Important Dates

Proposal submission open
6 Jul 2018

Proposals due
27 Aug 2018

Acceptance notification
10 Sept 2018

Final content due
1 Oct 2018

Click here to download Call for Exhibits

The Technology Expo at CASCON 2018 will provide an excellent opportunity to highlight your research results and generate interest in upcoming product and product areas. Exhibits on emerging technologies are encouraged. The format of an exhibit may be a demonstration, poster presentation, or a short talk. A booth is provided at no cost to the exhibitors. Each demonstration booth will be set up with power and wireless internet access. Exhibitors are responsible for bringing the necessary equipment such as machines, laptops, monitors, power cables, and security locks; and any printed materials including posters.

Accepted exhibits

Congratulations to the exhibit chairs and thank you for your submissions. See you all at CASCON 2018.

A Dynamic Knowledge Management System for Decision Support

Title: A Dynamic Knowledge Management System for Decision Support
Chair(s): Ftoon Kedwan , Farhana Zulkernine
Format: Short Talk
Description: We focus on; a) studying the state-of the art KMS to select the right system to store hybrid medical knowledge, b) define a multi-level graph structure with schema and metadata to dynamically and efficiently link knowledge sources and also to contain frequently accessed knowledge, and c) develop machine learning models to learn query patterns for efficient knowledge organization and retrieval. We are still at the very early phase of the research. Fig. 1 shows the conceptual architecture.
Product Name: Knowledge Management System
Web Site: https://github.com/isahqueens/multi-level_streaming_analytics/blob/master/KMS.jpeg
Research Team: Queen's University: Ftoon Kedwan 

A Hierarchical Rule-based Security Management System for Data-Intensive Applications

Title: A Hierarchical Rule-based Security Management System for Data-Intensive Applications
Chair(s): Marin Litoiu , Yar Rouf , Vio Onut
Format: Demo
Description: We propose a platform to support the SecDevOps framework, a hierarchical distributed architecture for security control that uses a Business Rules Engine. The Business Rules Engine simplifies security rules by allowing the teams to write them at an operational level rather than at the network level, which would require specialized knowledge.
Industry Company: IBM
Product Name: QRadar, Cloud Automation Manager
Research Team: York University: Yar Rouf  Marin Litoiu  Joydeep Mukherjee  Vio Onut 
IBM: Vio Onut 

A Machine Learning Approach to Test Case Prioritization

Title: A Machine Learning Approach to Test Case Prioritization
Chair(s): Francis Palma
Format: Poster
Description: We propose to use the defect-proneness value of source files as the essential measure for test case prioritization. Our proposed test case prioritization model comprises two stages: (1) Metric extraction and defect prediction and (2) Test case rescheduling.
Industry Company: IBM
Project Number Project SRG_2014_1
Product Name: Test case prioritization
Research Team: Ryerson University: Francis Palma and Ayse Bener 

A Multilevel Streaming Data Analytics Infrastructure for Predictive Analytics

Title: A Multilevel Streaming Data Analytics Infrastructure for Predictive Analytics
Chair(s): Haruna Isah , Farhana Zulkernine
Format: Short Talk
Description: We are developing a real-time multilevel streaming data analytics infrastructure using cutting-edge streaming data processing engines, in-memory data structure and analytics tools which will preprocess, filter and store necessary data in the memory for higher level processing. This talk is aimed at showcasing our preliminary framework and is targeted at individuals and organizations that are planning to optimize their current infrastructure to embrace multilevel streaming data analytics.
Industry Company: Gnowit Inc
Product Name: Multilevel Streaming Infrastructure for News Analytics
Web Site: https://github.com/isahqueens/multi-level_streaming_analytics
Research Team: Queens University: Haruna Isah  Farhana Zulkernine 

Ahead-Of-Time Compilation of WebAssembly Using Eclipse OMR

Title: Ahead-Of-Time Compilation of WebAssembly Using Eclipse OMR
Chair(s): Mark Thom
Format: Poster
Description: WebAssembly is an emerging assembly language that runs on the web. It has nearly native performance, and is already the compilation target of several programming languages. There is an ongoing project to create a JIT compiler for WebAssembly using Eclipse OMRs JITBuilder. The poster will describe our attempted contribution of an AOT compiler implemented with OMR. We hope it will expose any disparity between OMRs current support for AOT and what its completed AOT module should be capable of.
Industry Company: IBM
Project Number 1036
Product Name: Eclipse OMR
Web Site: http://www.unb.ca/research/casatlantic/index.html
Research Team: University of New Brunswick: Mark Thom 

Approximation Algorithms for Instruction Scheduling and Rapid Prototyping in Coconut

Title: Approximation Algorithms for Instruction Scheduling and Rapid Prototyping in Coconut
Chair(s): Curtis D'Alves , Bill O'Farrell
Format: Poster
Description: Instruction scheduling is an NP-complete problem which allows for altering the execution order of instructions in a function without altering the functions semantics. We present a continuous optimization based model for solving near-optimal schedules and a prototype implemented for the IBM z13 and z14 architectures and used to generate fast schedules for common cryptographic functions used in Blockchain operations implemented in GOLang assembly.
Industry Company: IBM
Project Number 1006
Product Name: Blockchain Cryptographic Functions in GO Assembly for Z
Research Team: McMaster University: Curtis D'Alves  Bill O'Farrell  Christopher Anand  Wolfram Kahl  James You 

API Connect 2018

Title: API Connect 2018
Chair(s): JISOO LEE , ivy ho , Peter El-koreh , Amirali Jafarian
Format: Poster
Description: Introducing the new API Connect 2018
Industry Company: IBM Canada Ltd.
Product Name: API Connect
Research Team: Epitome: Candy  Ivy  Amir  Jisoo  Peter 

Cold Object Identification, Sequestrating and Revitalization

Title: Cold Object Identification, Sequestrating and Revitalization
Chair(s): Abhijit Taware
Format: Poster
Description: Cold objects are alive and infrequently accessed. Such objects are an overhead for garbage collection. Identifying and moving them to a separate memory space would benefit the real memory footprint and cache coherence. Intel's NVRam technology will be used as a cold heap for this research. Application will be profiled to gather statistics about object temperature. This will be utilized for cold object sequestration and revitalization.
Industry Company: IBM
Project Number 1001
Product Name: Eclipse OMR, IBM Java SDK, OMR, J9
Web Site: http://www.unb.ca/research/casatlantic/projects/datatemp.html
Research Team: UNB: Abhijit Taware  Kenneth B. Kent  Gerhard W. Dueck 
IBM: Charlie Gracie 

Comparative Analysis of Cloud Console Systems

Title: Comparative Analysis of Cloud Console Systems
Chair(s): Julia Rubin , Satish Sarraf , John Steinbacher
Format: Poster
Description: Cloud Console systems feature a customizable presentation layer that allows users to manage their cloud resources, such as VMs, storage, etc. Designing such systems is challenging due to the demand to (a) handle massive amounts of data, (b) integrate with different resource providers, and (c) deliver data from these providers in near real-time. In this work, we analyze several Cloud Consoles, including that of IBM Cloud, and discuss design decisions and trade-offs made to satisfy these demands.
Industry Company: IBM
Project Number 1040
Product Name: IBM Cloud
Research Team: University of British Columbia: Julia Rubin  Satish Sarraf  Harshavardhan Kadiyala 
Seneca College: Mark Shtern 
IBM Canada: John Steinbacher 
IBM USA: Anthony W Erwin 

