The COMPaaS DLV: Composable Platform as a Service Instrument for Deep Learning & Visualization user community consists of faculty from Physics and four College of Engineering departments (Computer Science, Civil, Materials & Environmental Engineering, Mechanical and Industrial Engineering, and Electrical and Computer Engineering) - primarily run applications that are GPU-centric for compute, with significant variability in storage and networking around their data requirements. 
Computer Science applications primarily focus on security, data science, computer vision and Machine Learning (ML). Security projects explore the complexity of modern web applications and the intricacies of security mechanisms that often result in flaws that expose users to significant security and privacy threats. These projects try to develop methods and tools that enable users to understand and more effectively manage retrospective privacy in the context of modern, long-lived, online archives. Composable resources were used to develop Natural Language Processing (NLP)-based domain-specific classifiers that identified data practices stated in privacy policies. Adherence of corresponding applications were then adjusted based on this ground truth. 
Data-science applications have an intuitive framework that integrates state-of-the-art AI technologies with applications, workflows, smart visualizations and collaboration services to help users access, share, explore and analyze their data, whether local or remote, come to conclusions, and make decisions with greater speed, accuracy, comprehensiveness and confidence. One such project is developing and advancing tools that identify image data in biomedical literature to locate beneficial, targeted publications. This work involves training image classifiers, integrating classifiers into labeling pipelines, designing retrieval user interfaces, and identifying related visual representations. 
Computer vision projects include semantic segregation and 3D human pose estimation. Researchers are developing a novel network architecture, termed DependencyNet (dependency network), for semantic segmentation. They also achieved experimental results that demonstrate an effective approach for 3D human pose estimation. Over the past two years, they found COMPaaS to be consistently stable and efficient, with result output as expected, and the group&rsquo;s models achieved state-of-the-art performance on their respective benchmarks. 
ML applications include frameworks for many complex real-world reinforcement learning problems, such as the coordination of autonomous vehicles, network packet delivery, and distributed logistics. 
Civil, Materials & Environmental Engineering applications focus on simulation and modeling. Researchers run data-driven models on high-performance computers to develop an accurate and general neural network ML model that uses crystallographic data to study patterns of synthesizability. They also perform simulations of mass transport in alloys and ceramics. COMPaaS has been performing 1.5-10 times faster than comparable infrastructures they are familiar with. Additionally, they found our use of Jupyter Notebooks to be a significant asset. 
Mechanical and Industrial Engineering researchers use COMPaaS for three research projects: feature extraction in fluid flow using a Convolutional Neural Network (CNN); column height detection in metallic nanoparticles using a CNN; and, electric vehicle battery state-of-charge estimation using different ML methods. 
Electrical and Computer Engineering researchers recently started running mathematical models of ML algorithms and training language models using long-short term memory (LSTMs). 
In 2020, we began implementing a GPUoE (GPU over Ethernet) prototype. Using a GPU expansion chassis connected to compute nodes over Ethernet, we were able to compose remote GPUs into our existing composable infrastructure. Our APIs developed for composable infrastructure have been extended to support these remote GPUs. 
In 2021, we enhanced COMPaaS with a public-facing JupyterHub server providing secure web-based access to experiments, a modern PCIe fabric and a supporting expansion chassis, thereby making composable co-location available to our users. As familiarity with composable infrastructure grows, the idea of using composable co-location instead of traditional servers co-located in racks is possible. By composable co-location, we mean the integration of emerging hardware needs without changing the existing core infrastructure. Researchers can now co-locate their own accelerators or compute nodes and leverage existing COMPaaS resources natively.

