Data-driven Systems Research Articles

Unsupervised Learning for Product Use Activity Recognition: An Exploratory Study of a "Chatty Device".

To create products that are better fit for purpose, manufacturers require new methods for gaining insights into product experience in the wild at scale. “Chatty Factories” is a concept that explores the transformative potential of placing IoT-enabled data-driven systems at the core of design and manufacturing processes, aligned to the Industry 4.0 paradigm. In this paper, we propose a model that enables new forms of agile engineering product development via “chatty” products. Products relay their “experiences” from the consumer world back to designers and product engineers through the mediation provided by embedded sensors, IoT, and data-driven design tools. Our model aims to identify product “experiences” to support the insights into product use. To this end, we create an experiment to: (i) collect sensor data at 100 Hz sampling rate from a “Chatty device” (device with sensors) for six common everyday activities that drive produce experience: standing, walking, sitting, dropping and picking up of the device, placing the device stationary on a side table, and a vibrating surface; (ii) pre-process and manually label the product use activity data; (iii) compare a total of four Unsupervised Machine Learning models (three classic and the fuzzy C-means algorithm) for product use activity recognition for each unique sensor; and (iv) present and discuss our findings. The empirical results demonstrate the feasibility of applying unsupervised machine learning algorithms for clustering product use activity. The highest obtained F-measure is 0.87, and MCC of 0.84, when the Fuzzy C-means algorithm is applied for clustering, outperforming the other three algorithms applied.

Persona analytics: Analyzing the stability of online segments and content interests over time using non-negative matrix factorization

Personified big data and rapidly developing data science techniques enable previously unforeseen methodological developments for longitudinal analysis of online audiences. Applying data-driven persona generation on online customer statistics from a real organizational social media channel, we demonstrate how personas can be deployed to understand online customer patterns over time. We conduct 32 monthly rounds of data collection of customer demographics and content consumption patterns on the YouTube channel of a major publishing organization posting thousands of items of content and then algorithmically generate 15 personas monthly. We analyze the data-driven persona for changes monthly, yearly, and lifetime (period). Results show an average 40% change in the personas, and 78% of the personas experience more change than consistency for topic interests. The implications are that organizations frequently publishing online content should employ automatic data collection and periodic persona creation to ensure their customer understanding is current. For this, algorithmic data-driven systems that leverage methods for persona creation are recommended.

Collocation based training of neural ordinary differential equations.

The predictive power of machine learning models often exceeds that of mechanistic modeling approaches. However, the interpretability of purely data-driven models, without any mechanistic basis is often complicated, and predictive power by itself can be a poor metric by which we might want to judge different methods. In this work, we focus on the relatively new modeling techniques of neural ordinary differential equations. We discuss how they relate to machine learning and mechanistic models, with the potential to narrow the gulf between these two frameworks: they constitute a class of hybrid model that integrates ideas from data-driven and dynamical systems approaches. Training neural ODEs as representations of dynamical systems data has its own specific demands, and we here propose a collocation scheme as a fast and efficient training strategy. This alleviates the need for costly ODE solvers. We illustrate the advantages that collocation approaches offer, as well as their robustness to qualitative features of a dynamical system, and the quantity and quality of observational data. We focus on systems that exemplify some of the hallmarks of complex dynamical systems encountered in systems biology, and we map out how these methods can be used in the analysis of mathematical models of cellular and physiological processes.

MOMC-5. Systems pharmacogenomics identifies novel targets and clinically actionable therapeutics for medulloblastoma