Composing Web Services Using a Multi-Agent Framework

Title: Composing Web Services Using a Multi-Agent Framework
Chair(s): Yu Zhao
Format: Poster
Description: Software agents are autonomous entities that can proactively activate web services, and autonomously react upon the results to determine the next steps. However, programming agents are complex due to the the proactive, autonomous and reactive characteristics of agents. To reduce the workload of developers when programming agents, we propose an easy-to-understand semi-natural language syntax to program agents for service composition.
Industry Company: IBM
Project Number 920
Product Name: Building Intelligent Personal Agents
Research Team: Queen's University: Yu Zhao  Daniel Alencar da Costa  Ying Zou 
IBM: Joanna Ng 

Design Patterns for Machine Learning Solutions

Title: Design Patterns for Machine Learning Solutions
Chair(s): Soroosh Nalchigar
Format: Demo
Description: dvanced analytics and machine learning algorithms have a wide and rapidly growing range of applications in business contexts. In this project, we apply a design patterns approach to reduce the level of expertise and efforts needed to develop machine learning solutions to be incorporated into cognitive business operations. The approach is based on design knowledge catalogs encoded according to a meta-model. Implementation of the approach is being investigated in the context of a BPM suite.
Industry Company: IBM
Project Number 1030
Product Name: Business Process Management
Research Team: University of Toronto: Soroosh Nalchigar  Eric Yu 
IBM Canada: Sebastian Carbajales  Yazan Obeidi  John Green  Allen Chan 

Detecting Misuses of Crypto APIs

Title: Detecting Misuses of Crypto APIs
Chair(s): Kristen Newbury
Format: Poster
Description: This exhibit will present the integration of cryptography API misuse detection into a Just-in-Time context. My goal for a novel contribution to this research will be to enable a JIT compiler to perform security checks that have been generated by a pre-existing API misuse detection framework. Such an integration is expected to benefit the precision and performance of the security analyses, since runtime values can be exploited in this context.
Industry Company: IBM
Project Number 1064
Product Name: IBM SDK for Java, Testarossa, IBM JIT Compiler
Research Team: University of Alberta: Kristen 

Discovering business rules automatically from legacy applications

Title: Discovering business rules automatically from legacy applications
Chair(s): Arun Martin , David Budreau , Tala El Hallak
Format: Short Talk
Description: The BRM research and development team will demonstrate how our revised architecture focusing on machine learning and user research has transformed last years BRM preview in Application Discovery and Delivery Intelligence (ADDI) to enable automated discovery of terms and to attach them to relevant code snippets through a collaborative workflow between business and information technology teams resulting in increased knowledge, reach, and ability for members to make faster business decisions.
Product Name: Business Rules Discovery (ADDI)
Research Team: IBM: Arun Joseph  David Budreau  Tala Hallak 

Eclipse OMR on the Embedded AArch64 Platform

Title: Eclipse OMR on the Embedded AArch64 Platform
Chair(s): Jean-Philippe Legault , Aaron Graham
Format: Poster
Description: The AArch64 architecturea subset of the ARM architectural modelhas the characteristics to be driving the next generation of data centers. With low power and cost overhead, AArch64 helps move computation closer to the originating data; following the fog computing model. We develop support for AArch64 in the Eclipse OMR Toolkit to evaluate modern runtimes with new computing paradigms.
Industry Company: IBM
Project Number 1035
Product Name: Eclipse OMR
Web Site: http://www.unb.ca/research/casatlantic/index.html
Research Team: University of New Brunswick: Jean-Philippe Legault  Aaron G. Graham  Kenneth B. Kent 
IBM: Daryl Maier  James Kingdon 

Enabling a highly-scalable, hybrid cloud-based microservices architecture

Title: Enabling a highly-scalable, hybrid cloud-based microservices architecture
Chair(s): Anthony Kwan , Jonathon Wong
Format: Demo
Description: Traditional microservice scaling methods perform either horizontal or vertical scaling exclusively. When used in combination, however, these methods offer complementary benefits and compensate for each other's deficiencies. To leverage the high availability of horizontal scaling and the fine-grained resource control of vertical scaling, we developed two novel hybrid autoscaling algorithms and benchmarked them against Google's popular Kubernetes horizontal autoscaling algorithm.
Industry Company: IBM
Product Name: IBM Blueworks
Research Team: University of Toronto: Anthony Kwan, Jonathon Wong, Hans-Arno Jacob 
IBM: Allen Chan, Suzette Samoojh, Vinod Muthuswam 

Exploring Challenges In Adoption of Continuous Delivery

Title: Exploring Challenges In Adoption of Continuous Delivery
Chair(s): Trevor Rae
Format: Poster
Description: We look at the technical and cultural problems that exist in regards to adopting continuous delivery/integration into a company's infrastructure.
Industry Company: IBM
Product Name: IBM Cloud Continuous Delivery Service
Research Team: University Of Victoria: Trevor Rae  Dr.Damian 

Feasibility of Internal Object Pools to Reduce Memory Management Activity

Title: Feasibility of Internal Object Pools to Reduce Memory Management Activity
Chair(s): Konstantin Nasartschuk
Format: Poster
Description: Object pools is a software engineering pattern used to reuse object instances without the need of repeated allocation and instantiation. While the benefits of using object pool structures are still present when used in a garbage collected environment, it adds a memory management component to the development process. This work investigates automatically created object pools for predefined classes. Automatic object pools are implemented and discussed using the GenCon GC and Balanced GC policies.
Industry Company: IBM
Project Number 1019
Product Name: Object Locality (IBM SDK)
Research Team: University of New Brunswick: Konstantin Nasartschuk  Kenneth B. Kent  Stephen A. MacKay 
IBM Canada: Aleksandar Micic 

IBM Centers for Advanced Studies Worldwide

Title: IBM Centers for Advanced Studies Worldwide
Chair(s): Marcellus Mindel , Andy Rindos , Zoltan Szlavik
Format: Poster
Description: IBM Centers for Advanced Studies (CAS) is a system of engagement for academic collaboration that spans the globe. The strength of CAS is its diversity. There are centers based in IBM labs, and centers at academic institutions. Some centers focus on software and/or hardware, others on services. Drop by to learn more about how you can engage globally with CAS.
Industry Company: IBM
Product Name: IBM Centers for Advanced Studies
Research Team: IBM Centers for Advanced Studies: Marcellus Mindel  Andy Rindos  Zoltan Szlavik 

In-Depth Evaluation of Task Parallelism in Node.js

Title: In-Depth Evaluation of Task Parallelism in Node.js
Chair(s): Maria Patrou
Format: Poster
Description: Node.js is a framework for asynchronous I/O, event-driven, server-side Javascript. Computationally intensive tasks are bound to the performance of a single core. We evaluate the performance of some representative Node.js multi-processed and multi-threaded techniques used for scaling, running a CPU-intensive task. Computation, memory and garbage collection metrics, the effect of the execution environment and similarities between these modules are investigated and recommendations are provided.
Industry Company: IBM
Project Number 1002
Product Name: IBM SDK for Node.js
Web Site: http://www.unb.ca/research/casatlantic/index.html
Research Team: University of New Brunswick: Maria Patrou  Kenneth B. Kent 
IBM Canada: Michael Dawson 

Localizing Error Prone Code Using Software Analytics

Title: Localizing Error Prone Code Using Software Analytics
Chair(s): KONSTANTINOS KONTOGIANNIS , Sanjay Ghanathey
Format: Demo
Description: - Evaluation study into how software analytics and machine learning can be used to predict whether a source code file contains one or more errors that may contribute to a major system failure. - The approach is utilizing information extracted both from the system's source code, such as code metrics, and from a series of DevOps tools. - Results indicate that data obtained from repositories and used as training features performed equal to or better than source code metrics
Industry Company: IBM
Project Number 1048
Product Name: DevOps Services, DevOps Insights
Research Team: Western University: KONSTANTINOS KONTOGIANNIS  SANJAY GHANATHEY 
IBM US: ALBERTO GIAMMARIA  BRIAN COUNTRYMAN 
IBM CANADA: CHRIS BREALEY 

