Publications

2026

[S&P] Multiagent LLM Systems for Security Operations; IEEE Security & Privacy. Kim Hammar. IEEE Proceedings, bibtex.

[GC Wkshps] AI-Based Load Balancing Towards Autonomous Radio Access Networks; IEEE Globecom Workshops. Dilki Wijekoon, Hevish Cowlessur, Timothy J Hansen, Mohammad Javad Shokri, Kim Hammar, and Tansu Alpcan

[Submitted] Recovery Control in Replicated Systems through Autonomous Multiagent Rollout; submitted. Kim Hammar and Yuchao Li. ArXiv(preprint), bibtex.

[TMC] Causal Online Learning of Safe Regions in Cloud Radio Access Networks; TMC 2026: IEEE Transactions on Mobile Computing (IEEE TMC). Kim Hammar, Tansu Alpcan, and Emil C. Lupu. IEEE Proceedings, ArXiv(preprint), bibtex.

[GameSec] Multiagent Incident Response Planning with Code Models; Conference on Game Theory and AI for Security (GameSec), Ann Harbor, Michigan, USA, October 26th-28th 2026. Kim Hammar, Tansu Alpcan, and Emil C. Lupu. Springer book, bibtex.

[CDC] Optimal Stopping of Self-Refining Foundation Models; 65th IEEE Conference on Decision and Control, Honolulu, Hawaii, USA, Dec 15th-18th 2026. Kim Hammar, Tansu Alpcan, and Emil C. Lupu. To appear. ArXiv(preprint), bibtex.

[ESORICS] Agentic Incident Response through Digital Twin-Enhanced Multiscale Planning; 31st European Symposium on Research in Computer Security (ESORICS) 2026. Yiran Gao, Tao Li, and Kim Hammar. To appear. ArXiv(preprint), bibtex.

[MLSys] CSLE: A Reinforcement Learning Platform for Autonomous Security Management; Ninth Annual Conference on Machine Learning and Systems (MLSys 2026) 2026, Bellevue, WA, USA, May 18th-22nd 2026. Kim Hammar. Proceedings, presentation, arXiv(preprint), bibtex.

[AAAI symposium] In-Context Autonomous Network Incident Response: An End-to-End Large Language Model Agent Approach; AAAI summer symposium 2026, Seoul, South Korea, June 22-24 2026. Yiran Gao, Kim Hammar, and Tao Li. AAAI proceedings, arXiv(preprint), bibtex.

[NOMS] Hallucination-Resistant Security Planning with a Large Language Model; IEEE/IFIP Network Operations and Management Symposium (NOMS) 2026, Rome, Italy, May 18-22 2026. Kim Hammar, Tansu Alpcan, and Emil C. Lupu. IEEE Proceedings, DEMO, ArXiv(preprint), bibtex.

[NOMS] Optimal Incident Response in IT Systems; IEEE/IFIP Network Operations and Management Symposium (NOMS) 2026, Rome, Italy, May 18-22 2026. Kim Hammar. IEEE Proceedings, bibtex.

[NDSS] Incident Response Planning Using a Lightweight Large Language Model with Reduced Hallucination; Network and Distributed System Security (NDSS) Symposium 2026, San Diego, USA, February 23-27 2026. Kim Hammar, Tansu Alpcan, and Emil C. Lupu. Proceedings, arXiv(preprint), NDSS Video, news article, bibtex.

[AAAI] Scalable Solutions to Zero-Sum Partially Observable Stochastic Games Through Belief Aggregation with Approximation Guarantees; 40th AAAI Conference on Artificial Intelligence 2026, Singapore, January 20-27 2026. Kim Hammar and Tansu Alpcan. AAAI Proceedings, AAAI Video, Poster, bibtex.

2025

[TNSM] Online Identification of IT Systems through Active Causal Learning; TNSM 2026: IEEE Transactions on Network and Service Management (IEEE TNSM). Kim Hammar and Rolf Stadler. IEEE Proceedings, ArXiv(preprint), bibtex.

