Ising machines are emerging as specialized hardware solvers for computationally hard optimization problems. This review examines five major platforms—digital CMOS, analog CMOS, emerging devices, coherent optics, and quantum systems—highlighting physics‐rooted advantages and shared bottlenecks in scalability and connectivity.
Hyunjun Lee, Joon Pyo Kim, Sanghyeon Kim
wiley +1 more source
DynamicFU: Contribution-Aware Dynamic Federated Unlearning for Industrial IoT. [PDF]
Wu Z, He B, Si Z, Liao X, Su C.
europepmc +1 more source
Robust and Compatible Ferroelectric Memories with Polycrystalline TiO2 Channel for 3D Integration
Robust and monolithic 3D compatible ferroelectric memories are realized using the polycrystalline TiO2 channel‐based FeFET. The review covers physical mechanisms of the TiO2 channel FeFET, quantitative benchmarking, and advanced planar/vertical architectures for monolithic 3D integration based on HfO2‐TiO2 gate stack, offering a roadmap for reliable ...
Xujin Song +10 more
wiley +1 more source
Evaluation Protocol Sensitivity in Frequency-Constrained Truss Optimization: A Comparative Study of Adaptive and Evolutionary Metaheuristics. [PDF]
Ghiaskar A, Taheri F.
europepmc +1 more source
This work electrically characterizes sixteen logic gates built from three‐independent‐gate reconfigurable transistors fabricated on full‐scale 300 mm wafers using the industrial 22 nm fully depleted FDSOI process of GlobalFoundries. Static and time‐resolved measurements confirm correct operation, including a 1‐bit adder and reconfigurable AOI/OAI ...
Juan P. Martinez +12 more
wiley +1 more source
Class-adaptive oracle-free metamorphic test case prioritization framework for vision-based deep neural networks. [PDF]
Li J, Mohamad R, Ahmad J.
europepmc +1 more source
Through the introduction of a niobium oxide layer into a hafnia ferroelectric capacitor stack, we build a memory device with a strong imprint effect. This imprint leads to a millisecond retention loss that can be tuned by the programming conditions that can be utilized as a scalable, analog hardware time constant for bio‐inspired temporal computing ...
Luca Fehlings +3 more
wiley +1 more source
Thermodynamic natural gradient descent (NGD-T) regulates natural-gradient steps by a geometric speed-cost bound. [PDF]
You BJ.
europepmc +1 more source
Emerging Memory and Device Technologies for Hardware‐Accelerated Model Training and Inference
This review investigates the suitability of various emerging memory technologies as compute‐in‐memory hardware for artificial intelligence (AI) applications. Distinct requirements for training‐ and inference‐centric computing are discussed, spanning device physics, materials, and system integration.
Yoonho Cho +6 more
wiley +1 more source
Phengite-mediated fluorine and chlorine fluxes from subduction zones to the deep mantle. [PDF]
Liu Y, Wang Y, Chen C, Gao M, Xu YG.
europepmc +1 more source

