Search

In this paper, we prove a conjecture of Schnell in the surface case., Comment: comments welcome!

Few-shot class-incremental learning (FSCIL) aims to mitigate the catastrophic forgetting issue when a model is incrementally trained on limited data. While the Contrastive Vision-Language Pre-Training (CLIP) model has been effective in addressing 2D few/zero-shot learning tasks, its direct application to 3D FSCIL faces limitations. These limitations arise from feature space misalignment and significant noise in real-world scanned 3D data. To address these challenges, we introduce two novel components: the Redundant Feature Eliminator (RFE) and the Spatial Noise Compensator (SNC). RFE aligns the feature spaces of input point clouds and their embeddings by performing a unique dimensionality reduction on the feature space of pre-trained models (PTMs), effectively eliminating redundant information without compromising semantic integrity. On the other hand, SNC is a graph-based 3D model designed to capture robust geometric information within point clouds, thereby augmenting the knowledge lost due to projection, particularly when processing real-world scanned data. Considering the imbalance in existing 3D datasets, we also propose new evaluation metrics that offer a more nuanced assessment of a 3D FSCIL model. Traditional accuracy metrics are proved to be biased; thus, our metrics focus on the model's proficiency in learning new classes while maintaining the balance between old and new classes. Experimental results on both established 3D FSCIL benchmarks and our dataset demonstrate that our approach significantly outperforms existing state-of-the-art methods.

3D generation has raised great attention in recent years. With the success of text-to-image diffusion models, the 2D-lifting technique becomes a promising route to controllable 3D generation. However, these methods tend to present inconsistent geometry, which is also known as the Janus problem. We observe that the problem is caused mainly by two aspects, i.e., viewpoint bias in 2D diffusion models and overfitting of the optimization objective. To address it, we propose a two-stage 2D-lifting framework, namely DreamControl, which optimizes coarse NeRF scenes as 3D self-prior and then generates fine-grained objects with control-based score distillation. Specifically, adaptive viewpoint sampling and boundary integrity metric are proposed to ensure the consistency of generated priors. The priors are then regarded as input conditions to maintain reasonable geometries, in which conditional LoRA and weighted score are further proposed to optimize detailed textures. DreamControl can generate high-quality 3D content in terms of both geometry consistency and texture fidelity. Moreover, our control-based optimization guidance is applicable to more downstream tasks, including user-guided generation and 3D animation. The project page is available at https://github.com/tyhuang0428/DreamControl., Comment: Accepted by CVPR 2024

The neutrinophilic two Higgs doublet model is one of the simplest models to explain the origin of tiny Dirac neutrino masses. This model introduces a new Higgs doublet with eV scale VEV to naturally generate the tiny neutrino masses. Depending on the same Yukawa coupling, the neutrino oscillation patterns can be probed with the dilepton signature from the decay of charged scalar $H^\pm$. For example, the normal hierarchy predicts BR$(H^+\to e^+\nu)\ll$ BR$(H^+\to \mu^+\nu)\approx$ BR$(H^+\to \tau^+\nu)\simeq0.5$ when the lightest neutrino mass is below 0.01 eV, while the inverted hierarchy predicts BR$(H^+\to e^+\nu)/2\simeq$ BR$(H^+\to \mu^+\nu)\simeq$ BR$(H^+\to \tau^+\nu)\simeq0.25$. By precise measurement of BR$(H^+\to \ell^+\nu)$, we are hopefully to probe the lightest neutrino mass and the atmospheric mixing angle $\theta_{23}$. Through the detailed simulation of the dilepton signature and corresponding backgrounds, we find that the 3 TeV CLIC could discover $M_{H^+}\lesssim1220$ GeV for NH and $M_{H^+}\lesssim1280$ GeV for IH. Meanwhile, the future 100 TeV FCC-hh collider could probe $M_{H^+}\lesssim1810$ GeV for NH and $M_{H^+}\lesssim2060$ GeV for IH., Comment: 18 pages, 9 figures

Pressure-induced magnetic phase transition is attracting interest due to its ability to detect superconducting behaviour at high pressures in diamond anvil cells. However, detection of the local sample magnetic properties is a great challenge due to the small sample chamber volume. Recently, optically detected magnetic resonance (ODMR) of nitrogen vacancy (NV) centers in diamond have been used for in-situ pressure-induced phase transition detection. However, owing to their four orientation axes and temperature-dependent zero-field-splitting, interpreting the observed ODMR spectra of NV centers remain challenging. Here, we study the optical and spin properties of implanted silicon vacancy defects in 4H-SiC, which is single-axis and temperature-independent zero-field-splitting. Using this technique, we observe the magnetic phase transition of Nd2Fe14B at about 7 GPa and map the critical temperature-pressure phase diagram of the superconductor YBa2Cu3O6.6. These results highlight the potential of silicon vacancy-based quantum sensors for in-situ magnetic detection at high pressures.

Carbon disulfide, CS$_2$, is an archetypal double-bonded molecular system belonging to the rich class of group IV-group VI, AB$_2$ compounds. It is widely and since long time believed that upon compression at several GPa a polymeric chain of type (-(C=S)-S-)$_n$ named Bridgman's black polymer will form. By combining optical spectroscopy and synchrotron X-ray diffraction data with ab initio simulations, we demonstrate that the structure of the Bridgman's black polymer is remarkably different. Solid molecular CS$_2$ undergoes a pressure-induced structural transformation at around 10-11 GPa, developing a disordered polymeric system. The polymer consists of 3-fold and 4-fold coordinated carbon atoms with an average carbon coordination continuously increasing upon further compression to 40 GPa. Polymerization also gives rise to some C=C double bonds. Upon decompression, the structural changes are partially reverted, a very small amount of molecular CS$_2$ is recovered, while the sample undergoes partial chemical disproportionation. Our work uncovers the non-trivial high-pressure structural evolution in one of the simplest molecular systems exhibiting molecular as well as polymeric phases., Comment: 8 pages, 5 figures, supplementary materials

From the Feynman path integration theory of view, the Hanbury Brown--Twiss effect would not be observed for one definite two-photon propagation path, as well as the superbunching effect. Here, temporal and spatial superbunching effects are measured from a pair of modulated distinguishable classical light. These interesting phenomena are realized by passing two orthogonal polarized laser beams through two rotating ground glass plates in sequence. To understand the underlying physical process, the intensity fluctuation correlation theory is developed to describe the superbunching effect in the temporal and spatial domain, which agrees with experimental results well. Such experimental results are conducive to the study of superbunching effect which plays an important role in improving the performance in related applications, such as the contrast of ghost imaging., Comment: 12 pages, 4 figures

In this paper we consider the Galois covers of algebraic surfaces of degree 6, with all associated planar degenerations. We compute the fundamental groups of those Galois covers, using their degeneration. We show that for 8 types of degenerations the fundamental group of the Galois cover is non-trivial and for 20 types it is trivial. Moreover, we compute the Chern numbers of all the surfaces with this type of degeneration and prove that the signatures of all their Galois covers are negative. We formulate a conjecture regarding the structure of the fundamental groups of the Galois covers based on our findings. With an appendix by the authors listing the detailed computations and an appendix by Guo Zhiming classifying degree 6 planar degenerations., Comment: Main paper is 21 pages and contains 7 figures. The appendix is 65 pages and contains 42 figures. arXiv admin note: substantial text overlap with arXiv:1610.09612