RISE Research Radar

Adversarial Malware Generation using AI and ML

Research on adversarial malware generation using GANs (MalGAN) and Reinforcement Learning techniques. Studies how AI can be used to generate malware v...

AI for material

Applies AI and machine learning techniques to accelerate materials research and discovery. Uses computational methods to predict material properties a...

AI for solar cells

Developing self-driving lab (SDL) to explore new optoelectronic materials synthesized by PVD using active learning with Gaussian processes. Uses magne...

Sanna Jarl, Anders Holst, Jens Sjölund +1 more

AI SAXS

Applies AI to analyze Small Angle X-ray Scattering (SAXS) data for materials characterization. SAXS is a powerful technique for studying nanostructure...

Benchmarking LLMs on SweSAT

New benchmark dataset from Swedish university entrance exam (Högskoleprovet) to assess LLMs' ability to solve problems in Swedish. Contains 567 multip...

Murathan Kurfali, Sheroog Zahra, Evangelia Gogoulou +3 more

BIOPOL: Biologically Derived Principles for Online Learning

Develops online learning algorithms inspired by biological neural principles. Part of Center for Applied Cybernetics (CAC) research on bio-inspired co...

Climate AI Nordics

Network of researchers dedicated to tackling climate change through advanced AI solutions. Founded October 2024 with 133 affiliates from ~90 organizat...

Olof Mogren, Aleksis Pirinen

ClimateEval: A Comprehensive Benchmark for NLP Tasks Related to Climate Change

Presents unified benchmark aggregating climate-related NLP evaluation datasets. Introduces Guardian Climate News Corpus with 6 climate-related categor...

Murathan Kurfali, Sherong Zuhre, Johan Neiver +1 more

Combining GNNs and LLMs for Advanced Reasoning

Integrates implicit knowledge of pre-trained LLMs with structural reasoning capabilities of Graph Neural Networks for novel multi-modal AI models. App...

Daniel Perez-Ramirez, Lodovico Giaretta, Joakim Eriksson +1 more

COP-Pilot

Open platform integrating edge computing, 5G, and decentralized processing for security, automation, and cross-sector applications. Supports industrie...

Jonas Gustafsson, Nicolas Tsiftes, Josef Filerini +2 more

Curling Analytics

Provides data-driven decision support for Swedish national curling team using real-time data and advanced analytics. Collects data from Jönköping Nati...

Laura Meriama Giovannetti, Jia Fu, Sepideh Pashami +2 more

DARE: Digital Autonomy with RISC-V in Europe

Large-scale European supercomputing project developing next-generation European processors and computing systems with RISC-V. Features 3 RISC-V chiple...

DC Sustainability

Focuses on datacenter sustainability through life cycle assessment, heat-to-power conversion, and waste heat optimization. Evaluates environmental imp...

Parvathi Maya Thampi

Detection of Disinformation Campaigns

Content-agnostic detection of coordinated disinformation campaigns in social networks addressing FIMI (Foreign Information Manipulation and Interferen...

Dalia Fano Yela, Sheila Riazi

Do we really need matrix multiplication in NNs?

Hub for all quantum technology activities at RISE implementing the Swedish Quantum Strategy. Expertise includes semiconductor materials and devices fo...

Linda Johansson, Erik Wallman

Eavesdropping on Seabirds: AI Uncovers the Acoustic World of Guillemots

Uses AI to decode complex vocal language of Guillemots for ocean health indicators. Eavesdropping in unique lab (Stora Karlsö) with rich multi-sensor ...

John Martinsson, Delia Fano Yela, Olof Mogren

ED5GE

Virtual developer community combining 5G speed and coverage with edge computing's real-time processing capabilities. Goal to establish 10,000 edge nod...

Enabling Cyber Threat Intelligence Sharing for Resource Constrained IoT

Introduces tinySTIX, a compressed CTI data format optimized for constrained IoT environments. Extends STIX and TAXII protocols for resource-constraine...

August Karlsson, Rikard Höglund, Han Wang +2 more

Explaining NLP models for Drug Safety

Explores methods to explain predictions of machine learning classifiers assessing adverse event report seriousness for regulatory compliance in pharma...

Luise Dürlich, Erik Bergman, Maria Larsson +4 more

FAST: Rapid Prototyping Framework for Cybersecurity Applications on Heterogeneous Computing

Framework that simplifies and accelerates prototyping of secure IoT applications on FPGA-based edge platforms. Provides comprehensive HW/SW co-design ...

Hendra Irawan, Shahid Raza

Formal Testing of AI Systems (Citcom.AI)

Introduces formal testing approach for structured AI system evaluation under the AI Act regulatory framework. Establishes context definition across AI...

Nishat I. Mowla, Kabir Fahriq

Generalizable Representation for Wireless Networks through Native Graph Topology

Uses graph representation learning and transformer architectures for wireless network optimization in 5G/6G. Award-winning solution in ITU AI/ML in 5G...

Yifei Jin, Sarunas Girdzijauskas, Aristides Gionis

Heatwise

HEATWISE focuses on holistic energy management for data centers and buildings with high IT load. Targets 100% waste heat recovery, PUE below 1.05, 10%...

Higher

HIGHER aims to prototype the first all-European data center-ready processor and management modules integrated into cloud and edge infrastructures usin...

Manolis Marazakis

How Trustworthiness Can Destroy AI

Analyzes how AI ethical principles in the EU AI Act, when taken to extremes, can be destructive. Total human oversight removes AI usefulness, complete...

Sepideh Pashami, Laurynas Adomaitis

ICE Connect

Developing machine learning-based intrusion detection systems for vehicle networks including CAN and Automotive Ethernet. The project focuses on data ...

Nishat I. Mowla

In-Distribution and Out-of-Distribution Data in Logit Space

Investigates out-of-distribution (OOD) detection in practice using Cat vs Dog classifier as toy example. Shows OOD samples cluster around center of lo...

