Sicheng Yu
YOU?
Author Swipe
View article: An Uncertainty-aware DETR Enhancement Framework for Object Detection
An Uncertainty-aware DETR Enhancement Framework for Object Detection Open
View article: Unposed 3DGS Reconstruction with Probabilistic Procrustes Mapping
Unposed 3DGS Reconstruction with Probabilistic Procrustes Mapping Open
3D Gaussian Splatting (3DGS) has emerged as a core technique for 3D representation. Its effectiveness largely depends on precise camera poses and accurate point cloud initialization, which are often derived from pretrained Multi-View Stere…
View article: Outdoor Monocular SLAM with Global Scale-Consistent 3D Gaussian Pointmaps
Outdoor Monocular SLAM with Global Scale-Consistent 3D Gaussian Pointmaps Open
3D Gaussian Splatting (3DGS) has become a popular solution in SLAM due to its high-fidelity and real-time novel view synthesis performance. However, some previous 3DGS SLAM methods employ a differentiable rendering pipeline for tracking, l…
View article: Sparse-to-Dense: A Free Lunch for Lossless Acceleration of Video Understanding in LLMs
Sparse-to-Dense: A Free Lunch for Lossless Acceleration of Video Understanding in LLMs Open
Due to the auto-regressive nature of current video large language models (Video-LLMs), the inference latency increases as the input sequence length grows, posing challenges for the efficient processing of video sequences that are usually v…
View article: Sparse-to-Dense: A Free Lunch for Lossless Acceleration of Video Understanding in LLMs
Sparse-to-Dense: A Free Lunch for Lossless Acceleration of Video Understanding in LLMs Open
View article: Frame-Voyager: Learning to Query Frames for Video Large Language Models
Frame-Voyager: Learning to Query Frames for Video Large Language Models Open
Video Large Language Models (Video-LLMs) have made remarkable progress in video understanding tasks. However, they are constrained by the maximum length of input tokens, making it impractical to input entire videos. Existing frame selectio…
View article: OVFoodSeg: Elevating Open-Vocabulary Food Image Segmentation via Image-Informed Textual Representation
OVFoodSeg: Elevating Open-Vocabulary Food Image Segmentation via Image-Informed Textual Representation Open
In the realm of food computing, segmenting ingredients from images poses substantial challenges due to the large intra-class variance among the same ingredients, the emergence of new ingredients, and the high annotation costs associated wi…
View article: ELM of ELM-WD: An Extremely-low-mass Hot Star Discovered in LAMOST Survey
ELM of ELM-WD: An Extremely-low-mass Hot Star Discovered in LAMOST Survey Open
The extremely-low-mass white dwarfs (ELM WDs) and pre-ELM WDs are helium-core white dwarfs with mass <∼ 0.3 M ⊙ . Evolution simulations show that a lower mass limit for ELM WDs exists at ≈0.14 M ⊙ and no ELM WD is proposed by observation t…
View article: Discovery of One Neutron Star Candidate from Radial-velocity Monitoring
Discovery of One Neutron Star Candidate from Radial-velocity Monitoring Open
We report the discovery of one possible neutron star binary ( P orb = 0.8666 days) by using LAMOST low-resolution spectroscopic data. The visible companion is a late A-type dwarf ( T eff = 7900 ± 200 K; log g = 4.3 ± 0.2; M = 1.7 ± 0.1 M ⊙…
View article: Discovery of one neutron star candidate from radial velocity monitoring
Discovery of one neutron star candidate from radial velocity monitoring Open
We report the discovery of one possible neutron star binary ($P_{\rm orb} =$ 0.8666 day) by using the LAMOST low-resolution spectroscopic data. The visible companion is a late A-type dwarf ($T_{\rm eff} = 7900 \pm 200$ K; log$g$ $=$ 4.3$\p…
View article: ELM of ELM-WD: An extremely low mass hot star discovered in LAMOST survey
ELM of ELM-WD: An extremely low mass hot star discovered in LAMOST survey Open
The Extremely Low Mass White Dwarfs (ELM WDs) and pre-ELM WDs are helium core white dwarfs with mass $<\sim 0.3M_{\odot}$. Evolution simulations show that a lower mass limit for ELM WDs exists at $\approx0.14M_{\odot}$ and no one is propos…
View article: Translate-Train Embracing Translationese Artifacts
Translate-Train Embracing Translationese Artifacts Open
Translate-train is a general training approach to multilingual tasks. The key idea is to use the translator of the target language to generate training data to mitigate the gap between the source and target languages. However, its performa…
View article: Interventional Training for Out-Of-Distribution Natural Language Understanding
Interventional Training for Out-Of-Distribution Natural Language Understanding Open
Out-of-distribution (OOD) settings are used to measure a model’s performance when the distribution of the test data is different from that of the training data. NLU models are known to suffer in OOD. We study this issue from the perspectiv…
View article: The first data release of LAMOST low-resolution single-epoch spectra
The first data release of LAMOST low-resolution single-epoch spectra Open
LAMOST Data Release 5, covering ∼17 000 deg 2 from –10° to 80° in declination, contains 9 million co-added low-resolution spectra of celestial objects, each spectrum combined from repeat exposure of two to tens of times during Oct 2011 to …
View article: Analysis of Phase Noise Performance in Spatially Separated Backscatter Systems
Analysis of Phase Noise Performance in Spatially Separated Backscatter Systems Open
Backscatter radio systems are generally considered to be suitable only for short range transmission limited by the two-way path losses and receiver sensitivity. Using a simple model of the receiver noise due to the leakage of the carrier s…
View article: Context Modeling with Evidence Filter for Multiple Choice Question Answering
Context Modeling with Evidence Filter for Multiple Choice Question Answering Open
Multiple-Choice Question Answering (MCQA) is a challenging task in machine reading comprehension. The main challenge in MCQA is to extract "evidence" from the given context that supports the correct answer. In the OpenbookQA dataset, the r…
View article: Data processing and data products from 2017 to 2019 campaign of astronomical site testing at Ali, Daocheng and Muztagh-ata
Data processing and data products from 2017 to 2019 campaign of astronomical site testing at Ali, Daocheng and Muztagh-ata Open
Based on previous site testing and satellite cloud data, Ali, Daocheng and Muztagh-ata have been selected as candidate sites for the Large Optical/Infrared Telescope (LOT) in China. We present the data collection, processing, management an…