Session Chair: Derya Malak (EURECOM, France)
08:55am-09:00am (Opening)
Chair
9:00am-10:00am (Keynote Speech)
Multi-User Separable Function Computation, and the Deep Connections with Coding Theory, Topology and Compressed Sensing
Prof. Petros Elia (Eurecom, France)
Abstract: In this work, we investigate the problem of multi-user linearly separable computation, where various servers help compute the desired functions of various users while each desired function can be written as a linear combination of several (generally non-linear) sub-functions. Each server computes some of the subtasks, and communicates a linear combination of its computed outputs (files) to some of the users, where each user then linearly combines its received data in order to recover its desired function. We explore the classical problem of the tradeoff between computation and communication cost, and we establish novel relationships with coding theory and compressed sensing. Joint work with Ali Khalesi and Sajad Daei.
Bio: Petros Elia received the B.Sc. degree from the Illinois Institute of Technology, and the M.Sc. and Ph.D. degrees in electrical engineering from the University of Southern California (USC), Los Angeles, in 2001 and 2006 respectively. He is now a professor with the Department of Communication Systems at EURECOM in Sophia Antipolis, France. His latest research deals with the intersection of coded caching and feedback-aided communications in multiuser settings. He has also worked in the area of complexity-constrained communications, MIMO, queueing theory and cross-layer design, coding theory, information theoretic limits in cooperative communications, and surveillance networks. He is a Fulbright scholar, the co-recipient of the NEWCOM++ distinguished achievement award 2008-2011 for a sequence of publications on the topic of complexity in wireless communications, the recipient of the ERC Consolidator Grant 2017-2022 on cache-aided wireless communications, and the recipient of the ERC-PoC 2022-2024.
10:00am-10:30am (Invited Talk)
Privacy Issue in Coded Caching
Prof. Kai Wan (Huazhong University of Science and Technology, China)
Abstract: Caching is an efficient way to reduce network traffic congestion during peak hours by storing some content at the user’s local cache memory without knowledge of later demands. The seminal coded caching strategy proposed by Maddah-Ali and Niesen (MAN) was proved to be order optimal within a factor of 2. However, in the MAN coded caching scheme, each user can obtain the information about the demands of other users, i.e., the MAN coded caching scheme is inherently prone to tampering and spying the activity/demands of other users. This talk presents the recent results (based on the strategies of introducing virtual users, pre-storing some private keys into users’ caches, etc.) to preserve the privacy of the users’ demands against the other users, with the objective to minimize the transmission load and reduce the subpacketization level. Furthermore, if the information of the users’ caches is leaked in one-round transmission, the demand privacy is vulnerable in the next rounds. On this motivation, we then formulae the model with privacy constraint on both users’ demands and caches. For this model, a new private caching scheme which leverages existing Private Information Retrieval (PIR) protocols is introduced.
Bio: Dr. Kai Wan received the B.E. degree in Optoelectronics from Huazhong University of Science and Technology, China, in 2012, the M.Sc. and Ph.D. degrees in Communications from Université Paris-Saclay, France, in 2014 and 2018. He was a post-doctoral researcher with the Communications and Information Theory Chair (CommIT) at Technische Universität Berlin, Berlin, Germany. He is currently a Professor in the School of Electronic Information and Communications, at the Huazhong University of Science and Technology, China. His research interests include information theory, coding techniques, and their applications on coded caching, index coding, distributed storage, distributed computing, wireless communications, privacy and security. He has served as an Associate Editor of IEEE Communications Letters from Aug. 2021.
10:30am-11:00am
Coffee Break
11:00am-11:30am (Invited Talk)
Computing-aided Adaptive Wireless Streaming for Immersive Video
Prof. Ying Cui (Hong Kong University of Science and Technology, P.R. China)
Abstract: Immersive video, such as 360 video and multi-view video, provides immersive experiences and is rapidly gaining popularity. However, wireless streaming of a popular immersive video to multiple users simultaneously, which can be viewed as general multicast, is exceptionally challenging due to the large video size and heterogeneous user channel conditions. This talk presents a transcoding-based multicast mechanism for 360° video and a view synthesis-based multicast mechanism for multi-view video. Both mechanisms can effectively create multicast opportunities to significantly reduce heavy traffic load by utilizing video characteristics and computing resources. Furthermore, this talk illustrates cross-layer optimization methods for efficient wireless streaming of 360 video and multi-view video. These methods achieve joint optimization of the video encoding rate, view selection, and computing and communications resource allocation. Finally, this talk numerically demonstrates the notable gains of computing-aided adaptive wireless streaming techniques.