Vangjush Komini, Sarunas Girdzijauskas

Iroute: Towards Routing a GPU on a Laptop

Hardware design research on routing VLSI circuits using GPUs on laptops. Addresses complex timing constraints in physical design automation. Achieves ...

Karl-Filip Faxén

LeakPro: Testing for information leakage from AI systems

Open-source platform to assess information leakage of trained ML models under realistic assumptions. Addresses privacy attacks including membership in...

Johan Östman

Lightweight Self-Adaptive Jamming Detection for IoT Networks

Online adaptation mechanism for jamming detection thresholds in response to environmental variations. Handles dynamic network conditions, heterogeneou...

Shuai Zhu, Fatemeh Rahimian, Thiemo Voigt +1 more

LiMNet: efficient real-time AI inference for high-frequency temporal interactions

Lightweight Memory Network for processing high-frequency temporal interactions and capturing event causality patterns. Highly scalable: processes 1000...

Lodovico Giaretta, Sarunas Girdzijauskas

LLM and MCP for controlling the real-world

Demonstrates how Model Context Protocol (MCP) connects large language models to tools, devices, and live data for real-world control. MCP servers conn...

Joakim Eriksson, Nicolas Tsiftes, Joel Höglund +1 more

PrivaGym: Dynamic Optimization of Privacy-Utility trade-offs in Data Analytics with MABs

Framework using Reinforcement Learning and Multi-Armed Bandits to dynamically optimize privacy-preserving defense mechanisms. Adapts to different data...

Reethika Ambatipudi, Apostolos Pyrgelis, Vasilios Mavroudis +1 more

RANDOL: Robust (Adversarial perturbation, Noisy label, Domain shift, Open set) Learning

Advances open-set domain generalization using Evidential Bi-Level Hardness Domain Scheduler. Proposes EBL-HADS to dynamically adjust order of domain p...

Jia Fu, Kunyu Peng, Sepideh Pashami +1 more

SARDIN: decentralized secure processing environment for health data

Platform enabling privacy and data security in healthcare data sharing through Health Data Bank network. Uses decentralized data analysis, cryptograph...

Anneli Niu, Hanna Svensson, Johan Kristiansson +1 more

Saving Lives

Uses machine learning to analyze multimodal datasets for suicide risk prediction. Funded by American Foundation for Suicide Prevention (AFSP) in colla...

Fehmi Ben Abdessalem, Christian Ruch, Magnus Bowin +1 more

Scalable & Automated Hardware Configuration for Radio Access Networks using Constraint Programming

Automates RAN hardware configuration using constraint programming to match complex carrier requirements. Handles hundreds of radios, diverse standards...

Mats Carlsson, Erik Oskar-Oskar, Marko Đuričić +1 more

SeamAI: Seamless Integration of AI for IoT

Addresses security challenges of IoT devices with limited battery and resources by developing lightweight TEE. Provides hardware-based security mechan...

Amir Abdollahi, Giuseppe Nebbione, Simon Bouget +1 more

SEMLA: Securing Enterprises with Machine Learning Automation

Uses LLMs for enterprise security automation including vulnerability detection, code scanning, correct code generation, and fine-tuning. Achieves impr...

Mariano Scazzariello, Changjie Wang

SICStus Prolog: Powering Innovations for Decades

Premier commercial Prolog development platform trusted globally by businesses, governments, and academia. Features include database queries, constrain...

Mats Carlsson, Per Mildner

Similarity Learning for Spectral Clustering

Proposes novel modification of RBF kernel to mitigate need for hyperparameters in spectral clustering. Uses end-to-end data-driven similarity learning...

Vangjush Komini, Nashicha Kuriakose, Debashiya Roy +1 more

Talking with a robot

Demonstrates hybrid AI systems combining speech, vision, and language models to control a physical robot arm. Integrates speech recognition (whisper),...

Anders Holst, Nicolas Farre, Nicolas Tsiftes +1 more

The Contiki-NG OS for Resource-Constrained IoT Devices

Open-source operating system for severely resource-constrained embedded devices supporting secure, energy-efficient communication. Features low-power ...

Nicolas Tsiftes, Nicolas Farre, Peter A. Jonasen +1 more

The EDHOC and OSCORE Profile of the ACE Framework for Access Control

Combines EDHOC (lightweight key exchange) with OSCORE (end-to-end CoAP security) for efficient access control in constrained environments. Client obta...

Marco Tiloca, Rikard Höglund

TinyTor: Enabling Onion Routing for the IoT

Enables onion routing for IoT using OSCORE, CoAP, and EDHOC protocols. Introduces OSCORE-capable proxies and nested OSCORE for privacy. Uses EDHOC key...

Rikard Höglund, Marco Tiloca

Tracking Cancer with 3D Segmentation

Uses machine learning to identify organs in CT scan images to help clinicians find cancer metastases. Funded by Precision Medicine Centre at Karolinsk...

Fehmi Ben Abdessalem, Vitali Stromov, Magnus Bowin

Training AI assistants – build datasets, finetune, evaluate, repeat

Research on algorithms and tools for creating AI assistants: dataset building with human feedback, faster training algorithms, better testing methods ...

Lars Rasmusson, Fredrik Carlsson, Karl-Filip Faxén +2 more

When Facts Collide: Investigating LLMs in Conflicting Knowledge Scenarios

Investigates how LLMs retrieve information when exposed to conflicting or redundant facts using Needles in a Haystack benchmark. Analyzes variables in...

Murathan Kurfali, Robert Östling

Why the way we regulate AVs needs to change

Legal analysis critiquing EU automated vehicle regulations. Identifies crucial differences between AV behavior logic and EU legal duty concepts. Argue...

5G Forge – private 5G infrastructure

5G Forge: Private 5G and Edge Compute for industry & society. Private 5G infrastructure on licensed frequency bands with edge computing. Enables digit...

