Loading...

Welcome to DeepSpatial 2020

1st ACM SIGKDD Workshop on Deep Learning for Spatiotemporal Data, Applications, and Systems
Augest 24, 2020
KDD-organized Virtual Conference

Zoom Link to Attend the Online Workshop in 1:00pm-4:50pm (PDT timezone) on Aug 24, 2020


Aims and Scope

Deep learning has exhibited outstanding performance in handling data in space and time in specific domains such as image, audio, and video. Meanwhile, the novel applications such as location-based social media, data-driven climate and Earth science, and ride-sharing have enabled and accumulated large scale of spatiotemporal data over the years, which in turn has led to unprecedented opportunities and prerequisites for the discovery of macro- and micro- spatiotemporal phenomena accurately and precisely. Further developments of spatial/spatiotemporal computing and deep learning call for the synergistic techniques and the collaborations between different communities, as evidenced by the recent momentum in both domains. On one hand, fast-increasing large-scale and complex-structured spatiotemporal data requires the investigation and extension toward more scalable and powerful models than traditional ones in domains such as computational geography and spatial statistics. One the other hand, deep learning techniques are evolving beyond regular grid-based (e.g., images), tree-based (e.g., texts), and sequence-based (e.g., audio) data to more generic or irregular data in space and time (e.g., in transportation, geomorphology, and protein folding), which calls for the expertise in the domains such as spatial statistics, geodesy, geometry, graphics, and geography. Consequently, the aforementioned complementary strengths and challenges between spatiotemporal data computing and deep learning in recent years suggest urgent needs to bring together the experts in these two domains in prestigious venues, which is still missing until now.

This workshop will provide a premium platform for researchers from both academia and industry to exchange ideas on opportunities, challenges, and cutting-edge techniques of deep learning for spatiotemporal data, applications, and systems. Papers will be accepted under the topics including, but not limited to, the following three broad categories:

Novel Deep Learning Techniques for Spatial and Spatio-Temporal Data:
Spatial representation learning and deep neural networks for spatio-temporal data and geometric data
Interpretable deep learning for spatial-temporal data
Deep generative models for spatio-temporal data
Deep reinforcement learning for spatio-temporal decision making problems

Novel Applications of Deep Learning Techniques to Spatio-temporal Computing Problems. :
Geo-imagery and point cloud analysis (for remote sensing, Earth science, etc.)
Deep learning for mobility and traffic data analytics
Location-based social network data analytics, spatial event prediction and forecasting
Learning for biological data with spatial structures (bio-molecule, brain networks, etc.)

Novel Deep Learning Systems for Spatio-temporal Applications:
Real-time decision-making systems for traffic management, crime prediction, accident risk analysis, etc.
GIS systems using deep learning (e.g., mapping, routing, or Smart city)
Mobile computing systems using deep learning
Interpretable deep learning systems for spatio-temporal temporal data

In addition, we encourage submissions of spatiotemporal deep learning methods that address problems related to the COVID-19 pandemic.

Workshop Co-Chairs

Image
Liang Zhao

Assistant Professor
George Mason University

Image
Xun Zhou

Associate Professor
University of Iowa

Image
Feng Chen

Associate Professor
University of Texas, Dallas

Image
Jieping Ye

Vice President, Didi Chuxing
Professor, University of Michigan

Image
Shashi Shekhar

McKnight Distinguished University Professor
University of Minnesota

Program Committee
Arnold Boedihardjo, DigitalGlobe
Manzhu Yu, PSU
Wei Wang, Microsoft Research
Chao Zhang, Georgia Tech
Yanjie Fu, UCF
Xuchao Zhang, NEC Lab North America
Ray Dos Santos, Army Corps of Engineers
Yanhua Li, WPI
Lingfei Wu, IBM Research
Yinghui Wu, WSU
Zhe Jiang, University of Alabama
Kangkook Jee, UT Dallas
Jing Dai, Google

Webmaster
Yuanqi Du, George Mason University

Paper Submission

Important Dates: (all due Midnight Pacific Time).
Paper Submission: June 1, 2020 June 15, 2020 (Firm deadline)
Notification of Acceptance: June 15, 2020 July 5, 2020
Camera-ready Papers: June 20, 2020 July 10, 2020
Workshop Date: August 24, 2020

The workshop welcomes the two types of submissions

  • Full research papers – up to 9 pages (8 pages at most for the main body and the last page can only hold references)

  • Vision papers and short system papers - up to 5 pages (4 pages at most for the main body and the last page can only hold references)

All manuscripts should be submitted in a single PDF file including all content, figures, tables, and references, following the format of KDD conference papers. Paper submissions need to include author information (review not double blinded).

Papers should be submitted at: https://easychair.org/my/conference?conf=deepspatial2020
Concurrent submissions to other journals and conferences are acceptable. Accepted papers will be presented as posters during the workshop and posted on the website. Besides, a small number of accepted papers may be selected to be presented as contributed talks.