[TIFS] Adaptive Network Security Policies via Belief Aggregation and Rollout; TIFS 2025: IEEE Transactions on Information Forensics and Security (IEEE TIFS). Kim Hammar, Yuchao Li, Tansu Alpcan, Emil C. Lupu, and Dimitri Bertsekas. IEEE Proceedings, arXiv(preprint), bibtex.

[Submitted] Feature-Based Belief Aggregation for Partially Observable Markov Decision Problems; submitted. Yuchao Li, Kim Hammar, and Dimitri Bertsekas. ArXiv(preprint), bibtex.

[TIFS] Adaptive Security Response Strategies through Conjectural Online Learning; TIFS 2025: IEEE Transactions on Information Forensics and Security (IEEE TIFS). Kim Hammar, Tao Li, Rolf Stadler, and Quanyan Zhu. IEEE Proceedings, arXiv(preprint), bibtex.

[AiSec] Online Incident Response Planning under Model Misspecification through Bayesian Learning and Belief Quantization; ACM CCS AiSec 2025: 18 th ACM Workshop on Artificial Intelligence and Security, Taipei, Taiwan, Oct 17. Kim Hammar and Tao Li. ACM Proceedings, arXiv(preprint), Poster, bibtex.

2024

[TNSM] Optimal Defender Strategies for CAGE-2 using Causal Modeling and Tree Search; TNSM 2026: IEEE Transactions on Network and Service Management (IEEE TNSM). Kim Hammar, Neil Dhir, and Rolf Stadler. IEEE Proceedings, ArXiv(preprint), bibtex.

[GameSec] Intrusion Tolerance as a Two-Level Game; GameSec 2024: International Conference on Decision and Game Theory for Security, New York, USA, Oct 16-18. Kim Hammar and Rolf Stadler. Springer book chapter, bibtex.

[CDC] Conjectural Online Learning with First-order Beliefs in Asymmetric Information Stochastic Games; 63rd IEEE Conference on Decision and Control (CDC’24), Milan, Italy, December 16-19, 2024. Tao Li, Kim Hammar, Rolf Stadler, and Quanyan Zhu. IEEE Proceedings, arXiv(preprint), bibtex.

[DSN] Intrusion Tolerance for Networked Systems through Two-Level Feedback Control; 54th IEEE/IFIP Dependable Systems and Networks Conference (DSN’24), Brisbane, Australia, June 24-27, 2024. Kim Hammar and Rolf Stadler. IEEE Proceedings, arXiv(preprint), bibtex, Poster.

[NOMS] Online Policy Adaptation for Networked Systems using Rollout; NOMS 2024: IEEE/IFIP Network Operations and Management Symposium, Seoul, South Korea May 6-10. Forough Shahab Samani, Kim Hammar and Rolf Stadler. IEEE Proceedings, PDF, bibtex.

[Poster] Intrusion Tolerance for Networked Systems through Two-Level Feedback Control; CDIS spring conference 2024, Stockholm, Sweden May 22. Kim Hammar and Rolf Stadler. PDF.

[PhD thesis] Optimal Security Response to Network Intrusions in IT Systems; Doctoral thesis in electrical engineering, KTH, School of Electrical Engineering and Computer Science (EECS). Kim Hammar. The thesis was defended for the Degree of Doctor of Philosophy on Thursday, the 5th December 2024, at 14:00 in F3, Lindstedtsvägen 26, KTH Royal Institute of Technology, Stockholm. Author: Kim Hammar. Supervisor: Professor Rolf Stadler, KTH. Opponent: Professor Tansu Alpcan, The University of Melbourne. Grading committee: Professor Emil Lupu, Imperial College London; Professor Alina Oprea, Northeastern University; and Professor Karl H. Johansson, KTH. Reviewer: Professor Henrik Sandberg, KTH. ArXiv version, publication, PDF, bibtex, ai_generated_summary.

2023

[TNSM] Learning Near-Optimal Intrusion Responses Against Dynamic Attackers; TNSM 2023: IEEE Transactions on Network and Service Management (IEEE TNSM). Kim Hammar and Rolf Stadler. IEEE Proceedings, arXiv(preprint), PDF(preprint), bibtex.

