Wei Liu

Shounian Deng, Weiyin Fei, Wei Liu et al.

In this paper, we use the truncated EM method to study the finite time strong convergence for the SDEs with Poisson jumps under the Khasminskii-type condition. We establish the finite time $ \mathcal L ^r (r \ge 2) $ convergence rate when the drift and diffusion coefficients satisfy super-linear condition and the jump coefficient satisfies the linear growth condition. The result shows that the optimal $\mathcal L ^r$-convergence rate is close to $ 1/ (1 + γ)$, where $γ$ is the super-linear growth constant. This is significantly different from the result on SDEs without jumps. When all the three coefficients of SDEs are allowing to grow super-linearly, the $ \mathcal L^r (0<r<2)$ strong convergence results are also investigated and the optimal strong convergence rate is shown to be not greater than $1/4$. Moreover, we prove that the truncated EM method preserve nicely the mean square exponentially stability and asymptotic boundedness of the underlying SDEs with Piosson jumps. Several examples are given to illustrate our results.

Qinzhuo Wu, Weikai Xu, Wei Liu et al.

Recently, mobile AI agents based on VLMs have been gaining increasing attention. These works typically utilize VLM as a foundation, fine-tuning it with instruction-based mobile datasets. However, these VLMs are typically pre-trained on general-domain data, which often results in a lack of fundamental capabilities specific to the mobile domain. Therefore, they may struggle to recognize specific UI elements and understand intra-UI fine-grained information. In addition, the current fine-tuning task focuses on interacting with the most relevant element for the given instruction. These fine-tuned VLMs may still ignore the relationships between UI pages, neglect the roles of elements in page transitions and lack inter-UI understanding. To address issues, we propose a VLM called MobileVLM, which includes two additional pre-training stages to enhance both intra- and inter-UI understanding. We defined four UI-based pre-training tasks, enabling the model to better perceive fine-grained elements and capture page transition actions. To address the lack of mobile pre-training data, we built a large Chinese mobile dataset Mobile3M from scratch, which contains 3 million UI pages, and real-world transition actions, forming a directed graph structure. Experimental results show MobileVLM excels on both our test set and public mobile benchmarks, outperforming existing VLMs.