DAIS 2019 - Workshop on Distributed Analytics InfraStructure and Algorithms for Multi-Organization Federations


0845: Opening remarks (Organizing Committee and TPC Chairs)

0900-1000: Keynote (Dr. Tien Pham, Army Research Lab)

Title: AI & ML in Complex Multi-Domain Environment

Abstract: Artificial Intelligence (AI) is envisioned by current and future force the ability to converge capabilities from across multiple domains at speeds and scales beyond human cognitive abilities. At the tactical edge, military operations will involve teams of highly-dispersed warfighters and agents (robotic and software) operating in distributed, dynamic, complex, cluttered environments. Military domains are frequently distinct from commercial applications because of: rapidly changing situations; limited access to real data to train AI and limited resources with SWaP-T (size, weight, power, time) constraints, and noisy incomplete, uncertain, and erroneous data inputs during operations; and peer adversaries that employ deceptive techniques to defeat algorithms. The Army Research Laboratory has a number of AI cross-cutting efforts within the Network & Information Sciences and Computational Sciences Core Competencies. The primary goal of these AI efforts is to research and develop artificially intelligent agents (heterogeneous & distributed) that rapidly learn, adapt, reason and act in contested, austere and congested and address key AI research gaps for Multi-Domain Operations (MDO):

•    Learning in Complex Data Environments
-    AI & ML with small samples, dirty data, high clutter
-    AI & ML with highly heterogeneous data
-    Adversarial AI & ML in contested, deceptive environment
•    Resource-constrained AI Processing at the Point-of-Need
-    Distributed AI & ML with limited communications
-    AI & ML computing with extremely low SWaPT
•    Generalizable & Predictable AI
-    Explainability & programmability for AI & ML
-    AI & ML with integrated quantitative models

Bio: Dr. Tien Pham is the current Acting Chief Scientist of the Computational & Information Sciences Directorate (CISD) at the Combat Capabilities Development Command, Army Research Laboratory (CCDC ARL) in Adelphi, MD. Dr. Pham is responsible for the planning, direction, management, and oversight of very complex basic and applied research programs associated with the ARL Core Competencies in Network & Information Sciences (NIS) and Computational Sciences (CS), and the cross-cutting research efforts in Artificial Intelligence and Machine Learning (A/ML) in Complex Environment. He serves as the scientific ambassador and technical advisor to senior-level administrative and technical management officials within ARL, Army, Department of Defense (DoD), other government agencies, and outside organizations such as academia and industry; and performs and manages research in Computational Modeling of Complex Systems, Data Processing & Data Analytics, and Intelligent & Autonomous Systems; Communications, Networks & Cyber. He has over 25 years of R&D experience and over 100 publications in wide ranging research areas from information sciences to networked sensing, multi- modal sensor fusion, and acoustics. He received his B.S., M.S. and Ph.D. degrees in Electrical Engineering at the University of Maryland, College Park, in 1988, 1991 and 2006 respectively.

1000-1030: Coffee Break I

Paper presentations: 25 mins per paper (20 minutes presentaton + 5 for question/answering)

1030-1210: Session I: AI at Edge (Session Chair: Dinesh Verma)
Neural Networks at the Edge 
Deboleena Roy, Gopalakrishnan Srinivasan, Priyadarshini Panda, Richard Tomsett, Nirmit Desai, Raghu Ganti and Kaushik Roy

Online Distributed Analytics at the Edge with Multiple Service Grades 
Victor Valls, Geeth de Mel, Heesung Know and Leandros Tassiulas

Jointly Compressing and Caching Data in Wireless Sensor Networks
Nitish K. Panigrahy, Jian Li, Faheem Zafari, Don Towsley and Paul Yu

On Data Summarization for Machine Learning in Multi-Organization Federations
Bong Jun Ko, Shiqiang Wang, Ting He and Dave Conway-Jones

1210-1320: Lunch

1320-1500: Session 2: Distributed Learning (Session Chair: Mudhakar Srivatsa)

Overcoming the Lack of Labeled Data: Training Intrusion Detection Models using Transfer Learning
Elisa Bertino, Ankush Singla and Dinesh Verma

Cooperative Learning for Multi-Perspective Image Classification
Nick Nordlund, Heesung Kwon and Leandros Tassiulas

NeuroMask: Explaining Predictions of Deep Neural Networks through Mask Learning
Moustafa Alzantot, Amy Widdicombe, Simon Julier and Mani Srivastava

DeepCEP: Deep Complex Event Processing Using Distributed Multimodal Information
Tianwei Xing, Marc Roig Vilamala, Luis Garcia, Federico Cerutti, Lance Kaplan, Alun Preece and Mani Srivastava