Math function performance acceleration via integrated HW and algorithm design

Title: Math function performance acceleration via integrated HW and algorithm design
Chair(s): Robert Enenkel
Format: Poster
Description: We present results from our CAS project to design HW instructions together with algorithms to work together to produce fast, accurate mathematical functions.
Project Number 1006
Product Name: POWER processors, MASS math libraries
Web Site: https://www.eng.mcmaster.ca/cas/people/faculty/christopher-anand
Research Team: IBM: Robert Enenkel 

Model based Interference-aware Management System for Microservice Architectures

Title: Model based Interference-aware Management System for Microservice Architectures
Chair(s): Joydeep Mukherjee , Marin Litoiu , Cornel Barna , Amar Patel
Format: Demo
Description: The performance of microservices deployed in containers on a cloud platform can often be impacted due to interference. Interference is caused by other applications that run concurrently with the microservice containers on the same host and compete for host level shared resources. We build a model based management system that takes into account the effect of interference and works in parallel with existing platforms such as IBM Cloud Automation Manager for efficient deployment of microservices.
Industry Company: IBM
Product Name: Cloud Automation Manager (CAM)
Research Team: York University: Joydeep Mukherjee  Marin Litoiu  Cornel Barna  Amar Patel 

Modeling AUTOSAR Implementations in Simulink

Title: Modeling AUTOSAR Implementations in Simulink
Chair(s): Jian Chen
Format: Poster
Description: AUTOSAR defines the automotive three-layered software architecture. Application layer specifies functional behaviours and are realized as a set of tasks. Simulink is employed to address AUTOSARs lack of support for modelling behaviours of tasks. Simulink simulations assume tasks are completed in zero execution time, while real executions require a finite execution time. We present a Simulink block that can schedule tasks with a non-zero simulation time to enable a more realistic simulation.
Industry Company: Queen's University
Product Name: Model Scheduler
Research Team: Queen's University: Jian Chen  Thomas Dean 
Ryerson University: Manar Alalfi 
General Motors R&D: Ramesh S 

Node.js Scalability Investigation in the Cloud

Title: Node.js Scalability Investigation in the Cloud
Chair(s): Jiapeng Zhu , Maria Patrou
Format: Poster
Description: Node.js has gained popularity in cloud development. However, scalability issues can limit the number of concurrent requests while achieving an acceptable level of performance. We design and implement Ibenchjs, a scalability-oriented benchmarking framework, and a set of sample test applications. We deploy Ibenchjs in a local and isolated cloud to report scalability-related measurements and issues of Node.js as well as performance bottlenecks to investigate Node.js scalability issues in the cloud.
Industry Company: IBM
Project Number 1002 Node.js
Product Name: IBM SDK for Node.js
Web Site: http://www.unb.ca/research/casatlantic/index.html
Research Team: University of New Brunswick: Jiapeng Zhu  Panagiotis Patros  Kenneth B. Kent 
IBM Runtime Technologies: Michael Dawson 

Optimization Driven Inlining

Title: Optimization Driven Inlining
Chair(s): Erick Ochoa
Format: Poster
Description: Inlining is a compiler optimization that should be selectively applied in JIT compilation. Inlining too much increases compilation time and inlining too little decreases throughput. Different measures (like invocation count and method size) have been used to decide which methods to inline. Our proposal is to use abstract interpretation to determine which optimizations are feasible and use this calculation to drive inlining.
Industry Company: IBM
Project Number 1047
Product Name: OpenJ9
Research Team: University of Alberta: Erick Ochoa  Karim Ali  Jose Nelson Amaral 
IBM Canada: Andrew Craik 

OptImatch: Semantic-Web System for Automated Query Tuning

Title: OptImatch: Semantic-Web System for Automated Query Tuning
Chair(s): Guilherme Damasio , alexandar mihaylov
Format: Poster
Description: Optimatch is a system which discovers problem patterns automatically by matching them to templates of plans from a previously populated knowledge base. The system extends upon our previous work, but now alleviating all need for manual expert input by introducing a customizable knowledge base capable of learning problem patterns from previously ran execution plans. The system output is a query with the embedded optimization guideline which enforces the optimal execution plan to be applied.
Industry Company: IBM
Project Number 1049
Product Name: Optimatch
Research Team: University of Ontario Institute of Tech.: Guilherme Fetter Damasio  Alexandar Mihaylov  Jaroslaw Szlichta 
York University: Parke Godfrey 
IBM: Calisto Zuzarte  Piotr Mierzejewski  Vincent Corvinelli 

Persistent Memory Storage of Cold Regions in the OpenJ9 Java Virtual Machine

Title: Persistent Memory Storage of Cold Regions in the OpenJ9 Java Virtual Machine
Chair(s): Scott Young
Format: Poster
Description: In this poster an optimization technique for language runtimes with automatic memory management is presented. The technique involves segregating objects into different memory areas, backed by different memory devices, on a per object basis. This technique is compared against operating system paging mechanisms with swap partitions. Two different schemes for determining which objects should be segregated into slower memory are tested and their results are discussed.
Industry Company: IBM
Project Number 1001
Product Name: Eclipse OMR, IBM Java SDK, OMR, J9
Web Site: http://www.unb.ca/research/casatlantic/projects/datatemp.html
Research Team: Centre for Advanced Studies - Atlantic: Scott Young  Kenneth B. Kent  Gerhard Dueck 
IBM: Charlie Gracie 

Predicting Cross-Sell Opportunities in Financial Services Using Deep Learning

Title: Predicting Cross-Sell Opportunities in Financial Services Using Deep Learning
Chair(s): Chux Ejieh , Diane Reynolds
Format: Poster
Description: In this exhibition, we will explore the design and implementation of the Product Propensity model (part of the IBM Client Insights for Wealth Management SaaS) and share how it impacts our clients. Cross-sell refers to the practice of selling an additional product to an existing client. Ideally, marketing campaigns in this domain are more targeted as they leverage information the business has about its customers.
Industry Company: IBM
Product Name: Client Insights
Web Site: https://www.ibm.com/us-en/marketplace/client-insight-for-wealth-management
Research Team: IBM Canada: Chux Ejieh - Data Scientist  Diane Reynolds - Chief Data Scientist 

Pressure Sensitive Mats for Patient Monitoring in the NICU

Title: Pressure Sensitive Mats for Patient Monitoring in the NICU
Chair(s): Shermeen Nizami , James R. Green
Format: Demo
Description: A research project that is evaluating the use of the pressure sensitive mat (PSM) technology as well RGB-D streaming video as continuous patient monitoring modalities. This research is conducted at the Neonatal Intensive Care Unit (NICU) at CHEO, Ottawa. We have developed and deployed two software apps at the patient bedside to collect physiologic and clinical event data from patients in real-time. We have acquired data from fifteen critically ill newborn babies up to date.
Product Name: Watson Analytics and Health
Research Team: Carleton University: Shermeen Nizami  James R. Green  Randy Giffen 

Privacy-aware Efficient Visual Recognition Services for Smart Laboratories

Title: Privacy-aware Efficient Visual Recognition Services for Smart Laboratories
Chair(s): Christianne Huber , Mahima Chaudhary , Joydeep Mukherjee , Marin Litoiu
Format: Demo
Description: With concerns about personal privacy on the rise, individuals may prefer that their images captured by visual recognition services have controlled privacy levels. IBM Watson Visual Recognition Service ensures efficiency but individuals may prefer not to send images to such a public service. We propose an intelligent privacy-aware framework that maintains efficiency while delivering a desired privacy level to its end users by dynamically directing images to a localized privacy guaranteed server.
Industry Company: IBM
Product Name: IBM Watson Visual Recognition Service
Research Team: York University: Christianne Huber  Mahima Chaudhary  Joydeep Mukherjee  Marin Litoiu 