COMPaaS DLV User Community

Project Period: October 1, 2018 - September 30, 2021 
September 9, 2018 - The National Science Foundation (NSF) awarded a three-year grant to the UIC Computer Science (CS) Department&rsquo;s Electronic Visualization Laboratory (EVL) to acquire the &ldquo;Composable Platform as a Service Instrument for Deep Learning &amp; Visualization (COMPaaS DLV)&rdquo;. The total project cost is $1,424,806, with NSF providing $997,363 for hardware and the balance provided by UIC cost share to cover staff and graduate student support to design, operate and maintain the system. Maxine Brown, EVL Director, is Principal Investigator (PI) and co-PIs are CS faculty and EVL affiliated faculty members Andrew Johnson, Robert Kenyon and G. Elisabeta Marai. 
COMPaaS DLV will be acquired and made available for Deep Learning and Visualization research and training. Initial users will be 17 faculty from 5 departments in the UIC College of Engineering (COE) who provided compelling use cases for this system, and then expanded to other COE faculty as availability and demand permits. Users include the above-mentioned PI and co-PIs, plus COE Senior Personnel: Tanya Berger-Wolf (CS), Sybil Derrible (Civil &amp; Materials Eng.), Barbara Di Eugenio (CS), Chris Kanich (CS), Alex Leow (BioEng/Psychiatry, and by courtesy CS), Bing Liu (CS), Farzad Mashayek (Mechanical &amp; Industrial Eng.), Roberto Paoli (Mechanical &amp; Industrial Eng.), Luc Renambot (CS), Philip Yu (CS), Milos Zefran (Electrical &amp; Computer Eng.), Xinhua Zhang (CS), and Brian D. Ziebart (CS). 
COMPaaS DLV is critical infrastructure on which UIC faculty can pursue fundamental research in deep learning (data mining, data analytics, computer vision, natural language processing, artificial intelligence), visualization (simulation, rendering, visual analytics, video streaming, image processing), and a combination of deep learning and visualization (that is, when data is so large that it cannot be easily visualized, deep learning is used to extract features of interest to be visualized). UIC faculty are investigating, and contributing to, solutions for societal issues in such areas as anthropology, biology, cybersecurity, data literacy, fraud detection, healthcare, manufacturing, urban sustainability, and cyber-physical systems (e.g., autonomous vehicles). 
Though EVL specializes in visualization and virtual reality, it has a longstanding reputation as a technology enabler for the campus. EVL will design and operate COMPaaS DLV as a highly flexible, scalable, high-throughput system for Big Data. Its flexible 4th-generation computer architecture, &ldquo;composable infrastructure,&rdquo; treats computer components - computer processing units (CPUs), graphical processing units (GPUs), and storage - as pools of resources. Each application defines what resources it needs, and the operating system composes, or combines, resources on the fly and infrastructure is provisioned &lsquo;right-sized&rsquo; and resized as needed. The system is easily scalable - new infrastructure is automatically integrated with existing infrastructure and becomes part of the pool of capacity. The system&rsquo;s components are all interconnected with a high-speed internal network fabric, so that Big Data can quickly move between CPU, GPU and storage at optimum speed, i.e., high-throughput, with little to no bottlenecks. 
COMPaaS DLV will complement, but be connected via UIC&rsquo;s research network, to campus and regional clusters, data storage, supercomputing and cloud testbed facilities managed by the UIC ACCC&rsquo;s Advanced Cyberinfrastructure for Education and Research (ACER) group, the National Center for Supercomputing Applications at University of Illinois at Urbana-Champaign, Argonne National Laboratory&rsquo;s Leadership Computing Facility, and University of Chicago. 
Access to COMPaaS DLV will enable faculty, students and postdocs to execute deep learning and visualization codes faster, apply more sophisticated models to large-scale problems, gain greater insights, accelerate discovery, and open new avenues of research. The instrument will enable UIC to recruit new faculty who are attracted to institutions that can support their computational needs, and to recruit new students and produce a better trained workforce - which in turn attracts industrial partnerships. 
See <a href="https://today.uic.edu/nsf-grant-to-fund-advanced-deep-learning-and-visualization-computing-platform">UIC Today News Release</a> for more information. 
PHOTO CAPTION: One of the applications that will benefit from COMPaaS DLV is &ldquo;SENSEI Panama,&rdquo; a UIC collaborative project with evolutionary anthropologist Meg Crofoot (UC Davis) to enable anthropologists to use virtual reality to &lsquo;walk inside their data&rsquo; and study animal behavior. Specifically Crofoot and UIC computer scientists Tanya Berger-Wolf, Andy Johnson and Maxine Brown, are applying Deep Learning and Visualization to process data from LIDAR, aerial photography, terrain maps, GPS collars, and accelerometers, in order to study how coatis, kinkajous, spider monkeys, and capuchin monkeys on Barro Colorado Island (a Smithsonian Tropical Research Institute Research Center) in Panama, forage for food. This project received some funding from the NSF SENSEI $3.5M award to UIC CS Maxine Brown (principal investigator) and UIC CS co-PIs Andy Johnson, Tanya Berger-Wolf, and Robert Kenyon. (This project/photo won Honorable Mention in the <a href="https://www.evl.uic.edu/entry.php?id=2242">UIC 2016 Image of Research Competition</a>).

