Search

Many recent relation extraction methods are proposed based on distantly supervised learning to address the issue of expensive human labelling. They automatically align relation instances from a knowledge base with unstructured text. Although these methods solve the problem of insufficient labelled data and are scalable, due to automatic labelling, they suffer from a large amount of noisy data. In this paper, we propose a relation extraction method based on a BiLSTM approach to alleviate the noise by automatic labelling. It firstly performs unsupervised pretraining for word embeddings, dependency path tokens and entity embeddings. The attention mechanism we employed in the pretraining module helps select relevant information. As pretraining denoise and encode the semantic information, the relation types can be more efficiently classified by supervised learning using less labelled training data. Experiments show that our approach significantly outperforms previous state-of-the-art baselines when labelled training data is insufficient.

Although matrix factorization and its variants have proved their effectiveness in learning user preferences, they violate the triangle inequality and may fail to capture the inner grained preference information. Thus, metric learning based models have attracted increasing interests in recommender systems. However, most of these models only measure the distance in Euclidean space, though it is known to limit the prediction performance. To address this problem, this paper proposes curvilinear collaborative metric learning (CCML) as a solution. Rather than learning from Euclidean distance, CCML measures the “path” (i.e. curvilinear distance) between the user and items. Therefore, it is able to better infer information (in geometrical representation) and user-item relations behind the observed data points. Under this setting, the relations between the user and items are not linearly reflected by the relation vector. In order to effectively learn the representation of items, we propose to fuse macro features of categories and micro representations of items to form their latent vectors. Our model can be less prone to be affected by the noise commonly presented in datasets by separately learning macro and micro features. Experimental results on eight public recommendation datasets demonstrate that CCML produces a outstanding performance compared with five competitive baselines. Finally, quantitative analysis is included to verify the efficiency and effectiveness of CCML and the macro-micro module.

Relation extraction is an important task for information extraction aiming to detect and extract semantic relationships between entity pairs in sentences. A lot of recently proposed methods for relation extraction have gained remarkable results using deep neural networks. However, a deep network usually requires expensive computational power and is time-consuming when training. To address this problem, we introduce a novel neural relation extraction model that jointly trains a seq2seq regulariser with a classification module. As the regulariser and the classification module share the encoder, the whole model is less complicated compared to novel approaches like BERT. We show that our model achieves the same or better performance than the pre-trained BERT model using sufficient labelled data in much less training time.

In future long-duration exploration missions (LDEMs), communication with mission control will be significantly delayed. Crews will often need to react to time-sensitive issues or hazards without relying on mission control for assistance. The need for a higher degree of crew autonomy without direct oversight from mission control introduces the implementation of virtual assistants (VAs) to aid the crew; however, we must first create standards and guidelines for VAs in this context. For this purpose, we have developed a VA called Daphne-AT (Anomaly Treatment) to investigate the interaction between astronauts and virtual assistants in the context of anomaly resolution related to the Environmental Control and Life Support System (ECLSS). A series of experiments will be conducted in a laboratory environment and at the Human Exploration Research Analog (HERA) at NASA Johnson Space Center (JSC) to study user's performance, situational awareness, cognitive workload, and trust in Daphne-AT. Subjects in a simulated LDEM will be given several ECLSS anomalies to identify and solve with and without the support of Daphne-AT. We based the scenarios we developed on existing ECLSS hardware and pre-existing anomaly resolution procedures from the HERA environment. Solving an anomaly requires the subject to complete the correct anomaly resolution procedure(s) for the appropriate anomaly, which could include various tasks such as swapping out component parts and activating or deactivating ECLSS systems. In experiment sessions without Daphne-AT, subjects will rely on telemetry feeds as well as background knowledge and training to solve anomalies. However, during experiment sessions with Daphne-AT, the VA will also assist subjects in detecting and diagnosing these anomalies. We conducted an initial set of experiments at Texas A&M University (TAMU) prior to those at HERA. However, the subject must complete anomaly resolution procedures during an experiment that require access to hardware and components not available at the TAMU location. To emulate the HERA ECLSS hardware elements in our laboratory at TAMU, we recreated a virtual 3D representation of HERA and its ECLSS systems using the game engine Unreal Engine 4 (UE4). This virtual model will act as an analog for subjects who are tested on TAMU's campus to complete relevant ECLSS procedures. All subsystems featured in the UE4 model are interactive and allow the user to perform steps in an anomaly resolution procedure similar to how the user would do it in HERA. The UE4 model allows the user to complete more than 20 anomaly resolution procedures, thus increasing the TAMU experiment's fidelity compared to those conducted at HERA. This paper describes the development of the UE4 model and how we used it to validate the experimental protocol implemented at HERA. The results of these experiments will inform future guidelines for VAs deployed on LDEMs.