Joakim Eriksson, Alan Jonsson

Aerial View Localization with Reinforcement Learning: Towards Emulating Search-and-Rescue

Aerial View Localization with Reinforcement Learning: Towards Emulating Search-and-Rescue. AirLoc agent for UAV visual goal localization using only vi...

Aleksis Pirinen, Anton Samuelsson, John Backmans +1 more

AI 4 Predictive maintenance

Advanced Machine Learning for Predictive Maintenance. FREEPORT (Federated Learning) and FA2M (Future AI-based Maintenance) projects. Thermal resistivi...

Sepideh Pashami, Anders Holst

AI Assurance: Prevent Digital Meltdowns And Keep AI Tantrums At Bay!

AI Assurance: Prevent Digital Meltdowns And Keep AI Tantrums At Bay! Seven requirements for trustworthy AI: human agency, technical robustness, privac...

Fatemeh Rahimian, Sverker Janson

AI embeddings: Matrices are better than vectors

AI-ARC and VIGIMARE: Detecting illegal activities at sea. Using AI to detect suspicious behaviors in maritime data. Hybrid AI combining machine learni...

Björn Bjurling, Jan Ekman, Anders Holst +2 more

Aqua Poster

Aqua: A Programming Language for Continuous Analytics. Combines query languages and dataflow libraries. Features safe, concise, efficient, language-in...

Klas Segeljakt, Seif Haridi, Frej Drejhammar +1 more

Brain-inspired algorithms for computing

AI embeddings: Matrices are better than vectors. Matrix embeddings provide richer operations than vector embeddings: comparison, partial ordering, log...

Martin Nilsson

Brain-inspired Algorithms for Computing

Brain-inspired Algorithms for Computing. Four topics: Similarity-preserving Sparse Randomized Embeddings (FlyHash), Distributed Codes with Residue Num...

Denis Kleyko

ColonyOS – A Meta operating system

ColonyOS: A Meta-Operating System for Next-Generation AI. Manage distributed computing across HPC, cloud, edge, and IoT. Features: zero-trust security...

Johan Kristiansson

Connecting Earth and Space Networks in 6G: Security

Connecting Earth and Space networks in 6G: what about security? Studying security implications of integrating Non-Terrestrial Networks (NTNs) with tra...

Arianna Avenali, Giuseppe Caso, Alfonso Iacovazzi +2 more

Controlled Image Synthesis

Controlled Image Synthesis with Disentangled Latent Spaces. Learning identity-separated GAN latent space with self-supervised methods. DIFFUSE project...

Creating and Leveraging a Synthetic Dataset for Cloud Detection in MSI

Creating and Leveraging a Synthetic Dataset of Cloud Optical Thickness Measures for Cloud Detection in MSI. 200,000 synthetic data points for training...

Olof Mogren, Aleksis Pirinen, Thomas Ohlson Timoudas

Cybercampus Sverige

Cybercampus Sverige: Swedish national initiative for agile and cutting-edge cybersecurity and cyber defence research, innovation and education. Planni...

Pelle Lunden, Shahid Raza, Katarina Bunakov

CyberNode Games as a Tool Working Group

Cybernoden Tema Group: Games as a Tool for Cybersecurity Training. Using games like Mega and Pico games for cybersecurity training. NIS2 compliance pr...

Cybernoden

Cybernoden: En nationell kraftsamling kring cybersäkerhet. Swedish National Coordination Centre for Cybersecurity (NCC-SE). Coordinates with MSB and E...

Martin Bergling

Cybernoden: National Cybersecurity Coordination

Cybernoden: En nationell kraftsamling kring cybersäkerhet. Swedish National Coordination Centre for Cybersecurity. NCC-SE coordinates with MSB and ECC...

Martin Bergling, Pelle Fredrikson, Shahid Raza

Detecting Attacks against IoT Networks using Machine Learning at the Edge

Detecting Attacks against IoT Networks using Machine Learning at the Edge. Master's thesis at KTH. Classification of packet headers using Contiki-NG i...

Viktor Friдεn, Niclas Finne

Digital Earth Sweden

Digital Earth Sweden: Easy access to Earth observation data for academia, public sector, and general public. Platform for Baltic Sea and Scandinavian ...

Jasper Dehls, Daniel Elfström, Peter A Jonsson +2 more

Disentangled Latent Spaces: Controlled Image Synthesis

Disentangled Latent Spaces: Controlled Image Synthesis. Learning an Identity Separated GAN Latent Space with Self-Supervised Methods. DIFFUSE project ...

RISE Research Institutes of Sweden, Uppsala University, Smart Eye

EBRAINS – European Brain ReseArch INfraStructure

EBRAINS: A state-of-the-art ecosystem for neuroscience. Data, Brain Atlases, Modelling & Simulation, Validation & Inference. Nordic Neuromorphs and Eu...

Anastasia Andrievskaia

Edge AI on Spatio-Temporal Data for Smart Mobility

Edge AI on Spatio-Temporal Data for Smart Mobility. AloTwin project: distributed training and inference of Graph Neural Networks at the edge. Addressi...

Ivan Arall, Lodovico Giaretta, Sarunas Girdzijauskas

Fighting Dengue Fever with Drones

Fighting Dengue Fever with Drones. Detect mosquito breeding places and still-standing water using drones with mmWave radio and multispectral imagery. ...

Thiemo Voigt, Markus Flierl, Luca Mottola +1 more

First Responder – Locally Hosted LLM Chatbots for Everyone

First Responder: Locally Hosted LLM Chatbots for Everyone. Data privacy and security via local hosting. No third party involved. Customizable for spec...

Leon Sütfeld

Graphene Enhanced Immersion Cooling for Data Centers

Graphene Enhanced Immersion Cooling for Data Centers. Vinnova-funded project identifying best thermal interface material for servers in RISE wind tunn...

