Search

Your search keyword '"Liu, Songlin"' showing total 1,279 results
Start Over Author "Liu, Songlin"
1,279 results on '"Liu, Songlin"'

5. From CLIP to DINO: Visual Encoders Shout in Multi-modal Large Language Models

Author

Jiang, Dongsheng, Liu, Yuchen, Liu, Songlin, Zhao, Jin'e, Zhang, Hao, Gao, Zhen, Zhang, Xiaopeng, Li, Jin, and Xiong, Hongkai

Subjects
Computer Science - Computer Vision and Pattern Recognition
Abstract

Multi-modal Large Language Models (MLLMs) have made significant strides in expanding the capabilities of Large Language Models (LLMs) through the incorporation of visual perception interfaces. Despite the emergence of exciting applications and the availability of diverse instruction tuning data, existing approaches often rely on CLIP or its variants as the visual branch, and merely extract features from the deep layers. However, these methods lack a comprehensive analysis of the visual encoders in MLLMs. In this paper, we conduct an extensive investigation into the effectiveness of different vision encoders within MLLMs. Our findings reveal that the shallow layer features of CLIP offer particular advantages for fine-grained tasks such as grounding and region understanding. Surprisingly, the vision-only model DINO, which is not pretrained with text-image alignment, demonstrates promising performance as a visual branch within MLLMs. By simply equipping it with an MLP layer for alignment, DINO surpasses CLIP in fine-grained related perception tasks. Building upon these observations, we propose a simple yet effective feature merging strategy, named COMM, that integrates CLIP and DINO with Multi-level features Merging, to enhance the visual capabilities of MLLMs. We evaluate COMM through comprehensive experiments on a wide range of benchmarks, including image captioning, visual question answering, visual grounding, and object hallucination. Experimental results demonstrate the superior performance of COMM compared to existing methods, showcasing its enhanced visual capabilities within MLLMs.

Published
2023

8. Investigation on Rectangular Tunnel Mechanical Response Under Groundwater Level Fluctuation Based on Model Test and Numerical Simulation

Author

Sun, Weixin, Liu, Hanlong, Zhang, Wengang, Li, Yongqin, Liu, Songlin, di Prisco, Marco, Series Editor, Chen, Sheng-Hong, Series Editor, Vayas, Ioannis, Series Editor, Kumar Shukla, Sanjay, Series Editor, Sharma, Anuj, Series Editor, Kumar, Nagesh, Series Editor, Wang, Chien Ming, Series Editor, Cui, Zhen-Dong, Series Editor, Wu, Wei, editor, Leung, Chun Fai, editor, Zhou, Yingxin, editor, and Li, Xiaozhao, editor

Published
2024
Full Text
View/download PDF

31. Research on Intelligent Optimisation Strategies for Interior Space Layout Design with the Aid of 3DsMax+AI

Author

Liu Songlin

Subjects
deep network, ai intelligent evolution, wave function collapse, space layout design, 97p10, Mathematics, QA1-939
Abstract

Indoor intelligent assistive technology has been widely used in space layout design optimization. This paper proposes a heuristic scene space layout generation method based on the smart evolution of scene space AI with WFC and deep network, joint wave function collapse algorithm, as well as two deep neural networks to generate the layout wireframe diagram while strangling reasonable boundaries and predicting functional semantic labels for the layout subspace. Applying the method of this paper to the office space layout optimization design, the results show that 95.68% of the area of the office and conference room has a lighting coefficient greater than 2%, and 85.78% of the area of the design room has a lighting coefficient greater than 3%. Ordinary white glass can be utilized in the design room to enhance the lighting quality, with a lighting coefficient of 2.87 and a compliance rate of 100%. The inset atrium space has a greater percentage of indoor thermal comfort area and a faster air flow rate. In this paper, 3DsMax+AI intelligent assistance and other technologies are used to simulate the layout design of the office space and realize the optimization of its layout design.

Published
2024
Full Text
View/download PDF

32. A Hybrid Recommender System for Recommending Smartphones to Prospective Customers

Author

Biswas, Pratik K. and Liu, Songlin

Subjects
Computer Science - Information Retrieval, Computer Science - Machine Learning, I.m
Abstract

Recommender Systems are a subclass of machine learning systems that employ sophisticated information filtering strategies to reduce the search time and suggest the most relevant items to any particular user. Hybrid recommender systems combine multiple recommendation strategies in different ways to benefit from their complementary advantages. Some hybrid recommender systems have combined collaborative filtering and content-based approaches to build systems that are more robust. In this paper, we propose a hybrid recommender system, which combines Alternating Least Squares (ALS) based collaborative filtering with deep learning to enhance recommendation performance as well as overcome the limitations associated with the collaborative filtering approach, especially concerning its cold start problem. In essence, we use the outputs from ALS (collaborative filtering) to influence the recommendations from a Deep Neural Network (DNN), which combines characteristic, contextual, structural and sequential information, in a big data processing framework. We have conducted several experiments in testing the efficacy of the proposed hybrid architecture in recommending smartphones to prospective customers and compared its performance with other open-source recommenders. The results have shown that the proposed system has outperformed several existing hybrid recommender systems., Comment: Expert Systems With Applications, 2022

Published
2021
Full Text
View/download PDF

Catalog

Books, media, physical & digital resources
See catalog results