The Global Exploration Roadmap provides a pathway for humans to leave lower Earth orbit and develop a sustained presence in cislunar space as well as the lunar surface before embarking on a long journey to Mars. During these crewed Long Duration Exploration Missions (LDEM), communication delays will limit mission control availability to support the crew during critical in-flight anomalies. In order to address this challenge, future human spaceflight operations will benefit from intelligent Virtual Assistants (VAs) capable of assisting the crew in detecting, diagnosing, and resolving emergency anomalies. This intelligent technology could increase the probability of mission success and crew safety in emergency scenarios and could improve overall crew performance while operating in the hostile environment of deep space. Daphne-AT is a VA to support crewed LDEM in the context of diagnosing and resolving in-flight anomalies related to the Environmental Control and Life Support System (ECLSS) of a spacecraft. The overall research objective is to investigate the effect of Daphne-AT on human performance, cognitive workload, situational awareness, and trust in autonomous systems. Daphne-AT will be validated using controlled laboratory experiments, and further evaluation will occur during the Human Exploration Research Analog (HERA) C6 research campaign, consisting of four 45-day missions in 2020–21. In this paper, we provide an overview of our first version of Daphne-AT (i.e., baseline version), and discuss the design of anomaly scenarios using NASA's ECLSS simulator and the experimental set up for laboratory experiments. A preliminary within-subjects (i.e., with and without Daphne-AT) experiment was conducted with five human subjects, and preliminary results indicate that the use of Daphne-AT increases performance and decreases cognitive workload in the context of ECLSS anomaly scenarios. The data also show that subjects exhibited trust in using Daphne-AT, while it improved certain aspects of their situational awareness. Our work is part of the Human Capabilities Assessments for Autonomous Missions (HCAAM) Virtual NASA Science Center of Research (VNSCOR) within the NASA Human Research Program, and the ultimate goal is to generate standards and guidelines about how to design such VAs for future LDEMs.

Latent variable models have been widely adopted by recommender systems due to the advancements of their learning scalability and performance. Recent research has focused on hybrid models. However, due to the sparsity of user and/or item data, most of these proposals have convoluted model architectures and objective functions. In particular, the latter are mostly tailored for sparse data from either user or item spaces. Although it is possible to derive an analogous model for both spaces, this makes a system overly complicated. To address this problem, we propose a deep learning based latent model called Distilled Hybrid Network (DHN) with a teacher-student learning architecture. Unlike other related work that tried to better incorporate content components to improve accuracy, we instead focus on model learning optimization. To the best of our knowledge, we are the first to employ teacher-student learning architecture for recommender systems. Experiment results show that our proposed model notably outperforms state-of-the-art approaches. We also show that our proposed architecture can be applied to existing recommender models to improve their accuracies.

Big data has been transformed into knowledge by information systems to add value in businesses. Enterprises relying on it benefit from risk management to a certain extent. The value, however, depends on the quality of data. The quality needs to be verified before any use of the data. Specifically, measuring the quality by simulating the real life situation and even forecast it accurately turns into a hot topic. In recent years, there have been numerous researches on the measurement and assessment of data quality. These are yet to utilize a scientific computational method for the measurement and prediction. Current methods either fail to make an accurate prediction or do not consider the correlation and time sequence factors of the data. To address this, we design a model to extend machine learning technique to business applications predicting this. Firstly, we implement the model to detect data noises from a risk dataset according to an international data quality standard from banking industry and then estimate their impacts with Gaussian and Bayesian methods. Secondly, we direct sequential learning in multiple deep neural networks for the prediction with an attention mechanism. The model is experimented with various network methodologies to show the predictive power of machine learning technique and is evaluated by validation data to confirm the model effectiveness. The model is scalable to apply to any industries utilizing big data other than the banking industry.