Reconstruct the genealogy of phishing kits source code modifications

Title: Reconstruct the genealogy of phishing kits source code modifications
Chair(s): Stphane Heudron
Format: Poster
Description: In this project, we will study the similarity between phishing kits belonging to a reference database and develop an approach to extract a plausible genealogy to explain source code modifications. Our initial approach is based on static analysis of PHP scripts. To investigate the similarity, we will explore several distances and algorithms responding to our constraints (incrementality, efficiency). We will present preliminary results of similarity between kits and some candidate genealogy trees.
Industry Company: IBM
Project Number 1059
Product Name: IBM Forensics
Research Team: Polytechnique Montral: Stephane Heudron  Ettore Merlo 
University of Ottawa: Guy-Vincent Jourdan 
IBM Canada: Vio Onut 

ROAD - Realtime Outlier and Anomaly Detection for IBM Cloud

Title: ROAD - Realtime Outlier and Anomaly Detection for IBM Cloud
Chair(s): William Pourmajidi , Lei Zhang , Andriy Miranskyy
Format: Poster
Description: We refer to logs when there are errors or anomalies. Hence, anomaly detection is a crucial part of the log analysis, and if anomalies are detected at an early stage, many issues can be prevented. In this poster, we discuss practical issues and potential solutions needed to create an anomaly detection solution tailored for a large-scale Cloud platform consisting of multiple software and hardware components.
Industry Company: Ryerson University
Product Name: IBM Cloud
Research Team: Ryerson University: William Pourmajidi, Lei Zhang, Andriy Miran 
IBM: Tony Erwin, John Steinbacher, David Godwin,  David Kalmuk, Aslam Nomani, Vivien Page, Nic 

Runtime Automatic Parallelization of JVM Applications

Title: Runtime Automatic Parallelization of JVM Applications
Chair(s): Akihiro Hayashi , Gita Koblents
Format: Poster
Description: In this IBM-Rice University collaboration, we push the bounds on integrating JVM applications with GPU accelerators. This work uses advanced code generation and performance prediction techniques to effectively and automatically offload JVM applications including Apache Spark to hardware accelerators.
Industry Company: IBM
Project Number 900
Product Name: OpenJ9, Spark-DK
Research Team: Rice University: Akihiro Hayashi 
IBM Canada: Gita Koblents  Jimmy Kwa 
Georgia Institute of Technology: Vivek Sarkar 
IBM Research Tokyo: Kazuaki Ishizaki 

Static Dataflow Analysis for OpenMP Code Sanitization

Title: Static Dataflow Analysis for OpenMP Code Sanitization
Chair(s): Prithayan Barua , Jun Shirako , Whitney T Tsang , Vivek Sarkar
Format: Poster
Description: OpenMP provides high-level abstractions and rich functionality for GPU programming. However, understanding the complex semantics of various combinations of pragmas is not trivial and leads to their wrong usage. We present a static analysis technique to detect incorrect usage of OpenMP target data mapping. Assuming the correctness of underlying sequential programs, our approach detects the OpenMP errors by comparing the data flow information of the original OpenMP version with the sequential.
Industry Company: IBM
Project Number CAS-896
Product Name: XL Compiler
Research Team: Georgia Institute of Technology: Prithayan Barua,  Jun Shirako,  Vivek Sarkar 
IBM Toronto: Whitney T Tsang,  Jeeva Paudel,  Wang Chen 

Susereum - A Blockchain Platform for Sustainable Software Systems

Title: Susereum - A Blockchain Platform for Sustainable Software Systems
Chair(s): Omar Badreddin
Format: Demo
Description: Software sustainability is a systematic challenge, and not an individual or team failings. This systematic challenge stems because engineers often take shortcuts to maximize immediate gains that become detrimental to software longevity. We demo Susereum, a blockchain platform that creates a balance between long term longevity and short term necessities. Susereum creates a permanent public ledger to monitory the sustainability of software code modifications and credit their authors.
Industry Company: Susereum
Product Name: Susereum
Web Site: https://susereum.com/
Research Team: University of Texas: Omar Badreddin 

Tracking Unhandled Rejections in Node.js

Title: Tracking Unhandled Rejections in Node.js
Chair(s): Maxim Uzun , Kenneth B. Kent , Michael Dawson
Format: Poster
Description: Node.js is an open source cross-platform runtime environment for executing server-side JavaScript. The framework is based on Google Chromes V8 and is limited in regard to post-mortem diagnostics and debugging for Promises; a design pattern for managing asynchronous code. We propose and experimentally evaluate an implementation of Node.js that aims to improve troubleshooting for unhandled Promise rejections. The goal is to produce accurate core dumps and track unhandled rejections.
Industry Company: IBM
Project Number 1002
Product Name: IBM SDK for Node.js
Web Site: http://www.unb.ca/research/casatlantic/index.html

Assessing Databases for GDPR Compliance

Title: Assessing Databases for GDPR Compliance
Chair(s): Devan Shah , Walid Rjaibi , Larry Lindsay , Josue Diaz
Format: Short Talk
Description: The General Data Protection Regulation (GDPR) requires organization to implement adequate controls to protect Personally Identifiable Information (PII). A critical step in that journey is impact assessment. That is, understanding where data subject to GDPR is located and how vulnerable it is. In this session, we will introduce IBM Security Guardium Analyzer, a new SaaS offering which intends to help organizations easily complete the impact assessment for their databases.
Industry Company: IBM
Business unit IBM Security
Product Name: IBM Security Guardium Analyzer
Web Site: https://www.ibm.com/marketplace/guardium-analyzer
Research Team: IBM: Devan Shah  Walid Rjaibi  Larry Lindsay  Josue Diaz 

Continuous Delivery with Open Toolchains

Title: Continuous Delivery with Open Toolchains
Chair(s): Chris Brealey , Curtis D'Entremont , Christophe Elek , Joel Cayne
Format: Demo
Description: Continuous Delivery (n) - 1. The subtle science and exact art of releasing high quality, bite-sized changes to software at high velocity without sacrificing control or the ability to pivot with the needs of our clients. - 2. a feature of the IBM Cloud, built on the Open Toolchain platform, that delivers industry leading development tools and tool integrations as a service. Drop by to see toolchains in action, and to get the creative juices flowing on ways to evolve them!
Business unit IBM Watson & Cloud Platform
Product Name: IBM Cloud Continuous Delivery
Web Site: https://bluemix.net/devops
Research Team: IBM DevOps Services: Colleen Anderson  Chris Brealey  Joel Cayne  Curtis D'Entremont  Christophe Elek  Mike Melick  Philippe Mulet  Simon Kotwicz Herniczek  Tom Schmidt  Herman Singh  and many more on Dave Thomson's tribe 

Create data-driven visual reports in Reactive Reports, a framework built for big data

Title: Create data-driven visual reports in Reactive Reports, a framework built for big data
Chair(s): Tereza Nedelescu , Mihai Nicolae , Samar Sajnani
Format: Demo
Description: This expo will showcase Reactive Reports, an open source framework for generating data-driven reports. The framework can use Scala and Java code to take input from mixed sources, such as database records, objects, or streams. The generated output is one or more PDF reports that can be used to summarize or visualize data. Reactive Reports scales well for large amounts of input data due to its memory efficient design, the use of Akka streaming, and its internal data storage architecture.
Industry Company: N/A
Product Name: Reactive Reports
Web Site: https://github.com/sysalto/ReactiveReports
Research Team: IBM Canada: Tereza Nedelescu  Mihai Nicolae  Samar Sajnani 