BackgroundMedulloblastoma (MB) is the most common malignant paediatric brain tumour and a leading cause of cancer-related mortality and morbidity. Existing treatment protocols are aggressive in nature resulting in significant neurological, intellectual and physical disabilities for the children undergoing treatment. Clearly, there is an urgent need for improved, targeted therapies that minimize these harmful side effects.MethodsWe identified candidate drugs for MB using a network-based systems-pharmacogenomics approach: based on results from a functional genomics screen, we identified a network of interactions implicated in human MB growth regulation. We then integrated drugs and their known mechanisms of action, along with gene expression data from a large collection of medulloblastoma patients to identify drugs with potential to treat MB.ResultsOur analyses identified drugs targeting CDK4, CDK6, and AURKA as strong candidates for MB; all of these genes are well validated as drug targets in other tumour types. We also identified non-WNT MB as a novel indication for drugs targeting TUBB, CAD, SNRPA, SLC1A5, PTPRS, P4HB and CHEK2. Based upon these analyses we subsequently demonstrated that one of these drugs, the new microtubule stabilizing agent, ixabepilone, blocked tumour growth in vivo in mice bearing Sonic Hedgehog and Group 3 patient-derived xenograft tumours, providing the first demonstration of its efficacy in MB.ConclusionsOur findings confirm that this data-driven systems pharmacogenomics strategy is a powerful approach for the discovery and validation of novel therapeutic candidates relevant to MB treatment, and along with data validating ixabepilone in PDX models of the two most aggressive subtypes of medulloblastoma, we present the network analysis framework as a resource for the field.

Privacy and Reproductive Health: Curtailing Rights and Choices

Digital identification (ID) and data-driven systems have become central to the delivery of welfare and health services in the global South. These policies are designed to monitor and control developmental indicators, with some negative repercussions for women. They also impact the exercise of reproductive rights and access to health and welfare, in addition to informational privacy. This paper aims to understand the various axes along which digitalisation and data collection systems impact the exercise of the right to privacy, including reproductive rights. It focuses on two specific services that have been a target of digital and analog monitoring over the past decade – conditional cash transfers tied to maternal health, and abortion services. Through interviews with women patients, their families, and health providers in New Delhi, it found that monitoring systems can restrict women’s access to critical services. Extensive procedural requirements introduced for better targeting of welfare schemes, such as Aadhaar linked bank accounts and income and caste certifications acted as significant barriers to access. These requirements particularly impacted women who carry stigma when accessing abortion services, including poor and unmarried women, and adolescents. Health providers were forced to prioritise data collection over inclusive delivery of services. Data collection for Aadhaar-linked databases without informed consent was rampant, with repercussions for women’s informational privacy. Reproductive rights and various dimensions of privacy, including informational privacy and decisional autonomy are thus intrinsically linked. Future research on digital health needs to further probe these interlinkages and broaden the definition of the right to privacy.

Big Data Dreams and Reality in Shenzhen: An Investigation of Smart City Implementation in China

Chinese cities are increasingly using digital technologies to address urban problems and govern society. However, little is known about how this digital transition has been implemented. This study explores the introduction of digital governance in Shenzhen, one of China's most advanced smart cities. We show that, at the local level, the successful implementation of digital systems faces numerous hurdles in long-standing data management and bureaucratic practices that are at least as challenging as the technical problems. Furthermore, the study finds that the digital systems in Shenzhen entail a creeping centralisation of data that potentially turns lower administrative government units into mere users of the city-level smart platforms rather than being in control of their own data resources. Smart city development and big data ambitions thereby imply shifting stakeholder relations at the local level and also pull non-governmental stakeholders, such as information technology companies and research institutions, closer to new data flows and smart governance systems. The findings add to the discussion of big data-driven smart systems and their implications for governance processes in an authoritarian context.

Detecting malicious behavior in social platforms via hybrid knowledge- and data-driven systems

Among the wide variety of malicious behavior commonly observed in modern social platforms, one of the most notorious is the diffusion of fake news, given its potential to influence the opinions of millions of people who can be voters, consumers, or simply citizens going about their daily lives. In this paper, we implement and carry out an empirical evaluation of a version of the recently-proposed NetDER architecture for hybrid AI decision-support systems with the capability of leveraging the availability of machine learning modules, logical reasoning about unknown objects, and forecasts based on diffusion processes. NetDER is a general architecture for reasoning about different kinds of malicious behavior such as dissemination of fake news, hate speech, and malware, detection of botnet operations, prevention of cyber attacks including those targeting software products or blockchain transactions, among others. Here, we focus on the case of fake news dissemination on social platforms by three different kinds of users: non-malicious, malicious, and botnet members. In particular, we focus on three tasks: (i) determining who is responsible for posting a fake news article, (ii) detecting malicious users, and (iii) detecting which users belong to a botnet designed to disseminate fake news. Given the difficulty of obtaining adequate data with ground truth, we also develop a testbed that combines real-world fake news datasets with synthetically generated networks of users and fully-detailed traces of their behavior throughout a series of time points. We designed our testbed to be customizable for different problem sizes and settings, and make its code publicly available to be used in similar evaluation efforts. Finally, we report on the results of a thorough experimental evaluation of three variants of our model and six environmental settings over the three tasks. Our results clearly show the effects that the quality of knowledge engineering tasks, the quality of the underlying machine learning classifier used to detect fake news, and the specific environmental conditions have on smart policing efforts in social platforms.