Many recently proposed relation extraction methods are based on distantly supervised learning. They use data from existing knowledge bases as training data. Although the methods solve the problem of insufficient labelled data and are highly scalable, they suffer from a large amount of incorrectly labelled data. Instead of using these distantly supervised approaches, this paper proposes an alternative relation extraction method. It firstly performs unsupervised learning to train relation embeddings by a neural network. As the relation embeddings encode the semantic information of the original sentences and their entity pairs, these embeddings can be efficiently classified by supervised learning. Since the relation embedding phase is based on unsupervised learning, labelled data is only required in the classification phase. Experiments show that our proposed approach significantly outperforms the state-of-the-art baselines when labelled training data is insufficient.

Statistical divergence is widely applied in multimedia processing, basically due to regularity and interpretable features displayed in data. However, in a broader range of data realm, these advantages may no longer be feasible, and therefore a more general approach is required. In data detection, statistical divergence can be used as a similarity measurement based on collective features. In this paper, we present a collective detection technique based on statistical divergence. The technique extracts distribution similarities among data collections, and then uses the statistical divergence to detect collective anomalies. Evaluation shows that it is applicable in the real world., Comment: 6 pages, 6 figures and tables, submitted to ICME 2020

Recently, a new algorithm named dynamic group optimization (DGO) has been proposed, which is developed to mimic the behaviors of animal and human group socializing. However, one of the major drawbacks of the DGO is the premature convergence. Therefore, in order to deal with this challenge, we introduce chaos theory into the DGO algorithm and come up with a new chaotic dynamic group optimization algorithm (CDGO) that can accelerate the convergence of DGO. Various chaotic maps are used to adjust the update of solutions in CDGO. Extensive experiments have been carried out, and the results have shown that CDGO can be a very promising tool for solving optimization algorithms. We also demonstrated good results based on real world data, where, in particular, solving multimedia data clustering problems.

Matrix factorization is a popular method for building recommendation models. On e-commerce platforms, this method makes predictions of product ratings for goods which have not been rated. Similarly, in service computing, service rating platforms have been proposed to help users to select services. The idea is constantly evolving and the proposed models are often only validated by synthetic data. Existing recommendation systems rarely consider the fact that while customer feedbacks are usually valuable, some are questionable. Hence, how objective the given ratings are is an important factor. By considering the contribution of each rating according to its helpfulness and its objectivity, this paper proposes a platform that can make precise and cross-category recommendations. We exploit the parallelism between service and product recommendations to validate our proposed model by real-world data.

Relation extraction is a critical task in building a knowledge base from unstructured text documents. Most works in automatic relation extraction have applied deep learning techniques such as Convolutional Neural Network (CNN) and Long Short-Term Memory (LSTM) in large text corpora. However, they require a large amount of human labelling data, which is labour intensive and is hardly applied in a new domain of document without human supervision. This paper proposes a novel framework to extract relations in multi-domain texts effectively. In particular, we construct the framework in three phases including preprocessing, feature embedding and relation extraction. We show that a small proportion of training data is sufficient to train our relation extraction framework and achieve a good accuracy in relation extraction works.

In complex design projects, engineering and program management disciplines often work in isolation leading to inconsistencies in product information, tracking of design changes, challenges in decision-making, and potential product and project failure. Moreover, the lack of information early in the design can postpone decision-making further into the design cycle when it becomes more expensive and difficult to make changes. Differences across disciplines in their respective practices and lack of a common platform for exchanging information have contributed to this gap. In bridging this gap, an analytical decision support system is presented that enhances engineering design change management and knowledge management early in the design. The human factors involved in embracing such a system and adopting a structured knowledge management framework are investigated through a review of knowledge management theory, conducting a social network analysis, and analyzing the results of a survey of a west coast shipyard design team. To understand enablers and barriers to design change management and knowledge management, a system dynamics causal loop diagram is presented. The decision support system and design change - knowledge management framework presented in the current study is a viable option to bring multiple disciplines together on a common platform and to provide a method to enable informed decisions early in the design cycle. The iterative and dynamic nature of this model is aligned with the principles of agile management as well as other best practices in managing programs and complex systems.

