Hamed Haddadi
Professor of Human-Centred Systems
Builds privacy-preserving, user-centred systems for edge and IoT devices, including attestable and private on-device machine learning.
Hamed Haddadi's lab at Oxford works on efficient training & inference and distributed, cloud & edge systems. 60 papers since 2023, 2 at top venues, mostly JAMIA and USENIX Security. 13 alumni. Applications close .
At a glance
- papers, 2023–2026
- 16/ year
- at T1 venues
- 3%
- PhD students
- 0
- MS students
- 0
- alumni listed
- 13
Papers by yearfrom the lab homepage, 2023–2026
- 14 papers2026 · T1 0 · collecting
- 20 papers2025 · T1 1
- 13 papers2024 · T1 0
- 13 papers2023 · T1 1
T1 venuesOtherVenues: JAMIA, USENIX Security
Research focuspaper titles that use each topic's words
- Efficient training & inference8 papers
- Distributed, cloud & edge systems8 papers
- private4 papers
- on-device2 papers
- privacy-preserving2 papers
Paper list
From the lab homepage.
202614 papers
- The Workshop on Machine Learning and Systems (EuroMLSys ), April 27, , Edinburgh, United KingdomAgenTEE: Confidential LLM Agent Execution on Edge Devices (opens in a new tab)
- ACM Conference on AI and Agentic Systems (CAIS ), San Jose, California, May 26-29AgentStop: Terminating Local AI Agents Early to Save Energy in Consumer Devices (opens in a new tab)
- IEEE European Symposium on Security and Privacy (EuroS&P ), Lisbon, July 6 - 10CAEC: Confidential, Attestable, and Efficient Inter‑CVM Communication with Arm CCA (opens in a new tab)
- IEEE Conference on Standards for Communications and Networking (CSCN ), 7 - 9 September , London, United KingdomCERTIoT-6G: Continuous Cybersecurity Certification for IoT Devices in 5G/6G Networks (opens in a new tab)
- The Privacy Enhancing Technologies Symposium (PETS ), July 20-25, , Calgary, CanadaDynamic Probabilistic Noise Injection for Membership Inference Defense (opens in a new tab)
- The Annual Meeting of the Association for Computational Linguistics (ACL ), San Diego, California, United States, July 2-7, . [How Adversarial Environments Mislead Agentic AI? (opens in a new tab)
- The Workshop on Operating Systems Design for AI Agents (Agentic OS ), September 29, , Prague, CzechiaPatient Bytes: A Reliability-Budgeted Scheduler for Agentic LLM Workflows (opens in a new tab)
- ACM/IEEE International Conference on Embedded Artificial Intelligence and Sensing Systems (SenSys ), May 11-14, Saint-Malo, FrancePhysioLite: Towards Real-Time Physiological Monitoring on Constrained Wearable μNPUs (opens in a new tab)
- The ACM SIGOPS Annual Technical Conference (ATC ), November 15-18, , Shatin, Hong KongSharing is caring: Attestable and Trusted Cloud Pipelines out of Distrustful Components (opens in a new tab)
- The Workshop on Machine Learning and Systems (EuroMLSys ), April 27, , Edinburgh, United KingdomSystems-Level Attack Surface of Edge Agent Deployments on IoT (opens in a new tab)
- The Workshop on Machine Learning and Systems (EuroMLSys ), April 27, , Edinburgh, United KingdomTowards Graph-Based Detection of Jailbreak and Prompt-Leakage Attacks in LLMs (opens in a new tab)
- AAAI WorkshopTrustRAG: Enhancing Robustness and Trustworthiness in Retrieval-Augmented Generation (opens in a new tab)
202520 papers
- IEEE International Conference on Web Services (ICWS ), July 7-12, Helsinki, FinlandACCESS-FL: Agile Communication and Computation for Efficient Secure Aggregation in Stable Networks for FLaaS (opens in a new tab)