[GameSec] Scalable Learning of Intrusion Responses through Recursive Decomposition; GameSec 2023: International Conference on Decision and Game Theory for Security, Avignon, France, Oct 18-20. Kim Hammar and Rolf Stadler. Springer book chapter, arXiv(preprint), bibtex.

[NOMS] Digital Twins for Security Automation; NOMS 2023: IEEE/IFIP Network Operations and Management Symposium, Miami, USA May 8-12. Kim Hammar and Rolf Stadler. IEEE Proceedings, PDF(preprint), bibtex.

[NOMS] Demonstrating a System for Dynamically Meeting Management Objectives on a Service Mesh; NOMS 2023: IEEE/IFIP Network Operations and Management Symposium, Miami, USA May 8-12. Forough Shahab Samani, Kim Hammar and Rolf Stadler. IEEE Proceedings, PDF(preprint), bibtex, Poster.

[Technical report] Optimal Observation-Intervention Trade-Off in Optimisation Problems with Causal Structure; 2023. Kim Hammar and Neil Dhir. ArXiv(preprint), bibtex.

[Poster] Learning Near-Optimal Intrusion Responses Against Dynamic Attackers; CDIS spring conference 2023, Stockholm, Sweden May 25. Kim Hammar and Rolf Stadler. PDF.

2022

[TNSM] Intrusion Prevention through Optimal Stopping; TNSM 2022: IEEE Transactions on Network and Service Management (IEEE TNSM), special issue on recent advances in network security management. Kim Hammar and Rolf Stadler. IEEE Proceedings, arXiv(preprint), PDF(preprint), bibtex.

[CNSM] An Online Framework for Adapting Security Policies in Dynamic IT Environments; CNSM 2022: International Conference on Network and Service Management, Thessaloniki, Greece October 31 - November 4. Kim Hammar and Rolf Stadler. IEEE Proceedings, IFIP Open Library Conference Proceedings, IFIP Open Library PDF, PDF(preprint), bibtex.

[ICML] Learning Security Strategies through Game Play and Optimal Stopping; ICML ML4Cyber Workshop 2022: International Conference on Machine Learning, Baltimore, USA July 17-23. Kim Hammar and Rolf Stadler. PDF(extended version preprint), Camera ready version.

[NOMS] A System for Interactive Examination of Learned Security Policies; (Best demonstration paper award); Diploma; NOMS 2022: IEEE/IFIP Network Operations and Management Symposium, Budapest, Hungary April 25-29. Kim Hammar and Rolf Stadler. Video, IEEE Proceedings, PDF(preprint), arXiv(preprint), bibtex.

[Poster] Intrusion Prevention through Optimal Stopping; KTH EECS Summer Conference 8 June 2022. Kim Hammar and Rolf Stadler. PDF.

[Poster] Intrusion Prevention through Optimal Stopping; CDIS Spring Conference 24 May 2022. Kim Hammar and Rolf Stadler. PDF.

[Poster] Intrusion Prevention through Optimal Stopping; Digital Futures Machine Learning Day 17 Jan 2022. Kim Hammar and Rolf Stadler. PDF.

2021

[CNSM] Learning Intrusion Prevention Policies through Optimal Stopping; CNSM 2021: International Conference on Network and Service Management, Izmir, Turkey October 25-29. Kim Hammar and Rolf Stadler. IFIP Open Library Conference Proceedings, IFIP Open Library PDF, IEEE Proceedings, PDF, arXiv(preprint), PDF(preprint), bibtex.

[Poster] Learning Intrusion Prevention Policies through Optimal Stopping; CIFAR Deep Learning + Reinforcement Learning (DLRL) Summer School 2021. Kim Hammar and Rolf Stadler. PDF.

[Report] An RCE Exploit of a Remote SLD Resolver in Prolog; 18 June 2021. Kim Hammar. PDF.