Domain Adversarial Neural Network Regression to design transferable soft sensor in a power plant

In this paper, a new approach is proposed for designing transferable soft sensors. Soft sensing is one of the significant applications of data-driven methods in the condition monitoring of plants. While hard sensors can be easily used in various plants, soft sensors are confined to the specific plant they are designed for and cannot be used in a new plant or even used in some new working conditions in the same plant. In this paper, a solution is proposed for this underlying obstacle in data-driven condition monitoring systems. Data-driven methods suffer from the fact that the distribution of the data by which the models are constructed may not be the same as the distribution of the data to which the model will be applied. This issue ultimately leads to the decline of models’ accuracy. We proposed a new transfer learning (TL) based regression method, called Domain Adversarial Neural Network Regression (DANN-R), and employed it for designing transferable soft sensors. We used data collected from the SCADA system of an industrial power plant to comprehensively investigate the functionality of the proposed method. The result reveals that the proposed transferable soft sensor can successfully adapt to new plants and new working conditions.

Systems pharmacogenomics identifies novel targets and clinically actionable therapeutics for medulloblastoma

BackgroundMedulloblastoma (MB) is the most common malignant paediatric brain tumour and a leading cause of cancer-related mortality and morbidity. Existing treatment protocols are aggressive in nature resulting in significant neurological, intellectual and physical disabilities for the children undergoing treatment. Thus, there is an urgent need for improved, targeted therapies that minimize these harmful side effects.MethodsWe identified candidate drugs for MB using a network-based systems-pharmacogenomics approach: based on results from a functional genomics screen, we identified a network of interactions implicated in human MB growth regulation. We then integrated drugs and their known mechanisms of action, along with gene expression data from a large collection of medulloblastoma patients to identify drugs with potential to treat MB.ResultsOur analyses identified drugs targeting CDK4, CDK6 and AURKA as strong candidates for MB; all of these genes are well validated as drug targets in other tumour types. We also identified non-WNT MB as a novel indication for drugs targeting TUBB, CAD, SNRPA, SLC1A5, PTPRS, P4HB and CHEK2. Based upon these analyses, we subsequently demonstrated that one of these drugs, the new microtubule stabilizing agent, ixabepilone, blocked tumour growth in vivo in mice bearing patient-derived xenograft tumours of the Sonic Hedgehog and Group 3 subtype, providing the first demonstration of its efficacy in MB.ConclusionsOur findings confirm that this data-driven systems pharmacogenomics strategy is a powerful approach for the discovery and validation of novel therapeutic candidates relevant to MB treatment, and along with data validating ixabepilone in PDX models of the two most aggressive subtypes of medulloblastoma, we present the network analysis framework as a resource for the field.

Data-driven adaptive control of wide-area non-linear systems with input and output saturation: A power system application

In this paper, a Data-Driven Adaptive Controller is proposed for wide-area Multi-Input and Multi-Output (MIMO) non-linear systems featuring input–output saturations. Complexity has always been a major concern in the modeling and control of real-world industrial processes. Traditionally, researchers look for simplified and enough precise models of these systems. However, with the significant advancement in sensor and communication technologies, a considerable amount of data is available from the process which may compensate for the lack of precise models. The emerging data-driven control systems perform both identification and control actions based on data collected from the process, making the system less dependent on a pre-identified model. In this context, an efficient MIMO Data-Driven Adaptive Control (MIMO-DDAC) is designed in this paper for stabilizing a class of unknown non-linear systems. A recursive modeling/control process is proposed for the MIMO systems in the presence of saturation in sensors and actuators, and its stability and performance are mathematically proved. This approach is applied to the load frequency control problem in an interconnected multi-area power grid in the presence of governor saturation, load and parametric uncertainties. Simulation results reveal the effectiveness of the proposed method.

