arXiv subjects

Zhenchao Wang

Publications and source records attributed to Zhenchao Wang.

4 recordsLinked to original sources

arXiv2026

HNR-DAC: Hard-Negative Reranking and Distribution-Aligned Classification for Scientific Claim Verification

Scientific claim verification over a cited paper requires predicting the claim--paper relation and identifying the paragraphs that justify that prediction. This setting poses two linked challenges: within-paper distractors often resemble genuine evidence, while a classifier trained on gold evidence must operate on retrieved evidence at inference. We present HNR-DAC, a two-stage framework that trains each stage on the cases it will actually encounter. Hard-Negative Reranking (HNR) quantifies evidence confusability using a base reranker's scores on non-gold paragraphs and contrasts gold evidence against the most confusable candidates. Distribution-Aligned Classification (DAC) trains on the Top-1 paragraph produced by the same frozen HNR used to construct inference inputs, while HNR's Top-3 paragraph identifiers provide the evidence output. On the NLPCC 2026 Task 10 Track 2, the final configuration obtains 97.21% Hit@3, 95.79% Macro-F1, 94.47% Joint@3, and an average score of 95.13%. The corresponding submission ranks third on the official Track 2 leaderboard while achieving the highest overall Macro-F1 of 93.05%, alongside 70.16% Joint@3 and an average score of 81.61%.

cs.CL↗

arXiv2026

LayersReg: A Layer-by-Layer Progressive Regressor for Reliable Intraoperative 3D/2D Registration

3D/2D registration serves as a cornerstone technique in surgical navigation. Traditional iterative optimization algorithms suffer from low efficiency and high failure rates in intraoperative settings. Deep learning-based methods reformulate registration from iterative optimization to a regression problem that maps image appearance features to spatial pose, typically achieving improved real-time performance and accuracy. However, such learnable methods are confined to memory-driven retrieval of specific pose features rather than understanding the task of image alignment itself, which limits their generalization in complex scenarios. We propose LayersReg, a pioneering regression paradigm that endows the model with 3D anatomical awareness and searches for the correct pose in a progressive, layer-by-layer manner. Inspired by the iterative pose-searching optimization criterion of classical registration, LayersReg searches for correlations between the moving and fixed images in feature space, capturing the trend of pixel flow and thereby converging iteratively toward the correct spatial pose transformation. We further design a coupling of node-wise regression with the progressive registration framework to enhance the model's perception of spatial pose changes. Experimental results demonstrate that under large offsets and multimodality conditions, LayersReg achieves high accuracy on both X-ray/CT registration (0.68°, 1.41 mm) and slice localization (0.73°, 1.55 mm) tasks, outperforming existing state-of-the-art methods while meeting the intraoperative demands for precision and real-time capability.

cs.CV↗

arXiv2018

Space weathering of the Moon from in situ detection

Space weathering is an important surface process occurring on the Moon and other airless bodies, especially those that have no magnetic field. The optical effects of the Moon's space weathering have been largely investigated in the laboratory for lunar samples and lunar analogues. However, duplication of the pristine regolith on Earth is not possible. Here we report the space weathering from the unique perspective of the Chang'E-3's (CE-3) "Yutu" rover, building on our previous work (Wang et al. 2017; Wu and Hapke 2018). Measurement of the visually undisturbed uppermost regolith as well as locations that have been affected by rocket exhaust from the spacecraft by the Visible-Near Infrared Spectrometer (VNIS) revealed that the returned samples bring a biased information about the pristine lunar regolith. The uppermost surficial regolith is much more weathered than the regolith immediately below, and the finest fraction is rich in space weathered products. These materials are very dark and attenuated throughout the visible and near-infrared (VNIR) wavelengths, hence reduce the reflectance and mask the absorption features. The effects on the spectral slope caused by space weathering are wavelength-dependent: the visible and near-infrared continuum slope (VNCS) increases while the visible slope (VS) decreases. In the visible wavelength, the optical effects of space weathering and TiO2 are identical: both reduce albedo and blue the spectra. This suggests that developing new TiO2 abundance algorithm is needed. Optical maturity indices are composition related and hence only locally meaningful. Since optical remote sensing can only sense the uppermost few microns of regolith and since this surface tends to be very weathered, the interpretation of surface composition using optical remote sensing data needs to be carefully evaluated. Sampling the uppermost surface is suggested.

astro-ph.EP↗

arXiv2018

Mineralogy of northern nearside mare basalts

The mineralogy of mare basalts reflects the chemical composition of the magma source, as well as the physical and chemical environment of the rock's formation. It is significant for understanding the thermal evolution of the Moon. In this study, the spatial distribution of the mineralogy of the lunar northern nearside basalts was mapped using the Moon Mineralogy Mapper (M3) data. The study area, which is an elongated mare, Mare Frigoris and northern Mare Imbrium, was mapped and characterized into 27 units based on multi-source data, including spectrum, terrain, and element abundance. We extracted 177 M3 spectra from fresh craters. Spectral parameters such as absorption center and band area ratio (BAR) were obtained through data processing. The variation of mafic mineral of this region was acquired by analyzing these parameters. The basaltic units in eastern Mare Frigoris, which are older, have been found to be dominated by clinopyroxene with lower CaO compared to the returned lunar samples; this is similar to older basaltic units in Mare Imbrium. The basaltic units of western Mare Frigoris and Sinus Roris which are younger have been found to be rich in olivine. The late-stage basalts in Oceanus Procellarum and Mare Imbrium show the same feature. These wide-spread olivine-rich basalts suggest the uniqueness in the evolution of the Moon. Geographically speaking, Mare Frigoris is an individual mare, but the magma source region have connections with surrounding maria in consideration of mineral differences between western and eastern Frigoris, as well as mineral similarities with maria at the same location.

astro-ph.EP↗