An application that will benefit from COMPaaS DLV is &ldquo;SENSEI Panama,&rdquo; which applies Deep Learning and Visualization to help anthropologists walk inside their data using virtual reality to study animal behavior.

COMPaaS DLV: Composable Platform as a Service Instrument for Deep Learning & Visualization

The SAGE Development Team at the Laboratory for Advanced Visualization &amp; Applications (LAVA) at the University of Hawaiʻi at Manoa and the Electronic Visualization Laboratory (EVL) at the University of Illinois at Chicago recently received one-year supplemental funding from the National Science Foundation (NSF) to their SAGE2 award, in the amount of $527,000, to focus on SAGE2 cybersecurity. The diversity and distributed nature of the SAGE2 user base provides a growing set of security concerns. Identity and access management procedures, in particular, provide unique challenges, given the variety of institutions using SAGE2 to collaborate using display walls. 
In July 2018, the SAGE2 Team began working with Trusted CI, the NSF Cybersecurity Center of Excellence at Indiana University, to develop an Engagement Plan that outlines the scope of a 6-month timeline for the engagement, followed by 6 months of implementation. Trusted CI is helping outline Identity and Access Management (IAM) procedures appropriate for SAGE2&rsquo;s distributed user base. For example, SAGE2 needs to identify and provide access to local and remote sites involved in a collaboration as well as the users at those sites; such as: who is allowed to collaborate, are they allowed to share information or simply read it, and are they allowed to download information, etc. A secondary goal of the engagement, time permitting, is to determine if there are any unidentified high-priority cybersecurity needs revealed during the fact-finding stage that the SAGE2 Team should pursue; such as: securing the web server, securing documents that are shared, etc. 
<a href="http://blog.trustedci.org/search?updated-max=2018-08-02T11:59:00-04:00&max-results=7">Trusted CI Blog</a>

The SAGE2 team at UIC/EVL remotely collaborated with the University of Chicago’s Research Computing Center during its fourth annual research computing symposium and exposition, Mind Bytes.

SAGE2 Team Engaging with the Trusted CI Center on SAGE2 Cybersecurity

Ableton Live is music software designed for computer-based composition and performance. <a href="https://en.wikipedia.org/wiki/Ambisonics">Ambisonics</a> is a full-sphere surround sound technique used for playing audio on sound systems with more than two speakers, providing the illusion that sound is coming from all directions. Combining these has allowed music producers to compose music inside of Ableton, in 3D. Programmed in the visual programming language <a href="https://cycling74.com/products/max/">Max</a>, is an ambisonic decoder for the Sound Dome built by former EVL post-doc Anil Camci. 
The decoder outputs sound sources through specific speakers in the sound system based on their specified Azimuth, Elevation, and Distance parameters. To integrate Ableton Live with the decoder, all of the audio channels in Ableton were streamed into the decoder via an application called Jack which allows users to route audio from one application to another. A Max utility was also built to run inside of Ableton Live and allow the user to conveniently adjust the azimuth, elevation, and distance parameters of a particular channel. 
The azimuth elevation and distance values are sent over a network back to the Max decoder allowing producers to change the location of playback for every channel in Ableton without leaving the application. Custom audio effects are also being programmed to exploit all of the surrounding speakers and take full effect of system.