The core of deep learning which is convolution neural network (CNN) has been widely adopted in image processing and object recognition areas. In particular, medical imaging requires very precise, accurate and fine recognition power. Numerous works in the literature have reported promising prospects of CNN applied in prognosis and radiology diagnosis. A common goal among those works, largely is to try achieving a most accurate deep learning model in analysing the insights from the finest details of the medical imaging. To this end, a novel machine learning model that is equipped with multi-view data-preprocessing and swarm-based hyperparameter optimization is proposed. The former is for provi ding additional training data in the hope that salient features could discovered; the latter is for finding the most optimal set of model parameters for the CNN. They both serve only one purpose – to enhance the object recognition power to the highest possible. Preliminary experiments over datasets related to heart disease detection and breast cancer classification over CTG and mammograms respectively indicate encouraging results.

Association Rule Mining by Aprior method has been one of the popular data mining techniques for decades, where knowledge in the form of item-association rules is harvested from a dataset. The quality of item-association rules nevertheless depends on the concentration of frequent items from the input dataset. When the dataset becomes large, the items are scattered far apart. It is known from previous literature that clustering helps produce some data groups which are concentrated with frequent items. Among all the data clusters generated by a clustering algorithm, there must be one or more clusters which contain suitable and frequent items. In turn, the association rules that are mined from such clusters would be assured of better qualities in terms of high confidence than those mined from the whole dataset. However, it is not known in advance which cluster is the suitable one until all the clusters are tried by association rule mining. It is time consuming if they were to be tested by brute-force. In this paper, a statistical property called prior probability is investigated with respect to selecting the best out of many clusters by a clustering algorithm as a pre-processing step before association rule mining. Experiment results indicate that there is correlation between prior probability of the best cluster and the relatively high quality of association rules generated from that cluster. The results are significant as it is possible to know which cluster should be best used for association rule mining instead of testing them all out exhaustively.

Although past research in finance showed that future investment returns are hardly predictable, predictions in volatility are found possible. Financial news is known to be making persuasive impacts on investor behavior, and hence injecting disturbances to markets. While there exists many prediction services, most of them do not factor in financial news directly. Previously proposed supervised topic model provides an avenue to associate volatility with news, yet it displays poor resolutions at extreme regions. To address this problem, we propose a novel extreme topic model to provide a better service for market alerts. By mapping extreme events into Poisson point processes, volatile regions are magnified to reveal their hidden volatility-topic relationships. Derivative approach is used to quantify the volatility of financial instruments. It captures immediate, collective responses from investors on market events, e.g., Brexit. By acquiring domain knowledge on how financial news influence investor behavior, accurate volatility predictions are made under extreme conditions as shown by the improved prediction accuracy in our experiments.

The authentication and verification of devices require lightweight and cost-effective solutions such as the use of fingerprints of the components contained in the device. The power-up state of Static Random Access Memory (SRAMs) has a significant effect on some security and authentication applications, such as SRAM Physical Unclonable Function (PUFs) and Random Number Generation (RNG). In this paper, a detailed data collection scheme and corresponding statistical and physical analysis are presented. The results of tests on different models of SRAM chips show that the power-up state of SRAM need to be carefully collected in order to avoid increasing the volume of enrollment data. Otherwise, the probability of failure to verify will increase. Furthermore, this method provides a stable and reliable power-up state of SRAM chips and can be used as a guidance for data acquisition and analysis of similar applications. We illustrate the Discharge Inversion Effect (DIE) for SRAM chips from different manufacturers, and demonstrate how it may impact SRAM power-up applications unless our guidelines for proper data collection and use are followed.

