About the project
Objective
This project aims to develop a deep learning-based methodology to enhance the ability to model complicated dynamics for sequential data. With a special focus on the recent progress of transformer-based models, which have shown great potential in modelling very long sequences, we are inspired to integrate them with other state-of-the-art techniques, such as learning dynamic structures and self-supervised learning. By exploring such directions, we expect our results to be applicable to the sequence modelling research and provide good insights for other fundamental deep learning research areas.
Background
Sequence modelling is the fundamental problem of other time series related tasks, including future forecasting. Since being proposed in 2018, transformers have become the de facto choice for most sequence modelling tasks due to their superior performance over traditional RNN-based approaches. However, it appears that transformers usually need a significant amount of training data to achieve their full potential, making them an expensive and impractical option for many real-world scenarios. Thus, it becomes increasingly imperative to develop methods to effectively train transformers with limited labelled data, which is quite common for sequence modelling.
About the Digital Futures Postdoc Fellow
Hao Hu is a postdoc researcher at KTH RPL working with Hossein Azizpour. Before joining KTH, he worked as a research scientist in FX Palo Alto Laboratory (FXPAL), California, United States. Hao got his PhD in Computer Science from the University of Central Florida (UCF) in 2019. His research interests include various topics in machine learning and computer vision, with a special focus on temporal modelling and deep learning.
Main supervisor
Hossein Azizpour, Assistant Professor, Robotics, Perception and Learning, KTH.
Co-supervisor
Arne Elofsson, Professor in Bioinformatics, Stockholm University.
About the project
Objective
In the Deep Wetlands project, we are developing a machine learning platform to monitor water extent changes in wetlands by integrating multiple data sources from satellite images, altimetry radars, and other space sensors. Wetlands are vital ecosystems for the functioning of the Earth system and necessary to achieve sustainable development. Water availability determines whether wetlands can thrive and deliver services to humans. However, thick vegetation mostly covers water changes, impairing their remote detection from space. Wetlands are disappearing rapidly; approximately 70% have been lost in the last 120 years.
Despite the danger that wetlands are currently facing, there is no global high-resolution assessment of wetland changes. This limits the in-depth and temporal analysis of wetlands from space. Changes in wetlands are unnoticed as most space-based technologies cannot fully account for water below vegetation and are limited to large water bodies. Our grand challenge is quantifying the wetland surface area changes on existing wetlands.
About the Digital Futures Postdoc Fellow
Francisco J. Peña is a postdoctoral researcher working in the field of artificial intelligence and remote sensing. He works jointly at the Software and Computer Systems (SCS) division of KTH Royal Institute of Technology and the Department of Physical Geography of Stockholm University in Sweden. Francisco is also a member of the Distributed Computing at KTH (DC@KTH). Before that, he was a postdoctoral researcher at University College Dublin (2018-2020). He obtained his PhD from University College Cork in June 2019.
His research interests include:
- Natural Language Processing
- Computer Vision
- Deep Learning
- Recommender Systems
- Remote Sensing
Main supervisor
Fernando Jaramillo, Assistant Professor, Stockholm University.
Co-supervisor
Amir Payberah, Assistant Professor, Division of Software and Computer Science, KTH.
Watch the recorded presentation at Digitalize in Stockholm 2022 event.
About the project
Objective
It indeed consists of two sub-projects. Firstly, as the most promising technology in achieving 10 Gbs peak data rates, millimetre-wave (mmWave) communications have received remarkable attention from academia and industry. Thus, in the project of intelligent wireless communications, we aim to develop several machine learning-based beam tracking algorithms for mobile mmWave communications, which can work efficiently without relying on a priori knowledge of channel dynamics. While in the project of high-accuracy positioning systems, we aim to leverage mmWave signals and other techniques, such as intelligent reflecting surfaces, to achieve centimetre-level localization accuracy.
Background
Driven by the ever-increasing mobile data traffic, 5G-and-beyond (B5G) networks are envisioned as a key enabler to support a variety of novel use cases, such as autonomous cars, industrial automation, multisensory extended reality (XR), e-health, etc. Considering the emergence of these use cases and the more and more complicated network structure, artificial intelligence is expected to be essential to assist in making the B5G version conceivable.
With regard to high-accuracy localization, it will play a critical role in almost all use cases of the B5G networks. Specifically, depending on the usage scenarios, the requirement for localization accuracy ranges from 1 cm to 10 cm for smart factory applications. However, most current localization services can, at best, achieve meter-level localization accuracy and, therefore, cannot meet the centimetre-level localization accuracy requirements of the emerging use cases in the B5G era, which emphasizes the need for more advanced localization techniques.