Db2 on Cloud Free Forever Cloud Database

Title: Db2 on Cloud Free Forever Cloud Database
Chair(s): Saygin Yag , Simon Lightstone
Format: Short Talk
Description: In 2018, Db2 on Cloud introduced a lite plan that is completely free for as long as you need it. This plan provides users with 100MB of cloud database space on a multi-tenant system. This presents a perfect starting point for students, university classes, idea-stage startups and anyone who wants to have a world-class cloud database, for free!
Business unit IBM Analytics
Product Name: Db2 on Cloud
Web Site: https://www.ibm.com/cloud/db2-on-cloud
Research Team: IBM: Saygin Yag - Offering Manager, Db2 on Cloud  Simon Lightstone - Offering Manager, Db2 on  

Easy Cloud-Native Development with Microclimate

Title: Easy Cloud-Native Development with Microclimate
Chair(s): Elson Yuen , John Collier , Sakib Hasan , Stephanie Cao
Format: Poster
Description: Microclimate is a brand-new development environment that offers a end-to-end development experience for Microservices from local or IBM Cloud private cluster. With Microclimate, you can create or import Java, Node.js, or Swift applications, and using any editor of your choosing to develop applications. We will demonstrate Microclimates features on Iterative Development, DevOps pipeline and Intelligent Feedback, and will show how the tools will help developers to accelerate their development.
Business unit Hybrid Cloud
Product Name: Microclimate
Web Site: https://microclimate-dev2ops.github.io/
Research Team: IBM: Elson Yuen  Eric Peters  Erin Harris  Jingfu Wang  John Collier  John Pitman  Jonathan West  Joseph Kim  Laven Sathiyanathan  Leonard Theivendra  Keith Chong  Maysun Faisal  Rajiv Senthilnathan  Sakib Haasan  Stephanie Cao  Steven Hung  Tiaoyu Wang  Tim Etchells 

Installing features to the Open Liberty kernel with Maven and Gradle

Title: Installing features to the Open Liberty kernel with Maven and Gradle
Chair(s): Eric Lau
Format: Poster
Description: Open Liberty is a lightweight and modular server runtime for Java developers. It has features on Maven Central that can be installed to its kernel. We showcase the use of these features, the steps to add them into Maven and Gradle builds, and the benefits of doing so for developers. We explain how they provide relevant dependencies for compiling applications which simplifies the development process. We also provide high level details of the components used to make this solution possible.
Business unit IBM Hybrid Cloud
Product Name: Open Liberty
Web Site: https://openliberty.io/
Research Team: IBM: Eric Lau  Alvin So  Patrick Tiu 

Open Liberty and Eclipse MicroProfile: Foundation for Microservices

Title: Open Liberty and Eclipse MicroProfile: Foundation for Microservices
Chair(s): Eveline Cai , Gilbert Kwan , Yee-Kang Chang , Panagiotis Roubatsis
Format: Poster
Description: We will introduce Open Liberty, which is the largest open source contributions ever done by IBM, what it is, and why Open Liberty. We invite you to come see what Open Liberty is, collaborate with us, and contribute to this fast, dynamic, and simply amazing application server! Eclipse MicroProfile is the next step in the evolution of enterprise Java technologies for the microservices and cloud native world.
Business unit IBM Hybrid Cloud
Product Name: Open Liberty
Research Team: IBM: Eveline Cai  Gilbert Kwan  Yee-Kang Chang  Panagiotis Roubatsis 

REST APIs with LoopBack 4 and OpenAPI 3

Title: REST APIs with LoopBack 4 and OpenAPI 3
Chair(s): Taranveer Virk , Diana Lau , Biniam Admikew , Janny Hou
Format: Short Talk
Description: REST APIs power web applications, mobile applications, IoT devices & more. LoopBack 4 is an API creation framework that has been rewritten from the ground up and features a new programming interface, OpenAPI support, Dependency Injection, a new CLI and so much more. LoopBack makes it easy to create powerful OpenAPI 3.x (an industry standard) compliant REST APIs in minutes. It supports connecting to multiple SQL and NoSQL databases. Poster will show how to get started in minutes with LoopBack 4.
Business unit Hybrid Cloud
Product Name: LoopBack
Web Site: http://v4.loopback.io/
Research Team: IBM: Taranveer Virk  Diana Lau  Biniam Admikew  Janny Hou 

Space Sentry Challenge: A game built with IoT devices and Microservices

Title: Space Sentry Challenge: A game built with IoT devices and Microservices
Chair(s): Frank Ji , Prashanth Gunapalasingam
Format: Demo
Description: Space Sentry Challenge is a fun arcade game that allows players to take control of a 3D printed spaceship to shoot down pop-up targets. This game is built for the purpose of showcasing the application of Microservices in an IoT context. It serves as a good sample project for anyone who wants to build IoT projects with Microservices. It uses popular IoT kits like Arduino, ESP wifi modules and Raspberry Pi with Microservices Java applications running on Open Liberty.
Business unit IBM Cloud
Product Name: Open Liberty
Web Site: https://openliberty.io
Research Team: IBM: Frank Ji  Prashanth Gunapalasingam  King Lam 

Synthesizing Realistic Data for Machine Learning

Title: Synthesizing Realistic Data for Machine Learning
Chair(s): DAVID DCOSTA
Format: Poster
Description: This exhibit is to show several techniques to generate large volumes of synthetic data required by machine learning algorithms. The aspects explored will show the need and how to build proper relationships in attribute features using copulas to assist in testing machine learning algorithms. We will show ways to efficiently build and output large datasets in the order of 5 Million clients. In addition, a way to scale dataset creation in a parallel process across several virtual machines.
Product Name: Client Insights for Wealth Management
Web Site: https://www.ibm.com/us-en/marketplace/client-insight-for-wealth-management
Research Team: IBM Canada Lab: DAVID D'COSTA  DIANE REYNOLDS  DAVID XIE  DONNA LEE  PIA MEDINA  CONNOR OBRIEN  JOSH VALCHAR 

The Power of Accelerators for Cognitive Workloads

Title: The Power of Accelerators for Cognitive Workloads
Chair(s): Mel Bakhshi , Theresa Xu , Neil Graham
Format: Poster
Description: Moores Law is no longer operative. Yet, cognitive workloads demand ever-more computational resources to process ever-increasing volumes of data. What options does he industry have to move forward? In this poster session, well show how accelerators such as GPUs, FPGAs and high-bandwidth storage and network can help to address these challenges and show how POWER systems are uniquely designed to take advantage of these emerging technologies.
Product Name: Power 9 Systems
Research Team: IBM Power Systems Performance: Neil Graham  Mel Bakhshi  Theresa Xu 

Third Party Container Tests

Title: Third Party Container Tests
Chair(s): Mesbah Alam
Format: Poster
Description: This poster session depicts the '3rd party container tests' that we curate and run at the AdoptOpenJDK project. The IBM Runtime technologies team that produces the 'Eclipse OpenJ9' product uses many test materials to verify the SDK binaries. In case of 3rd party application tests, we run a variety of Java applications inside Docker containers. This poster session will describe the test automation techniques we use to execute, triage and debug application tests inside Docker.
Business unit IBM Runtime Technologies
Product Name: OpenJDK with Eclipse OpenJ9
Research Team: IBM: Mesbah 

(Semi)Automatic Construction of Access-Controlled Web Data Services

Title: (Semi)Automatic Construction of Access-Controlled Web Data Services
Chair(s): Kalvin Eng
Format: Demo
Description: A demo of a prototype system implementing a novel fitness application use case. This system uses model-driven methodology to (1) automatically abstract the representation of relational database schemas into a visual model and map them to existing ontologies, (2) use the mappings in order to create different role-based access-control views of APIs, and (3) automatically generate API endpoints and their responses, based on these mappings.
Industry Company: IBM
Research Team: University of Alberta: Kalvin Eng  Diego Serrano  Eleni Stroulia  Jacob Jaremko 

