EHR can be decrypted only by the cooperation of all parties. Based on the ElGamal threshold public key encryption scheme, we propose several EHR access control protocols where multiple parties cooperate to control clinicians' access to EHRs without actually knowing EHRs. Our solution can protect the patient data against the inside attacks as long as at least one party can be trusted. Because our solution is built on Public Key Infrastructure (PKI), it facilitates the clinician registration and revocation.

Visualised presentation of cognitive model can greatly promote the easy comprehension of the model, leading to the wide application of Fuzzy Cognitive Maps (FCM). This paper improves the visualisation of FCM from normal weighted digraph to graphs that can differentiate the strength of causal linkage and the strength of the concepts. The advantage of the new model is apparently illustrated through comparing the new model with the existing ones for the FCMs in the recent literature. The visualised FCM is also applied in word cloud analysis on text term frequency presentation, which improves the knowledge modeling through presenting the key linkage among text terms.

Graphical User Interfaces (GUIs) constitute a large proportion of today's software and are becoming more and more complex. Testing the correctness of GUIs and their underlying software is paramount for providing quality software products. Manual testing is extremely slow and unacceptably expensive. We present a new technique which enables the process of generating test cases and testing automation, based on an innovative model. Given a GUI based application, the set of GUI states and their running logic is modeled as a finite state machine (FSM). The efficiency of the model is formally analyzed and compared with event flow graph (EFG) model. The results show that our model is more efficient in storage.

Cognitive maps (CMs), fuzzy cognitive maps (FCMs), and dynamical cognitive networks (DCNs) are related tools for modeling the cognition of human beings and facilitating machine inferences accordingly. FCMs extend CMs, and DCNs extend FCMs. Domain experts often face the challenge that CMs/FCMs are not sufficiently capable in many applications and that DCNs are too complex. This paper presents a simplified DCN (sDCN) that extends the modeling capability of FCM/CM, yet maintains simplicity. Additionally, this paper proves that there exists a theoretical equivalence among models in the cognitive map family of CMs, FCMs, and sDCNs. It shows that every sDCN can be represented by an FCM or a CM, andvice versa; similarly, every FCM can be represented by a CM, andvice versa. The result shows that CMs, FCMs, and sDCNs are a family of cognitive models that differs from many extended models. This paper also provides a constructive approach to transforming one cognitive map model into other cognitive map models in the family. Therefore, domain experts are able to model applications with more descriptive sDCNs and leave theoretical analysis to the simpler CM forms. The existence of theoretical transformation links among the models provides strong support for their theoretical analysis and flexibility in their applications.

This paper proposes a software agent based simulation model for studying the behavior of complex, highly nonlinear, non-continuous, concurrent systems, especially those involving intelligent units or human factors. Such a system is normally very difficult, if at all possible, to be modeled with well formed differential equations or statistical models. The normal system modelling and analysis methods were not designed for such systems and could not handle them well. This paper attempts to provide insights into such systems through a software agent based simulation model. The model represents each important factors or units in the system as a software agent. One agent can have impacts on other agents. Each agent independently makes its own decision on the impacts from other agents, and/or observations of other agents. The hidden patterns of the complex system are then revealed through simulations. An simulation economic problem is used to illustrate the model.

This paper attempts to include dynamic links in game design to encourage players conduct deeper thinking, instead of making immediate, direct response. In such a design, users' input is through one or more dynamic links, which will generate the effect gradually, or accumulatively, or in an oscillated manner. It will also help the play to build patience through the game playing.

Learning activity negotiation will involve learner's commitment, engage the learner and improve learning quality. However, in current learning systems, learning activities are either decided by the system or decided by the learner. This paper proposes an intelligent negotiation mechanism, where the system and the learner can discuss about their needs and negotiate on learning activities. Because the learner and the system all participate in the decision making, the learning activities selected may meet both the learnerpsilas learning style and experience, as well as the system's curriculum requirement.

Many people find that they are able to gain deeper understanding if the wrong answer is corrected through debating. This paper proposes an intelligent tutoring system based on automated interest based negotiation dialogues. The system is able to automatically generate questions, present correct answer and justifications, attack the arguments of the learner's incorrect ideas, and provide alternative solutions if the original solution is disagreed by the learner.

Cognitive map (CM), fuzzy cognitive map (FCM) are two related tools for modeling human being's cognition and facilitating machine deduction over the cognitive model accordingly. FCM has extended CM by modeling the strength of the causal relationship. To acquire the ability of modeling the strength of cause and effect, multi-value FCM (MVFCM) is formally presented in this paper, which uses mutli-value concepts. Not surprisingly, FCM is more complex than CM and MVFCM is more complex than FCM. This paper proves that there exists a theoretical equivalence among MVFCM, FCM and CM. It shows that for every MVFCM, there exists a FCM, or a CM that represents the MVFCM. This result allows domain experts to model applications with more descriptive MVFCM form and perform theoretical analysis with the simpler CM form. The equivalence among the models also provides strong support to the theoretical analysis of the models and flexibility to their applications.

Goal selection and action selection are important topics in agent research. This paper first describes an agent model, Goal Net, which supports multiple goal selection methods and action selection mechanisms for an agent. Then a goal selection algorithm and action selection mechanisms supported by Goal Net are presented. Examples from real system development are also given to illustrate the algorithm and mechanisms.