Jon Summers, Jonas Gustafsson, Henrik Berkenhead +2 more

HEATWISE

HEATWISE: Holistic Energy management And Thermal Waste Integrated System for Energy optimization. Digital twin for data centers and buildings to optim...

High Level and Low Level Program Specifications for Safe Composability

High Level and Low Level Program Specifications for Safe Composability. Programs call other programs, are distributed, use parallel processing. Formal...

Lars Rasmusson

ICE Connect

ICE Connect Cloud Platform: Single sign-on portal for RISE ICE Data Center services. Kubernetes cluster with multi-GPU servers, S3 storage, OpenStack ...

ICE Datacenter

ICE Data Center - Your Digital Infrastructure Research Partner. Testbed for digitalization and IT infrastructure. Support for efficiency, sustainabili...

Tor Björn Minde, Izabella Aamla

ImpactDB: tracking environmental impact with LLMs

ImpactDB: Tracking extreme climate events with LLMs. Database of climate events and their impacts using LLM prompting for extraction. Standardized loc...

Joakim Nivre, Shuvam Datta, Chenjuan Peng +1 more

Inhibition: Efficient Attention for Transformers

The Inhibitor: ReLU and Addition-Based Attention for Computationally Efficient Quantized Transformers. Quantized Transformers without dot-product and ...

Rickard Brännvall, Andrei Stoian

LEAKPRO: Leakage Profiling and Risk Oversight for Machine Learning Models

Open-source platform to assess information leakage of trained ML models, federated learning, and synthetic data. Models privacy attacks including memb...

Rickard Brännvall, Henrik Forsgren

LeakPro: ML Privacy Profiling

LeakPro: Leakage Profiling and Risk Oversight for Machine Learning Models. Open-source platform to assess information leakage from trained ML models. ...

Johan Östman, Rickard Brännvall, Henrik Forsgren

Lightweight Key Update Procedure for the OSCORE Protocol

KUDOS: Lightweight Key Update Procedure for the OSCORE Protocol. Lightweight procedure to renew shared keying material. Robust against peer rebooting....

Rikard Höglund, Marco Tiloca

LLMs and Foreign Information Manipulation & Interference

Foreign Information Manipulation & Interference: A Large Language Model Perspective. LLMs can be exploited for creating malicious content. Training am...

Björn Bjurling, Stella Nowak

Membership Inference Attack

Protecting IoT Data against Membership Inference Attack with Robust Synthetization in ML-based IDS. RBMSparaAE differential-private approach for synth...

Han Wang, Seonghyun Kim, Luis Barriga +2 more

Mid-sized Haptic Interactions – The Pelvic Chair and Autonomous Driving

Mid-sized Haptic Interaction Design: The Pelvic Chair & Semi-autonomous Driving. Research on haptic interactions for larger body parts. Air pocket cha...

Anna Ståhl, Madeline Bayes, Caroline Yan Zheng

On-device Learning for Secure Internet of Things

On-device Learning for Secure Internet of Things. Perform model training on IoT devices for autonomy and security. JamSense use case for self-adaptive...

Shuai Zhu, Thiemo Voigt, Fatemeh Rahimian +1 more

Partner program

RISE ICE Data Center Partner Program. Extended opportunity for long-term collaboration. Three tiers: platinum, gold, silver. Early access to research ...

Izabella Aamla, Tor Björn Minde, Jonas Gustafsson

Representations of smells: The next frontier for language models?

Representations of smells: The next frontier for language models? Investigating how LLMs represent olfactory concepts. Testing on odor-perceptual rati...

Murathan Kurfali

RISC-V Demo

Building a Secure Open-Source Hardware Architecture with Trusted Execution Environments (TEEs). Physical Memory Protection (PMP) and Trusted Platform ...

Aamir Aaslam, Abir Yusuf Sfanks, Shahid Raza

RISE ICE Data Center Partner Program

RISE ICE Data Center Partner Program. Extended opportunity for long-term collaboration. Three tiers: platinum, gold, silver. Early access to research ...

Izabella Aamla, Tor Björn Minde, Jonas Gustafsson

Safe Application Execution on Resource-Constrained IoT Devices using WebAssembly

Safe Application Execution on Resource-Constrained IoT Devices using WebAssembly. WAMR (WebAssembly Micro Runtime) for embedded devices. Contiki-NG in...

Fredrik Engstrand

SARDIN: Decentralized platform for secure processing of health data

SARDIN: Secure and Accountable Results-sharing with Data-minimization and Integrity-protected Network. Health Data Bank platform using ColonyOS for pr...

Anneli Nou, Hanna Svensson, Johan Kristiansson +1 more

Secure Open-Source Hardware with TEEs

Building a Secure Open-Source Hardware Architecture with Trusted Execution Environments (TEEs). Physical Memory Protection (PMP) and Trusted Platform ...

Aamir Aaslam, Abir Yusuf Sfanks, Shahid Raza

Secure System Monitoring in TrustZone-Enabled IoT Devices

Secure System Monitoring in TrustZone-Enabled IoT Devices. Using ARM TrustZone TEE for mission-critical IoT applications. Contiki-NG running in secure...

Niclas Finne, Nicolas Tsiftes, Thiemo Voigt

Serverless CRDTs at the Edge

COGNIT: A Cognitive Serverless Framework for the Cloud-Edge Continuum. AI-enabled adaptive serverless for performance, security, resource management, ...

Thomas Ohlson Timoudas

Serverless CRDTs at the Edge

Serverless CRDTs at the Edge. Stateful Serverless with conflict-free replicated data types. State mutations are consistent and reliable. Low latency f...

Adam Hasselberg, Gyozo Dan, Paris Carbone

SICStus Prolog: Powering Innovations for Decades

SICStus Prolog: Powering Innovations for Decades. Premier commercial Prolog platform for businesses, governments, academia. Applications: database que...

