This innovative practice WIP paper considers the remarkable growth of Artificial Intelligence (AI), Machine Learning (ML) and Data Science (DS) in recent years. This surge has sparked an urgent need for skilled workers in these highly-related areas, subject to growing calls for a diverse, inclusive technology workforce. This WIP paper presents an innovative approach in the rapid design and development of interdisciplinary undergraduate degree programs that are computing-based, with heavy emphasis in AI, ML, and DS while providing foundations for applications in specific domains. Located at the center of Silicon Valley, the Department of Computer Science (CS) at SJSU (San Jose State University) has been constantly bombarded by the speedy changes of technology and the corresponding need for skilled tech workers, and more recently, calls for a diverse tech workforce that matches the demographic of our community. As both an HSI (Hispanic Serving Institution) and AANAPISI (Asian American and Native American Pacific Islander Serving Institution), we understand that underrepresented groups face significant challenges. Prevailing stereotypes about computer scientists tend to reinforce the notion that women and minority do not have a place in this field. Trying to break through these stereotypes, we have reached out and develop interdisciplinary programs with majors that traditionally attract more women and minorities (such as majors in biology or the humanities). This paper first describes the success of two existing interdis-ciplinary, joint master degree programs: MS in Bioinformatics and MS in Data Science. Then, it illustrates how we rapidly developed a computing-based BS in Data Science (in one year), and concurrently a BS in Computer Science and Linguistics (in another year) that is jointly offered with the Department of Linguistics and Language Development (LLD). The paper presents the curriculum development approaches, ethical and social awareness components of the curricula, and uniqueness and success factors. Preliminary data show a significant increase in enrollment of women and low-income students in the first major core course of BS-DS. We believe that this paper will serve as a model initiative for rapidly developing a diverse technology workforce that meets the needs of emerging AI, ML and DS industries.

Accurately analyzing dialogue in multi-party meetings is crucial for enhancing communication and decision-making. The informal and dynamic nature of these discussions introduces challenges for computational analysis, including non-standard language, interruptions, and rapid topic changes. To address these challenges, we developed two methods: Argument Classification and Relation Classification. Argument Classification employs machine learning models like Gradient Boosting to identify and categorize key points made by participants. This method improves conversation traceability and insight extraction, achieving an F1 score of 0.79, a 28% improvement over previous methods. Relation Classification uses advanced neural network architectures, such as BERT, to decode interactions and map relationships within discussions. This method captures conversational nuances, achieving an overall accuracy of 0.86 and an F1 score of 0.83. Our models effectively transform raw meeting transcripts into structured data, enhancing the understanding and analysis of dialogues. This approach not only clarifies discussions but also lays a foundation for better decision-making processes. The results demonstrate the potential for deploying these models in various domains requiring nuanced dialogue analysis.

Deduplication is the process of removing replicated data content from storage facilities like online databases, cloud datastore, local file systems, etc. It is commonly performed as part of data preprocessing to eliminate redundant data that requires extra storage spaces and computing power and is crucial for data storage management in cloud computing. Deduplication is essential for file backup systems since duplicated files will presumably consume more storage space, especially with a short backup period such as daily. A common technique in this field involves splitting files into chunks whose hashes can be compared using data structures or techniques like clustering. This paper explores the possibility of performing such file chunk deduplication leveraging an innovative reinforcement learning approach to achieve a high deduplication ratio. The proposed system is named SegDup, which achieves 13% higher deduplication ratio than Extreme Binning, a state-of-the art deduplication algorithm.

As society grows increasingly more online with each passing year, the problem of cyberbullying becomes more and more prominent, with such incidents having the capacity to negatively impact mental health in a major way, especially among children and teenagers. The proposed approach builds on our previous work that established multi-modal detection of cyberbullying on Twitter, and restructures the multi-modal approach by incorporating social media features such as time-related features and social network information. As a result, the new models reach a classification accuracy between 94.4% and 94.6%, from the previous accuracy of 93%. The proposed new approach affirms the use of context-based data in addition to more directly-related features when analyzing cyberbullying and other interactions with promising improvements. We believe that this work contributes significantly to the study of cyberbullying detection, which is an imminent problem with growing importance in the post-COVID society.

