Tutorial Sessions/Invited Talks

Industrial data science and black swans

Terrorists, Hackers, and Criminals: Understanding the Darknet

The Darknet is the area of the internet that terrorists, child pornographers, hackers, and insider traders call home. Utilizing technologies like TOR, I2P, and ZeroNet, the darknet’s anonymity and distributed nature make law enforcement operations all but impossible. Likewise, the covert and secretive nature of most sites means that most sites are known only to their users. For those not a part of the underworld, the darknet represents an interesting research opportunity. Otherwise hard-to-find underworld groups roam freely, and there are many opportunities to generate interesting data sets. 

In this tutorial, we introduce the audience to the different technologies and explore what type of security features are used by the different underworld groups. We will also briefly cover different attacks and potential attacks that could be used to break the security features provided by the different darknet networks. We will also cover some new research that attempts to recognize terrorist content from benign content, and its use as a tool for finding terrorist content on both the darknet and the regular internet. We will discuss how similar models can be made using text, image, and graph mining techniques.

This tutorial is suitable for a general  audience, but is especially recommended for data scientists, researchers, and cybersecurity practitioners who have an interest in cultivating data from criminal and terrorist enterprises, exploiting anonymity networks, or data mining on such networks. 

Conformal Prediction - Predicting with Confidence

How good is your prediction? In risk-sensitive applications, it is crucial to be able to assess the quality of a prediction, but traditional classification and regression models don't provide their users with any information regarding prediction trustworthiness.

Conformal predictors, on the other hand, are predictive models that associate each of their predictions with a precise measure of confidence. Given a user-defined significance level E, a conformal predictor outputs, for each test pattern, a multivalued prediction region (class label set or real-valued interval) that, under relatively weak assumptions, contains the test pattern’s true output value with probability 1-E. In other words, given a significance level E, a conformal predictor makes an erroneous prediction with probability E. The conformal prediction framework allows any traditional classification or regression model to be transformed into a confidence predictor with little extra work, both in terms of implementation and computational complexity.

Some key properties of conformal prediction are:
• We obtain probabilities/error bounds per instance
• Probabilities are well-calibrated: 95% means 95%
• We don't need to know the priors
• We make a single assumption - that the data is exchangeable ~ i.i.d.
• We can apply it to any machine learning algorithm
• It is rigorously proven and straightforward to implement
• There is no magic involved – only mathematics and algorithms

Hence, confidence predictors is an important tool that every data scientist should carry in their toolboxes, and conformal prediction represents a straight-forward way of associating the predictions of any predictive machine learning algorithm with confidence measures.

This tutorial aims to provide an introduction and an example-oriented exposition of the conformal prediction framework, directed at machine learning researchers and professionals. A publicly available Python library, developed by one of the authors of the tutorial, will be used for the running examples.

The goal of the tutorial is to provide attendees with the knowledge necessary for implementing functional conformal predictors, and to highlight current research on the subject.

eMaintenance and Industry 4.0: Knowledge extraction in Industry and Transportation

Information is key to successful and cost effective management of maintenance process. Therefore the integration of Information and Communication Technologies(ICT) into the maintenance process is essential to achieve excellence in maintenance performance. e-Maintenance solutions provide platform for seamless integration of ICT with maintenance. In short, development and implementation of eMaintenance solutions can be considered as a fusion of various technologies and methodologies.

The general opinion among the asset managers is that the application of information technology brings dramatic results in machine reliability and O&M (operations and maintenance) process efficiency, however only few operations and maintenance managers can show or calculate the benefits of such applications. Technology providers are trying to develop more and more advanced tools while the maintenance departments seem to struggle with the daily problems of implementing, integrating and operating such systems. The users combine their experience and heuristics in defining maintenance policies and in the usage of condition monitoring systems. The resulting maintenance systems seem to be a heterogeneous combination of methods and systems in which the integrating factor between the information and business processes is the personnel. The information of the assets goes through these human minds forming an organizational information system and creating a high reliance on the expertise of the as-set management staff.

Indeed, the contribution of operation and maintenance data to the different asset management stages has been triggered by the emergence of intelligent sensors for measuring and monitoring the health state of a component, the gradual implementation of Information and Communication Technologies (ICT), and the conceptualization and implementation of e-maintenance.

During the last decade, the global competition and advancement of ICT have forced Production and Process Industries through a continuous transformation and improvement process. The business scenario is focusing more on e-business intelligence to perform transactions with a focus on customers’ needs for enhanced value and improvement in asset management. Such business requirements compel the organizations to minimize the production and service downtime by reducing the machine performance degradation. The above organizational requirements stated in ISO 55000 necessitate the development of proactive maintenance strategies to provide optimized and continuous process performance with minimized system breakdowns and maintenance. With these changing systems of the busi-ness world in the 21st century, a new era of data e-services coming from asset data has emerged.

The goal of this tutorial is to provide a timely review of research efforts on the topic of e-Maintenance covering both theoretical and applied research that will contribute towards understanding the strategic role of eMaintenance solutions in the business and create a volume of recent work on this subject. We hope that this tutorial will stimulate new areas of thinking in management of complex maintenance processes which are increasing the relevance in the organizations.

Indeed organizations are under pressure to continually enhance their ability to create value for their customers and improve the cost effectiveness of their operations. In this regard, the maintenance of large-investment assets, once thought to be a necessary evil, is now considered a key function for improving the cost effectiveness of operations and creating additional value by delivering better and more innovative services to customers.

With the change in strategic thinking of organizations, the increasing amount of outsourcing and the separation of OEMs and asset owners, it is crucial to measure, control and improve the asset maintenance performance. Today, with the advances in technology, various maintenance strategies have evolved, including condition based maintenance, predictive maintenance, remote-maintenance, preventive maintenance, e-maintenance etc. A main challenge faced by most organizations is choosing the most efficient and effective strategies to enhance and continually improve operational capabilities, reduce maintenance costs and achieve competitiveness in the industry.