Data-driven smart sustainable cities of the future: urban computing and intelligence for strategic, short-term, and joined-up planning

Sustainable cities are quintessential complex systems—dynamically changing environments and developed through a multitude of individual and collective decisions from the bottom up to the top down. As such, they are full of contestations, conflicts, and contingencies that are not easily captured, steered, and predicted respectively. In short, they are characterized by wicked problems. Therefore, they are increasingly embracing and leveraging what smart cities have to offer as to big data technologies and their novel applications in a bid to effectively tackle the complexities they inherently embody and to monitor, evaluate, and improve their performance with respect to sustainability—under what has been termed “data-driven smart sustainable cities.” This paper analyzes and discusses the enabling role and innovative potential of urban computing and intelligence in the strategic, short-term, and joined-up planning of data-driven smart sustainable cities of the future. Further, it devises an innovative framework for urban intelligence and planning functions as an advanced form of decision support. This study expands on prior work done to develop a novel model for data-driven smart sustainable cities of the future. I argue that the fast-flowing torrent of urban data, coupled with its analytical power, is of crucial importance to the effective planning and efficient design of this integrated model of urbanism. This is enabled by the kind of data-driven and model-driven decision support systems associated with urban computing and intelligence. The novelty of the proposed framework lies in its essential technological and scientific components and the way in which these are coordinated and integrated given their clear synergies to enable urban intelligence and planning functions. These utilize, integrate, and harness complexity science, urban complexity theories, sustainability science, urban sustainability theories, urban science, data science, and data-intensive science in order to fashion powerful new forms of simulation models and optimization methods. These in turn generate optimal designs and solutions that improve sustainability, efficiency, resilience, equity, and life quality. This study contributes to understanding and highlighting the value of big data in regard to the planning and design of sustainable cities of the future.

Teaching the Old Dog New Tricks: Supervised Learning with Constraints

Adding constraint support in Machine Learning has the potential to address outstanding issues in data-driven AI systems, such as safety and fairness. Existing approaches typically apply constrained optimization techniques to ML training, enforce constraint satisfaction by adjusting the model design, or use constraints to correct the output. Here, we investigate a different, complementary, strategy based on "teaching" constraint satisfaction to a supervised ML method via the direct use of a state-of-the-art constraint solver: this enables taking advantage of decades of research on constrained optimization with limited effort. In practice, we use a decomposition scheme alternating master steps (in charge of enforcing the constraints) and learner steps (where any supervised ML model and training algorithm can be employed). The process leads to approximate constraint satisfaction in general, and convergence properties are difficult to establish; despite this fact, we found empirically that even a naive setup of our approach performs well on ML tasks with fairness constraints, and on classical datasets with synthetic constraints.

Integrating Pre-trained Model into Rule-based Dialogue Management

Rule-based dialogue management is still the most popular solution for industrial task-oriented dialogue systems for their interpretablility. However, it is hard for developers to maintain the dialogue logic when the scenarios get more and more complex. On the other hand, data-driven dialogue systems, usually with end-to-end structures, are popular in academic research and easier to deal with complex conversations, but such methods require plenty of training data and the behaviors are less interpretable. In this paper, we propose a method to leverages the strength of both rule-based and data-driven dialogue managers (DM). We firstly introduce the DM of Carina Dialog System (CDS, an advanced industrial dialogue system built by Microsoft). Then we propose the "model-trigger" design to make the DM trainable thus scalable to scenario changes. Furthermore, we integrate pre-trained models and empower the DM with few-shot capability. The experimental results demonstrate the effectiveness and strong few-shot capability of our method.

Managing bias and unfairness in data for decision support: a survey of machine learning and data engineering approaches to identify and mitigate bias and unfairness within data management and analytics systems

The increasing use of data-driven decision support systems in industry and governments is accompanied by the discovery of a plethora of bias and unfairness issues in the outputs of these systems. Multiple computer science communities, and especially machine learning, have started to tackle this problem, often developing algorithmic solutions to mitigate biases to obtain fairer outputs. However, one of the core underlying causes for unfairness is bias in training data which is not fully covered by such approaches. Especially, bias in data is not yet a central topic in data engineering and management research. We survey research on bias and unfairness in several computer science domains, distinguishing between data management publications and other domains. This covers the creation of fairness metrics, fairness identification, and mitigation methods, software engineering approaches and biases in crowdsourcing activities. We identify relevant research gaps and show which data management activities could be repurposed to handle biases and which ones might reinforce such biases. In the second part, we argue for a novel data-centered approach overcoming the limitations of current algorithmic-centered methods. This approach focuses on eliciting and enforcing fairness requirements and constraints on data that systems are trained, validated, and used on. We argue for the need to extend database management systems to handle such constraints and mitigation methods. We discuss the associated future research directions regarding algorithms, formalization, modelling, users, and systems.