Advances in technology and the need to meet many interdependent competing requirements have led to complex information intensive systems that can be best managed by integrating multi-discipline, multi-attribute models for better informed decision-making. Attempts to bridge the gap between systems engineering and project management have been sought through project management methods and through imposing practices from one discipline onto another. It is proposed that a more rigorous approach to bridging this gap be investigated to better enable the optimization of objectives, collaboration, and cooperation amongst multiple disciplines. The disciplines of interest in this study are plant modeling, systems engineering, and program management. In bridging the gap, both leveraging models from the different disciplines and adopting six sigma tools and visualization techniques are investigated. Common variables and outcome-based attributes provide a mechanism to integrate multi-discipline models. Options to bring together models from each discipline include multi-disciplinary system design optimization (MSDO). To demonstrate the feasibility of integrating multi-discipline attributes on a common platform, a case study on design of a notional future marine Integrated Power System (IPS) is investigated using a modified Multi-Attribute Tradespace Exploration (MATE) — Capability model. This case study highlights the need to integrate other multi-discipline models to help validate specific system attributes of interest.

Infrastructure failures have severe consequences which often have a negative impact on the society and the economy. In this paper, we propose a machine learning model to assist in risk management to minimise the cost of infrastructure maintenance. Due to the vast volume and complexity of infrastructure datasets, such problem is often computationally expensive to compute. A Bayesian nonparametric approach has been selected for this problem, as it is highly scalable. We propose a two-stage approach to model failures, such as water pipe failures. The first stage uses an Infinite Gamma-Poisson Mixture Model to group water pipes with similar characteristics together based on the number of failures. The second stage uses the groups created in the first stage as an input to the Hierarchical Beta Process (HBP) to rank water pipes based on their probability of failure. The proposed method is applied to a metropolitan water supply network of a major city. The experiment results have shown that the proposed approach is able to adapt to the complexity of tge large multivariate dataset and there is a double-digit improvement from the grouping created by domain experts.

Memes are topics that spread among people in a community. They could be ideas, behaviors, or styles, etc. Recent research found that the spreading of memes on microblogs displays epidemic patterns, while their popularity is predictable. Aiming for high accuracy, complex models that consider the structural details of community graphs are common. On the contrary, some models simply assume a community is evenly connected to sacrifice accuracy for efficiency. In this paper, we propose a prediction service that delivers the benefits of both worlds. Our service considers the properties of community graphs yet without the burden of their structural details to reduce complexity for online consumption. By comparing the viralities of memes against a community threshold, memes that would lead to population-wide coverage are identified. We also carefully examine sampling biases. Hence, reliable thresholds and accurate predictions are made as shown in our experiments.

Several technical and management disciplines are involved with complex system development and product life cycle management. Often they work in isolation leading to inconsistencies in product specifications, information, and challenges in decision-making. Differences across disciplines in their respective practices, techniques, tools, system metrics, and lack of a common platform for exchanging information have contributed to this gap. Attempts to bridge the gap between systems engineering and project management have been sought through project management methods and through imposing practices from one discipline onto another. It is proposed that a more rigorous approach to bridging this gap be investigated to better enable optimization of objectives, collaboration and cooperation amongst the multiple disciplines. The disciplines of interest in this study are plant modeling, systems engineering, and program management. In bridging the gap, both leveraging techniques and management tools from the different disciplines and adopting six sigma statistical methods are investigated. Common and unique system performance measurements are investigated in working toward a common language. Shared outcome-based measurements such as system and component operational capability levels, predicted and achieved, are also proposed. Options to bring together models from each discipline include multi-disciplinary system design optimization. The different management platforms and visualization techniques are explored with the goal to adopt a common platform to share information. The governance structure that would enable sharing of information across the disciplines forms part of this research.

Due to the dynamic nature of commerce, it is important to capture useful external information as a support to build appropriate strategy and make decision in cur-rent competitive market with large volume of data. A review and summary of the current competitive intelligence development, mainly new concepts and tools, is conducted in this report for proposing a system for food and beverage industry that aims at obtaining competitive advantage in Macau market.