Taipei, Taiwan – China is facing a steeper climb to overtake the United States and its allies in semiconductors as Washington ramps up measures to restrict Beijing's ability to produce advanced chips and secure dominance over the strategic technology. On Wednesday, Washington restricted the sale to China of select Nvidia and AMD advanced graphic processor units (GPUs) used in artificial intelligence applications and supercomputers. The move followed the US Commerce Department's announcement last month of a ban on exports to China of electronic design automation (EDA) software used in the production of next-generation chips. Meanwhile, Washington has been nudging East Asian partners Taiwan, South Korea, and Japan to form a "Chip 4" industry alliance to isolate China from the international tech ecosystem, and bolstered efforts to develop its homegrown industry with the passage of the CHIPS Act, offering $52bn in subsidies to firms that make chips on US soil. "The US is trying to reinforce its central role in the world's semiconductor ecosystem and ensure that China is unable to produce the most cutting edge chips," Chris Miller, author of the upcoming book Chip War: The Fight for the World's Most Critical Technology, told Al Jazeera.

While I complain publically about the lack of a governing global body for A.I. regulation that's independent, most BigTech firms pretend like they regulate themselves. Nobody actually trusts that they are doing this properly. The most impressive A.I. and tech regulation I've seen is actually coming out of China. The American narrative on this is that they are anti-capitalistic. I find that attitude interesting.

This paper focuses on the research of cross-database micro-expression recognition, in which the training and test micro-expression samples belong to different microexpression databases. Mismatched feature distributions between the training and testing micro-expression feature degrade the performance of most well-performing micro-expression methods. To deal with cross-database micro-expression recognition, we propose a novel domain adaption method called Transfer Group Sparse Regression (TGSR). TGSR learns a sparse regression matrix for selecting salient facial local regions and the corresponding relationship of the training set and test set. We evaluate our TGSR model in CASME II and SMIC databases. Experimental results show that the proposed TGSR achieves satisfactory performance and outperforms most state-of-the-art subspace learning-based domain adaption methods.

Washington – The United States added dozens of Chinese companies, including the country's top chipmaker SMIC and Chinese drone manufacturer SZ DJI Technology Co. Ltd., to a trade blacklist on Friday as U.S. President Donald Trump's administration ratchets up tensions with China in his final weeks in office. Reuters first reported the addition of SMIC and other companies earlier on Friday. The move is seen as the latest in Republican Trump's efforts to burnish his tough-on-China image as part of lengthy fight between Washington and Beijing over trade and numerous economic issues. The U.S. Commerce Department said the action against SMIC stems from Beijing's efforts to harness civilian technologies for military purposes and evidence of activities between SMIC and Chinese military industrial companies of concern. The Commerce Department will "not allow advanced U.S. technology to help build the military of an increasingly belligerent adversary," Secretary Wilbur Ross said in a statement.

Semi-supervised clustering aims to introduce prior knowledge in the decision process of a clustering algorithm. In this paper, we propose a novel semi-supervised clustering algorithm based on the information-maximization principle. The proposed method is an extension of a previous unsupervised information-maximization clustering algorithm based on squared-loss mutual information to effectively incorporate must-links and cannot-links. The proposed method is computationally efficient because the clustering solution can be obtained analytically via eigendecomposition. Furthermore, the proposed method allows systematic optimization of tuning parameters such as the kernel width, given the degree of belief in the must-links and cannot-links. The usefulness of the proposed method is demonstrated through experiments.

Information-maximization clustering learns a probabilistic classifier in an unsupervised manner so that mutual information between feature vectors and cluster assignments is maximized. A notable advantage of this approach is that it only involves continuous optimization of model parameters, which is substantially easier to solve than discrete optimization of cluster assignments. However, existing methods still involve non-convex optimization problems, and therefore finding a good local optimal solution is not straightforward in practice. In this paper, we propose an alternative information-maximization clustering method based on a squared-loss variant of mutual information. This novel approach gives a clustering solution analytically in a computationally efficient way via kernel eigenvalue decomposition. Furthermore, we provide a practical model selection procedure that allows us to objectively optimize tuning parameters included in the kernel function. Through experiments, we demonstrate the usefulness of the proposed approach.