Search

Your search keyword '"LIAO Ming"' showing total 3,242 results
Start Over Author "LIAO Ming"
3,242 results on '"LIAO Ming"'

1. CoSafe: Evaluating Large Language Model Safety in Multi-Turn Dialogue Coreference

Author

Yu, Erxin, Li, Jing, Liao, Ming, Wang, Siqi, Gao, Zuchen, Mi, Fei, and Hong, Lanqing

Subjects
Computer Science - Computation and Language, Computer Science - Artificial Intelligence
Abstract

As large language models (LLMs) constantly evolve, ensuring their safety remains a critical research problem. Previous red-teaming approaches for LLM safety have primarily focused on single prompt attacks or goal hijacking. To the best of our knowledge, we are the first to study LLM safety in multi-turn dialogue coreference. We created a dataset of 1,400 questions across 14 categories, each featuring multi-turn coreference safety attacks. We then conducted detailed evaluations on five widely used open-source LLMs. The results indicated that under multi-turn coreference safety attacks, the highest attack success rate was 56% with the LLaMA2-Chat-7b model, while the lowest was 13.9% with the Mistral-7B-Instruct model. These findings highlight the safety vulnerabilities in LLMs during dialogue coreference interactions., Comment: Submitted to EMNLP 2024

Published
2024

2. BirdNeRF: Fast Neural Reconstruction of Large-Scale Scenes From Aerial Imagery

Author

Zhang, Huiqing, Xue, Yifei, Liao, Ming, and Lao, Yizhen

Subjects
Computer Science - Computer Vision and Pattern Recognition
Abstract

In this study, we introduce BirdNeRF, an adaptation of Neural Radiance Fields (NeRF) designed specifically for reconstructing large-scale scenes using aerial imagery. Unlike previous research focused on small-scale and object-centric NeRF reconstruction, our approach addresses multiple challenges, including (1) Addressing the issue of slow training and rendering associated with large models. (2) Meeting the computational demands necessitated by modeling a substantial number of images, requiring extensive resources such as high-performance GPUs. (3) Overcoming significant artifacts and low visual fidelity commonly observed in large-scale reconstruction tasks due to limited model capacity. Specifically, we present a novel bird-view pose-based spatial decomposition algorithm that decomposes a large aerial image set into multiple small sets with appropriately sized overlaps, allowing us to train individual NeRFs of sub-scene. This decomposition approach not only decouples rendering time from the scene size but also enables rendering to scale seamlessly to arbitrarily large environments. Moreover, it allows for per-block updates of the environment, enhancing the flexibility and adaptability of the reconstruction process. Additionally, we propose a projection-guided novel view re-rendering strategy, which aids in effectively utilizing the independently trained sub-scenes to generate superior rendering results. We evaluate our approach on existing datasets as well as against our own drone footage, improving reconstruction speed by 10x over classical photogrammetry software and 50x over state-of-the-art large-scale NeRF solution, on a single GPU with similar rendering quality.

Published
2024

3. Evolution and Antigenic Differentiation of Avian Influenza A(H7N9) Virus, China

Author

Liu, Yang, Chen, Yuhua, Yang, Zhiyi, Lin, Yaozhong, Fu, Siyuan, Chen, Junhong, Xu, Lingyu, Liu, Tengfei, Niu, Beibei, Huang, Qiuhong, Liu, Haixia, Zheng, Chaofeng, Liao, Ming, and Jia, Weixin

Subjects
Glycoproteins -- Health aspects, Medical research, Medicine, Experimental, Influenza -- Genetic aspects -- Prevention, Viral vaccines -- Testing, Health
Abstract

From early 2013 through October 2017, a total of 5 outbreaks of avian influenza A(H7N9) virus infection occurred, resulting in 616 human deaths (1). In particular, the fifth wave of [...]