- Workshop on System Software for Trusted Execution SysTEX, Venice, Italy, June 30An Early Experience with Confidential Computing Architecture for On-Device Model Protection (opens in a new tab)
- The ACM Annual International Conference On Mobile Computing And Networking (MobiCom 25), Nov. 4-8, , Hong Kong, ChinaBenchmarking Ultra-Low-Power μ NPUs (opens in a new tab)
- IEEE Annual Congress on Artificial Intelligence of Things (AIoT ), 3-5 Dec , Osaka, JapanClient Clustering Meets Knowledge Sharing: Enhancing Privacy and Robustness in Personalized Peer-to-Peer Learning (opens in a new tab)
- The Conference on Empirical Methods in Natural Language Processing (EMNLP), November 5-9 , Suzhou, ChinaContext-Aware Membership Inference Attacks against Pre-trained Large Language Models (opens in a new tab)
- IEEE European Symposium on Security and Privacy, Venice, Italy, June 30 - July 4Deep Unlearn: Benchmarking Machine Unlearning for Image Classification (opens in a new tab)
- The Network and Distributed System Security (NDSS) Symposium , 23-28 February , San Diego, California. DistinguishedDiStefano: Decentralized Infrastructure for Sharing Trusted Encrypted Facts and Nothing More (opens in a new tab)
- International Workshop on Traffic Measurements for Cybersecurity WTMC , Venice, Italy, July 4Intelligent Detection of Non-Essential IoT Traffic on the Home Gateway (opens in a new tab)
- ACM Conference on Computer and Communications Security (CCS), October 13-17, , Taipei, TaiwanLocal Frames: Exploiting Inherited Origins to Bypass Content Blockers (opens in a new tab)
- Privacy Enhancing Technologies Symposium PETS , July 14-19, , Washington, DC and OnlineMeasuring the Accuracy and Effectiveness of PII Removal Services (opens in a new tab)
- The Conference on Empirical Methods in Natural Language Processing (EMNLP), November 5-9 , Suzhou, ChinaMembership and Memorization in LLM Knowledge Distillation (opens in a new tab)
- ACM Conference on Computer and Communications Security (CCS), October 13-17, , Taipei, TaiwanNebula: Efficient, Private and Accurate Histogram Estimation (opens in a new tab)
202413 papers
- ACM Internet Measurement Conference , November , Madrid, SpainA First Look at Related Website Sets (opens in a new tab)
- ICASSP WorkshopA Study on the Impact of Self-Supervised Learning on Automatic Dysarthric Speech Assessment (opens in a new tab)
- Artificial Intelligence in Medicine, ,102821, ISSN 0933-3657, https://doi.org/10.1016/j.artmed. .102821Analyzing entropy features in time-series data for pattern recognition in neurological conditions
- The Twelfth International Conference on Learning Representations ICLR , May , Vienna, Austria. SpotlightConfidential-DPproof: Confidential Proof of Differentially Private Training (opens in a new tab)
- IEEE AccessEfficient Anomaly Detection for Edge Clouds: Mitigating Data and Resource Constraints (opens in a new tab)
- The Workshop on Machine Learning and Systems EuroMLSys , co-located with EuroSys’24, April , Athens, GreeceGuaranTEE: Towards Attestable and private ML with CCA (opens in a new tab)
- The ACM/IEEE Symposium on Edge Computing (SEC ), December 4-7, Rome, ItalyLow-Energy On-Device Personalization for MCUs (opens in a new tab)
- ACM Computing SurveysMachine Learning with Confidential Computing: A Systematization of Knowledge (opens in a new tab)
- The ACM Annual International Conference On Mobile Computing And Networking (MobiCom 24), Nov. 18-22, , Washington, D.C., USAMELTing point: Mobile Evaluation of Language Transformers (opens in a new tab)
