Search

Your search keyword '"Zhang, WenCong"' showing total 516 results
Start Over Author "Zhang, WenCong"
516 results on '"Zhang, WenCong"'

1. Apple Intelligence Foundation Language Models

Author

Gunter, Tom, Wang, Zirui, Wang, Chong, Pang, Ruoming, Narayanan, Andy, Zhang, Aonan, Zhang, Bowen, Chen, Chen, Chiu, Chung-Cheng, Qiu, David, Gopinath, Deepak, Yap, Dian Ang, Yin, Dong, Nan, Feng, Weers, Floris, Yin, Guoli, Huang, Haoshuo, Wang, Jianyu, Lu, Jiarui, Peebles, John, Ye, Ke, Lee, Mark, Du, Nan, Chen, Qibin, Keunebroek, Quentin, Wiseman, Sam, Evans, Syd, Lei, Tao, Rathod, Vivek, Kong, Xiang, Du, Xianzhi, Li, Yanghao, Wang, Yongqiang, Gao, Yuan, Ahmed, Zaid, Xu, Zhaoyang, Lu, Zhiyun, Rashid, Al, Jose, Albin Madappally, Doane, Alec, Bencomo, Alfredo, Vanderby, Allison, Hansen, Andrew, Jain, Ankur, Anupama, Anupama Mann, Kamal, Areeba, Wu, Bugu, Brum, Carolina, Maalouf, Charlie, Erdenebileg, Chinguun, Dulhanty, Chris, Moritz, Dominik, Kang, Doug, Jimenez, Eduardo, Ladd, Evan, Shi, Fangping, Bai, Felix, Chu, Frank, Hohman, Fred, Kotek, Hadas, Coleman, Hannah Gillis, Li, Jane, Bigham, Jeffrey, Cao, Jeffery, Lai, Jeff, Cheung, Jessica, Shan, Jiulong, Zhou, Joe, Li, John, Qin, Jun, Singh, Karanjeet, Vega, Karla, Zou, Kelvin, Heckman, Laura, Gardiner, Lauren, Bowler, Margit, Cordell, Maria, Cao, Meng, Hay, Nicole, Shahdadpuri, Nilesh, Godwin, Otto, Dighe, Pranay, Rachapudi, Pushyami, Tantawi, Ramsey, Frigg, Roman, Davarnia, Sam, Shah, Sanskruti, Guha, Saptarshi, Sirovica, Sasha, Ma, Shen, Ma, Shuang, Wang, Simon, Kim, Sulgi, Jayaram, Suma, Shankar, Vaishaal, Paidi, Varsha, Kumar, Vivek, Wang, Xin, Zheng, Xin, Cheng, Walker, Shrager, Yael, Ye, Yang, Tanaka, Yasu, Guo, Yihao, Meng, Yunsong, Luo, Zhao Tang, Ouyang, Zhi, Aygar, Alp, Wan, Alvin, Walkingshaw, Andrew, Lin, Antonie, Farooq, Arsalan, Ramerth, Brent, Reed, Colorado, Bartels, Chris, Chaney, Chris, Riazati, David, Yang, Eric Liang, Feldman, Erin, Hochstrasser, Gabriel, Seguin, Guillaume, Belousova, Irina, Pelemans, Joris, Yang, Karen, Vahid, Keivan Alizadeh, Cao, Liangliang, Najibi, Mahyar, Zuliani, Marco, Horton, Max, Cho, Minsik, Bhendawade, Nikhil, Dong, Patrick, Maj, Piotr, Agrawal, Pulkit, Shan, Qi, Fu, Qichen, Poston, Regan, Xu, Sam, Liu, Shuangning, Rao, Sushma, Heeramun, Tashweena, Merth, Thomas, Rayala, Uday, Cui, Victor, Sridhar, Vivek Rangarajan, Zhang, Wencong, Zhang, Wenqi, Wu, Wentao, Zhou, Xingyu, Liu, Xinwen, Zhao, Yang, Xia, Yin, Ren, Zhile, and Ren, Zhongzheng

Subjects
Computer Science - Artificial Intelligence, Computer Science - Computation and Language, Computer Science - Machine Learning
Abstract

We present foundation language models developed to power Apple Intelligence features, including a ~3 billion parameter model designed to run efficiently on devices and a large server-based language model designed for Private Cloud Compute. These models are designed to perform a wide range of tasks efficiently, accurately, and responsibly. This report describes the model architecture, the data used to train the model, the training process, how the models are optimized for inference, and the evaluation results. We highlight our focus on Responsible AI and how the principles are applied throughout the model development.

