In practical industrial scenarios， the significant divergence in bearing state evolution caused by complex operating conditions imposes dual constraints on bearing life prediction： incomplete mechanism understanding and distribution shift in monitoring data. Although the existing data-driven methods perform well under stable conditions， they suffer from strong reliance on annotated data and limited generalization capability under rappidly changing conditions. To address these issues，a method for Bearing Life Prediction based on Dynamic Knowledge embedding （DK-BLP） was proposed in this paper. The method dynamically integrated data-driven features and domain prior knowledge. Firstly， a dynamic knowledge graph was constructed， the prior knowledge of bearing degradation was encoded into computable triplets， and a sliding window confidence mechanism was introduced， so as to realize adaptive updating of domain knowledge. Second， relational graph convolutional networks were used to extract physically meaningful embedding vectors， which were then fused cross-modally with time-frequency features of vibration signals extracted by hierarchical convolutional networks. Finally， dynamic interactions between features and knowledge were modeled by using a multi-head self-attention Transformer， thereby allowing the model to adaptively balance the contributions of data features and mechanistic knowledge. Experimental results on the PHM2012 and XJTU datasets demonstrate that the proposed method significantly improves the accuracy of cross-condition bearing life prediction.

Estimating ambiguous Three-Dimensional （3D） human poses from monocular videos accurately is a current research challenge. Though the existing methods can estimate 3D joint coordinates using deep learning models， most of them fail to consider the multi-solution nature of this inverse problem adequately. Some multi-hypothesis estimation methods address multi-solution problems， but they suffer from insufficient cross-level feature fusion. To address these issues， a 3D human pose estimation model based on Temporal-SPatial Feature Pyramid Network （TSP-FPN） and multi-hypothesis interaction mechanism， called TSP-FPN-MHFormer （Temporal-SPatial Feature Pyramid Network-Multi-Hypothesis Transformer）， was proposed. Firstly， based on Transformer encoder， the multi-possibility distribution of human poses was captured by using multi-head self-attention mechanism， thereby generating multiple initial hypothesis features. Then， a TSP-FPN was designed， and a gated adaptive fusion strategy was employed to achieve dynamic weighted integration of multi-level skeleton sequence features， thereby balancing the fusion of local details and global temporal information effectively. Finally， based on Multi-Hypothesis Transformer （MHFormer）， a multi-hypothesis optimization module that combined joint Relative Position Bias （RPB） with a cross-attention mechanism was implemented， thereby facilitating cross-hypothesis communication and feature aggregation to enhance the model’s long-range reasoning capability to human topology for high-precision 3D joint coordinate estimation. Experimental results on the Human3.6M dataset demonstrate that the proposed model achieves a Mean Per Joint Position Error （MPJPE） of 42.3 mm， and reduces the estimation error by 1.6% compared to the state-of-the-art method MHFormer， indicating substantial progress obtained by the proposed model in addressing multi-solution challenge of monocular 3D pose estimation.

Internet of Things （IoT） device manufacturers often reuse a large number of open-source components compiled from open-source code in firmware development， with each firmware typically comprising hundreds of such components. If these components are not updated promptly， they may carry unpatched vulnerabilities to integrate into the firmware， thereby posing significant security risks to IoT devices. Therefore， identifying binary components in IoT firmware is crucial for ensuring the security of IoT devices. To address the difficulty of the existing methods in identifying binary components on a large scale， a large-scale IoT binary component identification method based on Named Entity Recognition （NER） was proposed. Firstly， internal binary components were extracted from firmware through decompression. Then， semantic information of the component was obtained through two ways： extraction of readable strings and execution of the component. Finally， the RoBERTa-BiLSTM-CRF’s NER model was utilized to identify component names and version numbers. Experimental results on 6 575 firmware samples released by 12 popular IoT manufacturers demonstrate that the proposed method achieves an F1 value of 87.67%， and identifying 163 binary components successfully. It can be seen that this method effectively expands the identification range of binary components in IoT firmware， enhancing firmware security from the perspective of software supply chain.

It is crucially important for researchers especially scientists to judge the correctness and timeliness of data and experiments according to provenance. Regarding the technologies about view materialization and data annotation, provenance has emerged to be a new research topic. Appropriate provenance data set is the foundation for verifying the accuracy and functionality of new techniques and/or algorithms on provenance management, meanwhile, the synthetic provenance data set is also of importance for verification and improvement of algorithms before gleaning the real provenance data to some expected extent. In this paper, one novel provenance database generator, ProGen was proposed, which was able to generate a provenance database, according to the input data schema and provenance annotation, with the specific data volume. The evaluation indicates that our design and implementation is efficient and scalable.