Al2O3-supported Cu and Ni monometallic as well Cu–Ni bimetallic catalysts were synthesized using a coprecipitation method studied for the in-situ hydrogenation of furfural (FAL) with isopropanol solvent hydrogen donor. The showed improved activity toward production 2-methylfuran (2-MF) 2-methyltetrahydrofuran (2-MTHF) over that catalysts. results indicated exhibited better performance than methanol FAL to produce 2-MF 2-MTHF under same conditions. reaction conditions such copper–nickel...

Single-image super-resolution (SISR) is an important task in image processing, which aims to enhance the resolution of imaging systems. Recently, SISR has made a huge leap and achieved promising results with help deep learning (DL). In this survey, we give overview DL-based methods group them according their targets, such as reconstruction efficiency, accuracy, perceptual accuracy. Specifically, first introduce problem definition, research background, significance SISR. Secondly, some...

Large language models (LLMs) achieve impressive performance by scaling model parameters, but this comes with significant inference overhead. Feed-forward networks (FFNs), which dominate LLM exhibit high activation sparsity in hidden neurons. To exploit this, researchers have proposed using a mixture-of-experts (MoE) architecture, where only subset of parameters is activated. However, existing approaches often require extensive training data and resources, limiting their practicality. We...

Quantization is a critical technique employed across various research fields for compressing deep neural networks (DNNs) to facilitate deployment within resource-limited environments. This process necessitates delicate balance between model size and performance. In this work, we explore knowledge distillation (KD) as promising approach improving quantization performance by transferring from high-precision low-precision counterparts. We specifically investigate feature-level information loss...

Abstract Background Dyslipidemia is a key factor causing cardio cerebrovascular diseases, and the total cholesterol (TC) an important lipid indicator among them. Studies have shown that environmental factors strong association with TC levels. Previous studies only focused on seasonal variation of level short-term effects some over time, few explored geographical distribution quantified impact in space. Methods Based blood test data which was from China Health Retirement Longitudinal Study...

Lithography is fundamental to integrated circuit fabrication, necessitating large computation overhead. The advancement of machine learning (ML)-based lithography models alleviates the trade-offs between manufacturing process expense and capability. However, all previous methods regard system as an image-to-image black box mapping, utilizing network parameters learn by rote mappings from massive mask-to-aerial or mask-to-resist image pairs, resulting in poor generalization In this paper, we...

Recent years have seen rising research in logic synthesis recipe generation to improve the Quality-of-Result (QoR). However, existing approaches typically low efficiency and are stuck at local optima. In this work, we propose a optimization framework, AlphaSyn, that incorporates domain-specific Monte Carlo tree search (MCTS) algorithm. AlphaSyn enables exploration across entire space while optimizing sampling points utilization. We further develop synthesis-specific upper confidence bound...

The ever-growing complexity of modern VLSI circuits brings about a substantial increase in the design cycle. As for logic synthesis, how to efficiently obtain physical characteristics subsequent space exploration emerges as critical issue. In this paper, we propose ${\mathsf{LSTP}}$, an ML-based synthesis predictor, which can rapidly predict post-synthesis timing broad range circuit designs. Specifically, explicitly take optimization sequences into consideration so that comprehend synergy...

Generative Pre-trained Transformers (GPTs) have demonstrated remarkable performance across diverse domains through the extensive scaling of model parameters. Recent works observe redundancy transformer blocks and develop compression methods by structured pruning unimportant blocks. However, such straightforward elimination will always provide irreversible degradation. In this paper, we propose FuseGPT, a novel methodology to recycle pruned further recover performance. Firstly introduce new...

Hotspot detection is an essential step in the physical verification flow to identify layout patterns that are sensitive process variations. Recent advances machine learning have enabled deep neural networks (DNNs) achieve good performance for hotspot detection; however, these models require a high computational complexity and memory footprint. To reduce such costs, quantization provides promising solution by compressing DNNs into low-bit inference schemes. In this paper, we propose several...

Lithography is fundamental to integrated circuit fabrication, necessitating large computation overhead. The advancement of machine learning (ML)-based lithography models alleviates the trade-offs between manufacturing process expense and capability. However, all previous methods regard system as an image-to-image black box mapping, utilizing network parameters learn by rote mappings from massive mask-to-aerial or mask-to-resist image pairs, resulting in poor generalization In this paper, we...