<?xml version="1.0" encoding="UTF-8"?><rss version="2.0"><channel><title>Roozbeh Siyadatzadeh: notes and publications</title><description>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.</description><link>https://roozbeh.nl/</link><language>en</language><item><title>P2W: From Power Traces to Weights Matrix - An Unconventional Transfer Learning Approach (ACM SAC 2026)</title><link>https://roozbeh.nl/publications/p2w-power-traces-to-weights/</link><guid isPermaLink="true">https://roozbeh.nl/publications/p2w-power-traces-to-weights/</guid><description>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.</description><pubDate>Mon, 23 Mar 2026 00:00:00 GMT</pubDate><category>publication</category><category>Embedded machine learning</category><category>Power side channels</category><category>Transfer learning</category><category>Model extraction</category><category>System-on-chip</category></item><item><title>A Security Analysis of CNN Partitioning Strategies for Distributed Inference at the Edge (SPACE 2025 (LNCS))</title><link>https://roozbeh.nl/publications/security-analysis-cnn-partitioning-edge/</link><guid isPermaLink="true">https://roozbeh.nl/publications/security-analysis-cnn-partitioning-edge/</guid><description>Compares vertical, horizontal, and data partitioning of CNNs across edge devices against an eavesdropper who reconstructs the input image from intercepted intermediate data.</description><pubDate>Thu, 12 Feb 2026 00:00:00 GMT</pubDate><category>publication</category><category>Distributed edge AI</category><category>CNN security</category><category>Model partitioning</category><category>Input reconstruction</category><category>Privacy</category></item><item><title>TraceFormer: A Transformer-Based Method for Weight Extraction from AIMC Tiles (Euromicro DSD 2025)</title><link>https://roozbeh.nl/publications/traceformer-weight-extraction-aimc/</link><guid isPermaLink="true">https://roozbeh.nl/publications/traceformer-weight-extraction-aimc/</guid><description>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.</description><pubDate>Wed, 10 Sep 2025 00:00:00 GMT</pubDate><category>publication</category><category>Analog in-memory computing</category><category>Hardware security</category><category>Power side channels</category><category>Transformers</category><category>Weight extraction</category></item><item><title>Security Challenges in Analog In-Memory Computing</title><link>https://roozbeh.nl/blog/imc-security/</link><guid isPermaLink="true">https://roozbeh.nl/blog/imc-security/</guid><description>How ADC power leakage can expose weights stored in analog in-memory-computing accelerators, and what that means for secure edge AI.</description><pubDate>Sun, 15 Dec 2024 00:00:00 GMT</pubDate><category>note</category><category>hardware-security</category><category>aimc</category><category>side-channel-analysis</category></item><item><title>From Power Traces to Model Weights: P2W and TraceFormer</title><link>https://roozbeh.nl/blog/weight-extraction/</link><guid isPermaLink="true">https://roozbeh.nl/blog/weight-extraction/</guid><description>Two research methods for learning about embedded neural-network weights through physical power side channels.</description><pubDate>Thu, 28 Nov 2024 00:00:00 GMT</pubDate><category>note</category><category>hardware-security</category><category>embedded-ml</category><category>side-channel-analysis</category></item><item><title>Reinforcement Learning for Reliable Real-Time Fog Systems</title><link>https://roozbeh.nl/blog/rl-realtime/</link><guid isPermaLink="true">https://roozbeh.nl/blog/rl-realtime/</guid><description>How ReLIEF assigns primary and backup tasks while balancing deadlines, failures, communication delay, and fog-node workload.</description><pubDate>Sat, 12 Oct 2024 00:00:00 GMT</pubDate><category>note</category><category>reinforcement-learning</category><category>real-time-systems</category><category>fog-computing</category></item><item><title>Phase-Change Memory for Analog In-Memory Computing</title><link>https://roozbeh.nl/blog/pcm-future/</link><guid isPermaLink="true">https://roozbeh.nl/blog/pcm-future/</guid><description>A practical introduction to PCM conductance, neural-network acceleration, engineering trade-offs, and physical security.</description><pubDate>Thu, 05 Sep 2024 00:00:00 GMT</pubDate><category>note</category><category>phase-change-memory</category><category>aimc</category><category>hardware-security</category></item><item><title>My Journey from Bushehr to Leiden</title><link>https://roozbeh.nl/blog/academic-journey/</link><guid isPermaLink="true">https://roozbeh.nl/blog/academic-journey/</guid><description>Reflections on my academic path from Bushehr to the Netherlands.</description><pubDate>Thu, 18 Jul 2024 00:00:00 GMT</pubDate><category>note</category><category>personal</category><category>academia</category></item><item><title>ReLIEF: A Reinforcement-Learning-Based Real-Time Task Assignment Strategy in Emerging Fault-Tolerant Fog Computing (IEEE IoT Journal)</title><link>https://roozbeh.nl/publications/relief-rl-task-assignment-fog/</link><guid isPermaLink="true">https://roozbeh.nl/publications/relief-rl-task-assignment-fog/</guid><description>Reinforcement learning assigns primary and backup copies of real-time tasks to fog nodes, balancing communication delay against the workload on each device.</description><pubDate>Thu, 15 Jun 2023 00:00:00 GMT</pubDate><category>publication</category><category>Reinforcement learning</category><category>Fog computing</category><category>Real-time systems</category><category>Fault tolerance</category><category>Resource allocation</category></item><item><title>ATLAS: Aging-Aware Task Replication for Multicore Safety-Critical Systems (IEEE RTAS 2023)</title><link>https://roozbeh.nl/publications/atlas-aging-aware-task-replication/</link><guid isPermaLink="true">https://roozbeh.nl/publications/atlas-aging-aware-task-replication/</guid><description>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.</description><pubDate>Tue, 09 May 2023 00:00:00 GMT</pubDate><category>publication</category><category>Real-time systems</category><category>Safety-critical systems</category><category>Task replication</category><category>Processor aging</category><category>Thermal management</category></item><item><title>Passive Primary/Backup-Based Scheduling for Simultaneous Power and Reliability Management on Heterogeneous Embedded Systems (IEEE T-SUSC)</title><link>https://roozbeh.nl/publications/passive-primary-backup-scheduling-heterogeneous/</link><guid isPermaLink="true">https://roozbeh.nl/publications/passive-primary-backup-scheduling-heterogeneous/</guid><description>Passive primary/backup scheduling on heterogeneous multicores that mixes task mapping across core types and runs backups after primaries to remove execution overlap.</description><pubDate>Sun, 01 Jan 2023 00:00:00 GMT</pubDate><category>publication</category><category>Heterogeneous multicore systems</category><category>Real-time scheduling</category><category>Fault tolerance</category><category>Energy efficiency</category><category>Reliability</category></item></channel></rss>