Mats Carlsson, Per Mildner

Sound matters: machine learning for sound analysis

Sound Matters: Machine Learning for Sound Analysis. Soundscape analysis for biodiversity (30 by 30 goal). Few-shot bioacoustic event detection with ad...

John Martinsson, Martin Willbo, Marcus Pirinen +1 more

Supporting waste-room stewardship using sensor networks

Supporting waste-room stewardship using sensor networks. Digital environmental stewardship (DES) for recycling rooms. Sensors track manual work and en...

Mattias Jacobsson

Supporting waste-room stewardship using sensor networks

Supporting waste-room stewardship using sensor networks. Digital environmental stewardship (DES) for recycling rooms. Sensors to track manual work and...

Mattias Jacobsson

Sustainability in Data Centers

Sustainability in Data Centers. E2P2 fuel cell assessment, EECONE (second life of data center electronics). Life cycle assessment following ISO 14040/...

Saravana Maya Thampi

Towards GenAI on Wireless RayTracing

Towards an Offline, Differentiable, Fully-Trainable Wireless Neural Ray-Tracing Surrogate. Deep reinforcement learning for 3D wireless channel modelin...

Yifei Jin, Ali Moutaki, Marios Daoutis +1 more

Towards Robust ML: Defending Against Adversarial Attacks

Towards robust machine learning: defending against adversarial attacks. DIPPAD: Denoising Diffusion-based Adversarial Patch Decontamination. Methods f...

Jia Fu, Sepideh Pashami, Yifei Zhang +2 more

Using the OSCORE Security Protocol at Proxies and for CoAP Secure Tunneling

Using OSCORE Security Protocol at Proxies and for CoAP Secure Tunneling. OSCORE-capable proxies for transport chain security. One-to-many notification...

Marco Tiloca, Rikard Höglund

Verification of AI models

Verification of AI models. Using formal methods to verify neural networks. Applications include pendulum control systems, reinforcement learning, and ...

David Eklund, Jan Ekman, Koen Hoejmans

<a title of your poster or demo>

Testing and certifying AI systems on national and European Level - partnering for an AI Testbed. Looking for interest and funding to develop testing a...

Lars Rasmusson, Sverker Jansson, Petra Daluader +1 more

3D Segmentation for CT Scan Images

3D Segmentation for CT Scan Images. Using Machine Learning to identify organs in CT scan images and help clinicians find cancer metastases. Funded by ...

Fehmi Ben Abdesslem, Yilei Diesmos, Magnus Boman

A Cognitive Serverless Framework for the Cloud-Edge Continuum

A Cognitive Serverless Framework for the Cloud-Edge Continuum. COGNIT: AI-enabled adaptive serverless framework for performance, security, resource ma...

Thomas Ohlson Timoudas

ACES: Translation Accuracy Challenge Sets for Evaluating Machine Translation Metrics

ACES: Translation Accuracy Challenge Sets for Evaluating Machine Translation Metrics. Contrastive challenge sets based on MQM Ontology for Translation...

Chantal Amrhein, Nikita Moghe, Liane Guillou

Advanced Security Solutions for the IoT

Advanced Security Solutions for the IoT. Four topics: KUDOS (Key Update for OSCORE), OSCORE-Capable Proxies, One-to-Many Notification Responses, and O...

Marco Tiloca, Rikard Höglund

Battery-free Neighbor Discovery

Battery-free Neighbor Discovery for bootstrapping battery-free wireless networks. Challenge: intermittent and unpredictable energy availability. Appro...

Sapientia Hayde-Hardardóttir, Carlos Perez Penichet, Thiemo Voigt

Better Sawing for Future Sawmills

Better Sawing for Future Sawmills. Using CT scanning and constraint programming optimization to maximize value yield from logs. Compares 2D vs 3D opti...

Mats Carlsson

Causal Possible Worlds

Causal Possible Worlds: Finding the belief behind causal arguments. Cluster arguments based on implied beliefs, bridge opposing arguments based on com...

Björn Bjurling, Jan Ekman, Fabian Sinzker +2 more

Collborate with RISE in Advanced and Innovative Digitalisation?

Stateful Serverless at the Edge: Function-as-a-Service with cloud computing abstraction. Features stateless functions, rapid auto-scaling, pay-as-you-...

Adam Hasselberg, Gyuho Dae, Paris Carbone

ColonyOS

ColonyOS: Unleashing computational power everywhere. Meta-operating system for distributed compute engines across HPC, K3S, edge, and IoT. Features ze...

Johan Kristiansson

Compiler Expertise

Compilers, Toolchains, Languages: Computer Systems unit expertise. Includes LLVM/MLIR work, RISC-V compilers and toolchains, Erlang runtime optimizati...

Frej Drejhammar, Björn Forsberg

Confidential Evaluation of Software Trustworthiness

CUSTODES: Leveraging Trusted-Execution Environment for Confidential Software Certification. CEST platform uses TEE (Trusted Execution Environment) for...

Simon Bouget, Arash Vahidi, Jakob Palmkvist +1 more

Cyber Range Training Environment

RISE Cyber Range: Cybersecurity training and education platform. Real software and real hacking in a controlled environment. Virtualization platform f...

Gustav Mideus

Cyber Threat Intelligence for IoT

STACK: Smart, Attack-Resistant Internet of Things Networks. Making IoT networks maintain functionality in challenging situations. Robust IoT communica...

Joakim Eriksson, Niclas Finne, Aadarsha Mard +3 more

Data Driven Analysis of Sailing Olympic Strategy

Data-driven Decision Support for Olympic Sailing Team. AI-based prediction and visualization for Swedish Olympic Sailing Team training and competition...

Laura Marlene Siravegna, Björn Forsberg, Federico Perossi +4 more

Decentralized Training of GNNs for Edge/IoT Devices

Fully-decentralized, scalable, privacy-preserving architecture to train Graph Neural Networks directly on edge/IoT devices without central aggregation...

