Search

Your search keyword '"Chen, Yihui"' showing total 1,162 results
Start Over Author "Chen, Yihui"
1,162 results on '"Chen, Yihui"'

1. Low Latency Transformer Inference on FPGAs for Physics Applications with hls4ml

Author

Jiang, Zhixing, Yin, Dennis, Chen, Yihui, Khoda, Elham E, Hauck, Scott, Hsu, Shih-Chieh, Govorkova, Ekaterina, Harris, Philip, Loncar, Vladimir, and Moreno, Eric A.

Subjects
Computer Science - Machine Learning
Abstract

This study presents an efficient implementation of transformer architectures in Field-Programmable Gate Arrays(FPGAs) using hls4ml. We demonstrate the strategy for implementing the multi-head attention, softmax, and normalization layer and evaluate three distinct models. Their deployment on VU13P FPGA chip achieved latency less than 2us, demonstrating the potential for real-time applications. HLS4ML compatibility with any TensorFlow-built transformer model further enhances the scalability and applicability of this work. Index Terms: FPGAs, machine learning, transformers, high energy physics, LIGO

Published
2024

2. FPGA Deployment of LFADS for Real-time Neuroscience Experiments

Author

Liu, Xiaohan, Chen, ChiJui, Huang, YanLun, Yang, LingChi, Khoda, Elham E, Chen, Yihui, Hauck, Scott, Hsu, Shih-Chieh, and Lai, Bo-Cheng

Subjects
Quantitative Biology - Neurons and Cognition, Computer Science - Machine Learning, Computer Science - Neural and Evolutionary Computing
Abstract

Large-scale recordings of neural activity are providing new opportunities to study neural population dynamics. A powerful method for analyzing such high-dimensional measurements is to deploy an algorithm to learn the low-dimensional latent dynamics. LFADS (Latent Factor Analysis via Dynamical Systems) is a deep learning method for inferring latent dynamics from high-dimensional neural spiking data recorded simultaneously in single trials. This method has shown a remarkable performance in modeling complex brain signals with an average inference latency in milliseconds. As our capacity of simultaneously recording many neurons is increasing exponentially, it is becoming crucial to build capacity for deploying low-latency inference of the computing algorithms. To improve the real-time processing ability of LFADS, we introduce an efficient implementation of the LFADS models onto Field Programmable Gate Arrays (FPGA). Our implementation shows an inference latency of 41.97 $\mu$s for processing the data in a single trial on a Xilinx U55C., Comment: 6 pages, 8 figures

Published
2024

6. Diversity and composition changes of intestinal Fungi in a mouse model of total parenteral nutrition

Author

XIE Huichao, CHEN Yihui, and LIU Xingyu

Subjects
total parental nutrition, fungi, its dna, Medicine (General), R5-920
Abstract

Objective To analyze the diversity and composition changes of gut fungal communities between mouse model of total parental nutrition (TPN) and normal control mice. Methods After mouse model of TPN was constructed, fresh feces were collected from TPN mice (n=5) and normal control mice (n=5). Internal transcribed spacer (ITS) DNA sequencing was applied to determine intestinal fungi, and then bioinformatics analysis was conducted to identify the differences in fungal diversity, structure, and functional properties between the 2 groups of mice. Results There were significant differences in Alpha diversity (P < 0.05) and Beta diversity (P < 0.01) of intestinal fungi between the 2 groups. In the TPN model group, the relative abundances of Candida, Penicillium, Aspergillus and Talaromyces were obviously reduced (all P < 0.01). LEfSe analysis indicated that the above 4 strains were notably enriched in the normal control mice. Conclusion TPN mice exhibit characteristic changes in the composition of gut fungal flora compared to normal control mice. Dysfunction of gut fungal community may promote the occurrence of TPN related complications, and regulating the balance of gut fungal community may become a new strategy for preventing TPN related complications.