1500-1530: Coffee Break II

1530-1710: Session 3: Distributed Systems and Control (Session Chair: Seraphin Calo)

Magnalium: Highly Reliable SDC Networks using Multiple Control Plane Composition
Geng Li, Akrit Mudvari, Kerim Gokarslan, Patrick Baker, Sastry Kompella, Franck Le, Kelvin Marcus, Jeremy Tucker, Y. Richard Yang and Paul Yu

Using an ASG based Generative Policy to Model Human Rules
Graham White, John Ingham, Mark Law and Alessandra Russo

On the Impact of Generative Policies on Security Metrics
Dinesh Verma, Elisa Bertino, Geeth De Mel and John Melrose

Hybrid SDN Control in Mobile Ad Hoc Networks
Konstantinos Poularakis, Qiaofeng Qin, Kelvin M. Marcus, Kevin S. Chan, Kin K. Leung and Leandros Tassiulas. 

Call for Papers

DAIS 2019 will be co-held with SMARCOMP 2019 at Washington DC, USA between June 12-15, 2019

As the processing power and storage capacity of client-side devices increases, the intelligent management of cyber-physical systems will leverage and benefit from this wide-spread pervasive computational capacity that is available. The computational capacity, however, is usually spread across many different organizations and administrative domains.  In many domains, new research activities which explore how the resources across multiple federated systems can be leveraged have started to emerge. These domains include the field of Smarter Cities, Internet of Things, HealthCare, public safety, military coalition systems, and enterprise business alliances. Another set of domains include environments where sensors belong to many different organizations with split administrative domains, e.g. in home Internet of Things environments, devices like Smart Thermostats may be controlled partially by the equipment manufacturer, cable modem and entertainment systems be controlled by the Internet service Provider, the smart car have partial control from manufacturer and insurance companies, while other devices be controlled by the home owner. It is imperative that intelligent management of emerging cyber-physical systems considers multi-organization federations as a norm rather than the exception.

In such distributed environments, new approaches for leveraging the power of distributed processing systems are needed. In many of these environments, bandwidth may be limited, and connectivity be disrupted and intermittent. At the same time, the different intelligent devices need to work together with each other to perform distributed analytics, including pattern extraction, machine learning and situational awareness leveraging the collective processing capacity of all the devices. Policy and Security mechanisms need to be supported to mitigate against actions taken by organizations in the federation that may not be fully trusted.

Distributed analytics requires rethinking of the infrastructure as well as algorithms that are used for processing data. In this workshop, we will bring together the researchers from different countries who are conducting research in distributed analytics. The goal of the workshop is to bring together researchers active in this area and facilitate exchange of information among them.

Topics of interest to the workshop include but are not limited to:

  • Applying paradigms like Software Defined Networking across multiple organizations
  • Distributed edge computing across multiple organizations
  • Fog Computing across multiple organizations
  • Agile code architectures for analytics in multi organization federations
  • Human-Agent collaborations in multi organization federations
  • Cultural, Ethical and Legal issues in multi organization federations
  • Distributed machine learning across organizational boundaries
  • Situational Awareness in multi-organization federations
  • Self Organization of Services in multi-organization federations

Paper Submission and Guidelines

Paper submissions must be no longer than 6 pages and formatted according to the two-column IEEE proceedings template. IEEE provides corresponding formatting templates at IEEE conference template. Make sure to use the conference mode of the template, i.e., LaTeX users must use the conference option of the IEEEtran document class.
Papers must be submitted electronically as a single PDF file on US Letter size paper (not A4), with all fonts embedded (the PDF-A standard complies with that). Prior to submission, ensure that any running headers/footers, page numbering, as well as blue underlining for URLs and email addresses has been removed. Submissions will be handled via easychair. Once you login as an author then you can select a "track" - one of the tracks is Workshop on Distributed Analytics Infrastructure and Algorithms for Multi-Organization Federations (DAIS).

Important Dates

  • Paper submission deadline: 22 March 2019
  • Notification of paper acceptance: 10 April 2019
  • Camera-ready submission deadline: 28 April 2019

Program Chair:

  • Mudhakar Srivatsa, IBM Research, U.S.A.
  • Seraphin Calo, IBM Research, USA

Organizing Committee:

  • Ananthram Swami, Army Research Labs, U.S.A.
  • John Melrose, Defense Science Technology Laboratories, UK
  • Dinesh Verma, IBM Research, U.S.A.
  • Gavin Pearson, Defense Science Technology Laboratories, UK
  • Tien Pham, Army Research Laboratories, U.S.A.