Feng Ji

Chao Lu, Xueyou Lv, Yiming Lin, Dejian Li, Lihua Chen, Feng Ji, Youming Li, Chaohui Yu

AbstractConventional forceps biopsy (CFB) is the most popular way to screen for gastric epithelial neoplasia (GEN) and adenocarcinoma of gastric epithelium. The aim of this study was to compare the diagnostic accuracy between conventional forceps biopsy and endoscopic submucosal dissection (ESD).Four hundred forty-four patients who finally undertook ESD in our hospital were enrolled from Jan 1, 2009 to Sep 1, 2015. We retrospectively assessed the characteristics of pathological results of CFB and ESD.The concordance rate between CFB and ESD specimens was 68.92% (306/444). Men showed a lower concordance rate (63.61% vs 79.33%; P = 0.001) and concordance patients were younger (P = 0.048). In multivariate analysis, men significantly had a lower concordance rate (coefficient −0.730, P = 0.002) and a higher rate of pathological upgrade (coefficient −0.648, P = 0.015). Locations of CFB did not influence the concordance rate statistically.The concordance rate was relatively high in our hospital. According to our analysis, old men plus gastric fundus or antrum of CFB were strongly suggested to perform ESD if precancerous lesions were found. And young women with low-grade intraepithelial neoplasia could select regular follow-up.

Qi Hu, Lijun Fu, Fan Ma, Feng Ji

Similar large signal synchronizing instability that is common in traditional power system also exists in voltage source converter (VSC) dominated power system, which is increasingly reported and investigated. In this paper, the large signal instability of phase locked loop (PLL) synchronized VSC connected to weak ac grid is investigated. First, the influence of high grid impedance on PLL's dynamics is explored and an additional feedback loop is found to be introduced, which deteriorates the performance of PLL. Then an analysis model that is similar as the rotor motion model of synchronous generator (SG) is developed. Based on the developed model, equal-area method is employed to carry out the large-signal stability analysis. The large signal instability is found to be mainly resulted by the following two factors. One is the nonexistence of equilibrium point, which is similar as that there does not exist intersection between the mechanical power input curve and electromagnetic restoring force curve in SG grid connected system. The other is related with the transition process resulted by insufficient decelerating area, which leads to that the system cannot stably go through the transition from initial point to the existing equilibrium point. The proposed analysis not only contributes to revealing the physical mechanism but also provides guidance for the future improvement measures.

F. Ji, Zhaoxu Sun, Tianyi Hang, Jiajia Zheng, Xiping Li, Gaigai Duan, Chunmei Zhang, Yiming Chen

Ji Liang, Feng Ji, Anisah Lee Binti Abdullah, Wei Qin, Tian Zhu, Y. Tay, Yiming Li, Mingming Han

The widespread presence of microplastics and nanoplastics (MPs/NPs) in the environment has become a critical public health issue due to their potential to infiltrate and affect various biological systems. Our review is crucial as it consolidates current data and provides a comprehensive analysis of the cardiovascular impacts of MPs/NPs across species, highlighting significant implications for human health. By synthesizing findings from studies on aquatic and terrestrial organisms, including humans, this review offers insights into the ubiquity of MPs/NPs and their pathophysiological roles in cardiovascular systems. We demonstrated that exposure to MPs/NPs is linked to various cardiovascular ailments such as thrombogenesis, vascular damage, and cardiac impairments in model organisms, which likely extrapolate to humans. Our review critically evaluated methods for detecting MPs/NPs in biological tissues, assessing their toxicity, and understanding their behaviour within the vasculature. These findings emphasise the urgent need for targeted public health strategies and enhanced regulatory measures to mitigate the impacts of MP/NP pollution. Furthermore, the review underlined the necessity of advancing research methodologies to explore long-term effects and potential intergenerational consequences of MP/NP exposure. By mapping out the intricate links between environmental exposure and cardiovascular risks, our work served as a pivotal reference for future research and policymaking aimed at curbing the burgeoning threat of plastic pollution.

Xiangxiang Gao, Weisheng Xie, Yiwei Xiang, Feng Ji

Striking an optimal balance between minimal drafting latency and high speculation accuracy to enhance the inference speed of Large Language Models remains a significant challenge in speculative decoding. In this paper, we introduce Falcon, an innovative semi-autoregressive speculative decoding framework fashioned to augment both the drafter's parallelism and output quality. Falcon incorporates the Coupled Sequential Glancing Distillation technique, which fortifies inter-token dependencies within the same block, leading to increased speculation accuracy. We offer a comprehensive theoretical analysis to illuminate the underlying mechanisms. Additionally, we introduce a Custom-Designed Decoding Tree, which permits the drafter to generate multiple tokens in a single forward pass and accommodates multiple forward passes as needed, thereby boosting the number of drafted tokens and significantly improving the overall acceptance rate. Comprehensive evaluations on benchmark datasets such as MT-Bench, HumanEval, and GSM8K demonstrate Falcon's superior acceleration capabilities. The framework achieves a lossless speedup ratio ranging from 2.91x to 3.51x when tested on the Vicuna and LLaMA2-Chat model series. These results outstrip existing speculative decoding methods for LLMs, including Eagle, Medusa, Lookahead, SPS, and PLD, while maintaining a compact drafter architecture equivalent to merely two Transformer layers.