Speech separation is a special form of blind source separation in which the objective is to decouple two or more sources such that they are distinct. The need for such an ability grows as speech activated device usage increases in our every day life. These systems, however, are susceptible to malicious actors. In this work, we repurpose proven adversarial attacks and leverage them against a combination speech separation and speech recognition system. The attack adds adversarial noise to a mixture of two voices such that the two outputs of the speech separation system are similarly transcribed by the speech recognition system despite hearing clear differences in the speech. Against ConvTasNet, degradation of separation remains low at 0.34 decibels, allowing the speech recognition system to still work. When testing against automatic speech recognition, the attack achieves a 64.07% word error rate (WER) against Wav2Vec2, compared to 4.22% for unmodified samples. Against Speech2Text, the WER is 84.55%, compared to 10% WER for unmodified samples. For similarity to the target transcript, the attack achieves 24.77% character error rate (CER), reduced from 113% CER. This indicates relatively high similarity between the target transcription and the resulting transcription.

The growth in both frequency and severity of gun violence in the United States has necessitated increased research into prevention, despite the lack of funding. Comprising more than 60k gun violence media articles with a total data size of 520 MB, the gun violence database (GVDB) was developed to assist natural language processing researchers in developing and testing prevention methods. Original research based on the GVDB utilized a span-selection model to extract shooter and victim information, but their works might potentially trim out important span candidates. We proposed a new approach to improve identification accuracy and recognize every token in a sentence using a sequence tagging technique. We implemented a BIO sequence tagging model at the token-level using BERT, then further classified each token using LSTM, BiLSTM, and CRF. We found that utilizing BERT as an embedding layer, and decoding word representation as a sequence tagging task, improved shooter/victim identification compared to a span-selection model. We believe that if this improved model is combined with gun violence related keywords, automated techniques could be implemented to identify precursors/risks to gun violence on social media, allowing for intervention by law enforcement or community agencies before escalation to deaths.

We applied the state-of-art performance Deep Reinforcement Learning (DRL) algorithm, Proximal Policy optimization (PPO), to the minimal robot-legs locomotion for the challenging multi-agent, continuous and high-dimensional state-space environments. The main contribution of this work is identifying the potential factors/hyperparameters and their effects on the performance of the multi-agent settings by varying the number of agents. Based on the comprehensive experiments with 2-10 multi-walkers environments, we found that 1) A minibatch size and a sampling reuse ratio (experience replay buffer size containing multiple minibatches) are critical hyperparameters to improve performance of the PPO; 2) Optimal neural network size depends on the number of walkers in the multi-agent environments; and 3) Parameter sharing among multi-agent is a better training strategy than fully independent learning in terms of comparable performance and improved efficiency with reduced parameters consuming less memory.

With recent advancements in the field of artificial intelligence, deep learning has created a niche in the technology space and is being actively used in autonomous and IoT systems globally. Unfortunately, these deep learning models have become susceptible to adversarial attacks that can severely impact its integrity. Research has shown that many state-of-the-art models are vulnerable to attacks by well- crafted adversarial examples. These adversarial examples are perturbed versions of clean data with a small amount of noise added to it. These adversarial samples are imperceptible to the human eye yet they can easily fool the targeted model. The exposed vulnerabilities of these models raise the question of their usability in safety-critical real-world applications such as autonomous driving and medical applications. In this work, we have documented the effectiveness of six different gradient-based adversarial attacks on ResNet image recognition model. Defending against these adversaries is challenging. Adversarial re-training has been one of the widely used defense technique. It aims at training a more robust model capable of handling the adversarial examples attack by itself. We showcase the limitations of traditional adversarial-retraining techniques that could be effective against some adversaries but does not protect against more sophisticated attacks. We present a new ensemble defense strategy using adversarial retraining technique that is capable of withstanding six adversarial attacks on cifar10 dataset with a minimum accuracy of 89.31%.

The paper presents research outcomes of classifying music into moods and provides an end-to-end, open source pipeline for mood classification using lyrics. It explores techniques that classify music using audio features and lyrics using various natural language processing methods and machine learning. The paper performs a comparative study across different classification models and mood frameworks. The linguistic aspects of lyrics are explored and are used as features for classification methods to understand what model classifies mood in the most adequate manner. The results show how lyrics are a valuable information source for classification of music. Term-frequency/inverse-document frequency and word embeddings are explored to connect words to mood classes. Various machine learning and deep learning classifiers are tested across different arrangements of the mood labels. The paper demonstrates that models which learn from lyrics using current methods of natural language processing using deep learning demonstrate higher levels of accuracy. Our final model achieves an accuracy of 71%.