A novel approach to generational garbase collection

Title: A novel approach to generational garbase collection
Chair(s): Kim Briggs
Format: Poster
Description: Presenting a new algorithm for evacuating live objects from the nursery during generational collection cycles. This algorithm uses a stack to follow object reference chains and selectively copies small objects close to referring pointers, thereby improving collector throughput and object locality within the generational heap. This algorithm is being developed within IBMs open source OMR framework, which provides language-agnostic components to provide common components for language runtimes.
Industry Company: IBM
Web Site: https://github.com/eclipse/omr
Research Team: IBM Canada: Kim Briggs 

A semi-automated Framework for Migrating Web applications from SQL to NoSQL database

Title: A semi-automated Framework for Migrating Web applications from SQL to NoSQL database
Chair(s): Rahma Al Mahruqi
Format: Poster
Description: We propose a semi-automated approach to migrate highly dynamic SQL-based web applications to ones that uses NoSQL database. There are two parts to this approach: the migration of schema and data, and the migration of the actual application code with embedded queries. Our approach provides contributions to migrating and optimizing the embedded SQL queries to interact with the new database system and changing the application code to use the translated queries.
Industry Company: IBM
Research Team: Queen's University: Rahma Al Mahruqi  Manar Alalfi  Thomas Dean 

Accessibility in the DevOps Era

Title: Accessibility in the DevOps Era
Chair(s): Ali Asghar , Larry Lindsay , Thomas Brunet , Babz Kukoyi
Format: Demo
Description: As enterprises adopt continuous deployment practices, the DevOps testing practice must adapt accessibility testing tools and processes or user experience will suffer. These tools must work hand in hand with your application DevOps methodology - both required by & enabling those practices. In this hands-on demo, the audience will learn about IBMs cutting-edge accessibility DevOps tooling and see how simple it is to integrate the tools into an existing continuous integration pipeline (CI).
Web Site: https://ibm.biz/a11yDashboard
Research Team: IBM: Ali  Jason  Babz Kukoyi 

Attention models for gene promoter sequence characterization

Title: Attention models for gene promoter sequence characterization
Chair(s): Raul Ivan Perez Martell , Alison Ziesel , Alex Thomo , Ulrike Stege
Format: Poster
Description: Augmented recurrent neural networks are well-suited to the analysis of eukaryotic promoter sequences. DNA sequences, like other biological sequences, are related to other sequential problems such as natural language processing (NLP). In this work we will investigate the efficacy of attention-augmented RNN networks on molecular biology data including DNA sequence, ChIP, and protein binding assays. We also investigate different biological data types and their relevance for a well-trained network.
Industry Company: IBM
Research Team: University of Victoria: R. I. Perez Martell  A. Ziesel  A. Thomo  U. Stege 

Automating the Detection of Third-Party Java Library Migration At The Function Level

Title: Automating the Detection of Third-Party Java Library Migration At The Function Level
Chair(s): Mohamed Wiem Mkaouer
Format: Poster
Description: This poster complements our accepted paper entitled "Automating the Detection of Third-Party Java Library Migration At The Function Level". In this poster, we show the design of algorithms related to mining all the manually-performed function replacements for a given library migration. We further show the detailed results of the comparative study between state-of-art approaches for detecting migration traces at the function level.
Industry Company: IBM
Web Site: http://migrationlab.net/
Research Team: Rochester Institute of Technology: Mohamed Wiem 

Container Technologies! Which One Best Suits Eclipse IoT Deployments?

Title: Container Technologies! Which One Best Suits Eclipse IoT Deployments?
Chair(s): Mohab Aly , Foutse Khomh
Format: Poster
Description: The increasing apex in the IoT paradigm requires specific platforms to allow different components to communicate. Our research dives into assessing the performance implications of the Open-Sourced Eclipse IoT, including Hono, to provide guidelines while deploying them in production. In our work, we setup loads in container technologies, such as Kubernetes and OpenShift, and explore how to effectively scale them out using EnMasse while analyzing the CPU usage, memory consumption and network I/O.
Industry Company: IBM
Research Team: Polytechnique Montral: Mohab Aly  Foutse Khomh 

Detection of Feature Interaction in Dynamic Scripting Languages

Title: Detection of Feature Interaction in Dynamic Scripting Languages
Chair(s): Omar Al Harthi
Format: Poster
Description: Our research indicates that duplicate names, from published plugins, can be a potential cause of conflict due to overlapping. Such feature interaction may create a challenge for smooth compatibility of JavaScript plugin. To detect duplicate names we need first to understand how the objects and properties are created, used and changed in JavaScript.This work aims to detect such conflict using static​ analysis approach.
Industry Company: IBM
Research Team: Queens University: Omar Al Harthi  Manar Alalfi  Thomas Dean 

DevOps Round-trip Engineering: Traceability from Dev to Ops and Back

Title: DevOps Round-trip Engineering: Traceability from Dev to Ops and Back
Chair(s): Miguel Jimenez
Format: Demo
Description: DevOps engineers follow an iterative and incremental process to develop Deployment and Configuration specifications. Such process is likely to involve manual modifications to the running environment. Failing to appropriately update the specifications leads to technical debt. In this demo, we present a two-way Continuous Integration framework that automatically updates Terraform templates when the target OpenStack deployment changes, enabling bi-directional traceability of the modifications.
Industry Company: IBM
Web Site: https://github.com/RigiResearch/jachinte-DevOps2018-evaluation
Research Team: University of Victoria: Miguel Jimenez  Hausi Mller 
Universidad Icesi: Norha M. Villegas  Gabriel Tamura 
IBM Canada: Ian Watts  Joe Wigglesworth 

Distribution and Severity of Test Smells: An exploratory study

Title: Distribution and Severity of Test Smells: An exploratory study
Chair(s): Mohamed Wiem Mkaouer
Format: Short Talk
Description: Similar to production code, unit tests are also susceptible to bad programming practices, which can, in turn, have a negative impact not only on the quality of the software system but also on maintenance activities. In this talk, I showcase various test smell types, demonstrate their existence in Java projects, analyze their impact on software maintenance in terms of code vulnerability to changes and bugs.
Industry Company: IBM
Web Site: https://testsmells.github.io/
Research Team: Rochester Institute of Technology: Mohamed Wiem 

Document Signing on Blockchain

Title: Document Signing on Blockchain
Chair(s): Asic Chen , Arno Jacobsen
Format: Demo
Description: Our project provides a fast, secure, and decentralized solution for document signing based on blockchain technology. Document endorsements are immutably recorded and reliably time-stamped on the blockchain, easily verifiable by third parties that neither need to authenticate the signatories, nor need to access the content of the document. User identities are verified without the need for a centralized certificate authority. Our prototype is built on Hyperledger Fabric with an Angular front end.
Industry Company: IBM
Web Site: http://www.msrg.org/
Research Team: University of Toronto: Asic Chen  Arno Jacobsen  Yuxi (Eric) Zheng  Chen (Jackie) Gao  Yuxuan (Eric) Zheng  Bogdan Scaunasu 

Dynamic Large-Scale Graph Processing over Data Streams

Title: Dynamic Large-Scale Graph Processing over Data Streams
Chair(s): Tariq Abughofa
Format: Short Talk
Description: Processing time-evolving graphs is still a challenging problem. In this work, we studied approaches to dynamic graph processing with Spark. The study helped us design and develop Sprouter, a dynamic graph processing framework that enables storing enormous graphs, real-time updates, and efficient OLTP and analytics. We selected community detection as a case study of incremental analytics with Sprouter and proposed IDWCC, a novel community detection algorithm for dynamic graphs.
Industry Company: IBM
Research Team: Queen's University: Tariq 