Published
2024
Full Text
View/download PDF

4. DFR: Depth from Rotation by Uncalibrated Image Rectification with Latitudinal Motion Assumption

Author

Zhang, Yongcong, Xue, Yifei, Liao, Ming, Zhang, Huiqing, and Lao, Yizhen

Subjects
Computer Science - Computer Vision and Pattern Recognition
Abstract

Despite the increasing prevalence of rotating-style capture (e.g., surveillance cameras), conventional stereo rectification techniques frequently fail due to the rotation-dominant motion and small baseline between views. In this paper, we tackle the challenge of performing stereo rectification for uncalibrated rotating cameras. To that end, we propose Depth-from-Rotation (DfR), a novel image rectification solution that analytically rectifies two images with two-point correspondences and serves for further depth estimation. Specifically, we model the motion of a rotating camera as the camera rotates on a sphere with fixed latitude. The camera's optical axis lies perpendicular to the sphere's surface. We call this latitudinal motion assumption. Then we derive a 2-point analytical solver from directly computing the rectified transformations on the two images. We also present a self-adaptive strategy to reduce the geometric distortion after rectification. Extensive synthetic and real data experiments demonstrate that the proposed method outperforms existing works in effectiveness and efficiency by a significant margin.

Published
2023

5. Compressive Sensing in High-resolution 3D SAR Tomography of Urban Scenarios

Author

Liao Ming-sheng, Wei Lian-huan, Wang Zi-yun, Timo Balz, and Zhang Lu

Subjects
SAR tomography, Compressive Sensing (CS), Sparse reconstruction, Basis pursuit, TWo-step Iterative Shrinkage/Thresholding (TWIST), Super resolution, Electricity and magnetism, QC501-766
Abstract

In modern high resolution SAR data, due to the intrinsic side-looking geometry of SAR sensors, layover and foreshortening issues inevitably arise, especially in dense urban areas. SAR tomography provides a new way of overcoming these problems by exploiting the back-scattering property for each pixel. However, traditional non-parametric spectral estimators, e.g. Truncated Singular Value Decomposition (TSVD), are limited by their poor elevation resolution, which is not comparable to the azimuth and slant-range resolution. In this paper, the Compressive Sensing (CS) approach using Basis Pursuit (BP) and TWo-step Iterative Shrinkage/Thresholding (TWIST) are introduced. Experimental studies with real spotlight-mode TerraSAR-X dataset are carried out using both BP and TWIST, to demonstrate the merits of compressive sensing approaches in terms of robustness, computational efficiency, and super-resolution capability.

Published
2015
Full Text
View/download PDF

9. Measurable residual disease testing by next generation sequencing is more accurate compared with multiparameter flow cytometry in adults with B-cell acute lymphoblastic leukemia

Author

Yan, Nan, Wang, Zi-Long, Wang, Xiao-Juan, Gale, Robert Peter, Zhou, Ya-Lan, Zhao, Ming-Yue, Wu, Li-Xin, Liao, Ming-Yue, Yang, Jie, Wang, Chun-Yang, Zhu, Jian-Hua, Jiang, Hao, Jiang, Qian, Liu, Yan-Rong, Chang, Ying-Jun, Xu, Lan-Ping, Zhang, Xiao-Hui, Ma, Tong-Hui, Huang, Xiao-Jun, and Ruan, Guo-Rui

Published
2024
Full Text
View/download PDF

21. Curvature of matrix and reductive Lie groups

Author

Gan, Luyining, Liao, Ming, and Tam, Tin-Yau

Subjects
Mathematics - Differential Geometry, Mathematics - Algebraic Geometry, Mathematics - Representation Theory, 53B20, 14L35, 51N30
Abstract

In this paper, we give a simple formula for sectional curvatures on the general linear group, which is also valid for many other matrix groups. Similar formula is given for a reductive Lie group. We also discuss the relation between commuting matrices and zero sectional curvature., Comment: 11 pages; the final version is published in Journal of Lie Theory