Lodovico Giaretta, Šarunas Girdzijauskas

Digital Earth Sweden - OpenEO

Digital Earth Sweden: Providing easy access to Earth observation data via openEO API. Collaboration with Swedish National Space Agency. Features clear...

Joel Sundholm, Daniell Elfström, Peter A. Jonsson +1 more

DIgital Earth Sweden -mapviewer

Digital Earth Sweden Mapviewer. Easy access to data and compute resources for Earth observation over Baltic Sea and Scandinavian Peninsula. Collaborat...

Joel Sundholm, Daniell Elfström, Peter A. Jonsson

Digital Futures Drone Arena

A novel aerial drone testbed exploring human-drone relationships through competitions. Teams program nano drones for gesture-based piloting and aesthe...

Luca Mottola, Airi Lampinen, Fehmi Ben Abdessem +6 more

Digital Säkerhet

Cybernoden: En nationell kraftsamling kring cybersäkerhet för att accelerera forskning och innovation inom industri och offentlig sektor. Swedish nati...

Martin Bergling, Göran Holm, Shahid Raza

EDHOC - Lightweight Key Establishment for the Internet of Things

EDHOC - Lightweight Key Establishment for the Internet of Things. Ephemeral Diffie-Hellman Over COSE based on SIGMA-I MAC-Then-Sign. Three main messag...

Marco Tiloca, Rikard Höglund

Efficient Prediction with Recurrent Circuits of Sigma-Pi Neurons

Efficient Prediction with Recurrent Circuits of Sigma-Pi Neurons. Distributed representation approach for reservoir computing that outperforms NG-RC. ...

Denis Kleyko, Christopher Kymn, Bruno A. Olshausen

EMERGENCE 2.0: Securing Edge Networks

EMERGENCE 2.0: Securing Edge Networks with a Programmable Intelligent Architecture. Disaggregated architecture for securing edge networks using P4/FPG...

Nicolas Tsiftes

ENCCS

ENCCS - EuroCC National Competence Centre Sweden. What you can do with supercomputers: Train AI models, simulate molecules and materials, automate ind...

EU-HYBNET: Empowering a Pan-European Network to Counter Hybrid Threats

EU-HYBNET: Empowering a Pan-European Network to Counter Hybrid Threats. Horizon 2020 project with 25 EU member states. Addresses hybrid threats includ...

Rolf Blom

Federated Learning for Cybersecurity

SparSFA: Towards robust and communication-efficient peer-to-peer federated learning. Sparsified Federated Aggregation for intrusion detection in IoT. ...

Han Wang, Luis Muñoz-González, Muhammad Zaid Hameed +2 more

Group OSCORE - Secure Group Communication for the Internet of Things

Group OSCORE - Secure Group Communication for the Internet of Things. End-to-end security for CoAP with ACE framework. Supports one-to-many group requ...

Marco Tiloca, Rikard Höglund

ICE Datacenter

ICE Connect Cloud Platform: Web portal for single sign-on access to RISE ICE Data Center services. Kubernetes cluster with multi-GPU servers, S3 objec...

Machine Learning for Combinatorial Resource Allocation

Finding Small Rectilinear Steiner Minimal Trees Quickly. VLSI routing optimization using search with pruning. Initial transformation to Hanan grid, th...

Karl-Filip Faxén

Machine Learning for Combinatorial Resource Allocation

Machine Learning for Combinatorial Resource Allocation (ML4COP). Using ML to reduce optimality gap in resource allocation for networked systems. DeepG...

Daniel F. Perez-Ramirez, Dejan Kostic, Magnus Boman +1 more

Management of Flexibility in Energy Communities with Distributed Resources

RESCHOOL: Strategies and tools for incentivization and management of flexibility in Energy Communities with distributed resources. Catalyzes creation ...

Ahmad Al-Shishtawy

MAS-CTI: ML Assisted System for Cyber Threat Intelligence

MAS-CTI: Machine Learning Assisted System for Cyber Threat Intelligence. Elevates CTI processing using ML and Federated Learning. Addresses threat ran...

Han Wang, Alfonso Iacovazzi, Seonghan Shin +1 more

Mid-sized Haptic Interactions - The Pelvic Chair and Autonoumous Driving

Mid-sized Haptic Interaction Design: The Pelvic Chair and Semi-autonomous Driving. Research on haptic interactions for larger body parts. Pelvic Chair...

Anna Ståhl

Multimodal Dance Improvisation Recognition

Component attention network for multimodal dance improvisation recognition. Uses Unico-Maastricht Dance dataset with 13 joints, 6 IMUs, accelerometers...

Jia Fu, Jianui Tan, Wenjie Yin +2 more

Opportunities and shortcomings of Large-Scale Generative Language Models

What is Wrong with Large Language Models? Covers generation issues (false information, harmful content, hallucination, bias), limited language coverag...

NLP Group @ RISE

Poster RISC-V initiatives

RISC-V open-specification and extensible ISA and eco-system. RISE research and innovation in RISC-V for digital sovereignty. Includes R&D platforms, O...

Joakim Eriksson, Björn Forsberg

Regulatory Risks Associated with Cybersecurity

Regulatory Risks Associated with Cybersecurity. Understanding, identifying, and managing regulatory risks in cybersecurity. Covers compliance with evo...

Mazaher Kianpour, Shahid Raza

SCAIA: Societal Collaborative Artificial Intelligence Awareness

SCAIA: Societal Collaborative Artificial Intelligence Awareness. Initiative by AI researchers at RISE to share basic understanding of AI technologies,...

Scalable Privacy-Preserving Federated Machine Learning

Scalable Privacy-Preserving Federated Machine Learning Analytics. Using multi-party homomorphic encryption (MHE) and federated learning for data confi...

