Workshop Program

8:15 – 8:30
[Slides] Opening Session
Chairs: Laura Galluccio, Giovanni Schembra

8:30 – 10:00
S1: ML for Network design, management and orchestration

Chair: Giovanni Schembra, University of Catania, Italy

[Slides] Optimizing Degree Distributions of LT-based Codes with Deep Reinforcement Learning  
Yehor Savchenko and Yi Liu (Beihang University, P.R. China)

[Slides] PIQoS: A Programmable and Intelligent QoS Framework  
Udaya Bhanu Lekhala and Israat Haque (Dalhousie University, Canada)

[Slides] Sparse Control and Data plane Telemetry features for BGP anomaly detection  
Jose Cordova Garcia (ESPOL, Ecuador & Stony Brook University, USA)

[Slides] Network Traffic Prediction based on Diffusion Convolutional Recurrent Neural Networks  
Davide Andreoletti (University of Applied Sciences of Southern Switzerland, Switzerland); Sebastian Troia and Francesco Musumeci (Politecnico di Milano, Italy); Silvia Giordano (University of Applied Sciences and Arts of Southern Switzerland (SUPSI), Switzerland); Guido Maier (Politecnico di Milano, Italy); Massimo Tornatore (Politecnico di Milano & University of California, Davis, Italy)

10:00 – 10:30
Coffee Break

10:30 – 11:15
Keynote Session: Prof. T. Znati

Abstract. With the advent of new computing paradigms and the proliferation of ubiquitous technology, Distributed Denial of Services (DDoS) attacks have been growing dramatically in frequency, sophistication and impact. Traditional security measures to defend against DDoS attacks focused on stochastic analysis of network traffic and on exploiting its entropy to identify anomalous network behavior. Recently, the trend to mitigate the impact of DDoS attacks is to incorporate “intelligence” into the defense architecture, leveraging artificial intelligence and machine learning techniques to detect malicious traffic. In this keynote, I will explore the new threat landscape, including the complexity of the security infrastructure and how disruptive technology is enabling not only the development of intelligent security strategies but also the proliferation of “smarter”, more devastating multi-vector attacks. In this cat-and-mouse battle between defenders and attackers, I will discuss how artificial intelligence and machine learning based strategies are impacting cyber-security, including the opportunities they present and the challenges they face, as the sophistication of the attacks continue to increase. The keynote will conclude with the discussion of possible areas of research to enable intelligent security frameworks in ubiquitous and pervasive environments.

Short Bio. Dr. Znati is a Professor and Chair of the Department of Computer Science, School of Computing and Information at the University of Pittsburgh. He also served as the Director of the Computer and Network Systems Division and led the Information Technology Research Initiative at NSF. Dr. Znati’s main research interests are in building secure and reliable networked systems, focusing on intelligent infrastructure for DDoS mitigation, and on designing fault-tolerant computational models for energy-aware resiliency in extreme-scale systems. Dr. Znati has served as the General Chair of several main conferences, including GlobeCom 2010, INFOCOM 2005, and SECON 2004.  He served or currently serves as a member of Editorial Boards of a number of networking, distributed system and security journals and transactions. He also co-chaired the Network and Information Research and Development – Large Scale Networking and served in the Internet2 Research Advisory Council, for Developing Strategies for Excellence in Internet2 in Support of Research.

11:15 – 12:45
S2: ML for Cellular and Wireless Networks Planning

Chair: Ben Liang (University of Toronto, Canada)

[Slides] Node Centrality Metrics for Hotspots Analysis in Telecom Big Data
Emil Mededovic (RWTH Aachen University, Germany); Vaggelis G. Douros (Institute for Networked Systems, RWTH Aachen University, Germany); Petri Mähönen (RWTH Aachen University, Germany)

[Slides] Self-Organizing Cellular Radio Access Network with Deep Learning  
Wuyang Zhang (Rutgers University, USA); Russell Ford (Samsung Research America, USA); Joonyoung Cho (Samsung Electronics Co., Ltd., Korea); Jianzhong Zhang (Samsung, USA); Yanyong Zhang and Dipankar Raychaudhuri (Rutgers University, USA)

[Slides] α-OMC: Cost-Aware Deep Learning for Mobile Network Resource Orchestration  
Dario Bega (IMDEA Networks, Spain); Marco Gramaglia (Universidad Carlos III de Madrid, Spain); Marco Fiore (National Research Council of Italy, Italy); Albert Banchs (Universidad Carlos III de Madrid, Spain); Xavier Costa-Perez (NEC Laboratories Europe, Germany)