Fall as an irregular pattern in IoT streaming data

Title: Fall as an irregular pattern in IoT streaming data
Chair(s): Sazia Mahfuz
Format: Poster
Description: Detecting patterns in real time streaming data has been an interesting and challenging data analytics problem. In this work, we address the problem of learning an irregular human activity pattern, fall, from streaming IoT data using wearable sensors. We present a deep neural network model for detecting fall based on accelerometer data using an online dataset "MobiAct" by Vavoulas et al. The initial model was developed using IBM Watson studio and then later transferred and deployed on IBM Cloud.
Industry Company: IBM
Research Team: Queen's University: Sazia Mahfuz  Dr. Farhana Zulkernine 
IBM Canada: Peter Nicholls 

Fear and Disgust during the 2015 Canadian Election: Automated Analysis with Watson

Title: Fear and Disgust during the 2015 Canadian Election: Automated Analysis with Watson
Chair(s): Trevor Deley
Format: Poster
Description: There is a tremendous opportunity for automated content analysis solutions in communications research as well as the media industry at large. This poster presents a proof of concept for automated content analysis in a media studies context. Currently there is a gap between the long standing field of content analysis in media studies and the field of sentiment analysis that arose as a subset of natural language processing. This exhibit hopes to begin to bring the two fields closer together.
Industry Company: IBM
Research Team: University of Ottawa: Trevor Deley 

Feature engineering in Big Data for the detection of information systems misuse

Title: Feature engineering in Big Data for the detection of information systems misuse
Chair(s): Eduardo Lopez , Kamran Sartipi
Format: Poster
Description: This cybersecurity research project explores the Apache Spark-enabled feature engineering activities conceptualization and generation in machine learning detection of an insider's threat. Our experiment uses a very large, 1.6 billion records dataset of information systems logs, collected for more than 12,000 users over a period of 58 days. We test the feature set using classification models, validating feature relevance and significance using Receiver Operating Characteristics curves.
Industry Company: IBM
Research Team: McMaster University,East Carolina University: Eduardo Lopez (IS at McMaster)  Kamran Sartipi (CS at ECU and IS at McMaster 

Finding Missing Labels for Large Industrial Datasets

Title: Finding Missing Labels for Large Industrial Datasets
Chair(s): Mona Nashaat
Format: Poster
Description: Obtaining labelled training datasets is a common challenge in the business domain. It costs a considerable amount of time and domain experience. We present a new approach to generate labelled datasets, in which we employ active learning in the data programming process. The proposed method aims at improving the accuracy of the training data. It optimizes the user engagement with the labelling process along with the annotation cost. It helps the users to trust the predictions of the final model.
Industry Company: IBM
Research Team: University of Alberta: Mona Nashaat  Aindrila Basak  James Miller 
IBM Canada: Shaikh Quader 
IBM US: Chad Marston 

FLASH: Multithreaded Fault Localization

Title: FLASH: Multithreaded Fault Localization
Chair(s): Luisa Rojas Garcia , Jeremy Bradbury
Format: Poster
Description: Multithreaded programs can have considerable performance benefits over sequential programs. However, this often comes at a cost with respect to program understandability as well as testing and debugging. To address these challenges we have developed an automatic fault localization tool for multithreaded Java programs (FLASH). Our tool utilizes a combination of code injection and a heuristic search algorithm to identify high probability code blocks.
Industry Company: IBM
Research Team: University of Ontario IT: Luisa Rojas Garcia  Jeremy Bradbury 

Foodie: Conversational AI for the smart kitchen

Title: Foodie: Conversational AI for the smart kitchen
Chair(s): Prashanti Priya Angara , Miguel Jimenez , Hausi Muller , Ulrike Stege
Format: Demo
Description: Foodie is a cognitive text-and-voice based conversational agent that augments the capabilities of home cooks by incorporating health-related information to aid one's eating habits. This exhibit demonstrates the capabilities and limitations of existing conversational frameworks via a hands-on demo of Foodie which is built with IBM's Natural Language Understanding and voice services. On the back-end, Foodie is connected to food databases such as Spoonacular (for recipes).
Industry Company: IBM
Research Team: University of Victoria: Prashanti Priya Angara,  Hausi A. Mller,  Ulrike Stege 

Interactive Visualization to Engage End-users in Prescriptive Analysis

Title: Interactive Visualization to Engage End-users in Prescriptive Analysis
Chair(s): Aindrila Basak
Format: Poster
Description: The research proposes a Visual Analytics System that can help end-users of Data Programming to visualize and understand patterns in input datasets and to write better labeling functions for generating labels. The proposed approach aims at enhancing the end-users' trust in predictive models and helping with the conversion of the predictions into actions. The design requirements of this work were identified by performing semi-structured interviews with end-users, data analysts, and domain experts.
Industry Company: IBM
Research Team: University of Alberta: Aindrila Basak  Mona Nashaat  James Miller 
IBM Canada: Shaikh Quader 
IBM US: Chad Marston 

Machine Learning to Provide Intelligent Digital Health for Underserved Regions

Title: Machine Learning to Provide Intelligent Digital Health for Underserved Regions
Chair(s): Kamran Sartipi , Jing Chong
Format: Poster
Description: This project provides advanced information intelligence for physicians in underserved regions to utilize mined knowledge of medical specialties. A pilot project is ongoing with collaboration of Gastroenterologists and Family Physicians at McMaster University. Resources are: anonymized specialty patient data; research datasets; expert decisions of specialists. Concept lattice, machine learning will produce RDF knolwlegebase. Intelligent consultant service navigates RDF and interacts with doctors.
Industry Company: IBM
Web Site: http://profs.degroote.mcmaster.ca/ads/sartipi/HTMLfiles/Research/ResearchProj2.php
Research Team: McMaster and East Carolina Universities: Kamran Sartipi (CS at ECU and IS at McMaster  David Armstrong (Gastroenterology at McMaste  Smita Halder (Gastroenterology at McMaster U  Henry Siu (Family Medicine at McMaster Unive  Jing Chong (Information Systems at McMaster 

Performance and Security Enhanced Containerized Analytics for IoT Systems

Title: Performance and Security Enhanced Containerized Analytics for IoT Systems
Chair(s): Simon Nadeau , Nathaniel Holeksa , Marios Fokaefs
Format: Poster
Description: Our research delves into the pseudo dichotomy that is to be secure or to be performant, in a cloud based IoT architecture. We explore the two in the context of a sensor network in smart buildings, a novel technology that hopes to improve our environmental footprint. In our work, we deploy Spark streaming jobs in Docker containers and explore how to effectively, and automatically, scale them using Docker Compose pursuant to the CPU and memory usage of the standard and the security analytics.
Industry Company: IBM
Research Team: Polytechnique Montreal: Simon Nadeau  Nathaniel Holeksa  Marios Fokaefs 

Plasma Focused Ion Beam Curtaining Artefact Correction by Fourier-Based Linear Model

Title: Plasma Focused Ion Beam Curtaining Artefact Correction by Fourier-Based Linear Model
Chair(s): Christopher Schankula , Christopher K. Anand
Format: Poster
Description: This exhibit describes one example of a multidisciplinary partnership between the McMaster Computing & Software and Materials Science & Engineering departments to better solve modern Materials Science problems computationally. Specifically, we describe a software package for reducing artefacts in focused ion beam scanning electron microscope (FIB-SEM) images. We will describe and provide a live demonstration of the linear optimization model used to reduce the curtaining artefacts.
Industry Company: IBM
Research Team: McMaster University: Christopher W. Schankula  Christopher K. Anand  Nabil D. Bassim 