CAPTCHA is a web-based authentication method used by websites to distinguish between humans (valid users) and bots (attackers). Audio captcha is an accessible captcha meant for the visually disabled section of users such as color-blind, blind, near-sighted users. Firstly, this paper analyzes how secure current audio captchas are from attacks using machine learning (ML) and deep learning (DL) models. Each audio captcha is made up of five, seven or ten random digits[0-9] spoken one after the other along with varying background noise throughout the length of the audio. If the ML or DL model is able to correctly identify all spoken digits and in the correct order of occurance in a single audio captcha, we consider that captcha to be broken and the attack to be successful. Throughout the paper, accuracy refers to the attack model's success at breaking audio captchas. The higher the attack accuracy, the more unsecure the audio captchas are. In our baseline experiments, we found that attack models could break audio captchas that had no background noise or medium background noise with any number of spoken digits with nearly 99% to 100% accuracy. Whereas, audio captchas with high background noise were relatively more secure with attack accuracy of 85%. Secondly, we propose that the concepts of adversarial examples algorithms can be used to create a new kind of audio captcha that is more resilient towards attacks. We found that even after retraining the models on the new adversarial audio data, the attack accuracy remained as low as 25% to 36% only. Lastly, we explore the benefits of creating adversarial audio captcha through different algorithms such as Basic Iterative Method (BIM) and deepFool. We found that as long as the attacker has less than 45% sample from each kinds of adversarial audio datasets, the defense will be successful at preventing attacks.

Computer systems utilize logging to record events of interest. These logs are a rich source of information, and can be analyzed to detect attacks, failures, and many other issues. Due to the automated generation of logs by computer processes, the volume and throughput of logs can be extremely large, limiting the effectiveness of manual analysis. Rule-based systems were introduced to automatically detect issues based on rules written by experts. However, these systems can only detect known issues for which related rules exist in the rule-set. On the other hand, anomaly detection (AD) approaches can detect unknown issues. This is achieved by looking for unusual behaviors significantly different from the norm. In this paper, we target the problem of semi-supervised log anomaly detection, where the only training data available are normal logs from a baseline period. We propose a novel hybrid model called "CausalConvLSTM" for modeling log sequences that takes advantage of Convolutional Neural Network's (CNN) ability to efficiently extract spatial features in a parallel fashion, and Long Short-Term Memory (LSTM) network's superior ability to capture sequential relationships. Another major challenge faced by anomaly detection systems is concept drift, which is the change in normal system behavior over time. We proposed and evaluated concrete strategies for retraining neural-network (NN) anomaly detection systems to adapt to concept drift.

Online job search and talent procurement have given rise to challenging match and search problems in the e-recruitment domain. Existing systems perform direct keyword matching of technical skills which misses out a closely matching candidate on account of it not having the exact skills. This results in substandard results which ignores the relationships between technical skills. In an attempt to improve relevancy, this paper proposes a semantic similarity measure between IT skills using a knowledge based approach. The approach builds an ontology using DBpedia and uses it to derive a similarity score using feature based similarity measures. The proposed approach performs better than the Resumatcher system in finding the similarity between skills.

Object detection plays a vital role in many real-world computer vision applications such as self-driving cars, human-less stores and general purpose robotic systems. Convolutional Neural Network(CNN) based Deep Learning has evolved to become the backbone of most computer vision algorithms, including object detection. Most of the research has focused on detecting objects that differ significantly e.g. a car, a person, and a bird. Achieving fine-grained object detection to detect different types within one class of objects can be crucial in tasks like automated retail checkout. This research has developed deep learning models to detect 200 types of similar birds. The models were trained and tested on CUB-200-2011 dataset. To the best of our knowledge, by attaining a mean Average Precision (mAP) of 71.5% we achieved an improvement of 5 percentage points over the previous best mAP of 66.2%.

