arXiv subjects

Sicheng Yu

Publications and source records attributed to Sicheng Yu.

26 records · Page 2Linked to original sources

arXiv2025

RGB-Only Gaussian Splatting SLAM for Unbounded Outdoor Scenes

3D Gaussian Splatting (3DGS) has become a popular solution in SLAM, as it can produce high-fidelity novel views. However, previous GS-based methods primarily target indoor scenes and rely on RGB-D sensors or pre-trained depth estimation models, hence underperforming in outdoor scenarios. To address this issue, we propose a RGB-only gaussian splatting SLAM method for unbounded outdoor scenes--OpenGS-SLAM. Technically, we first employ a pointmap regression network to generate consistent pointmaps between frames for pose estimation. Compared to commonly used depth maps, pointmaps include spatial relationships and scene geometry across multiple views, enabling robust camera pose estimation. Then, we propose integrating the estimated camera poses with 3DGS rendering as an end-to-end differentiable pipeline. Our method achieves simultaneous optimization of camera poses and 3DGS scene parameters, significantly enhancing system tracking accuracy. Specifically, we also design an adaptive scale mapper for the pointmap regression network, which provides more accurate pointmap mapping to the 3DGS map representation. Our experiments on the Waymo dataset demonstrate that OpenGS-SLAM reduces tracking error to 9.8\% of previous 3DGS methods, and achieves state-of-the-art results in novel view synthesis. Project Page: https://3dagentworld.github.io/opengs-slam/

cs.CV↗

arXiv2024

OVFoodSeg: Elevating Open-Vocabulary Food Image Segmentation via Image-Informed Textual Representation

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 with large food segmentation datasets. Existing approaches primarily utilize a closed-vocabulary and static text embeddings setting. These methods often fall short in effectively handling the ingredients, particularly new and diverse ones. In response to these limitations, we introduce OVFoodSeg, a framework that adopts an open-vocabulary setting and enhances text embeddings with visual context. By integrating vision-language models (VLMs), our approach enriches text embedding with image-specific information through two innovative modules, eg, an image-to-text learner FoodLearner and an Image-Informed Text Encoder. The training process of OVFoodSeg is divided into two stages: the pre-training of FoodLearner and the subsequent learning phase for segmentation. The pre-training phase equips FoodLearner with the capability to align visual information with corresponding textual representations that are specifically related to food, while the second phase adapts both the FoodLearner and the Image-Informed Text Encoder for the segmentation task. By addressing the deficiencies of previous models, OVFoodSeg demonstrates a significant improvement, achieving an 4.9\% increase in mean Intersection over Union (mIoU) on the FoodSeg103 dataset, setting a new milestone for food image segmentation.

cs.CV↗

arXiv2024

GliDe with a CaPE: A Low-Hassle Method to Accelerate Speculative Decoding

Speculative decoding is a relatively new decoding framework that leverages small and efficient draft models to reduce the latency of LLMs. In this study, we introduce GliDe and CaPE, two low-hassle modifications to vanilla speculative decoding to further improve the decoding speed of a frozen LLM. Specifically, GliDe is a modified draft model architecture that reuses the cached keys and values from the target LLM, while CaPE is a proposal expansion method that uses the draft model's confidence scores to help select additional candidate tokens for verification. Extensive experiments on different benchmarks demonstrate that our proposed GliDe draft model significantly reduces the expected decoding latency. Additional evaluation using walltime reveals that GliDe can accelerate Vicuna models up to 2.17x and further extend the improvement to 2.61x with CaPE. We will release our code, data, and the trained draft models.

cs.CL↗

arXiv2022

ELM of ELM-WD: An extremely low mass hot star discovered in LAMOST survey

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 proposed by observation to be less massive than that. Here we report the discovery of a binary system, LAMOST J224040.77-020732.8 (J2240 in short), which consists of a very low mass hot star and a compact companion. Multi-epoch spectroscopy shows an orbital period $P_{orb} =$0.219658$\pm0.000002$ days and a radial velocity semi-amplitude $K1=318.5\pm3.3km/s$, which gives the mass function of 0.74$M_{\odot}$, indicating the companion is a compact star. The F-type low resolution spectra illustrate no emission features, and the temperature ($\sim 7400K$) is consistent with that from Spectral Energy Distribution fitting and multi-color light curve solution. The optical light curves, in ZTF g, r and i bands and Catalina V band, show ellipsoidal variability with amplitudes $\sim30\%$, suggesting that the visible component is heavily tidally distorted. Combining the distance from Gaia survey, the ZTF light curves are modeled with Wilson-Devinney code and the result shows that the mass of the visible component is $M1=0.085^{+0.036}_{-0.024}M_{\odot}$, and the mass of the invisible component is $M2=0.98^{+0.16}_{-0.09}M_{\odot}$. The radius of the visible component is $R1=0.29^{+0.04}_{-0.03}R_{\odot}$. The inclination angle is approximately between 60$^{\circ}$ and 90$^{\circ}$. The observations indicate the system is most likely a pre-ELM WD + WD/NS binary, and the mass of pre-ELM is possibly lower than the $0.14M_{\odot}$ theoretical limit.