- The Second Workshop on Generative Information Retrieval (Gen-IR@SIGIR24), July 18, - Washington DC, USAMitigating Hallucinations in Large Language Models via Self-Refinement-Enhanced Knowledge Retrieval (opens in a new tab)
- Passive and Active Measurement Conference PAM , March , VirtualSunBlock: Cloudless Protection for IoT Systems (opens in a new tab)
- The Workshop on Machine Learning and Systems EuroMLSys , co-located with EuroSys’24, April , Athens, GreeceTowards Low-Energy Adaptive Personalization for Resource-Constrained Devices (opens in a new tab)
202313 papers
- ACM Internet Measurement Conference , October , Montreal, CanadaA First Look at the Privacy Harms of the Public Suffix List (opens in a new tab)
- UbiComp/ISWC Workshop on Fairness and Robustness in Machine Learning for Ubiquitous Computing (FairComp ), October 9, , Cancún, MexicoAnalysing Fairness of Privacy-Utility Mobility Models (opens in a new tab)
- JAMIA (T1)Federated and Distributed Learning Applications for Electronic Health Records and Structured Medical Data: A Scoping Review (opens in a new tab)
- IEEE International Conference on Distributed Computing Systems (ICDCS ), 18 July - 21 July, , Hong Kong, ChinaGT-TSCH: Game-Theoretic Distributed TSCH Scheduler for Low-Power IoT Networks (opens in a new tab)
- —Information Theory Inspired Pattern Analysis for Time-series Data (opens in a new tab)
- the USENIX Security Symposium USENIX Security ‘23, AUGUST 9-11, , Anaheim, CA, USAPool-Party: Exploiting Browser Resource Pools as Side-Channels for Web Tracking
- The ACM International Conference on Mobile Systems, Applications, and Services (MobiSys ), June 18 - June 22, , Helsinki, FinlandPoster: Towards Battery-Free Machine Learning Inference and Model Personalization on MCUs (opens in a new tab)
- IEEE International Conference on ICT Solutions for eHealth (ICTS4eHealth ), in conjunction with the IEEE Symposium on Computers and Communications (ISCC) 9 -12 July, - Tunis, TunisiaPRISM: Privacy Preserving Healthcare Internet of Things Security Management (opens in a new tab)
- The Privacy Enhancing Technologies Symposium (PETS ), July 10-14, , Lausanne, SwitzerlandPrivacy-Aware Adversarial Network in Human Mobility Prediction (opens in a new tab)
- IEEE Symposium on Security and Privacy (Oakland ), May 22-25, , San Francisco, CA, & OnlineProtected or Porous: A Comparative Analysis of Threat Detection Capability of IoT Safeguards
- IFIP/IEEE Network Traffic Measurement and Analysis Conference (TMA ), June 26 - June 29, , Napoli, ItalyProtecting against IP Tracking using NAT Interfaces (opens in a new tab)
- DecemberThe Boomerang protocol: A Decentralised Privacy-Preserving Verifiable Incentive Protocol (opens in a new tab)
Where alumni went13 listed
Apply to Oxford2027-28
DPhil in Computer Science
- Round
- 2027-28 entry (January round)
- Opens
- —
- GRE
- See program page
- English
- See program page
- Fee
- See program page
Not yet confirmed for this intake.
Similar labsnearest by research summary
- CNSJTU
Guohao Dai
- Efficient training & inference
- USGeorgia Tech
Alexey Tumanov
- Efficient training & inference
- CNSJTU
Jieru Zhao
22 T1 - CNSJTU
Xiaofeng Hou
- Accelerators, chips & architecture
- USGeorgia Tech
Joy Arulraj
8 T1 - CNSJTU
Yu Feng
- Efficient training & inference
Your notes
Save this lab to keep notes.
Data
Checked against the lab homepage.
- Collected
- Verified
- Papers
- The lab homepage
- Members
- Lab homepage
- Alumni
- Lab homepage
Tiers are venue classifications, not lab rankings.
Oxfordcloses