Core Viewpoint - The article discusses the introduction of a new generative model paradigm called Drifting Models, proposed by He Kaiming's team, which shifts the distribution evolution process from the inference stage to the training stage, enabling one-step generation of high-quality samples [1][4][36]. Summary by Sections Introduction of Drifting Models - The Drifting Model represents a significant innovation in generative modeling by introducing the "Drifting Field" mechanism, which aligns the prior distribution with the real data distribution during training, eliminating common instabilities in GANs and avoiding reliance on multi-step ODE/SDE solutions [5][12][19]. Mechanism of Drifting Models - The core of the Drifting Model is to learn a mapping function that transforms a simple prior distribution (like Gaussian noise) into a pushforward distribution that matches real data [9][10]. - Unlike traditional models that require multiple iterations during inference, the Drifting Model allows for single-step generation by leveraging the iterative nature of neural network training as the driving force for distribution evolution [14][18]. Training Process - The training process involves calculating a drift vector for each sample based on the distribution of positive and negative samples, guiding the model to align its output distribution with the target distribution [21][26]. - The model's training trajectory is essentially equivalent to the path of distribution evolution, allowing for high-quality generation with only a single forward pass during inference [18][36]. Experimental Results - In the ImageNet 256x256 benchmark, the Drifting Model achieved a FID score of 1.54 in latent space and 1.61 in pixel space during one-step inference, outperforming many traditional diffusion models that require hundreds of iterations [32][33]. - The model also demonstrated strong generalization capabilities in embodied intelligence control tasks, matching or exceeding the decision quality of diffusion policies that require significantly more inference steps [34][35]. Conclusion - The Drifting Model successfully transfers the generative pressure from the inference stage to the training stage, providing a new perspective on generative modeling that reinterprets the training process as a mechanism for distribution evolution [36][37].

Core Insights - The technology and manufacturing sectors have seen significant financing activity, particularly in artificial intelligence, semiconductors, and biomedicine, indicating strong investor interest in these areas [1] - The overall financing scale in the domestic primary market from January 5 to January 11 included 35 events, with a total amount of approximately 154.27 billion RMB [1] Financing Overview - The technology and manufacturing sectors led in financing activity, with notable performances in smart vehicles, semiconductors, and advanced technologies [1] - The biomedicine sector completed four financing rounds totaling around 5 billion RMB, while the artificial intelligence sector had three rounds amounting to approximately 0.9 billion RMB [3][4] Regional Distribution - The majority of financing events occurred in Beijing, Zhejiang, and Guangdong, with 9, 6, and 6 events respectively [5][6] Active Investment Institutions - Shunxi Fund and Zhongke Chuangxing were particularly active, each completing two financing rounds focused on technology and manufacturing [7] Notable Company Financing - Huixin Yigu completed over 100 million RMB in Series A financing, led by Jingneng Green Fund, to advance clinical research in cell therapy for neurological diseases [9][10] - Anlong Bio secured nearly 100 million RMB in Series B+ financing, supported by municipal and district-level industry funds, to develop its gene therapy pipeline [11] - Shanghai Ruizhou Bio raised 200 million RMB in Series B financing, led by Ruile Synthetic Biology Fund, to support clinical research for its pneumonia vaccine [12] - Thunderbird Innovation received over 1 billion RMB in financing from China Mobile and China Unicom for its AR smart glasses [14] - Zhizhan Technology completed nearly 300 million RMB in Series C financing, led by Zhejiang State-owned Assets Fund, to enhance its market share in the electric vehicle sector [15] - Mingxin Qirui raised over 100 million RMB in Pre-A financing to advance RRAM technology for AI and data center applications [16] - Zhixing Technology secured 400 million RMB in strategic financing from Huangshi State-owned Capital Investment Group for its autonomous driving technology [17] - Jiukexin completed over 100 million RMB in B2 financing to expand its AI-driven automation solutions for state-owned enterprises [18] - Zhidong Dalu raised nearly 200 million USD in financing to accelerate the development of its advanced intelligent driving solutions [19]