Bio: Ying Cui received her B.Eng degree in Electronic and Information Engineering from Xi’an Jiao Tong University, China, in 2007, and her Ph.D. degree from the Hong Kong University of Science and Technology, Hong Kong SAR, China, in 2012. She held visiting positions at Yale University, US, in 2011 and Macquarie University, Australia, in 2012. From June 2012 to December 2014, she was a postdoctoral research associate, first at Northeastern University, US, and then at Massachusetts Institute of Technology, US. From January 2015 to July 2022, she was an associate professor at Shanghai Jiao Tong University, China. Since August 2022, she has been an associate professor with the IoT Thrust at The Hong Kong University of Science and Technology (Guangzhou), China, and an affiliate associate professor with the Department of ECE at The Hong Kong University of Science and Technology, Hong Kong SAR, China. Her current research interests include optimization, learning, IoT communications, mobile edge caching and computing, and multimedia transmission. She was selected to the Thousand Talents Plan for Young Professionals of China in 2013. She received Best Paper Awards from IEEE ICC 2015 and IEEE GLOBECOM 2021. She serves as an Editor for the IEEE Transactions on Wireless Communications.
11:30am-11:50am (Invited Paper)
A Lower Bound on Load of Coded Caching Schemes for Finite Subpacketizations
Minquan Cheng (Guangxi Normal University, China); Youlong Wu (ShanghaiTech University, China)
11:50am-12:10pm (Invited Paper)
On the Regret of Online Edge Service Hosting
R Sri Prakash (IIT Bombay, India); Nikhil Karamchandani (Indian Institute of Technology Bombay, India); Sharayu Moharir (Indian Institute of Technology Bombay, India)
12:10pm-12:30pm (Invited Paper)
Cache-Aided Multi-Access Multi-User Private Information Retrieval
Kanishak Vaidya (Indian Institute of Science Bangalore, India); B. Sundar Rajan (Indian Institute of Science, India)
12:30pm-02:00pm
(Lunch Break)
Session Chair: Minseok Choi (Kyung Hee University)
2:00pm-3:00pm (Keynote Speech)
Adaptive Control for Federated Learning at the Edge
Dr. Shiqiang Wang (IBM T.J. Watson Research Center, USA)
Abstract: Federated learning (FL) is an emerging technique for model training from decentralized data. Compared to learning from data in a central storage, FL has benefits of privacy preservation and communication bandwidth reduction. A challenge in FL is that data and model characteristics can vary largely across different tasks, and an FL task with improper configuration could waste a lot of computation/communication resources and may cause the trained model to diverge from the optimal result. In this talk, I will present adaptive FL methods that learns near-optimal configurations (e.g., synchronization interval, compressed model size) over time during the FL process, to reach the best model accuracy with the smallest amount of training time. These adaptive FL algorithms are derived from convergence analysis, online learning, and related analytical techniques. The performance of these algorithms is evaluated both theoretically and empirically. Some open problems will be also outlined.
Bio: Shiqiang Wang is a Research Staff Member at IBM T. J. Watson Research Center, NY, USA. He received his Ph.D. from Imperial College London, United Kingdom, in 2015. His current research focuses on the intersection of distributed computing, machine learning, networking, and optimization, with a broad range of applications including data analytics, edge-based artificial intelligence (Edge AI), Internet of Things (IoT), and future wireless systems. He received the IEEE Communications Society (ComSoc) Leonard G. Abraham Prize in 2021, IEEE ComSoc Best Young Professional Award in Industry in 2021, IBM Outstanding Technical Achievement Awards (OTAA) in 2019, 2021, and 2022, multiple Invention Achievement Awards from IBM since 2016, Best Paper Finalist of the IEEE International Conference on Image Processing (ICIP) 2019, and Best Student Paper Award of the Network and Information Sciences International Technology Alliance (NIS-ITA) in 2015. For more details, please visit his homepage at: https://shiqiang.wang
3:00pm-3:30pm (Invited Talk)
Distributed General Function Computation
Prof. Derya Malak (Eurecom, France)
Abstract: Large-scale distributed computing systems, such as MapReduce, Spark, or distributed deep networks, are critical for parallelizing the execution of computational tasks. Nevertheless, a struggle between computation and communication complexity lies at the heart of distributed computing. There has been recently substantial effort to address this problem for a class of functions, such as distributed matrix multiplication, distributed gradient coding, linearly separable functions, etc. The optimal cost has been achieved under some constraints, based mainly on ideas of linear separability of the tasks and linear space intersections. Motivated by the same challenge, we propose a novel distributed computing framework where a master seeks to compute an arbitrary function of distributed datasets in an asymptotically lossless manner. Our approach exploits notions of distributed source and functional compression using characteristic graphs each source builds. These graphs have been widely utilized by Shannon, Korner, and ¨Witsenhausen to derive the rate lower bounds for computation, and later by Alon-Orlitsky, Orlitsky-Roche, Doshi-Shah-Medard, and Feizi-Medard, to resolve some well-known distributed coding and communication problems, allowing for lowered communication complexity and even for a) correlated data, b) a broad class of functions, and c) well-known topologies. The novelty of our approach lies in accurately capturing the communication-computation cost tradeoff by melding the notions of characteristic graphs and distributed placement, to provide a natural generalization of distributed linear function computation, thus elevating distributed gradient coding and distributed linear transform to the realm of distributed computing of any function. This approach is well suited to one-shot or multi-shot computations, under uniform or skewed data distributions, and for general placement models. In toy scenarios, we demonstrate gains up to %70 over fully distributed solutions and an approximation ratio of 2 within the optimal centralized rate. This work is joint with Prof. Petros Elia, Dr. Berksan Serbetci, and Federico Brunero in the Communication Systems Department at EURECOM.