The online apparel retail market size in the United States is worth about seventy-two billion US dollars. Recommender systems on retail websites generate a lot of this revenue. Thus, improving recommender systems can increase their revenue. Traditional recommendations for clothes consisted of lexical methods. However, visual-based recommendations have gained popularity over the past few years. This involves processing a multitude of images using different image processing techniques. In order to handle such a vast quantity of images, deep neural networks have been used extensively. With the help of fast Graphics Processing Units, these networks provide results which are extremely accurate, within a small amount of time. However, there are still ways in which recommendations for clothes can be improved. We propose an event-based clothing recommender system which uses object detection. We train a model to identify nine events/scenarios that a user might attend: White Wedding, Indian Wedding, Conference, Funeral, Red Carpet, Pool Party, Birthday, Graduation and Workout. We train another model to detect clothes out of fifty-three categories of clothes worn at the event. Object detection gives a mAP of 84.01. Nearest neighbors of the clothes detected are recommended to the user.

A solar panel layout is a diagram of a roof, with the roof edges and obstacles marked. Currently, the user has to manually draw boundary over each obstacle in a tedious and meticulous manner. In this work, we have built a framework using the existing object detection models. We have leveraged the power of traditional edge detection algorithms, fusing with the cutting-edge machine learning based object detection frameworks. This fusion results in a framework capable of detecting objects to their exact edges. Thus, the boundary of each obstacle in a solar panel can be generated automatically with the edge pixel count variation of less than 25% compared to the ground truth.

Music is a complicated form of communication, where creators and cultures communicate and expose their individualities. Thanks to music digitalization, recommendation systems and other online services have become indispensable in the field of Music Information Retrieval (MIR). Classification of music is essential for music recommendation systems. In this paper, we propose an approach for finding similarity between music. Our approach is based on mid-level attributes like pitch, midi value, interval, contour, and duration, and applying text-based classification techniques. Performance evaluation has been done using the accuracy score of scikit-learn. As a preliminary study, our system first predicted jazz, metal, and ragtime for western music. The genre prediction system has been tested on 476 music files with a maximum accuracy of 95.8% across different n-grams. Then, we have analyzed and classified the Indian classical Carnatic music based on their raga. Our system has predicted Sankarabharam, Mohanam, and Sindhubhairavi ragas. The raga prediction system was tested on 68 music files with a maximum accuracy of 90.14% across different n-grams.

Despite clinical trials by pharmaceutical companies as well as current FDA reporting systems, there are still drug side effects that have not been caught. To find a larger sample of reports, a possible way is to mine online social media. With its current widespread use, social media such as Twitter has given rise to massive amounts of data, which can be used as reports for drug side effects. To process these large datasets, Apache Spark has become popular for fast, distributed batch processing. In this work, we have improved on previous pipelines in sentimental analysis-based mining, processing, and extracting tweets with drug-caused side effects. We have also added a new ensemble classifier using a combination of sentiment analysis features to increase the accuracy of identifying drug-caused side effects. In addition, the frequency count for the side effects is also provided. Furthermore, we have also implemented the same pipeline in Apache Spark to improve the speed of processing of tweets by 2.5 times, as well as to support the process of large tweet datasets. As the frequency count of drug side effects opens a wide door for further analysis, we present a preliminary study on this issue, including the side effects of simultaneously using two drugs, and the potential danger of using less-common combination of drugs. We believe the pipeline design and the results present in this work would have great implication on studying drug side effects and on big data analysis in general.

One challenge in text classification is that it is difficult to make feature reductions based on the definition of the features. An ineffective feature reduction may even worsen the classification accuracy. Word2Vec, a word embedding method, has recently been gaining popularity due to its high precision rate of analyzing the semantic similarity between words at relatively low computational cost. However, there is limited research about feature reduction using Word2Vec. In this project, we developed a method using Word2Vec to reduce the feature size while increasing the classification accuracy. We achieved feature reduction by loosely clustering similar features using graph search techniques. Similarity thresholds above 0.5 were used in our method to pair and cluster the features. Finally, we utilized Multinomial Naïve Bayes classifier, Support Vector Machine, K Nearest Neighbor and Random Forest classifier to evaluate the effect of our method. Four datasets with dimensions up to 100,000 feature size and 400,000 document size were used to evaluate the result of our method. The result showed that around 4-10% feature reduction was achieved with up to 1-4% improvement of classification accuracy in terms of different datasets and classifiers. Meanwhile, we also succeeded in improving feature reduction and classification accuracy by combining our method with other classic feature reduction techniques such as chi-square and mutual information.