EVL Sound Dome developed by Anil Camci

Ableton Live Integration & Audio Effects for a 32 Speaker Ambisonic Sound System

Pi: Barbara Di Eugenio; Co-I: Debaleena Chattopadhyay 
A growing body of research has found that patients who actively manage their health have better outcomes, and patient engagement has become a central tenet in recent healthcare approaches. However, it has also been shown that many patients lack knowledge of their disease and medications, particularly as concerns chronic health conditions, whose exacerbation causes about two thirds of hospitalizations every year in the USA. Discharge instructions after hospitalization are a critical tool to engage patients in their self-care, but are often unsuccessful. Our hypothesis is that, to engage patients at discharge, a personalized hospitalization summary should be provided, that integrates physician and nurse documentation, and is informed by the patient's level of engagement in self-care and their socio-cultural perspective. 
Our transformative pilot work on physician and nurse documentation has identified an additional obstacle to the effective flow of information from professionals to patients: vocabulary differences between physicians and nurses. By leveraging the Unified Medical Language System Metathesaurus, we found no common synonyms between physician and nurse documentations for 75% of patients. Further, 86% of terms patients use are simple compared to only 15% for physicians. The lack of common language among patients, physicians and nurses motivates including all perspectives to make instructions meaningful for patients. Our hypothesis is that a better informed patient population who understands the connections between their illness, hospital experience, and discharge instructions will be more empowered and more engaged in their chronic disease management. Our innovation is to automatically create an easy-to-understand, personalized hospitalization summary that integrates events described in the EHR; and that is responsive to the patient's needs and preferences, and their activation framework (PAF) level. We pursue four specific aims: 1. To investigate socio-cultural patterns of patient perceptions of illness via a socio linguistic analysis of patient interviews; 2. To integrate multi-faceted information about a patient, as provided by physicians and nurses; 3a. To develop MyPHA, a novel application program, which will automatically generate the personalized summaries and deliver them to the patient on a tablet. MyPHA will be informed by aims 1 and 2, and will be grounded in Natural Language Generation technology; 3b. To evaluate MyPHA with multiple cycles of formative evaluation with all stakeholders, and a summative evaluation with patients. 
Whereas our pilot work has focused on heart failure, the techniques we will develop can easily be adapted to the management of other chronic health conditions, in particular for patients with cancer and cancer survivors. In addition, cancer and heart failure can co-occur, especially in older patients, and several cancer treatments can lead to the development of heart failure. The software program MyPHA will be released ooen source to the research communitv to enable raoid customization to other chronic conditions. 
Awarded Amount: $1,495,213

NIH: R01: SCH: INT: MyPHA: Automatically generating personalized accounts of inpatient hospitalization

The Continuum environment is a high-resolution computer-enhanced group collaboration workspace, consisting of large format 2D touch-enabled and 3D tiled displays, and instrumented with arrays of ambient sensors, enabling users to collaborate with local and remote colleagues and their data more effectively by giving the room itself the ability to anticipate their needs, improve their interaction, and focus their time on discovery rather than the technology. 
EVL&rsquo;s SAGE2 middleware enables teams of users to manage their data, documents, images and media on the Continuum display walls as if the room were one seamless canvas. SAGE2 lets users post and manipulate information, visualizations and animations across the Continuum displays. EVL REU Sean Walker developed a spatial audio system that ties audio playback from SAGE2 media content directly to the media&rsquo;s (i.e. video) location on the display wall. 
Additionally, any audio playback in the Continuum environment can be played back from any of the X number of speakers in the space. The spatial audio functionality was implemented using AVB (Audio Video Bridging) to connect all of the speakers positioned throughout the room coupled with SuperCollider (a functional programming language for audio) to pan the sound around the space.