Apostolos Pyrgelis

Secure Sharing of Health Data

Secure Sharing of Health Data. Research on VINTER, DELFIN, and HEIDA projects using Fully Homomorphic Encryption (FHE) for privacy-preserving cloud se...

Helena Linge, Rickard Brännvall

SICStus Prolog: Powering Innovations for Decades

Commercial Prolog development platform used globally for business rule systems, product configuration, railway specification, and NLP. Features declar...

Mats Carlsson, Per Mildner

STACK: Smart, Attack-Resistant Internet of Things Networks

STACK: Smart, Attack-Resistant Internet of Things Networks. ITEA4 project enabling IoT networks to maintain functionality under attack or harsh radio ...

Thiemo Voigt

Tackling climate change with AI

Tackling climate change with AI across multiple projects: Soundscape analysis for biodiversity monitoring, Wetland estimation using satellite data, Co...

Mia Ekstääf, Axel Ersson, John Martinsson +2 more

The Brain's Assembly Language

The Brain's Assembly Language. Mammalian nervous system organized as networks of communicating neuron populations. Using mechanistic model with plasti...

Martin Nilsson

The Portals Framework - Enabling Flexible Stateful Serverless Applications

Portals: A Showcase of Multi-Dataflow Stateful Serverless. Programming framework for stateful serverless with multi-dataflow composition using atomic ...

Jonas Spenger, Chengyang Huang, Philipp Haller +1 more

Titeln är decentralized adaptive clustering is beneficial for client collaboration.

Decentralized adaptive clustering of deep nets is beneficial for client collaboration. DAC (Decentralized Adaptive Clustering) helps clients with simi...

Edvin Listo Zec, Ebba Ekblom, Martin Willbo +2 more

Training Machine Learning Models for Cloud Optical Thickness Estimation using Synthetic Data

Training Machine Learning Models for Cloud Optical Thickness Estimation using Synthetic Data. 200,000 data points simulating how atmosphere distorts l...

Olof Mogren, Aleksis Pirinen, Thomas Ohlson Timoudas +1 more

When the Electric Car Moves In

When the Electric Car Moves In! Social and technical potential with coordinated control of heat pumps and electric car charging in single-family homes...

Jan Ekman, Carolina Hollem

5G Edge Testbed

Research testbed for device-edge-cloud compute continuum with 5G connectivity. Features edge systems hardware with micro grid, batteries, solar/fuel-c...

AI & Cloud platforms/infrastructure, where are you running your AI, HPC, Quantum?

Overview of AI platforms and infrastructure options for running AI, HPC, and quantum computing workloads.

Martin Körling

AI for drug influence detection using recorded eye scans

Detecting drug influence by tracking eye movements and pupil dilation using deep learning. Collaboration with startup Eyescanner and Sahlgrenska Hospi...

Aleksis Pirinen, Martin Willbo, Olof Mogren

AI to learn about porcelain

ML model to help people learn about porcelain from Rörstrand museum. Uses pretrained ResNet to predict object type, artist, and production year from p...

AI-Ecosystems – ENCCS

ENCCS (EuroCC National Competence Centre Sweden) provides training and support for industry, academia, and public administration in accessing and usin...

AI@Edge – AI for network automation

AI@Edge: A Secure and Reusable Artificial Intelligence Platform for Edge Computing in Beyond 5G Networks. Building platform and tools for secure autom...

AutoPKI: automated and scalable transfer of trust

Pre-training Transformers for molecular property prediction using reaction prediction. Uses SMILES representation of chemical reactions. Essential for...

Johan Broberg

Big Data Analytics and AI for a Smarter Society

Three projects: BADA (Big Data Analytics of Road Traffic Datasets for Swedish automotive industry), IntraBod (Smart Intra-Body Network with sensors an...

Ahmad Al-Shishtawy

colonyOS – a meta operating system

ColonyOS: Orchestrate processes inside containers across platforms and infrastructures. Features batch processing, distributed RPC, pull-based orchest...

Cross-lingual transfer of monolingual models

Research on using existing language models for training in new languages to save computational resources and benefit low-resource languages. Method: r...

Evangelia Gogoulou, Ariel Ekgren, Tim Isbister +1 more

Data trace generation with Multi-Trace

Multi-Trace: Tools for security research on low-power mesh networks. Features multi-level data traces, global timestamps, and utilities for generating...

Data-science for Earth Observation (EO) Data

Building a knowledge and data hub for Earth Observation to help researchers and decision makers understand events and measure development over time us...

Nuria Agues Paszkowsky

Decentralized Graph Neural Networks

Efficient communication in decentralized deep learning. Research on collaborative learning in P2P networks when data sharing is prohibited. Addresses ...

Edvin Listo Zec

Designing Touch Interaction

Mid-sized haptic interactions for autonomous vehicles. The Pelvic Chair project (with Volvo Cars and KTH) uses air pockets to create awareness of pelv...

Anna Ståhl

Digital Earth Sweden

Providing machine learning ready Earth Observation data. Built with open source from the ground up, running in containers on Kubernetes at RISE ICE. U...

Empowering a Pan-European Network to Counter Hybrid-Threats

EU-HYBNET: Horizon 2020 project with 25 partners from 14 EU member states addressing hybrid threats. Workshops, training, knowledge exchange to counte...

End-To-End Secure Group Communication for the IoT

IETF standardization work on secure group communication for IoT using CoAP and OSCORE protocols. Includes Group OSCORE security protocol for one-to-ma...

Marco Tiloca

Environmental impact of data centres

Life Cycle Assessment (LCA) study on data centers to understand environmental impact of building and operating a DC. Scope: Boden Type Data Center (BT...

Borisova

Federated Learning for IoT Security

Sec4Factory: Securing the next generation of Industrial Control Systems for Industry 4.0. Research topics include cloud control, intrusion detection i...

