Search for: All records

Abstract Realizing increasingly complex artificial intelligence (AI) functionalities directly on edge devices calls for unprecedented energy efficiency of edge hardware. Compute-in-memory (CIM) based on resistive random-access memory (RRAM) 1 promises to meet such demand by storing AI model weights in dense, analogue and non-volatile RRAM devices, and by performing AI computation directly within RRAM, thus eliminating power-hungry data movement between separate compute and memory 2–5 . Although recent studies have demonstrated in-memory matrix-vector multiplication on fully integrated RRAM-CIM hardware 6–17 , it remains a goal for a RRAM-CIM chip to simultaneously deliver high energy efficiency, versatility to support diverse models and software-comparable accuracy. Although efficiency, versatility and accuracy are all indispensable for broad adoption of the technology, the inter-related trade-offs among them cannot be addressed by isolated improvements on any single abstraction level of the design. Here, by co-optimizing across all hierarchies of the design from algorithms and architecture to circuits and devices, we present NeuRRAM—a RRAM-based CIM chip that simultaneously delivers versatility in reconfiguring CIM cores for diverse model architectures, energy efficiency that is two-times better than previous state-of-the-art RRAM-CIM chips across various computational bit-precisions, and inference accuracy comparable to software models quantized to four-bit weights across various AI tasks, including accuracy of 99.0 percent on MNIST 18 and 85.7 percent on CIFAR-10 19 image classification, 84.7-percent accuracy on Google speech command recognition 20 , and a 70-percent reduction in image-reconstruction error on a Bayesian image-recovery task. 

Solid oxide cell (SOC) based energy conversion systems have the potential to become the cleanest and most efficient systems for reversible conversion between electricity and chemical fuels due to their high efficiency, low emission, and excellent fuel flexibility. Broad implementation of this technology is however hindered by the lack of high-performance electrode materials. While many perovskite-based materials have shown remarkable promise as electrodes for SOCs, cation enrichment or segregation near the surface or interfaces is often observed, which greatly impacts not only electrode kinetics but also their durability and operational lifespan. Since the chemical and structural variations associated with surface enrichment or segregation are typically confined to the nanoscale, advanced experimental and computational tools are required to probe the detailed composition, structure, and nanostructure of these near-surface regions in real time with high spatial and temporal resolutions. In this review article, an overview of the recent progress made in this area is presented, highlighting the thermodynamic driving forces, kinetics, and various configurations of surface enrichment and segregation in several widely studied perovskite-based material systems. A profound understanding of the correlation between the surface nanostructure and the electro-catalytic activity and stability of the electrodes is then emphasized, which is vital to achieving the rational design of more efficient SOC electrode materials with excellent durability. Furthermore, the methodology and mechanistic understanding of the surface processes are applicable to other materials systems in a wide range of applications, including thermo-chemical photo-assisted splitting of H 2 O/CO 2 and metal–air batteries.