Spatial Audio in the Continuum Environment

Caterpillar Inc. is working with the UIC Electronic Visualization Laboratory (EVL) to develop a SAGE2&trade; application to enable distributed teams to collaborate on project planning and scheduling. 
Planning and scheduling the design, development, engineering and manufacturing processes of construction equipment with worldwide teams and suppliers is an extraordinarily complex task. Caterpillar, inspired by &ldquo;StickySchedule,&rdquo; an interactive SAGE2 multi-user application for conference scheduling, asked whether EVL could help automate their project management workflow, which is currently a very manual process. 
The goal is to develop an innovative SAGE2 application to enable people from different Caterpillar units to schedule interdependent tasks over several months or years using electronic post-it notes on a large tiled display. The resulting schedule would be translated into a spreadsheet, which would then be input to project management software. SAGE2 would also enable groups to work collaboratively over distance, whereas now, everyone must be in the same room at the same time. Today, these management plans are updated only a couple times per year due to the high labor and cost of travel. This tool will enable these planning sessions to occur for significantly less labor and cost, resulting in a more regular update by the extended team. 
Additional information: 
<a href="http://sage2.sagecommons.org">SAGE2 Website</a> 
<a href="http://sage2.sagecommons.org/download/2117/">StickySchedule Paper</a> 
<a href="http://apps.sagecommons.org/">StickySchedule App</a> 
Project team: 
Lisa Creger, Lean Product Development Advisor, Innovation & Technology Development Division (ITDD), Caterpillar 
Peter E. Cooper, Product Development Manager, Medium Wheel Loaders, Caterpillar 
Luc Renambot, Computer Science Research Professor and EVL SAGE2 Project Manager, UIC 
Tommaso Massari, Computer Science MS student and EVL Research Assistant, UIC

Example of Caterpillar 'manual' project management workflow session.

SAGE2 Project Management Application

The Shirley Ryan AbilityLab of Chicago worked with the Electronic Visualization Laboratory (EVL) at the University of Illinois at Chicago to use virtual reality to help people needing rehabilitation be more enthusiastic and energetic when doing their exercises. 
Using the VIVE head-mounted display, EVL researchers develop virtual reality worlds to augment the rehabilitation patients&rsquo; physical therapy spaces and sessions. The goal of this project is to increase patient interest and engagement, by supplementing physical therapy tasks and exercises with tasks incorporated into a number of different virtual worlds. Unity3D is used to build virtual environments, that are &lsquo;mapped&rsquo; to the physical therapy areas (e.g. a physical balance beam is represented by a very narrow bridge over a stream in a forest; the patient steps onto the actual beam, but are stepping onto a virtual bridge in the VR world). Data about the patient&rsquo;s movements and activities in these virtual worlds are visualized in real-time, allowing AbilityLab personnel to tailor the experience interactively. The enhanced rehabilitation experience intends to foster patient recovery and treatment success.

EVL director Maxine Brown explores a virtual world designed to enhance rehab patient treatment options.

Virtual Reality Enhanced Physical Therapy

In the past decades, science has become increasingly interdisciplinary and scientific discovery has experienced a paradigm shift from the classical scientific method to data-intensive scientific discovery. SAGE2 provides these interdisciplinary collaborators with a powerful, shared workspace. However, there remains a high cost of parallel work integration for data-intensive work. VisSnippets is a system which allows collaborators to perform expressive, impromptu data analysis and exploration within SAGE2. 
VisSnippets acts as a collaborative interface to exploratory programming, allowing users to write modular &ldquo;snippets&rdquo; of code for data retrieval, processing/aggregation, and visualization. Users may then compose modular analysis blocks from the shared snippet pool into branching data pipelines. The snippets may be edited or &ldquo;cloned&rdquo; by collaborators, and reused/applied multiple times across a collaborative session. 
To further facilitate collaborative work, the VisSnippets system provides further abstraction onto the expressive foundation of exploratory programming. VisSnippets reduces the overhead to scaffold a responsive web application and allows users to walk away from the collaborative analysis with a packaged web application containing all of the analysis.