Search for: All records

Processing-in-memory (PIM) based architecture shows great potential to process several emerging artificial intelligence workloads, including vision and language models. Cross-layer optimizations could bridge the gap between computing density and the available resources by reducing the computation and memory cost of the model and improving the model’s robustness against non-ideal hardware effects. We first introduce several hardware-aware training methods to improve the model robustness to the PIM device’s nonideal effects, including stuck-at-fault, process variation, and thermal noise. Then, we further demonstrate a software/hardware (SW/HW) co-design methodology to efficiently process the state-of-the-art attention-based model on PIM-based architecture by performing sparsity exploration for the attention-based model and circuit architecture co-design to support the sparse processing. 

Neural network models have demonstrated outstanding performance in a variety of applications, from image classification to natural language processing. However, deploying the models to hardware raises efficiency and reliability issues. From the efficiency perspective, the storage, computation, and communication cost of neural network processing is considerably large because the neural network models have a large number of parameters and operations. From the standpoint of robustness, the perturbation in hardware is unavoidable and thus the performance of neural networks can be degraded. As a result, this paper investigates effective learning and optimization approaches as well as advanced hardware designs in order to build efficient and robust neural network designs. 

The electrochemical CO₂reduction reaction (CO₂RR) is a promising approach to achieving sustainable electrical‐to‐chemical energy conversion and storage while decarbonizing the emission‐heavy industry. The carbon‐supported, nitrogen‐coordinated, and atomically dispersed metal sites are effective catalysts for CO generation due to their high activity, selectivity, and earth abundance. Here, we discuss progress, challenges, and opportunities for designing and engineering atomic metal catalysts from single to dual metal sites. Engineering single metal sites using a nitrogen‐doped carbon model was highlighted to exclusively study the effect of carbon particle sizes, metal contents, and M−N bond structures in the form of MN₄moieties on catalytic activity and selectivity. The structure‐property correlation was analyzed by combining experimental results with theoretical calculations to uncover the CO₂to CO conversion mechanisms. Furthermore, dual‐metal site catalysts, inheriting the merits of single‐metal sites, have emerged as a new frontier due to their potentially enhanced catalytic properties. Designing optimal dual metal site catalysts could offer additional sites to alter the surface adsorption to CO₂and various intermediates, thus breaking the scaling relationship limitation and activity‐stability trade‐off. The CO₂RR electrolysis in flow reactors was discussed to provide insights into the electrolyzer design with improved CO₂utilization, reaction kinetics, and mass transport.

 

The electrochemical CO₂reduction reaction (CO₂RR) is a promising approach to achieving sustainable electrical‐to‐chemical energy conversion and storage while decarbonizing the emission‐heavy industry. The carbon‐supported, nitrogen‐coordinated, and atomically dispersed metal sites are effective catalysts for CO generation due to their high activity, selectivity, and earth abundance. Here, we discuss progress, challenges, and opportunities for designing and engineering atomic metal catalysts from single to dual metal sites. Engineering single metal sites using a nitrogen‐doped carbon model was highlighted to exclusively study the effect of carbon particle sizes, metal contents, and M−N bond structures in the form of MN₄moieties on catalytic activity and selectivity. The structure‐property correlation was analyzed by combining experimental results with theoretical calculations to uncover the CO₂to CO conversion mechanisms. Furthermore, dual‐metal site catalysts, inheriting the merits of single‐metal sites, have emerged as a new frontier due to their potentially enhanced catalytic properties. Designing optimal dual metal site catalysts could offer additional sites to alter the surface adsorption to CO₂and various intermediates, thus breaking the scaling relationship limitation and activity‐stability trade‐off. The CO₂RR electrolysis in flow reactors was discussed to provide insights into the electrolyzer design with improved CO₂utilization, reaction kinetics, and mass transport.

 

With the recent demand of deploying neural network models on mobile and edge devices, it is desired to improve the model's generalizability on unseen testing data, as well as enhance the model's robustness under fixed-point quantization for efficient deployment. Minimizing the training loss, however, provides few guarantees on the generalization and quantization performance. In this work, we fulfill the need of improving generalization and quantization performance simultaneously by theoretically unifying them under the framework of improving the model's robustness against bounded weight perturbation and minimizing the eigenvalues of the Hessian matrix with respect to model weights. We therefore propose HERO, a Hessian-enhanced robust optimization method, to minimize the Hessian eigenvalues through a gradient-based training process, simultaneously improving the generalization and quantization performance. HERO enables up to a 3.8% gain on test accuracy, up to 30% higher accuracy under 80% training label perturbation, and the best post-training quantization accuracy across a wide range of precision, including a > 10% accuracy improvement over SGD-trained models for common model architectures on various datasets. 

As a new frontier, the rapid development of single‐atom catalysts (SACs) in heterogeneous catalysis has attracted extensive attention. However, a fundamental understanding of the dynamic formation process of active metal atom sites is lacking. The rational design of high‐performance SACs now becomes possible by explicitly addressing the conundrum of constructing high‐density SACs and elucidating the evolution of SACs during the synthesis.