Published
2021

31. Cryo-EM structure of severe fever with thrombocytopenia syndrome virus

Author

Du, Shouwen, Peng, Ruchao, Xu, Wang, Qu, Xiaoyun, Wang, Yuhang, Wang, Jiamin, Li, Letian, Tian, Mingyao, Guan, Yudong, Wang, Jigang, Wang, Guoqing, Li, Hao, Deng, Lingcong, Shi, Xiaoshuang, Ma, Yidan, Liu, Fengting, Sun, Minhua, Wei, Zhengkai, Jin, Ningyi, Liu, Wei, Qi, Jianxun, Liu, Quan, Liao, Ming, and Li, Chang

Published
2023
Full Text
View/download PDF

36. Towards Expectation-Maximization by SQL in RDBMS

Author

Zhao, Kangfei, Yu, Jeffrey Xu, Rong, Yu, Liao, Ming, and Huang, Junzhou

Subjects
Computer Science - Databases
Abstract

Integrating machine learning techniques into RDBMSs is an important task since there are many real applications that require modeling (e.g., business intelligence, strategic analysis) as well as querying data in RDBMSs. In this paper, we provide an SQL solution that has the potential to support different machine learning modelings. As an example, we study how to support unsupervised probabilistic modeling, that has a wide range of applications in clustering, density estimation and data summarization, and focus on Expectation-Maximization (EM) algorithms, which is a general technique for finding maximum likelihood estimators. To train a model by EM, it needs to update the model parameters by an E-step and an M-step in a while-loop iteratively until it converges to a level controled by some threshold or repeats a certain number of iterations. To support EM in RDBMSs, we show our answers to the matrix/vectors representations in RDBMSs, the relational algebra operations to support the linear algebra operations required by EM, parameters update by relational algebra, and the support of a while-loop. It is important to note that the SQL'99 recursion cannot be used to handle such a while-loop since the M-step is non-monotonic. In addition, assume that a model has been trained by an EM algorithm, we further design an automatic in-database model maintenance mechanism to maintain the model when the underlying training data changes.We have conducted experimental studies and will report our findings in this paper., Comment: 12 pages

Published
2021

41. Rapid detection of newly isolated Tembusu-related Flavivirus by reverse-transcription loop-mediated isothermal amplification assay

Author

Wang Youling, Yuan Xiaoyuan, Li Yufeng, Yu Kexiang, Yang Jinxing, Xu Huaiying, Zhang Yuxia, Yu Kangzhen, Liao Ming, and Qin Zhuoming

Subjects
Tembusu-related Flavivirus, newly-isolated virus, RT-PCR, RT-LAMP, Infectious and parasitic diseases, RC109-216
Abstract

Abstract Background From April 2010 to January 2011, a severe new viral disease had devastated most duck-farming regions in China. This disease affected not only laying ducks but also meat ducks, causing huge economic losses for the poultry industry. The objective of this study is to develop a one-step reverse transcription loop-mediated isothermal amplification (RT-LAMP) assay for the detection of the new virus related to Tembusu-related Flavivirus. Results The RT-LAMP assay is very simple and rapid, and the amplification can be completed within 50 min under isothermal conditions at 63°C by a set of 6 primers targeting the E gene based on the sequences analysis of the newly isolated viruses and other closely related Flavivirus.The monitoring of gene amplification can also be visualized by using SYBR green I fluorescent dye. In addition, the RT-LAMP assay for newly isolated Tembusu-related Flavivirus showed higher sensitivity with an RNA detection-limit of 2 copies/μL compared with 190 copies/μL of the conventional RT-PCR method. The specificity was identified without cross reaction to other common avian pathogens. By screening a panel of clinical samples this method was more feasible in clinical settings and there was higher positive coincidence rate than conventional RT-PCR and virus isolation. Conclusion The RT-LAMP assay for newly isolated Tembusu-related Flavivirus is a valuable tool for the rapid and real-time detection not only in well-equipped laboratories but also in general conditions.