Panel: The Future of AI for Spatiotemporal Data Science

Image
Wei Ding

National Science Foundation

Image
Wei-Shinn Ku

National Science Foundation

Image
Amarda Shehu

National Science Foundation

Keynotes Speakers

Image
Jing Dai

Google

Program

(All times are in PDT time zone)

1:00pm-1:05pm Opening and Welcome
1:05pm-1:35pm Keynote 1 by Dr. Zhenhui (Jessie) Li, PSU
  • Title: Reinforcement Learning for Traffic Signal Control
  • 1:35pm-2:05pm Keynote 2 by Dr. Jing Dai, Google
  • Title: Building Maps using ML - Applications and Challenges
  • 2:05pm-2:29pm Paper Splotlight Presentation (All the presentation videos are available here):

  • Modeling Spatiotemporal Geographic-Semantic Dynamics for Urban Hotspots Prediction (3 mins)

  • Temporal Interpolation of Geostationary Satellite Imagery with Task Specific Optical Flow (3 mins)

  • Fusion Recurrent Neural Network (3 mins)

  • DeepSampling: a Bounded-error System for Approximate Geospatial Query Processing (3 mins)

  • Generate Street Map Images from Satellite Images and Crowd-sourced Geographic data using GAN (3 mins)

  • Transformer Hawkes Process (3 mins)

  • Towards Spatial Variability Aware Deep Neural Networks (SVANN): A Summary of Results (3 mins)

  • Deep Multi-Sensor Domain Adaptation on Active and Passive Satellite Remote Sensing Data (3 mins)

  • 2:29pm-2:45pm Poster Breakout Room
    2:45pm-3:15pm Keynote 3 by Dr. Ranga Raju Vatsavai, NCSU
  • Title: Recent Advances in Machine Learning for Remote Sensing
  • 3:15pm-4:45pm Panel “The future of AI for spatiotemporal data science”

    Panelists: Dr. Wei Ding (NSF), Dr. Wei-Shinn Ku (NSF), Dr. Shashi Shekhar (UMN), and Dr. Amarda Shehu (NSF).

  • Panel presentations (35 mins)
  • Panel discussions (30 mins)
  • Panel Q&A. (25 mins)

  • 4:45pm-4:50pm Close

    Keynotes

    Keynote 1: Reinforcement Learning for Traffic Signal Control

    Abstract:

    This talk presents how to utilize mobility data and advanced learning methods for traffic signal control. First, I will examine the existing traffic signal control system and discuss why today we have the opportunity for a potential breakthrough in traffic signal control. Second, the talk presents our recent research results in traffic signal control via reinforcement learning which are published in recent KDD, CIKM, and AAAI conferences. Finally, I would like to discuss the open challenges in this research topic and its implications for smart city applications.

    Speaker Bio:

    Dr. Zhenhui (Jessie) Li is a tenured associate professor of Information Sciences and Technology at the Pennsylvania State University. She is Haile family early career endowed professor. Her research has been focused on mining spatial-temporal data with applications in transportation, ecology, environment, social science, and urban computing. She is a passionate interdisciplinary researcher and has been actively collaborating with cross-domain researchers. To learn more, please visit her homepage: https://faculty.ist.psu.edu/jessieli



    Keynote 2: Building Maps using ML - Applications and Challenges

    Abstract:

    The Maps industry has been the frontline of applied machine learning techniques, especially for generating maps data from imagery. This talk is going to provide an overview to the Maps features and data at Google that rely on deep learning techniques, to introduce the recent efforts on improving Maps for better user experiences for navigation and assistant driving, and to discuss the challenges that the engineering teams have faced when applying deep learning to build Maps.

    Speaker Bio:

    Dr. Jing (David) Dai is an experienced researcher and senior engineer on spatial and spatiotemporal data management and data mining. He is currently an Engineering Lead at Google where he joined in 2011, leading the efforts to generate Maps data for Auto. He received his Ph.D. degree in Computer Science from Virginia Tech in 2009. Dr. Dai was a Research Scientist at IBM T.J. Watson Research Center from 2009 to 2011, where he worked on building the first Smart City in the US. He has served on the committee for several international conferences and workshops, as well as the NSF panels.



    Keynote 3: Recent Advances in Machine Learning for Remote Sensing

    Abstract:

    Global earth observations with constellation of more than 100 operational satellites are providing unprecedented spatiotemporal data coverage, which can be exploited to continuously monitor key resources. Earth is a dynamical system continually changing due to both natural and human induced factors. Recent decade has witnessed major changes on the Earth, for example, deforestation, varying cropping and human settlement patterns, and crippling damages due to disasters. Monitoring this dynamic phenomenon is critical for human wellbeing. In this talk we will explore recent advances in AI and machine learning for monitoring natural and as well as man-made structures at regional and global scales.

    Speaker Bio:

    Dr. Raju Vatsavai is a Chancellor’s Faculty Excellence Program Cluster Associate Professor in the Computer Science department at the North Carolina State University. He works at the intersection of spatial and temporal big data management, analytics, and high-performance computing with applications in the national security, geospatial intelligence, natural resources, climate change, location-based services, and human terrain mapping. As the Associate Director of the Center for Geospatial Analytics (CGA), Raju plays a leadership role in the center’s strategic vision for spatial computing research. He has published more than 100 peer-reviewed articles in conferences and journals and edited two books on “Knowledge Discovery from Sensor Data.” He served on program committees of leading international conference including ACM KDD, ACM GIS, ECML/PKDD, SDM, CIKM, IEEE BigData, and co-chaired several workshops including ICDM/SSTDM, ICDM/KDCloud, ACM SIGSPATIAL BigSpatial, Supercomputing/BDAC, KDD/LDMTA, KDD/Sensor-KDD, and SDM/ACS. He holds MS and PhD degrees in computer science from the University of Minnesota. At present, he is on leave from the NCSU and working as Distinguished Research Fellow at the Lirio.

    Contact Information

    Liang Zhao, lzhao9@gmu.edu, Tel: (703) 993 5910
    Xun Zhou, xun-zhou@uiowa.edu, Tel: (319) 384-3335
    Feng Chen, feng.chen@utdallas.edu, Tel: (571) 265-7769