Resources
Demonstrating how a ReRAM-based platform can support on-chip customized learning in real-time at the edge (Weebit, CEA-Leti and Université Paris-Saclay; 2025 IEEE Symposium on VLSI Technology and Circuits)
Proposing an optimized programming strategy to mitigate the impact of write termination (WT) on relaxation effects in ReRAM, particularly relevant for neuromorphic and in-memory computing applications (IMW 2025)
An analysis of the memristor industry, focusing on products that are commercially available and prototypes that might affect the market in the near future (Nature, April 2025)
How ReRAM’s physical properties can be controlled and how the device can be operated to align with AI and in-memory computing (IMC) requirements (IEEE SOCC 2024)
A comprehensive assessment of the impact of various design assist techniques on the inherent performance and reliability of native ReRAM on silicon (IEEE Journal of Solid State Circuits)
ReRAM for neuromorphic today and over time – from embedded memory for synaptic weight storage, to in-memory computing, to new concepts like Bayesian networks (International Memory Workshop IMW 2024)
Using Weebit ReRAM devices to demonstrate two non-stateful logic techniques to explore Processing in Memory (PIM) possibilities for various applications (IEEE, with CEA-Leti and The Technion)
The benefits of using Weebit ReRAM to provide PUF (Physically Unclonable Function) capability for enhanced on-chip security and cyber-security applications (Weebit, CEA-Leti and IIT Delhi, IMW 2023)
Results on the statistical and dynamic behavior of ReRAM technology and its impact on reliability (Weebit and CEA-Leti, IRPS 2023)
Proposing an event-based hardware based on ReRAM capable of self-adaptation for efficient neurocomputing (Weebit, CEA-Leti, Politechnico di Milano and NaMLab)
A fully-binarized XOR based IMSS (In-Memory Similarity Search) using RRAM (Resistive Random Access Memory) arrays (IEEE Transactions on Circuits and Systems II)
Details of the performance and reliability of Weebit ReRAM integrated in 28nm, highlighting the technology’s intrinsic high reliability (Weebit and CEA-Leti, IMW 2022)