Search

In recent years, the foundation models have swept the computer vision field and facilitated the development of various tasks within different modalities. However, it remains an open question on how to design an infrared foundation model. In this paper, we propose InfMAE, a foundation model in infrared modality. We release an infrared dataset, called Inf30 to address the problem of lacking large-scale data for self-supervised learning in the infrared vision community. Besides, we design an information-aware masking strategy, which is suitable for infrared images. This masking strategy allows for a greater emphasis on the regions with richer information in infrared images during the self-supervised learning process, which is conducive to learning the generalized representation. In addition, we adopt a multi-scale encoder to enhance the performance of the pre-trained encoders in downstream tasks. Finally, based on the fact that infrared images do not have a lot of details and texture information, we design an infrared decoder module, which further improves the performance of downstream tasks. Extensive experiments show that our proposed method InfMAE outperforms other supervised methods and self-supervised learning methods in three downstream tasks., Comment: 8 pages, 2 figures

Knowledge Base Question Answering (KBQA) aims to answer factoid questions based on knowledge bases. However, generating the most appropriate knowledge base query code based on Natural Language Questions (NLQ) poses a significant challenge in KBQA. In this work, we focus on the CCKS2023 Competition of Question Answering with Knowledge Graph Inference for Unmanned Systems. Inspired by the recent success of large language models (LLMs) like ChatGPT and GPT-3 in many QA tasks, we propose a ChatGPT-based Cypher Query Language (CQL) generation framework to generate the most appropriate CQL based on the given NLQ. Our generative framework contains six parts: an auxiliary model predicting the syntax-related information of CQL based on the given NLQ, a proper noun matcher extracting proper nouns from the given NLQ, a demonstration example selector retrieving similar examples of the input sample, a prompt constructor designing the input template of ChatGPT, a ChatGPT-based generation model generating the CQL, and an ensemble model to obtain the final answers from diversified outputs. With our ChatGPT-based CQL generation framework, we achieved the second place in the CCKS 2023 Question Answering with Knowledge Graph Inference for Unmanned Systems competition, achieving an F1-score of 0.92676., Comment: Runner up of the CCKS 2023 question answering with knowledge graph inference for unmanned systems evaluation task, accepted as an evaluation paper

We investigate the linear stability analysis of a pathway-based diffusion model (PBDM), which characterizes the dynamics of the engineered Escherichia coli populations [X. Xue and C. Xue and M. Tang, P LoS Computational Biology, 14 (2018), pp. e1006178]. This stability analysis considers small perturbations of the density and chemical concentration around two non-trivial steady states, and the linearized equations are transformed into a generalized eigenvalue problem. By formal analysis, when the internal variable responds to the outside signal fast enough, the PBDM converges to an anisotropic diffusion model, for which the probability density distribution in the internal variable becomes a delta function. We introduce an asymptotic preserving (AP) scheme for the PBDM that converges to a stable limit scheme consistent with the anisotropic diffusion model. Further numerical simulations demonstrate the theoretical results of linear stability analysis, i.e., the pattern formation, and the convergence of the AP scheme.

We study the well-posedness of the biological models with AHL-dependent cell mobility on engineered Escherichia coli populations. For the kinetic model proposed by Xue-Xue-Tang recently, the local existence for large initial data is proved first. Furthermore, the positivity and local conservation laws for density $\rho (t,x,z)$ and nutrient $n(t,x)$ with initial assumptions are justified. Based on these properties, it can be extended globally in time near the equilibrium $(0,0,0)$. Considering the asymptotic behaviors of faster response CheZ turnover rate (i.e.,$\varepsilon\rightarrow 0$), one formally derives an anisotropic diffusion engineered Escherichia coli populations model (in short, AD-EECP) for which we find a key extra a priori estimate to overcome the difficulties coming from the nonlinearity of the diffusion structure. The local well-posedness and the positivity and local conservation laws for density and nutrient of the AD-EECP are justified. Furthermore, the global existence around the steady state $(\varrho_a, h_a, 0)$ with $\varrho_a \in [0, \Lambda_b)$ is obtained., Comment: 35 pages