# Roozbeh Siyadatzadeh

> Roozbeh Siyadatzadeh is a PhD candidate at Leiden University working on hardware security for embedded AI: power side-channel analysis of analog in-memory computing, model extraction, FPGA and RISC-V systems, and dependable real-time computing.

Roozbeh Siyadatzadeh (ORCID 0000-0002-2258-4110) is a PhD candidate in the Systems and Security group of the Leiden Institute of Advanced Computer Science (LIACS), Leiden University, The Netherlands. Since 2022 the work has been part of NeuroSoC, a Horizon Europe project (grant 101070634) whose consortium includes IBM, STMicroelectronics, Bosch, Thales, Ubotica and 9 further partners. Role: side-channel security assessment of the project's PCM-based analog in-memory computing system-on-chip.

This file follows the llms.txt convention. Everything linked below is public. Static pages use canonical descriptive URLs on https://roozbeh.nl; Jarvis lives at https://roozbeh.app/. Content updated 2026-09-06.

## Start here

- [Home](https://roozbeh.nl/): who I am, research focus, the PhD project, selected work
- [Curriculum vitae](https://roozbeh.nl/cv/): education, positions, publications, teaching, skills, awards ([PDF](https://roozbeh.nl/cv.pdf))
- [Publications](https://roozbeh.nl/publications/): all 6 peer-reviewed papers with abstracts, PDFs, BibTeX, and AI-generated overviews
- [NeuroSoC project](https://roozbeh.nl/projects/neurosoc/): the PhD project, consortium, contributions, timeline
- [Projects](https://roozbeh.nl/projects/): all research, engineering, and course projects
- [Notes](https://roozbeh.nl/blog/): plain-language explanations of the research
- [Memory map](https://roozbeh.nl/map/): every page with its stable 16-bit address and short link
- [Jarvis](https://roozbeh.app/): assistant that answers questions about this work from the same records (20 messages per visitor per day)

## Publications

- [P2W: From Power Traces to Weights Matrix - An Unconventional Transfer Learning Approach](https://roozbeh.nl/publications/p2w-power-traces-to-weights/): Turns power traces captured from an embedded SoC into an approximate weight matrix that initialises a new model when the original model and its training data are out of reach. (ACM SAC 2026, 2026; DOI 10.1145/3748522.3779939; PDF https://roozbeh.nl/papers/P2W.pdf)
- [A Security Analysis of CNN Partitioning Strategies for Distributed Inference at the Edge](https://roozbeh.nl/publications/security-analysis-cnn-partitioning-edge/): Compares vertical, horizontal, and data partitioning of CNNs across edge devices against an eavesdropper who reconstructs the input image from intercepted intermediate data. (SPACE 2025 (LNCS), 2026; DOI 10.1007/978-3-032-16342-4_15; PDF https://roozbeh.nl/papers/SPACE_25.pdf)
- [TraceFormer: A Transformer-Based Method for Weight Extraction from AIMC Tiles](https://roozbeh.nl/publications/traceformer-weight-extraction-aimc/): Shows that the weights stored in an analog in-memory computing tile can be recovered from the power consumption of its ADCs, using a Transformer and an input-controlled weight-isolation technique. (Euromicro DSD 2025, 2025; DOI 10.1109/DSD67783.2025.00092; PDF https://roozbeh.nl/papers/TraceFormer.pdf)
- [ReLIEF: A Reinforcement-Learning-Based Real-Time Task Assignment Strategy in Emerging Fault-Tolerant Fog Computing](https://roozbeh.nl/publications/relief-rl-task-assignment-fog/): Reinforcement learning assigns primary and backup copies of real-time tasks to fog nodes, balancing communication delay against the workload on each device. (IEEE IoT Journal, 2023; DOI 10.1109/JIOT.2023.3240007; PDF https://roozbeh.nl/papers/ReLIEF.pdf)
- [ATLAS: Aging-Aware Task Replication for Multicore Safety-Critical Systems](https://roozbeh.nl/publications/atlas-aging-aware-task-replication/): Aging-aware task replication that keeps periodic hard real-time tasks at their reliability target over a multicore system’s lifetime while limiting the temperature that accelerates aging. (IEEE RTAS 2023, 2023; DOI 10.1109/RTAS58335.2023.00025)
- [Passive Primary/Backup-Based Scheduling for Simultaneous Power and Reliability Management on Heterogeneous Embedded Systems](https://roozbeh.nl/publications/passive-primary-backup-scheduling-heterogeneous/): Passive primary/backup scheduling on heterogeneous multicores that mixes task mapping across core types and runs backups after primaries to remove execution overlap. (IEEE T-SUSC, 2023; DOI 10.1109/TSUSC.2022.3186656; PDF https://roozbeh.nl/papers/tsusc.pdf)

Citation counts: 74 in total (Semantic Scholar Graph API, 2026-09-02).