2020

[Web Intelligence] Deep text classification of Instagram data using word embeddings and weak supervision; Web Intelligence 2020. Kim Hammar, Shatha Jaradat, Nima Dokoohaki and Mihhail Matskin. Journal article, PDF, bibtex.

[CNSM] Finding Effective Security Strategies through Reinforcement Learning and Self-Play; CNSM 2020: International Conference on Network and Service Management, Izmir, Turkey November 2-6. Kim Hammar and Rolf Stadler. IEEE Proceedings, IFIP Open Library Conference Proceedings, arXiv(preprint), PDF(preprint), bibtex.

[Report] Using Reinforcement Learning in Self-Driving Systems; 28 May 2020. A PoC. Forough Shahab and Kim Hammar. PDF.

2019

[COMPSAC] TALS: A Framework For Text Analysis, Fine-Grained Annotation, Localisation and Semantic Segmentation; COMPSAC 2019: Data Driven Intelligence for a Smarter World Hosted by Marquette University, Milwaukee, Wisconsin, USA July 15-19. Shatha Jaradat, Nima Dokoohaki, Ummal Wara, Mallu Goswami, Kim Hammar and Mihhail Matskin. IEEE Proceedings, bibtex.

[MLSys] Horizontally Scalable ML Pipelines with a Feature Store; Demo track, MLSys Conference, March 31 - April 2 2019, Stanford CA. Alexandru A. Ormenisan, Mahmoud Ismail, Kim Hammar, Robin Andersson, Ermias Gebremeskel, Theofilos Kakantousis, Antonios Kouzoupis, Fabio Buso, Gautier Berthou, Jim Dowling and Seif Haridi. Proceedings, PDF.

[Blog post] Guide to File Formats for Machine Learning: Columnar, Training, and Inferencing; 25 Oct 2019. Jim Dowling, Moritz Meister, and Kim Hammar. Blog post.

2018

[SocialCom] Dynamic CNN Models For Fashion Recommendation in Instagram; International Conference on Social Computing and Networking (SocialCom 2018), 11-13 Dec. 2018, Melbourne, Australia. Shatha Jaradat, Nima Dokoohaki, Kim Hammar, Ummul Wara and Mihhail Matskin. DiVA, bibtex.

[Web Intelligence] Deep Text Mining of Instagram Data Without Strong Supervision; Web Intelligence Conference 2018, Santiago Chile. Kim Hammar, Shatha Jaradat, Nima Dokoohaki and Mihhail Matskin. IEEE Proceedings, PDF, ArXiv, bibtex.

[Blog post] Feature Store: the missing data layer in ML pipelines?; 30 Dec 2018. Kim Hammar and Jim Dowling. Blog post, bibtex.

[Blog post] Goodbye Horovod, Hello CollectiveAllReduce; 22 Oct 2018. Robin Andersson, Jim Dowling, Ermias Gebremeskel and Kim Hammar. Blog post.

[MSc thesis] Deep Text Mining of Instagram Data Without Strong Supervision; Master’s Thesis, KTH School of Information and Communication Technology (ICT). Kim Hammar. Publication, PDF, bibtex.

[Tutorial] Programming the Semantic Web; 12 Jan 2018. A Tutorial. Kim Hammar. PDF.

2017

[Course project] Distributed Human Activity Recognition; Dec 2017. Scalable Deep Learning Course Project at KTH. (Best project award, awarded a graphics card). Kim Hammar and Konstantin Sozinov. Code.

[Report] Machine Learning for Failure Detection in Distributed Systems; 23 Nov 2017. A project report. Kim Hammar and Konstantin Sozinov. PDF.

[Report] Conflict free p2p replicated datatypes; 22 May 2017. A project report. Kim Hammar and Maxime Dufour. PDF.

[Report] Linearizable Key-Value Store; 12 Mar 2017. A project report. Kim Hammar and Konstantin Sozinov. PDF.

2016

[BSc thesis] Integrating Monitoring Systems - Pre-Study; Bachelor’s Thesis, KTH School of Information and Communication Technology (ICT). Kim Hammar and Marcus Blom. Publication, PDF, bibtex.