Extensive clinical trials are required before a drug is placed on the market; yet it is difficult to discover all the side effects for any approved drugs. The United States Food and Drug Administration actively monitors approved medications to identify adverse events. The FDA Adverse Event Reporting System contains a database of adverse drug events (ADE) reported by the healthcare providers and consumers. The pervasive online social networks, such as Twitter, can provide additional information ADE. Concurrently, advancements in social media technology have resulted in the booming of massive public data; the availability of these huge datasets offers numerous research opportunities for extracting ADEs. Towards this purpose, in this paper a simple, effective computation pipeline is proposed, which uses simple drug-related classification and sentiment analysis to extract ADEs on Twitter. The pipeline is described in detail, and is implemented into an automatic process. Experiments are carried out based on 4-months of Twitter data collected. Comparing with an existing pipeline, the new design is able to successfully capture 5 times more valid ADEs, among them 20% are new ADEs. The proposed method may be applied to other areas such as food, beverages, and other daily consumer products for identifying side effects and user opinions.

Sentiment analysis is considered to be a category of machine learning and natural language processing. It is used to extricate, recognize, or portray opinions from different content structures, including news, audits and articles and categorizes them as positive, neutral and negative. It is difficult to predict election results from tweets in different Indian languages. We used Twitter Archiver tool to get tweets in Hindi language. We performed data (text) mining on 42,235 tweets collected over a period of a month that referenced five national political parties in India, during the campaigning period for general state elections in 2016. We made use of both supervised and unsupervised approaches. We utilized Dictionary Based, Naive Bayes and SVM algorithm to build our classifier and classified the test data as positive, negative and neutral. We identified the sentiment of Twitter users towards each of the considered Indian political parties. The results of the analysis for Naive Bayes was the BJP (Bhartiya Janta Party), for SVM it was the BJP (Bhartiya Janta Party) and for the Dictionary Approach it was the Indian Nathional Congress. SVM predicted a 78.4% chance that the BJP would win more elections in the general election due to the positive sentiment they received in tweets. As it turned out, BJP won 60 out of 126 constituencies in the 2016 general election, far more than any other political party as the next party (the Indian National Congress) only won 26 out of 126 constituencies.

As social media has become increasingly popular in the modern world, people are using these platforms to express their opinions about products, businesses, and services. The need for categorizing these consumer reviews has been prominent. One effective solution is sentiment analysis (SA), which has been an active research topic. The goal of SA is to automatically extracting and classifying user opinions. Pervious research works however have not shown satisfied results. In this paper, a multilayer architecture is proposed to increase the performance of multiclass classification. The framework includes data-preprocessing, feature extraction and selection, and classifier building. The framework is a two-layer classification, choosing from Naïve Bayes, Support Vector Machine, Random Forest, and Logistic Regression as base models, and using a voting scheme to obtain the final predicted class. The proposed model is applied to more than 1.3 million restaurant reviews from the Yelp Challenge dataset. We have achieved a high accuracy of 86% for cross validation, and using real-world online review data as test data, we have achieved an accuracy of 80%. The results show that the proposed framework has greatly improved classification accuracy while comparing with those using single-layer architectures. We believe that the proposed method may be applied to, and would have significant contributions to other areas of opinion mining.

Efforts have been made to introduce an extra layer of security on mobile devices, including a good amount of research initiated in the behavioral biometrics domain. However, all prior research approaches for mobile gesture-based authentication has been carried out uni-directionally. Despite of the fact that there are many devices with their own configurations, the study of mobile authentication based on behavioral biometrics has been done only with the Android operating system and devices. In this paper, a novel approach to identifying the owner of a mobile device based on Behavioral Biometrics Mobile Gestures Recognition is presented. This research takes the first step towards implementing behavioral biometrics identification for iOS based iPhone devices. In this research work, it is shown that a user can be identified as the true owner or an imposter of such a device based on the interactive behavior and gestures of the user. In this way continuous identification or authentication of an owner can be done based on the interaction of the user and the device. It is shown that a continuous authentication mechanism can be established using a self-learning model based on machine learning classification approaches such as Random Forests, Gradient Boosting Machine, Deep Learning, and Naive Bayes. The results in this paper show that, with behavioral biometrics, automated user authentication mechanism, EER (Equal Error Rate) can be improved to around 27%, clearly demonstrating that the chances of authenticating the user are good.