In this paper, we present the probabilistic fuzzy cognitive map (PFCM) which is a novel extension of FCM theory. Each concept in PFCM is extended to a fuzzy event that models not only the fuzzy degree but also the fuzzy probability of both the cause and the effect concepts. PFCM enhances the capability of conventional FCMs to handle both randomness and fuzziness which are necessary to model the uncertainty involved in inference process of complex causal system. A formalized inference process of PFCMs is presented for adjustments on probability of fuzzy events and for dynamic update of causal weights. This enables PFCM to synthetically analyze the impacts of randomness and fuzziness on causal inference process. The simulation result shows a good match to the above features of PFCM. PFCM, as an initial attempt, provides a heuristic approach to model the uncertainty of complex causal systems and opens a collection of interesting research issues for further research.

The concept of application service provider (ASP) has been proposed at the end of last millennium. It has attracted comprehensive attention. However, it was not as successful as expected. It was thought that the security issue may be the one which could prevent ASP model from being widely accepted. However, it is proven that the availability issue is far more important than expected. As an attempt to address the problem, we have designed an agent based ASP model. Based on the model, the agent based application service providing system is able to provide wider types of application services, with better availability and less human intervention.

Fuzzy Cognitive Map (FCM) is a powerful and flexible framework for knowledge representation and causal inference. However, in most real applications, it is difficult to design and analyze FCMs due to their structural complexity. Simplification, merging, and division are the important operations on the structure of FCMs. In this paper we present approaches to simplifying FCMs. These approaches show how to clean up a FCM, how to divide a complex FCM into basic FCMs, and how to extract the eigen structure of these basic FCMs. Two improved methods for merging FCMs from different human experts are also proposed in this paper. We discuss difficulties in merging FCMs and present possible solutions.

This paper proposes an Agent Inference Model (AIM) for constructing intelligent software agents. AIM has the ability of representing various types of fuzzy concepts, temporal concepts, and dynamic causal relationships between concepts. It also has the ability of handling feedback and analyzing inference patterns over different causal impact models. Based on AIM, a new type of intelligent agent, Dynamic Inference Agent (DIA) is presented. A dynamic inference agent has the ability to model, infer and make decisions on behalf of human beings. It uses numeric representations and computation instead of symbolic representation and logic deduction to represent knowledge and to carry out the inferences respectively. Thus the construction of DIA is simplified and the implementation code is compact. The application of DIA to various areas, especially for electronic commerce over the Internet is exemplified.

We present the dynamic cognitive network (DCN) which is an extension of the fuzzy cognitive map (FCM). Each concept in the DCNs can have its own value set, depending on how precisely it needs to be described in the network. This enables the DCN to describe the strength of causes and the degree of effects that are crucial to conducting meaningful inferences. The arcs in the DCN define dynamic, causal relationships between concepts. Structurally, DNCs are scalable and more flexible as compared to FCMs. A DCN can be as simple as a cognitive map and FCM, or as complex as a nonlinear dynamic system. To demonstrate the potential applications of DCNs, we present some simulation results. This paper represents our first attempt to develop a dynamic fuzzy inference system using causal relationships. There are many interesting and challenging theoretical and practical issues in DCNs open to further research.

Fuzzy cognitive maps (FCM) is a powerful paradigm for representing human knowledge and causal inference. This paper formally analyzes the causal inference mechanism of FCM. We focus on binary concept states. It is known that given initial conditions, FCM is able to reach only certain states in its state space. We prove that the problem of finding whether a state is reachable in the FCM is nondeterministic polynomial (NP) hard, that we can divide fuzzy cognitive maps containing circles into basic FCM modules. The inference patterns in these basic modules can be studied individually in a hierarchical fashion. This paper also presents a recursive formula for computing FCM's inference patterns in terms of key vertices. The theoretical results presented in this paper provide a feasible and effective framework for the analysis and design of fuzzy cognitive maps in real-world large-scale applications.

Changeability is one of the major concerns in software development. OO (object-oriented) technology itself is not enough to solve this problem. In this paper, we first discuss how the changeability issue is handled in current software development methods and what is the most important prerequisite for solving the software changeability problem. Based on these analyses, we argue that most of current OO analysis methods have missed an important activity in their analysis phase, i.e. change analysis. Thus, OO software systems developed according to these methods often cannot exhibit the expected maintainability, extensibility and reusability. In order to improve current OO analysis methods, we have proposed a conceptual framework for conducting a systematic change analysis in the software analysis phase. A detailed case study of the application of this framework in a real software project has also been described in this paper.

We present a dynamic cognitive network (DCN) which is a systematic extension of the fuzzy cognitive map (FCM). Each concept in a dynamic cognitive network can have its own value set, depending on how precisely it needs to be described in the network. This enables the dynamic cognitive network to describe the strength of the cause and the degree of the effect. This ability also enables DCN to carry out inferences via numerical calculation instead of symbolic deduction. The arcs of DCN define dynamic relationships between concepts and describe the causal procedures. DCN overcomes the problems of FCM without losing the essential advantages.