Randomized Offloading Algorithms for Mobile Cloud Computing

Title: Randomized Offloading Algorithms for Mobile Cloud Computing
Chair(s): Haleh Shahzad , Majid Risman Kar Zadeh
Format: Poster
Description: Computational offloading can improve the energy efficiency of mobile devices, by executing some tasks of a mobile application in the cloud. In my research, a set of new algorithms are presented to find the best decision for each task. These algorithms iteratively improve an offloading decision vector, trying to find the optimal one that can minimize the total energy consumption of the mobile device when the application is running while satisfying the delay requirement of the application.
Industry Company: IBM
Research Team: McMaster University: Haleh Shahzad 
Rogers Communications: Majid Risman Kar Zadeh 

Requirements-based Framework for Evaluating Trust in Service-oriented Systems

Title: Requirements-based Framework for Evaluating Trust in Service-oriented Systems
Chair(s): KONSTANTINOS KONTOGIANNIS , KOSTAS TSIOUNIS
Format: Poster
Description: - We define trust as the degree of expectation that a service, resource, or system, will satisfy its predefined functional and non-functional requirements. - Users define their requirement expectations sought by a service or a resource (i.e. their goals) as expressions. - Fuzzy goal model reasoners, the TidalTrust algorithm, and information obtain from the system are used to compute the final trust a user associates that the service will be achieving the goals set in a given context.
Industry Company: IBM
Research Team: Western University: KOSTAS KONTOGIANNIS  KOSTAS TSIOUNIS 

Semantic-Aware Disambiguation for Entity Resolution

Title: Semantic-Aware Disambiguation for Entity Resolution
Chair(s): Yu Huang
Format: Poster
Description: As data is integrated, duplicates inevitably occur in the integrated instance. Existing similarity functions often rely on string similarity to identify duplicates, which is insufficient in capturing the semantic meaning of terms. We propose a semantic-aware deduplication framework which can identify the duplicates by extracting the semantic information to measure 'closeness' of entities.
Industry Company: IBM
Research Team: McMaster University: Yu Huang  Fei Chiang 
IBM Germany: Albert Maier  Yannick Saillet 
IBM Chicago: Damir Spisic 
IBM Toronto Lab: Calisto Zuzarte 

Supporting Software Variability in Eclipse OMR

Title: Supporting Software Variability in Eclipse OMR
Chair(s): Samer AL Masri
Format: Poster
Description: Eclipse OMR is a C++ library of language runt-ime components that supports multiple architectures and various languages. Aiming to support variability in OMR, we created OMRStatistics, a static analysis tool that helps developers collect statistics about the project and make development decisions about variability. I will be explaining how we use OMRStatistics to support variability in OMR and our experience switching from static polymorphism to dynamic polymorphism in OMR's implementation.
Industry Company: IBM
Web Site: https://github.com/samasri/omr/tree/master/tools/compiler/OMRStatistics
Research Team: University of Alberta: Samer 

User Research: Mapping the end-to-end experience of a System Administrators deployme

Title: User Research: Mapping the end-to-end experience of a System Administrators deployme
Chair(s): Arun Martin , Steve Murphy
Format: Short Talk
Description: In this talk, we share how design research helped the product stakeholders identify issues in the end-to-end experiences and various support channels accessed while deploying complex enterprise applications. During this design research journey, the product stakeholders understood the ecosystem of primary & secondary users, dependencies and handoffs and how the current customer journey impacted the organizational teams and processes. The findings from the study informed the creation of a MVP.
Research Team: IBM Canada Lab: Steve Murphy, UX Design Lead  Arun Joseph, Design Researcher 

Using AP-TED to Detect Phishing Attack Variations

Title: Using AP-TED to Detect Phishing Attack Variations
Chair(s): Sophie Le Page
Format: Poster
Description: Detection of phishing attack variations is accordingly currently used as a prevention tool against phishing. We evaluate here the feasibility of applying Pawlik and Augsten's recent implementation of Tree Edit Distance (AP-TED) calculations as a way to compare DOMs and identify similar phishing attack instances. We also compare this tree method with an existing method that uses the distance between tag vectors to quantify similarity between phishing sites.
Industry Company: IBM
Research Team: University of Ottawa: Sophie Le Page 

Variability Aware analysis using Clang

Title: Variability Aware analysis using Clang
Chair(s): Jacob Reckhard
Format: Poster
Description: Tool-based analysis is important to catch errors before the codebase becomes too dependent on the erroneous parts. One well known project for compilation and analysis of code is Clang. Here, we describe the changes required to coerce Clang into conducting variability-aware analysis of C++ code. Varibility-aware analysis is a technique in which the code base is analysed for all possible configurations of the code.
Industry Company: IBM
Research Team: University of Alberta: Jacob Reckhard  Dr. Sarah Nadi 

Watch Phishing Evolution Over time

Title: Watch Phishing Evolution Over time
Chair(s): Qian Cui , Sophie Le Page , Emad Badawi , Ali Almokhtar
Format: Poster
Description: We proposed a model to monitor and analyze the phishing evolution. By using our model, we could observe the evolution pattern over different phishing groups
Industry Company: IBM
Web Site: https://esorics2018.upc.edu/program.do
Research Team: University of Ottawa: Gregor v. Bochmann  Guy-Vincent Jourdan  Qian Cui  Emad Badawi  Ali Almokhtar  Sophie Le Page 
Polytechnique Montreal: Ettore Merlo  Stephane Heudron 
IBM-CAS: Iosif-Viorel (Vio) Onut 

Data Science 1 - Competency Model

Title: Data Science 1 - Competency Model
Chair(s): Marcellus Mindel , Dennis Buttera , Jennifer Collins , Ana Echeverri
Format: Poster
Description: In the first of three exhibits on Data Science Education, we will discuss applications of a Competency Model for Data Science originally developed by Ana Echeverri, IBM Data Science Growth Strategies Lead, to support the nontraditional education paths of new collar jobs. The model may also prove useful for postsecondary institutions that use competency matrices to manage the coverage of data and analytics skills in academic programs.
Industry Company: IBM
Research Team: IBM Advanced Studies: Marcellus Mindel  Dennis Buttera  Jennifer Collins 
IBM Data Science Growth Strategies: Ana Echeverri 

Data Science 2 - Bootcamp Materials

Title: Data Science 2 - Bootcamp Materials
Chair(s): Marcellus Mindel , Dennis Buttera , Jennifer Collins
Format: Poster
Description: In the second of three exhibits on Data Science Education, we will discuss uses for 12 hours of course materials that cover the data science workflow using IBM Watson Studio. These materials are suitable for inclusion as part of an introductory Data Science course. They could also be delivered as an intensive hands-on bootcamp, for example, to complement a business program, or to enable graduate students from different backgrounds to begin with a consistent practical foundation for data science.
Industry Company: IBM
Research Team: IBM Advanced Studies: Marcellus Mindel  Dennis Buttera  Jennifer Collins 

Data Science 3 - Student Success Projects

Title: Data Science 3 - Student Success Projects
Chair(s): Marcellus Mindel , Jennifer Collins , Dennis Buttera
Format: Poster
Description: In the third of three exhibits on Data Science Education, we will discuss how students at three different institutions are using Watson Studio to analyze institutional data to improve student retention and satisfaction. This approach is intended to create a pattern for supporting a wide variety of student capstone projects involving IBM tools and practices. Drop by to learn more about tools like Watson Studio and practices like Enterprise Design Thinking and Outcome Delivery.
Industry Company: IBM
Research Team: IBM Advanced Studies: Marcellus Mindel  Jennifer Collins  Dennis Buttera