Although rigorous clinical studies are required before a drug is placed on the market, it is impossible to predict all side effects for the approved medication. The United States Food and Drug Administration actively monitors approved drugs to identify adverse events. The FDA Adverse Event Reporting System (FAERS) contains a database of adverse drug events reported by the healthcare providers and consumers. The ubiquitous online social networks, such as Twitter, can provide complementary information about adverse drug events. Short Twitter postings, or tweets, are often used to express an opinion about drugs, as well as solicit and receive feedback from consumers of a drug. Thus, adverse drug events can be discovered by extracting from tweets users' opinions about drugs. Here, we developed a computational pipeline for collecting, processing, and analyzing tweets to find signals about adverse drug reactions, defined as drug side effects caused by a drug at a normal dose during normal use. Manual examination of processed tweets identified several known side effects of four drugs.

This paper introduces the concept of graphing the size of a level-set against its respective density threshold. This is used to develop a new recursive version of DBSCAN that successfully performs hierarchical clustering, called Level-Set Clustering (LSC).

The electrical power grid forms the functional foundation of our modern societies, but in the near future our aging electrical infrastructure will not be able to keep pace with our demands. As a result, nations worldwide have started to convert their power grids into smart grids that will have improved communication and control systems. A smart grid will be better able to incorporate new forms of energy generation as well as be self-healing and more reliable. This paper investigates a threat to wireless communication networks from a fully realized quantum computer, and provides a means to avoid this problem in smart grid domains. We discuss and compare the security aspects, the complexities and the performance of authentication using public-key cryptography and using Merkel Trees. As a result, we argue for the use of Merkle Trees as opposed to public key encryption for authentication of devices in wireless mesh networks (WMN) used in smart grid applications.

Blogs have become a major communication media, and have recently been extremely effective in making political and social changes. It is therefore vital to recognize powerful, persuasive bloggers in a web community. In this work we examine existing models for identifying influential bloggers, and construct an improved model based on two new factors: uniqueness and FacebookCount. The former measures the originality of a post, combining with outlink count they represent the novelty of the post. The latter reflects the influence of emerging social network platforms, and can be extended to include twitter share, G +1's, etc. The proposed model also adopts other effectual factors including the number of inlinks, outlinks and comments, the timing of posts and of comments, and the influence of commenters. In addition, to capture the true influence of a post we mine through each comment on a post to identify the sentiment, or the tone of the comment. The experiments show that, comparing with an existing model [3], the proposed approach is able to capture the true influence, and to give bloggers distinctive rankings. We believe that the two new factors are timely and are significant for identifying influential bloggers in the blogosphere.

We developed software that performs priority-based power management and downtime reduction for virtual machines running in data centers. The software deals with power management at the processor level. The software automatically performs load distribution among servers in data centers to save power. In addition, it reduces downtimes for critical virtual machines. Our experiments have achieved an energy saving of up to 35%. We therefore confirm that the proposed scheme has significantly decreased energy consumption while maintaining high runtime availability for the mission critical applications.

One major challenge in broadband wireless access networks is to provide fast, reliable services to time-sensitive communications. We propose an enhanced hybrid automatic repeat request (HARQ) scheme for the worldwide interoperability for microwave access (WiMAX). The new scheme follows the multi-channel stop-and-wait chase combining HARQ adopted by WiMAX, yet it enables the base station (BS) to proactively react to poor channel conditions. The BS would send multiple copies of the same data burst to the subscriber station (SS) over the multiple HARQ channels available to the SS. We design a novel method which would determine the number of multiple copies needed based on the channel feedback. The new HARQ scheme effectively reduces the total time needed for a successful transmission. Simulation results show that, comparing with the original WiMAX HARQ, the new scheme significantly reduces the waiting time for a successful transmission by up to 45%, while maintaining a comparable level of throughput. This work is most significant for time-critical applications that need fast, reliable services, such as VoIP and other interactive uses.

Wireless sensor network (WSN) is a technology that will become an integral part of everyday life. It has countless potential applications in ubiquitous and pervasive computing, including those in biomedical technology and healthcare industry. As the number and variety of wireless sensor device types increase, a middleware layer capable of linking various implementations into a single programming abstraction will be needed. This paper aims to build a platform that allows users to develop WSN applications with ease, and is robust enough to support complex system development. A flexible platform based on Java and JADE (Java Agent Based Development) is proposed to handle sensor network integration and application development. Design methodology is carefully presented. A prototype implementation is described, and illustrated by an application in the healthcare domain. We believe that this work would help bridging the gaps existed among WSN researchers, application developers, and users, and is significant in realizing the full potential of WSN.