Published
2011
Full Text
View/download PDF

42. Genetic characterization of H1N2 influenza a virus isolated from sick pigs in Southern China in 2010

Author

Kong Wei, Huang Liang, Qi Hai, Cao Nan, Zhang Liang, Wang Heng, Guan Shang, Qi Wen, Jiao Pei, Liao Ming, and Zhang Gui

Subjects
Infectious and parasitic diseases, RC109-216
Abstract

Abstract In China H3N2 and H1N1 swine influenza viruses have been circulating for many years. In January 2010, before swine were infected with foot and mouth disease in Guangdong, some pigs have shown flu-like symptoms: cough, sneeze, runny nose and fever. We collected the nasopharyngeal swab of all sick pigs as much as possible. One subtype H1N2 influenza viruses were isolated from the pig population. The complete genome of one isolate, designated A/swine/Guangdong/1/2010(H1N2), was sequenced and compared with sequences available in GenBank. The nucleotide sequences of all eight viral RNA segments were determined, and then phylogenetic analysis was performed using the neighbor-joining method. HA, NP, M and NS were shown to be closely to swine origin. PB2 and PA were close to avian origin, but NA and PB1were close to human origin. It is a result of a multiple reassortment event. In conclusion, our finding provides further evidence about the interspecies transmission of avian influenza viruses to pigs and emphasizes the importance of reinforcing swine influenza virus (SIV) surveillance, especially before the emergence of highly pathogenic FMDs in pigs in Guangdong.

Published
2011
Full Text
View/download PDF

43. A baculovirus dual expression system-based vaccine confers complete protection against lethal challenge with H9N2 avian influenza virus in mice

Author

Ye Yu, Cheng Xiaoliang, Fan Huiying, Lin Wenyao, Chen Xiaowei, Ren Tao, Qi Wenbao, and Liao Ming

Subjects
Avian influenza viruses, H9N2, vaccine, HA protein, Baculovirus dual expression system, immune response, Infectious and parasitic diseases, RC109-216
Abstract

Abstract Background Avian influenza viruses of H9N2 subtype have become highly prevalent in avian species. Although these viruses generally cause only mild to moderate disease, they can infect a wide variety of species, including chickens, quail, turkeys, ducks, geese, pheasant, partridge, and pigeon, even transmitted to mammalian species, including humans, accelerating the efforts to devise protective strategies against them. Results The results showed that stronger immune responses were induced in a mouse model immunized with BV-Dual-HA than in those vaccinated with a DNA vaccine encoding the same antigen. Moreover, complete protection against lethal challenge with H9N2 virus was observed in mice. Conclusion BV-Dual-HA could be utilized as a vaccine candidate against H9N2 virus infection.

Published
2011
Full Text
View/download PDF

44. SELDI-TOF MS profiling of serum for detection of nasopharyngeal carcinoma

Author

He Min, Deng Kai-Feng, Liao Ming, Zhang Bei-Bei, Xuan Chao, Huang Yuan-Jiao, and Zhao Jin-Min

Subjects
Neoplasms. Tumors. Oncology. Including cancer and carcinogens, RC254-282
Abstract

Abstract Background No satisfactory biomarkers are currently available to screen for nasopharyngeal carcinoma (NPC). We have developed and evaluated surface-enhanced laser desorption/ionization time-of-flight mass spectrometry (SELDI-TOF MS) for detection and analysis of multiple proteins for distinguishing individuals with NPC from control individuals. Methods A preliminary learning set and a classification tree of spectra derived from 24 patients with NPC and a group of 24 noncancer controls were used to develop a proteomic model that discriminated cancer from noncancer effectively. Then, the validity of the classification tree was challenged with a blind test set, which included another 20 patients with NPC and 12 noncancer controls. Results A panel of 3 biomarkers ranging m/z 3–20 k was selected to establish Decision Tree model by BPS with sensitivity of 91.66% and specificity of 95.83%. The ability to detect NPC patients was evaluated, a sensitivity of 95.0% and specificity of 83.33% were validated in blind testing set. Conclusion This high-flux proteomic classification system will provide a highly accurate and innovative approach for the detection/diagnosis of NPC.

Published
2009
Full Text
View/download PDF

Catalog

Books, media, physical & digital resources
See catalog results