## Projects

- [NeuroSoC: security assessment of an in-memory computing SoC](https://roozbeh.nl/projects/neurosoc/): Side-channel security assessment of a PCM-based analog in-memory computing system-on-chip developed by a Horizon Europe consortium of industry and academic partners.
- [Thermal management and task scheduling for multicore systems](https://roozbeh.nl/projects/#project-thermal-aware-task-scheduling): How task placement and scheduling shape the thermal behaviour of a multicore system, evaluated with a gem5, McPAT, and HotSpot workflow.
- [Complex-number ALU on FPGA](https://roozbeh.nl/projects/#project-complex-alu-fpga): A VHDL arithmetic logic unit for complex-number operations, synthesised and verified with Xilinx ISE.
- [Neural-network inference on FPGA](https://roozbeh.nl/projects/#project-deep-learning-on-fpga): An RTL implementation of neural-network inference for resource-constrained edge hardware.
- [Reinforcement-learning DVFS manager for multicore platforms](https://roozbeh.nl/projects/#project-rl-dvfs-manager): A reinforcement-learning controller for dynamic voltage and frequency scaling, evaluated on a simulated multicore with Sniper, McPAT, and HotSpot.
- [Distributed VSOP87 planetary positions with MPI](https://roozbeh.nl/projects/#project-distributed-vsop87): The VSOP87 planetary-position calculation parallelised across MPI workers in C.
- [FPGA CNN for ECG classification](https://roozbeh.nl/projects/#project-fpga-cnn-ecg-classification): A convolutional-neural-network pipeline for ECG classification, from a TensorFlow model to an RTL implementation.
- [Distributed Keras inference over gRPC](https://roozbeh.nl/projects/#project-distributed-keras-inference): A prototype prediction service that coordinates Keras model inference across networked Python processes with gRPC.
- [Facial attribute editing with a GAN](https://roozbeh.nl/projects/#project-facial-attribute-editing-gan): Controlled facial-attribute editing with a generative adversarial network in TensorFlow.
- [Sentiment analysis of social-media comments](https://roozbeh.nl/projects/#project-social-media-sentiment-analysis): Machine-learning sentiment classification of social-media comments in TensorFlow.

## Notes

- [Security Challenges in Analog In-Memory Computing](https://roozbeh.nl/blog/imc-security/): How ADC power leakage can expose weights stored in analog in-memory-computing accelerators, and what that means for secure edge AI.
- [From Power Traces to Model Weights: P2W and TraceFormer](https://roozbeh.nl/blog/weight-extraction/): Two research methods for learning about embedded neural-network weights through physical power side channels.
- [Reinforcement Learning for Reliable Real-Time Fog Systems](https://roozbeh.nl/blog/rl-realtime/): How ReLIEF assigns primary and backup tasks while balancing deadlines, failures, communication delay, and fog-node workload.
- [Phase-Change Memory for Analog In-Memory Computing](https://roozbeh.nl/blog/pcm-future/): A practical introduction to PCM conductance, neural-network acceleration, engineering trade-offs, and physical security.
- [My Journey from Bushehr to Leiden](https://roozbeh.nl/blog/academic-journey/): Reflections on my academic path from Bushehr to the Netherlands.

## Machine-readable

- [llms-full.txt](https://roozbeh.nl/llms-full.txt): the whole site as one Markdown document
- [Open Knowledge Format bundle](https://roozbeh.nl/okf/index.md): one Markdown record with YAML frontmatter per person, publication, project, and note (OKF v0.2)
- [publications.bib](https://roozbeh.nl/publications.bib): BibTeX for every publication (per paper: /publications/<slug>.bib)
- [publications.json](https://roozbeh.nl/publications.json): schema.org ScholarlyArticle records as JSON-LD
- [rss.xml](https://roozbeh.nl/rss.xml): feed of notes and publications
- [sitemap-index.xml](https://roozbeh.nl/sitemap-index.xml): XML sitemap, including paper PDFs
- Every HTML page embeds JSON-LD; publication pages also carry Google Scholar (Highwire Press) citation tags and link to their Markdown and BibTeX records.

## Profiles

- [LinkedIn](https://www.linkedin.com/in/roozbeh24/)
- [GitHub](https://github.com/Roozbeh-Sdtz)
- [Google Scholar](https://scholar.google.com/citations?user=4mz_3_YAAAAJ&hl=en)
- [ORCID](https://orcid.org/0000-0002-2258-4110)
- [ResearchGate](https://www.researchgate.net/profile/Roozbeh-Siyadatzadeh-2)
- [Leiden University](https://www.universiteitleiden.nl/en/staffmembers/roozbeh-siyadatzadeh)
- Email: s.r.siyadatzadeh@liacs.leidenuniv.nl

## Optional

- [robots.txt](https://roozbeh.nl/robots.txt): crawler policy (every crawler is welcome everywhere except the assistant API under /api/)