astro-ph.SR↗

arXiv2022

Discovery of one neutron star candidate from radial velocity monitoring

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$\pm$0.2; $M =$ 1.7$\pm$0.1 M$_{\odot}$; $R\ =\ 1.7\pm0.2$ R$_{\odot}$), at a distance of 1.11$\pm0.03$ kpc. No double-lined feature can be seen from the GTC/HORuS high-resolution spectra, thus the radial velocity variation indicates an invisible object hiding in the binary. The system's optical light curves show clear ellipsoidal variability, suggesting that the visible companion is tidal distorted. By fitting the multi-band light curves with the ELC and WD codes, we constrain the mass of the invisible star to be 1.1--1.3 M$_{\odot}$. Spectral disentangling shows no additional component with optical absorption spectra, supporting the system contains one compact object. No X-ray or UV emission are detected in the ROSAT archive observations. Therefore, we suspect the invisible object is more likely a neutron star rather than a white dwarf. Our finding suggests the ability of LAMOST spectroscopic survey to discover X-ray quiescent compact objects.

astro-ph.HE↗

arXiv2021

NOAHQA: Numerical Reasoning with Interpretable Graph Question Answering Dataset

While diverse question answering (QA) datasets have been proposed and contributed significantly to the development of deep learning models for QA tasks, the existing datasets fall short in two aspects. First, we lack QA datasets covering complex questions that involve answers as well as the reasoning processes to get the answers. As a result, the state-of-the-art QA research on numerical reasoning still focuses on simple calculations and does not provide the mathematical expressions or evidences justifying the answers. Second, the QA community has contributed much effort to improving the interpretability of QA models. However, these models fail to explicitly show the reasoning process, such as the evidence order for reasoning and the interactions between different pieces of evidence. To address the above shortcomings, we introduce NOAHQA, a conversational and bilingual QA dataset with questions requiring numerical reasoning with compound mathematical expressions. With NOAHQA, we develop an interpretable reasoning graph as well as the appropriate evaluation metric to measure the answer quality. We evaluate the state-of-the-art QA models trained using existing QA datasets on NOAHQA and show that the best among them can only achieve 55.5 exact match scores, while the human performance is 89.7. We also present a new QA model for generating a reasoning graph where the reasoning graph metric still has a large gap compared with that of humans, e.g., 28 scores.

cs.CL↗

arXiv2020

Counterfactual Variable Control for Robust and Interpretable Question Answering

Deep neural network based question answering (QA) models are neither robust nor explainable in many cases. For example, a multiple-choice QA model, tested without any input of question, is surprisingly "capable" to predict the most of correct options. In this paper, we inspect such spurious "capability" of QA models using causal inference. We find the crux is the shortcut correlation, e.g., unrobust word alignment between passage and options learned by the models. We propose a novel approach called Counterfactual Variable Control (CVC) that explicitly mitigates any shortcut correlation and preserves the comprehensive reasoning for robust QA. Specifically, we leverage multi-branch architecture that allows us to disentangle robust and shortcut correlations in the training process of QA. We then conduct two novel CVC inference methods (on trained models) to capture the effect of comprehensive reasoning as the final prediction. For evaluation, we conduct extensive experiments using two BERT backbones on both multi-choice and span-extraction QA benchmarks. The results show that our CVC achieves high robustness against a variety of adversarial attacks in QA while maintaining good interpretation ability.

cs.CL↗

arXiv2020

Context Modeling with Evidence Filter for Multiple Choice Question Answering

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 requirement of extracting "evidence" is particularly important due to the mutual independence of sentences in the context. Existing work tackles this problem by annotated evidence or distant supervision with rules which overly rely on human efforts. To address the challenge, we propose a simple yet effective approach termed evidence filtering to model the relationships between the encoded contexts with respect to different options collectively and to potentially highlight the evidence sentences and filter out unrelated sentences. In addition to the effective reduction of human efforts of our approach compared, through extensive experiments on OpenbookQA, we show that the proposed approach outperforms the models that use the same backbone and more training data; and our parameter analysis also demonstrates the interpretability of our approach.

cs.CL↗