The Solution for Temporal Sound Localisation Task of ICCV 1st Perception Test Challenge 2023

In this paper, we propose a solution for improving the quality of temporal sound localization. We employ a multimodal fusion approach to combine visual and audio features. High-quality visual features are extracted using a state-of-the-art self-supervised pre-training network, resulting in efficient video feature representations. At the same time, audio features serve as complementary information to help the model better localize the start and end of sounds. The fused features are trained in a multi-scale Transformer for training. In the final test dataset, we achieved a mean average precision (mAP) of 0.33, obtaining the second-best performance in this track.

The Solution for the CVPR 2023 1st foundation model challenge-Track2

In this paper, we propose a solution for cross-modal transportation retrieval. Due to the cross-domain problem of traffic images, we divide the problem into two sub-tasks of pedestrian retrieval and vehicle retrieval through a simple strategy. In pedestrian retrieval tasks, we use IRRA as the base model and specifically design an Attribute Classification to mine the knowledge implied by attribute labels. More importantly, We use the strategy of Inclusion Relation Matching to make the image-text pairs with inclusion relation have similar representation in the feature space. For the vehicle retrieval task, we use BLIP as the base model. Since aligning the color attributes of vehicles is challenging, we introduce attribute-based object detection techniques to add color patch blocks to vehicle images for color data augmentation. This serves as strong prior information, helping the model perform the image-text alignment. At the same time, we incorporate labeled attributes into the image-text alignment loss to learn fine-grained alignment and prevent similar images and texts from being incorrectly separated. Our approach ranked first in the final B-board test with a score of 70.9.

Knowledge Navigation: Inferring the Interlocking Map of Knowledge from Research Trajectories

"If I have seen further, it is by standing on the shoulders of giants," Isaac Newton's renowned statement hints that new knowledge builds upon existing foundations, which means there exists an interdependent relationship between knowledge, which, yet uncovered, is implied in the historical development of scientific systems for hundreds of years. By leveraging natural language processing techniques, this study introduces an innovative embedding scheme designed to infer the "knowledge interlocking map." This map, derived from the research trajectories of millions of scholars, reveals the intricate connections among knowledge. We validate that the inferred map effectively delineates disciplinary boundaries and captures the intricate relationships between diverse concepts. The utility of the interlocking map is showcased through multiple applications. Firstly, we demonstrated the multi-step analogy inferences within the knowledge space and the functional connectivity between concepts in different disciplines. Secondly, we trace the evolution of knowledge across domains, observing trends such as shifts from "Theoretical" to "Applied" or "Chemistry" to "Biomedical" along predefined functional directions. Lastly, by analyzing the high-dimensional knowledge network structure, we found that knowledge connects each other with shorter global pathways, and the interdisciplinary knowledge plays a critical role in accessibility of the global knowledge network. Our framework offers a novel approach to mining knowledge inheritance pathways in extensive scientific literature, which is of great significance for understanding scientific development patterns, tailoring scientific learning trajectories, and accelerating scientific progress.