Darkness, Datafication, and Provenance as an Illuminating Methodology

Darkness, Datafication, and Provenance as an Illuminating Methodology

In the Shadow of Platforms

In the Shadow of Platforms

Sampling Techniques Used in Big-Data Driven Applications

Data-driven systems process the data from various sources in multiple applications. Data retrieved from heterogeneous sources need to be available in an aggregate and unique format. This requirement gives rise to the process of the Big-data and proposed next-generation big-data processing systems. There are many applications based on contextual data useful for identifying the traffic intensity, changing users per application, weather conditions etc., and serve as next- generation business-specific systems. In such systems data abstraction and representation are the important tasks & granularity can be applied in the data processing. Granularity will process the data from low granularity to high granularity. Sampling plays an important role in the data processing.

Constructing Dependable Data-Driven Software With Machine Learning

Many data-driven software systems present dependability challenges caused by the distribution and dynamicity of their application environments. We investigate machine learning-driven construction techniques combined with pattern-based architecture, showing that system dependability is linked to data and function quality.

A metabolic data-driven systems pharmacology strategy for decoding and validating the mechanism of Compound Kushen Injection against HCC

Ethnopharmacological relevanceCompound Kushen Injection (CKI) is a widely used TCM formula for treatment of carcinomatous pain and tumors of digestive system including hepatocellular carcinoma (HCC). However, the potential mechanisms of CKI for treatment of HCC have not been systematically and deeply studied. Aim of studyA metabolic data-driven systems pharmacology approach was utilized to investigate the potential mechanisms of CKI for treatment of HCC. Materials and methodsBased on phenotypic data generated by metabolomics and genotypic data of drug targets, a propagation model based on Dijkstra program was proposed to decode the effective network of key genotype-phenotype of CKI in treating HCC. The pivotal pathway was predicted by target propagation mode of our proposed model, and was validated in SMMC-7721 cells and diethylnitrosamine-induced rats. ResultsMetabolomics results indicated that 12 differential metabolites, and 5 metabolic pathways might be involved in the anti-HCC effect of CKI. A total of 86 metabolic related genes that affected by CKI were obtained. The results calculated by propagation model showed that 6475 shortest distance chains might be involved in the anti-HCC effect of CKI. According to the results of propagation mode, EGFR was identified as the core target of CKI for the anti-HCC effect. Finally, EGFR and its related pathway EGFR-STAT3 signaling pathway were validated in vivo and in vitro. ConclusionThe proposed method provides a methodological reference for explaining the underlying mechanism of TCM in treating HCC.

Using community engagement to implement evidence-based practices for opioid use disorder: A data-driven paradigm & systems science approach

Community-driven responses are essential to ensure the adoption, reach and sustainability of evidence-based practices (EBPs) to prevent new cases of opioid use disorder (OUD) and reduce fatal and non-fatal overdoses. Most organizational approaches for selecting and implementing EBPs remain top-down and individually oriented without community engagement (CE). Moreover, few CE approaches have leveraged systems science to integrate community resources, values and priorities. This paper provides a novel CE paradigm that utilizes a data-driven and systems science approach; describes the composition, functions, and roles of researchers in CE; discusses unique ethical considerations that are particularly salient to CE research; and provides a description of how systems science and data-driven approaches to CE may be employed to select a range of EBPs that collectively address community needs. Finally, we conclude with scientific recommendations for the use of CE in research. Greater investment in CE research is needed to ensure contextual, equitable, and sustainable access to EBPs, such as medications for OUD (MOUD) in communities heavily impacted by the opioid epidemic. A data-driven approach to CE research guided by systems science has the potential to ensure adequate saturation and sustainability of EBPs that could significantly reduce opioid overdose and health inequities across the US.