About the Digital Futures Postdoc Fellow
Deyou Zhang is a Digital Futures Postdoc at the School of Electrical Engineering and Computer Science of KTH, supervised by Dr Ming Xiao, Prof. Lihui Wang, and Dr Zhibo Pang. Before joining KTH, he obtained his PhD at the University of Sydney, Australia. His research interests include millimetre-wave communications, intelligent reflecting surfaces, and wireless federated learning.
Main supervisor
Ming Xiao, Associate Professor, Division of ISE, EECS School, KTH.
Co-supervisor
Zhibo Pang, Senior Principal Scientist, Department of Automation Technology, ABB Corporate Research Sweden and Adjunct Professor, Department of Intelligent Systems, EECS, KTH.
Lihui Wang, Professor and Chair of Sustainable Manufacturing, KTH.
Watch the recorded presentation at Digitalize in Stockholm 2022 event.
About the project
Objective
Dragons seeks to support inclusive, safe, resilient, and sustainable urban development by merging computational methods to gain insights into urban SET systems with governance approaches to act on these insights. Hence, its guiding questions are: (Q1) How can we combine available data to monitor inclusiveness, safety, resilience, and sustainability in urban SET systems? (Q2) How can we understand the evolution and interaction of structures and processes related to these goals? (Q3) How can urban governance incorporate the findings from Q1 and Q2?
Background
With over half of the world’s population living in cities and most population growth projected to occur in urban areas, making cities inclusive, safe, resilient, and sustainable is a key policy concern expressed in the eleventh UN Sustainable Development Goal (SDG 11). To ensure these properties in urban development, policymakers need to navigate the complex interplay between social (including economic and political), ecological, and technological factors shaping and shaped by human urban activity. This requires an interdisciplinary approach to urban areas as Social-Ecological-Technological Systems (SET systems).
About the Digital Futures Postdoc Fellow
Corinna Coupette studied law at Bucerius Law School and Stanford Law School, completing their First State Exam in Hamburg in 2015. They obtained a PhD in law (Dr. iur.) from Bucerius Law School and a BSc in computer science from LMU Munich, both in 2018, as well as an MSc in computer science in 2020 and a PhD in computer science (Dr. rer. nat.) in 2023, both from Saarland University. Their legal dissertation was awarded the Bucerius Dissertation Award in 2018, and the Otto Hahn Medal of the Max Planck Society in 2020, and their interdisciplinary research profile was recognized by the Caroline von Humboldt Prize for outstanding female junior scientists in 2022.
The overarching goal of Corinna’s research is to understand how we can combine code, data, and law to better model, measure, and manage complex systems. To this end, they explore novel ways of connecting computer science and law, such as using algorithms to collect and analyze legal data as networks or formalizing and implementing legal and mathematical desiderata for responsible data-centric machine learning with graphs.
Main supervisor
Aristides Gionis, WASP Professor of Computer Science, EECS, KTH.
Co-supervisor
Örjan Bodin, Professor, Stockholm Resilience Center and Stockholm University.
About the project
Objective
With over 1 billion people over 60 worldwide, creating technology that supports the aged to live independently for longer by assisting them in everyday tasks became essential. While companion robots are aimed toward this need, current technology falls short in maintaining engagement over long-term interactions. Among the reasons is the inability to learn from users and adapt, known as lifelong learning, especially in open-domain dialogue that is not limited to any topic.
This project aims to develop a long‐term memory model for open‐domain dialogue such that a robot can learn and recall a person’s attributes, preferences, and shared history to provide personalized assistance in a variety of tasks, such as performing preferred activities, adaptive collaboration in chores, and providing reminders based on their schedule and needs.
About the Digital Futures Postdoc Fellow
Bahar Irfan is a Postdoctoral researcher at KTH Digital Futures. Her research focuses on creating personal robots that can continually learn and adapt to assist everyday life. Previously, she was a Research and Development Associate at Evinoks Service Equipment Industry and Commerce Inc., developing customizable software for industrial robots and smart buffets. Before that, she worked as an R&D Lab Associate at Disney Research Los Angeles on emotional language adaptation in multiparty interactions.
She has a diverse background in robotics, from personalization in long-term human-robot interaction during her PhD at the University of Plymouth and SoftBank Robotics Europe as a Marie Skłodowska-Curie Actions fellow to user-centred task planning for household robotics during her MSc in computer engineering, and building robots for BSc in mechanical engineering at Boğaziçi University.