Transmission control protocol (TCP) plays an important role in defining a network's performance. Its use in wireless networks has exposed several inadequacies in its operation. We propose an improvement over TCP for communication between fixed host (FH) senders and wireless, mobile host (MH) receivers that are interlinked by base stations (BS). We believe that this is becoming the main infrastructure supporting most mobile Internet traffic. Our new scheme is called TCP-ECN (explicit congestion notification). It aims to address the challenges of high bit error rate in wireless links and long disconnections due to mobility. It keeps the modification in BS and MH minimal without requiring any changes to FH. The BS uses one bit (ECN bit) to enable the MH in distinguishing between congestion and wireless losses, and thus, suppressing unnecessary duplicate acknowledgments (dupACK) due to wireless losses. In addition, zero window acknowledgment and triplicate dupACK schemes are adopted as part of the handoff procedure in BS. Both simulation and network test-bed results show that TCP-ECN performs significantly better than TCP-Reno and Snoop in the face of high wireless loss, high or low congestion loss, and mobility. Performance improvement is more significant when more wireless receivers are supported. We believe this work is significant to wireless carriers and vendors when designing protocols and devices that support mobile Internet access.

Search engines give so many results that a user cannot handle them all in a short period of time. Many approaches have been proposed to alleviate this problem. For example, some approaches try to add personalized features and some try to group the results into different categories. The latter one is called a clustering engine, which is the emphasis of this paper. It first reviews several existing approaches such as STC, SHOC, LINGO and SnakeT. It then gives a new approach called HICSEC (Hierarchical Clustering Search Engine with Correlation) to improve the accuracy of clustering by the Correlation calculated with the Singular Value Decomposition.

Wireless mesh network is a promising new direction resulting from the recent rapid progress made in wireless communication technologies. While multiple radio, multiple channel technologies have offered a great potential for increasing network throughput, they also create new research challenges. In this paper, we propose RingMesh, a token ring-based protocol for multi-channel routing. By applying delay-guaranteed rules for joining a ring and for creating new rings, the multiple-ring protocol bounds the end-to-end packet delay within the mesh network, from the source node to a gateway that connects with the wired network. The RingMesh protocol is described as a state machine; the analytical results on bounded delay are presented. We believe that both the protocol and analysis may be applied to other areas of delay-guaranteed wireless network research

This paper investigates the effectiveness of data gathering protocols for in-body biomedical sensor networks. We studied the performance of representatives from each of three major protocol categories: (1) low energy adaptive clustering hierarchy (LEACH), a cluster-based protocol, (2) power efficient gathering for sensory information systems (PEGASIS), a chain-based protocol, and (3) hybrid indirect transmissions (HIT), a hybrid of chains and clusters. First, the ability of each protocol to perform in-network source separation was judged. We consider a human-machine interaction application in which implanted bio-sensors communicate motor unit actions of human muscles to a remote computer. Motor unit action potentials (MUAPs) were modeled, and source separation and recovery at each sensor was simulated. We compare the performance of HIT and LEACH in terms of signal distortion ratios and the energy costs of fusion and communication. Second, we report on the efficiency of each protocol for in-body data collection, using Gupta et al's propagation loss model for biomedical applications (PMBA) - an accurate model of power loss due to signal absorption by the human body. We investigate the effectiveness of HIT, LEACH, and PEGASIS under this model, and compare their performance in terms of energy efficiency and network lifetime

Group mutual exclusion (GME) was first introduced by Joung in 1998 as a generalization of the n-process mutual exclusion problem, and subsequently modeled as the Congenial Talking Philosophers problem. GME allows processes requesting for the same resource to concurrently access the shared resource. However, a process requesting a resource that is different from the one currently being used will not be able to access the requested resource at the same time. At any time, only one resource may be in use.There have been two solutions to the GME problem for a ring network. The first solution proposed by Wu and Joung [2] had an unbounded message size problem, while the second solution proposed by Cantarell et. al. [1] attempted to provide an upper-bound for the message size. However, in doing so, the second solution generated an unbounded number of messages. Hence, in this work we present a GME algorithm for a token-passing network that solves both of the problems found in the two previous solutions by providing an upper bound both to the number of messages and to message size.