Martin Gunnarsson

Formal verification of communication protocol

Using formal methods and the Tamarin prover for mathematically proving security properties of communication protocols. Provides strong guarantees, can...

Simon Bouget

Homomorphic encryption for CPS

Using Fully Homomorphic Encryption (FHE) to protect private data while allowing computation in the cloud. Work on Cyber-Physical Systems with TenSEAL ...

ICE Test & Demo Datacenter

Service continuity for real-time edge-based applications. Use case: 5G-connected drone with object detection via video inferencing on the edge. Addres...

Hamid Reza Faragardi

Jamming Detection with JamSense

Extracting causal relations from Swedish government reports (SOU). Uses keyword filter and Swedish semantic textual similarity model (BERT-Base/Swe-CT...

Luise Dürlich

Key Update for OSCORE (KUDOS)

KUDOS: Lightweight procedure for updating shared keying material in CoAP/OSCORE IoT protocols. Two peers exchange nonces to generate fresh key materia...

Rikard Höglund

Large Language Models

Introduction to Large Language Models. LLMs model text distribution from corpora, useful for general language tasks. Explains prompting to steer model...

Amaru Cuba Gyllensten

Machine learning and acoustic monitoring for wildlife conservation

Research on automated analysis of years of audio data from natural habitats using machine learning. Focuses on few-shot learning for annotation-effici...

John Martinsson

Machine Learning for Causal Inference in Observational Studies

Using ML to address causal inference challenges in observational studies where randomized controlled trials aren't feasible. Methods include Random Ca...

Fatemeh Rahimian

Machine learning for predicting water flow intensity based on physical characteristics of catchment area

Data-driven AI model for predicting water flow intensity using spatial maps and temporal inputs (rainfall, temperature) to address flooding risks from...

Aleksis Pirinen, Olof Mogren

MiniZinc for Problem Solving

MiniZinc: A portable language for optimization problems using constraint programming. Developed at Monash University. Applications include configurati...

Mats Carlsson

Multilingual Image Search with M-CLIP

Resource-Efficient Multilingual NLP: Providing high-quality NLP for world's languages using massively multilingual pre-trained models (mBERT, XLM-R). ...

Natural Language Reasoning (NLR)

Natural Language Reasoning: Developing machines capable of human-level reasoning based on text. Uses NLP to extract reasoning elements and applies log...

Jan Ekman, Börn Bjurling

Powering the edge

ANIARA WP5 Power and Infrastructure EDGE: Solutions for high-performance services at the network edge. First generation demonstrator with outdoor-comp...

Privacy preservation by FHE enables data sharing in digital health management

Privacy-Preserving Computing with PODS (Secure Workflows). System for composing secure workflows for privacy-preserving computation with support for d...

Jonas Spenger, Paris Carbone, Philipp Haller

Radio Fingerprinting for IoT intrusion detection

Detecting cyber attacks on containerized applications using unsupervised graph-based ML algorithms. Built dataset with syscall traces and network traf...

Iacovazzi

Reducing Hazardous Chemicals in Society with AI and Patent Data

Using AI and NLP to proactively reduce hazardous chemicals in society via patent analysis. Patents indicate chemical use before products reach market....

Erik Ylipää, Olof Görnerup

Resource-efficient machine learning

Research on resource efficient machine learning covering both inference and training dimensions. Techniques include sparseness, advanced sparseness, m...

Karl-Filip Faxén

SAIS 2022 - Swedish AI Society Workshop

SAIS 2022: 34th annual workshop of the Swedish AI Society, June 13-14, 2022. Hybrid event hosted by RISE at Kista Science Tower, Stockholm. Keynote sp...

Åsa Rudström

Securing IoT Systems through Stateful Fuzz Testing

Secure group communication for IoT using CoAP and Group OSCORE protocol. Access control through ACE framework with Authorization Server, Resource Serv...

Marco Tiloca

Security Roadmap Towards 6G Systems

Security Roadmap towards 6G Systems. 6G expected by 2030 with 100-1000x performance over 5G. Covers security challenges including privacy, eavesdroppi...

Mohammed Ramadan

Semantic textual similarity for causality detection

Investigating how keyword filters and Swedish semantic textual similarity models can identify sentences containing causal relations and rank them in g...

Luise Dürlich

Semi-supervised Few-shot Learning for Sound Event Detection

Semi-supervised few-shot learning for sound event detection. Addresses challenge of annotating large audio datasets. Uses prototypical networks and me...

Willbo

Socio-economic effects of hard and soft digital infrastructure: from modelling to visualization

WeDistrict: EU-funded project with 21 partners demonstrating 100% fossil-free heating and cooling solutions. Swedish demonstrator integrates excess he...

Swedish AI Society annual workshop June 13-14 – hosted by RISE

JamSense: Detecting and classifying interference in low-power networks. Uses RSSI sampling for jamming detection, then classifies jamming type. High a...

Thiemo Voigt, Wenbo Xiao

Teaching a Computer About Beauty for Image-based Recommendations

Image-based recommendations using beauty assessment. CNN for feature extraction and MLP for rating prediction. Trained on social network photos labele...

Leon René Sütfeld

The Digital Futures Drone Arena

Digital Futures Drone Arena: A concrete and conceptual platform for interdisciplinary investigations at the intersection of mobile robotics, autonomou...

Thermal energy storage for data centers:

Thermal Energy Storage (TES) for edge data centers. Benefits: increased energy efficiency, better use of free cooling, operation optimization, UPS sup...

Stark, Mellgren

Trusted Execution Environments for Resource-constrained IoT

Research on Trusted Execution Environments (TEEs) using ARM TrustZone-M for resource-constrained embedded devices. Covers use cases including secure b...

Anum Khurshid

Vulnerability Analysis for Establishing More Secure Systems

Using Meta Attack Language (MAL) for vulnerability analysis and attack simulation. Enables practitioners to simulate attacks on their own organization...

Ismail Butun