Main supervisor
Gabriel Skantze, Professor in Speech Communication and Technology, KTH.
Co-supervisor
Sanna Kouppamäki, Assistant Professor, Division of Technology in Health Care, KTH.
Watch the recorded presentation at the Digitalize in Stockholm 2023 event.
About the project
Objective
Arzu’s future research focus is co-designing and developing gamified robot-enhanced interventions for children and adolescents with neurodevelopmental disorders (NDDs). The research will be based on an iterative design approach to develop interventions for and with the target user groups tailored to the individual to enhance the functional recovery of sensorimotor, social, or cognitive functions in children with NDDs. She aims to investigate what the best roles for robots are in different inclusive practices, including neurodivergent and neurotypical groups where they play and learn together, how to involve children in the design process of robot-mediated activities, and how to design inclusive gamified practices to enhance social interaction between the neurotypical and neurodivergent children as well as their families.
Background
Neurodevelopmental disorders (NDDs) result in different degrees of emotional, physical, social, academic and economic consequences for individuals and in turn, families and society [1, 2]. Upon diagnosis, families report significant delays in treatment initiation and unsatisfactory levels of treatment monitoring. [2, 3].There is a need to establish effective easy-to-access strategies for assessing, treating and monitoring NDD.
Rapid progress in the area of robotics offers excellent chances for innovation in the treatment of children with NDDs, thanks to robots allowing the execution of specific and repetitive tasks which can be tailored according to the particular needs of the individuals. Robots thus offer the opportunity to deliver automated and independent interventions that enable therapy to be delivered over a distance in inclusive and collaborative education environments [4,5] and personalise treatment procedures [6,7]. Combined with gamification, which improves the learning rate and ensures effective improvement in the pedagogical, social and behavioural sense [8,9], robot-enabled therapy becomes a highly promising avenue for research.
About the Digital Futures Postdoc Fellow
Arzu Guneysu Ozgur is a Postdoctoral researcher at Digital Futures. Arzu got a PhD in Robotics on “Designing Gamified Activities with Haptic-Enabled Tangible Robots for Therapy and Assistance” from EPFL in 2021. Her research interests include various topics in Human-Robot Interaction, Adaptive Robot-Enhanced Therapy, Iterative Design, Participatory Design, Neurodevelopmental Disorders, Gamified Therapeutic Technologies, Healthy Aging, Intergenerational Practices for Elderly and Children, and Special Education.
Main supervisor
Iolanda Leite, Associate professor, Department of Robotics, Perception and Learning, KTH.
Co-supervisor
Ali Reza Majlesi, Associate Professor, Department of Education, Stockholm University.
Watch the recorded presentation at Digitalize in Stockholm 2022 event.
About the project
Objective
The project aims to fill urban population data gaps in developing countries by harnessing the power of Earth Observation (EO) data and AI. An innovative framework will fuse high-resolution satellite information with ancillary sources, such as Volunteer Geographic Information data and machine learning. The long-term goal of POPAI is to understand better the synergy and potential of AI and EO towards scalable population mapping, help address the United Nations Sustainable Development Goals, support evidence-based policymaking and foster a better future for the cities of tomorrow.
Background
Accurate urban population distribution information is necessary prerequisites for a wide range of applications related to urban sustainability. The quality and quantity of population data in numerous countries are often inadequate due to the absence of detailed censuses or large temporal gaps between them. The disparaging effects of this lack of information are most evident in Sub-Saharan Africa (SSA) and the Global South. As estimated by the UN, more than 60% of the African population will reside in cities by 2050, which further emphasizes the need for accurate population information.
About the Digital Futures Postdoc Fellow
Stefanos Georganos is a research fellow at the Division of Geoinformatics, Royal Institute of Technology. He does research in quantitative human geography, remote sensing, spatial epidemiology and machine learning. He is particularly interested in the use of geo-information to help address the UN Sustainable Development Goals, with a geographical interest in sub-Saharan African cities. His latest research unravels the potential of Artificial Intelligence and Earth Observation to detect, measure and characterize socio-economic inequalities in deprived urban areas in support of the most vulnerable populations.
Main supervisor
Yifang Ban, Professor and Head of Division Geoinformatics at KTH.
Co-supervisor
Anders Wästfeldt, Professor at Stockholm University.
Watch the recorded presentation at Digitalize in Stockholm 2022 event.