One major challenge in sensor networks is to maximize network life under the constraint of extremely limited power supply. Thus, an important design issue of routing and data-gathering protocols is minimizing energy. This paper investigates the energy efficiency of two data gathering protocols, based on distributed versions of Shortest Path Tree (SPT) and Maximum Leaf Tree (MLT) algorithms [1]. The two have been extended to be dynamic by applying localized tree-reconstruction mechanism [2] to handle joining and leaving (death) of sensor nodes. Accurate energy consumption has been carefully modeled for both leaf-nodes and intermediate nodes, when sending and receiving data. Performance is evaluated through detailed simulation, including exchange of control messages among sensor nodes. Simulation results have shown that SPT, due to its simplicity and smaller number of control messages, achieves better energy efficiency and less delay in tree constructions, data transmissions, and dynamic tree reconstructions.

The most important contribution in the title paper by F.Y. Young et al. [ibid., vol. 20, pp. 687-692, 2001] is that they devised an efficient method to compute the shapes of soft modules to give the optimal packing, as described by the authors. We would like to clarify a misunderstanding that occurred when the authors referred to our paper [IEEE Trans. Circuits Syst. I, vol. 43, pp. 713-720, 1996].

The complexity of multimedia applications, which integrate a variety of information sources, such as audio, voice, graphics, images, animation, and full-motion video, into a wide range of applications, stresses all the components of computer and communication systems. Many new ideas have been proposed and implemented for advanced LAN (Local Area Network) protocols in order to support multimedia networking. In this paper we study three recently proposed high speed LAN protocols including the 100 Base-T Fast Ethernet, the Ethernet++, and the 100 VG-AnyLAN. We also propose a new enhancement for the 100 VG-AnyLAN. For these four protocol, we describe their design methodology, strengths, weaknesses, and compare their performance in supporting multimedia communications including both real-time and non-real-time traffic. We believe that the results and discussions presented here provide important information for network designers and engineers when designing or deciding network protocols for multimedia systems.

In a multiple access broadcast network, all network nodes share a single shared communication channel, and there is the possibility of a collision when two or more nodes transmit at overlapping times. We propose a dynamic prioritized conflict resolution algorithm in which, when a collision occurs, all colliding messages are retransmitted according to their priority. When a new message arrives, it is allowed to participate in the algorithm as soon as it finds its priority is higher than that of some broadcast message. Using a time-slotted model, for any arrival pattern and priority distribution, we show that our algorithm runs in expected linear time, i.e., O(r), where r is the total number of message transmitted. We also show that the expected waiting time for any message x is O(rank(x)+log s), where rank(x) is the expected number of messages with higher priority which transmit while x is waiting, and s is the minimum of the total number of messages participating in the algorithm and the total number of nodes in the network. The analysis presented in this work also includes an improved analysis of our original (static) prioritized conflict resolution algorithm. This work is applicable to multimedia communications where different priority values could be assigned to different kinds of traffic, and our algorithm ensures that the high-priority real-time traffic has the minimum delay.

Multimedia applications integrate a variety of media namely, audio, video, images, graphics, text, and data, each of which has different quality of service (QoS) requirements. To study the support of multimedia traffic on high speed protocols, we examine two significant metrics: delay fairness and worst-case delay performance. Four members of reservation-based high speed protocols are studied DQDB (distributed queue dual bus), CRMA (cyclic reservation multiple access), DQMA (distributed queue multiple access), and FDQ (fair distributed queue). The first part of this work presents delay fairness study of the protocols under various input traffic. Both access delay and message delay experienced by individual network nodes are measured to illustrate their fairness performance. The second part conducts worst-case delay experiments. Too worst-case scenarios of a test message are studied. The delay experienced by this message under varied inter-node distance is measured. The simulation results show that DQDB has the worst fairness performance. CRMA and DQMA present mixed results. FDQ proves itself to be a very fair protocol. The results also suggest that to there is no singe metric to justify a new protocol. Comparing our results with previous results on throughput evaluation of multimedia traffic support, we found that the two results are strongly correlated the fairer a protocol is, the better it is for supporting heterogeneous traffic under heavy network load.