Published
2024
Full Text
View/download PDF

8. Data-driven discovery and extrapolation of parameterized pattern-forming dynamics

Author

Nicolaou, Zachary G., Huo, Guanyu, Chen, Yihui, Brunton, Steven L., and Kutz, J. Nathan

Subjects
Nonlinear Sciences - Pattern Formation and Solitons, Condensed Matter - Disordered Systems and Neural Networks
Abstract

Pattern-forming systems can exhibit a diverse array of complex behaviors as external parameters are varied, enabling a variety of useful functions in biological and engineered systems. First-principles derivations of the underlying transitions can be characterized using bifurcation theory on model systems whose governing equations are known. In contrast, data-driven methods for more complicated and realistic systems whose governing evolution dynamics are unknown have only recently been developed. Here we develop a data-driven approach, the {\em sparse identification for nonlinear dynamics with control parameters} (SINDyCP), to discover dynamics for systems with adjustable control parameters, such as an external driving strength. We demonstrate the method on systems of varying complexity, ranging from discrete maps to systems of partial differential equations. To mitigate the impact of measurement noise, we also develop a weak formulation of SINDyCP and assess its performance on noisy data. We demonstrate applications including the discovery of universal pattern-formation equations, and their bifurcation dependencies, directly from data accessible from experiments and the extrapolation of predictions beyond the weakly nonlinear regime near the onset of an instability., Comment: 6 pages, 4 figures, plus supplement

Published
2023
Full Text
View/download PDF

34. CES2 sustains HNF4α expression to promote pancreatic adenocarcinoma progression through an epoxide hydrolase-dependent regulatory loop

Author

Chen, Yihui, Capello, Michela, Perez, Mayrim V Rios, Vykoukal, Jody V, Roife, David, Kang, Ya'an, Prakash, Laura R, Katayama, Hiroyuki, Irajizad, Ehsan, Fleury, Alia, Ferri-Borgogno, Sammy, Baluya, Dodge L, Dennison, Jennifer B, Do, Kim-Anh, Fiehn, Oliver, Maitra, Anirban, Wang, Huamin, Chiao, Paul J, Katz, Matthew HG, Fleming, Jason B, Hanash, Samir M, and Fahrmann, Johannes F

Subjects
Biochemistry and Cell Biology, Biological Sciences, Pancreatic Cancer, Cancer, Digestive Diseases, Rare Diseases, Biotechnology, Development of treatments and therapeutic interventions, 5.1 Pharmaceuticals, Adenocarcinoma, Animals, Carboxylesterase, Carcinoma, Pancreatic Ductal, Cell Line, Tumor, Epoxide Hydrolases, Humans, Mice, Pancreatic Neoplasms, PDAC, HNF4 alpha, CES2, Classical/progenitor subtype, Phospholipid catabolism, HNF4α, Physiology, Biochemistry and cell biology
Abstract

ObjectiveIntra-tumoral expression of the serine hydrolase carboxylesterase 2 (CES2) contributes to the activation of the pro-drug irinotecan in pancreatic ductal adenocarcinoma (PDAC). Given other potential roles of CES2, we assessed its regulation, downstream effects, and contribution to tumor development in PDAC.MethodsAssociation between the mRNA expression of CES2 in pancreatic tumors and overall survival was assessed using The Cancer Genome Atlas. Cell viability, clonogenic, and anchorage-independent growth assays as well as an orthotopic mouse model of PDAC were used to evaluate the biological relevance of CES2 in pancreatic cancer. CES2-driven metabolic changes were determined by untargeted and targeted metabolomic analyses.ResultsElevated tumoral CES2 mRNA expression was a statistically significant predictor of poor overall survival in PDAC patients. Knockdown of CES2 in PDAC cells reduced cell viability, clonogenic capacity, and anchorage-independent growth in vitro and attenuated tumor growth in an orthotopic mouse model of PDAC. Mechanistically, CES2 was found to promote the catabolism of phospholipids resulting in HNF4α activation through a soluble epoxide hydrolase (sEH)-dependent pathway. Targeting of CES2 via siRNA or small molecule inhibitors attenuated HNF4α protein expression and reduced gene expression of classical/progenitor markers and increased basal-like markers. Targeting of the CES2-sEH-HNF4α axis using small molecule inhibitors of CES2 or sEH reduced cell viability.ConclusionsWe establish a novel regulatory loop between CES2 and HNF4α to sustain the progenitor subtype and promote PDAC progression and highlight the potential utility of CES2 or sEH inhibitors for the treatment of PDAC as part of non-irinotecan-containing regimens.

Published
2022

Catalog

Books, media, physical & digital resources
See catalog results