Zhifeng Yang

Zhishu Sun, Zhifeng Shen, Luojun Lin et al.

Domain generalization (DG) is a fundamental yet very challenging research topic in machine learning. The existing arts mainly focus on learning domain-invariant features with limited source domains in a static model. Unfortunately, there is a lack of training-free mechanism to adjust the model when generalized to the agnostic target domains. To tackle this problem, we develop a brand-new DG variant, namely Dynamic Domain Generalization (DDG), in which the model learns to twist the network parameters to adapt the data from different domains. Specifically, we leverage a meta-adjuster to twist the network parameters based on the static model with respect to different data from different domains. In this way, the static model is optimized to learn domain-shared features, while the meta-adjuster is designed to learn domain-specific features. To enable this process, DomainMix is exploited to simulate data from diverse domains during teaching the meta-adjuster to adapt to the upcoming agnostic target domains. This learning mechanism urges the model to generalize to different agnostic target domains via adjusting the model without training. Extensive experiments demonstrate the effectiveness of our proposed method. Code is available at: https://github.com/MetaVisionLab/DDG

Qipeng Liu, Luojun Lin, Zhifeng Shen et al.

Source-free object detection (SFOD) aims to adapt the source detector to unlabeled target domain data in the absence of source domain data. Most SFOD methods follow the same self-training paradigm using mean-teacher (MT) framework where the student model is guided by only one single teacher model. However, such paradigm can easily fall into a training instability problem that when the teacher model collapses uncontrollably due to the domain shift, the student model also suffers drastic performance degradation. To address this issue, we propose the Periodically Exchange Teacher-Student (PETS) method, a simple yet novel approach that introduces a multiple-teacher framework consisting of a static teacher, a dynamic teacher, and a student model. During the training phase, we periodically exchange the weights between the static teacher and the student model. Then, we update the dynamic teacher using the moving average of the student model that has already been exchanged by the static teacher. In this way, the dynamic teacher can integrate knowledge from past periods, effectively reducing error accumulation and enabling a more stable training process within the MT-based framework. Further, we develop a consensus mechanism to merge the predictions of two teacher models to provide higher-quality pseudo labels for student model. Extensive experiments on multiple SFOD benchmarks show that the proposed method achieves state-of-the-art performance compared with other related methods, demonstrating the effectiveness and superiority of our method on SFOD task.

Zhifeng Yang, Peizong Wu

The steel structure demolition scheme needs to be compiled according to the specific engineering characteristics and the update results of the finite element model. The designers need to refer to the relevant engineering cases according to the standard requirements when compiling. It takes a lot of time to retrieve information and organize language, and the degree of automation and intelligence is low. This paper proposes an intelligent generation method of structural demolition suggestions based on multi-model collaboration, and improves the text generation performance of large language models in the field of structural demolition by Retrieval-Augmented Generation and Low-Rank Adaptation Fine-Tuning technology. The intelligent generation framework of multi-model collaborative structural demolition suggestions can start from the specific engineering situation, drive the large language model to answer with anthropomorphic thinking, and propose demolition suggestions that are highly consistent with the characteristics of the structure. Compared with CivilGPT, the multi-model collaboration framework proposed in this paper can focus more on the key information of the structure, and the suggestions are more targeted.