[Slides] Delay-optimal traffic engineering through multi-agent reinforcement learning
Pinyarash Pinyoanuntapong and Minwoo Lee (UNC Charlotte, USA); Pu Wang (University of North Carolina at Charlotte, USA)

12:45 – 14:00
Lunch Break

14:00 – 15:30
S3: ML for Applications

Chair: Imen Grida Ben Yahia, Orange Labs, France

Exploring Feature Relevance for Real-time Stalling Prediction of Encrypted Video Streaming Traffic  
Michael Seufert (AIT Austrian Institute of Technology GmbH, Austria); Pedro Casas (Austrian Institute of Technology (AIT), Austria); Nikolas Wehner (AIT Austrian Institute of Technology GmbH, Austria); Gang Li and Li Kuang (Huawei, P.R. China)

[Slides] An LSTM-based Approach for Overall Quality Prediction in HTTP Adaptive Streaming  
Huyen Tran and Duc V. Nguyen (The University of Aizu, Japan); Duong D. Nguyen (Hanoi University of Science and Technology, Vietnam); Pham Nam (Hanoi University of Science and Technology & School of Electronics and Telecoms, Vietnam); Truong Cong Thang (The University of Aizu, Japan)

[Slides] FlowPic : Encrypted Internet Traffic Classification is as Easy as Image Recognition  
Tal Shapira (Tel Aviv University, Israel) ;  Yuval Shavitt (Tel Aviv University, Israel)

[Slides] NFV Closed-loop Automation Experiments using Deep Reinforcement Learning  
Zhu Zhou (Intel, USA) ; Tong Zhang and Atul Kwatra (Intel Corporation, USA)

15:30 – 16:00
Coffee Break

16:00 – 16:45
Keynote Session: Dr. Angelo Corsaro (ADLINK Technology Inc.)
[Slides] The Making of Edge Intelligence

Abstract. In this talk we will motivate through real-world use cases the need for bringing intelligence closer to the source of data — the edge — and identify the challenges that need to be addressed in order to make this possible. The speaker will report on his research as well as the emerging trends and innovations in hardware and software geared toward making edge intelligence a reality.

Short Bio. Angelo Corsaro, Ph.D.,  is Chief Technology Officer (CTO) at ADLINK Technology Inc.  As CTO he leads the Advanced Technology Office and looks after corporate technology  strategy and innovation. Angelo is a world top expert in edge/fog computing and a well know researcher in the area of high performance and large scale distributed systems.  Angelo has over 100 publications on referred journal, conferences, workshops, and magazines. His research interests are on Fog/Edge Computing, Industrial and Consumer Internet of Things, Innovation and Innovation Management, Product Strategy, Open Source,  High Performance Computing, Large Scale Mission/Business Critical Distributed Systems,  Real-Time Systems, Software Patterns, Functional Programming Languages.

16:45 – 18:15
S4: ML at the Edge and in the Cloud

Chair: Laura Galluccio, University of Catania, Italy

[Slides] DeepViNE: Virtual Network Embedding with Deep Reinforcement Learning  
Mahdi Dolati and Seyedeh Bahereh Hassanpour (University of Tehran, Iran); Majid Ghaderi (University of Calgary, Canada); Ahmad Khonsari (University of Tehran, Iran)

[Slides] Multi-domain non-cooperative VNF-FG embedding: A deep reinforcement learning approach  
Tran Anh Quang Pham (INRIA, France); Abbas Bradai (XLIM Institute, University of Poitiers, France); Kamal Deep Singh (Telecom Saint Etienne / University Jean Monnet, France); Yassine Hadjadj-Aoul (University of Rennes 1, France)

[Slides] Task Dispatch through Online Training for Profit Maximization at the Cloud  
Sowndarya Sundar and Ben Liang (University of Toronto, Canada)

[Slides] DeePar: A Hybrid Device-Edge-Cloud Execution Framework for Mobile Deep Learning Applications  
Yutao Huang (Simon Fraser University, Canada); Feng Wang (University of Mississippi, USA); Fangxin Wang and Jiangchuan Liu (Simon Fraser University, Canada)

18:15 – 18:30
[Slides] Workshop Wrap up and Closing Remarks

Chairs: Laura Galluccio, Imen Grida Ben Yahia, Giovanni Schembra