Published
2024

2. Direct in-situ imaging of electrochemical corrosion of Pd-Pt core-shell electrocatalysts.

Author

Shi, Fenglei, Tieu, Peter, Hu, Hao, Peng, Jiaheng, Zhang, Wencong, Li, Fan, Tao, Peng, Song, Chengyi, Shang, Wen, Deng, Tao, Gao, Wenpei, Pan, Xiaoqing, and Wu, Jianbo

Abstract

Corrosion of electrocatalysts during electrochemical operations, such as low potential - high potential cyclic swapping, can cause significant performance degradation. However, the electrochemical corrosion dynamics, including structural changes, especially site and composition specific ones, and their correlation with electrochemical processes are hidden due to the insufficient spatial-temporal resolution characterization methods. Using electrochemical liquid cell transmission electron microscopy, we visualize the electrochemical corrosion of Pd@Pt core-shell octahedral nanoparticles towards a Pt nanoframe. The potential-dependent surface reconstruction during multiple continuous in-situ cyclic voltammetry with clear redox peaks is captured, revealing an etching and deposition process of Pd that results in internal Pd atoms being relocated to external surface, followed by subsequent preferential corrosion of Pt (111) terraces rather than the edges or corners, simultaneously capturing the structure evolution also allows to attribute the site-specific Pt and Pd atomic dynamics to individual oxidation and reduction events. This work provides profound insights into the surface reconstruction of nanoparticles during complex electrochemical processes.

Published
2024

4. Less is More: Removing Text-regions Improves CLIP Training Efficiency and Robustness

Author

Cao, Liangliang, Zhang, Bowen, Chen, Chen, Yang, Yinfei, Du, Xianzhi, Zhang, Wencong, Lu, Zhiyun, and Zheng, Yantao

Subjects
Computer Science - Computer Vision and Pattern Recognition, Computer Science - Artificial Intelligence
Abstract

The CLIP (Contrastive Language-Image Pre-training) model and its variants are becoming the de facto backbone in many applications. However, training a CLIP model from hundreds of millions of image-text pairs can be prohibitively expensive. Furthermore, the conventional CLIP model doesn't differentiate between the visual semantics and meaning of text regions embedded in images. This can lead to non-robustness when the text in the embedded region doesn't match the image's visual appearance. In this paper, we discuss two effective approaches to improve the efficiency and robustness of CLIP training: (1) augmenting the training dataset while maintaining the same number of optimization steps, and (2) filtering out samples that contain text regions in the image. By doing so, we significantly improve the classification and retrieval accuracy on public benchmarks like ImageNet and CoCo. Filtering out images with text regions also protects the model from typographic attacks. To verify this, we build a new dataset named ImageNet with Adversarial Text Regions (ImageNet-Attr). Our filter-based CLIP model demonstrates a top-1 accuracy of 68.78\%, outperforming previous models whose accuracy was all below 50\%., Comment: 10 pages, 8 figures

Published
2023

8. Automatic Segmentation of Liver Tumor from Multi-phase Contrast-Enhanced CT Images Using Cross-Phase Fusion Transformer

Author

Zhang, Wencong, Tao, Yuxi, Liang, Wei, Li, Junjie, Chen, Yingjia, Song, Tengfei, Ma, Xiangyuan, Zhang, Yaqin, Magjarević, Ratko, Series Editor, Ładyżyński, Piotr, Associate Editor, Ibrahim, Fatimah, Associate Editor, Lackovic, Igor, Associate Editor, Rock, Emilio Sacristan, Associate Editor, Wang, Guangzhi, editor, Yao, Dezhong, editor, Gu, Zhongze, editor, Peng, Yi, editor, Tong, Shanbao, editor, and Liu, Chengyu, editor

Published
2024
Full Text
View/download PDF

35. Dzhaparidze-van Zanten type inequalities for self-normalized martingales

Author

Zhang, WenCong

Subjects
Mathematics - Probability
Abstract

The Bernstein inequality is a tight upper bound on tail probabilities for independent random variables. Freedman extended the Bernstein inequality to martingales with differences bounded from above, and then Dzhaparidze and van Zanten generalized Freedman's result to non-bounded locally square integrable martingales. In this paper, we derive some Dzhaparidze-van Zanten type inequalities for self-normalized martingales with square integrable and non-square integrable differences.

Published
2020

36. Some deviation inequalities for sums of negatively associated random variables

Author

Zhang, WenCong

Subjects
Mathematics - Probability
Abstract

Let $\{X_i,i\geq1\}$ be a sequence of negatively associated random variables, and let $\{X_i^\ast,i\geq 1\}$ be a sequence of independent random variables such that $X_i^\ast$ and $X_i$ have the same distribution for each $i$. Denote by $S_k=\sum_{i=1}^{k}X_i$ and $S_k^\ast=\sum_{i=1}^{k}X_i^\ast$ for $k\geq 1$. The well-known results of Shao \cite{Shao2000} sates that $\mathbb{E}f(S_n)\leq \mathbb{E}f(S_n^\ast)$ for any nondecreasing convex function. Using this very strong property, we obtain a large variety of deviation inequalities for $S_n$

Published
2020

Catalog

Books, media, physical & digital resources
See catalog results