Bio: Derya Malak is a tenure track Assistant Professor (Maˆıtre de Conference) in the Communication Systems Department at Eurecom, France. Previously, she was an Assistant Professor in the Department of ECSE at Rensselaer Polytechnic Institute between 2019-2021, and a Postdoctoral Associate at MIT between 2017-2019. She received her Ph.D. in ECE at the University of Texas at Austin in 2017, B.S. in Electrical and Electronics Engineering (EEE) with a minor in Physics at Middle East Technical University, Ankara, Turkey, in 2010, and M.S. in EEE at Koc University, Istanbul, Turkey, in 2013. Dr. Malak has held visiting positions in INRIA and LINCS, Paris, France, and at Northeastern University, Boston, MA. She has held summer internships at Huawei Technologies, Plano, TX, and Bell Laboratories, Murray Hill, NJ. She was awarded the Graduate School fellowship by UT Austin between 2013-2017. She was selected to participate in the Rising Stars Workshop for women in EECS, MIT, Cambridge, MA, in 2018, and the 7th Heidelberg Laureate Forum, Germany, in 2019. Dr. Malak has expertise in information theory, communication theory, and networking areas. She has developed novel distributed computation solutions, and wireless caching algorithms by capturing the confluence of storage, communication, and computation aspects. Dr. Malak’s research has been funded by the Huawei Chair, NSF, the Rensselaer-IBM AI Research Collaboration, and the DARPA Dispersive Computing Programs.
3:30pm-4:00pm (Invited Talk)
Edge Caching and Computing for Video Services
Prof. Minseok Choi (Kyung Hee University, South Korea)
Abstract: Online video services account for a large portion of global wireless data traffic, and many users availing themselves of video streaming services make overlapping and repeated requests for content. This has given rise to the idea of a wireless caching network which reduces the network traffic and latency. Also, as the paradigm of edge computing where sufficient capabilities of proceeing computational tasks at the wireless edge has become popular, videos can be transcoded at the wireless edge. In this talk, challenges and recent studies on edge caching and computing for online video services in wireless networks will be presented. Particularly, adaptive video streaming has several characteristics, for example, a video can be encoded into multiple quality versions, a video stream consists of multiple segments, each segment can have different bitrates, and the video can be played with earlier chunks before receiving all chunks of the entire stream. We will see how edge caching and computing techniques can be improved for handling these features.
Bio: Minseok Choi is an Assistant Professor in Electronic Engineering with Kyung Hee University, Yongin, South Korea. He received the B.S., M.S., and Ph.D. degrees from the School of Electrical Engineering, Korea Advanced Institute of Science and Technology (KAIST), Daejeon, South Korea, in 2011, 2013, and 2018, respectively. He was an Assistant Professor in telecommunications engineering with Jeju National University, a Visiting Postdoctoral Researcher in electrical and computer engineering with the University of Southern California (USC), Los Angeles, CA, USA, and a Research Professor in electrical engineering with Korea University, Seoul, South Korea. He received the IEEE Communications Society (ComSoc) Multimedia Communications Technical Committee (MMTC) Best Paper Award, 2022. His research interests include wireless caching networks, federated learning, stochastic network optimization, wireless intelligent networks.
4:00pm-4:30pm
Coffee Break
4:30pm-4:50pm
Analysis of the LRU Cache StartUp Phase and Convergence Time and Error Bounds on Approximations by Fagin and Che
Gerhard Hasslinger (Deutsche Telekom, Germany); Konstantinos Ntougias (University of Cyprus, Cyprus); Frank Hasslinger (TUDa, Germany); Oliver Hohlfeld (Brandenburg University of Technology, Germany)
4:50pm-5:10pm
Cooperative Video Quality Adaptation for Delay-Sensitive Dynamic Streaming Using Adaptive Super-Resolution
Minseok Choi (Kyung Hee University, South Korea); Won Joon Yun (Korea University, South Korea); Joongheon Kim (Korea University, South Korea)
5:10pm-5:30pm
A New Design Framework for Heterogeneous Uncoded Storage Elastic Computing
Mingyue Ji (University of Utah, USA); Xiang Zhang (University of Utah, USA); Kai Wan (Huazhong University of Science and Technology, China)
05:30pm-05:35pm (Closing Remarks)
Chair