Searcharxiv⌕ Search

arXiv subjects

Qiyu Wang

Publications and source records attributed to Qiyu Wang.

4 recordsLinked to original sources

Marginal-Preserving Coupled Diffusion Composition for Long-Horizon Planning

Compositional diffusion planners extend short-horizon trajectory models to long-horizon goal-conditioned planning by generating and stitching multiple local trajectory chunks. However, in multimodal environments, local plausibility is not sufficient for global consistency: independently initialized chunks may denoise toward incompatible overlap modes, producing plans that are individually feasible but incoherent after stitching. Existing remedies typically add search, guidance, pruning, or graph-based structure at inference time, increasing computational cost or relying on task-specific machinery. We instead investigate graph-structured dependence among chunk-wise initial diffusion noises as a lightweight coordination mechanism for compositional trajectory planning. We propose \textbf{M}arginal-Preserving \textbf{C}oupled \textbf{D}iffusion \textbf{C}omposition (MCDC), a training-free inference method for pretrained compositional diffusion planners. Instead of sampling chunk noises independently, MCDC constructs coupled Gaussian initializations whose per-chunk marginals remain exactly standard Gaussian, as required by the pretrained denoiser, while neighboring chunks receive positively correlated noise components. This preserves the standard-Gaussian initialization marginal of each chunk while biasing adjacent chunks toward compatible overlap-mode choices. We prove exact marginal preservation and show, under a simple multimodal overlap model, that positive coupling increases the probability of compatible mode selection. On OGBench long-horizon stitch tasks, MCDC consistently improves over independent compositional diffusion sampling while retaining nearly the same inference time, and achieves a favorable success--cost trade-off compared with search-based composition methods.

cs.AI↗

Contactless cavity sensing of superfluid stiffness in atomically thin 4Hb-TaS$_2$

The exceptional tunability of two-dimensional van der Waals materials offers unique opportunities for exploring novel superconducting phases. However, in such systems, the measurement of superfluid phase stiffness, a fundamental property of a superconductor, is challenging because of the mesoscopic sample size. Here, we introduce a contact-free technique for probing the electrodynamic response, and thereby the phase stiffness, of atomically thin superconductors using on-chip superconducting microwave resonators. We demonstrate this technique on 4Hb-TaS$_2$, a van der Waals superconductor whose gap structure under broken mirror symmetry is under debate. In our cleanest few-layer device, we observe a superconducting critical temperature comparable to that of the bulk. The temperature evolution of the phase stiffness features nodeless behavior in the presence of broken mirror symmetry, inconsistent with the scenario of nodal surface superconductivity. With minimal fabrication requirements, our technique enables microwave measurements across wide ranges of two-dimensional superconductors.

cond-mat.supr-con↗

Dynamic Proxy Domain Generalizes the Crowd Localization by Better Binary Segmentation

Crowd localization targets on predicting each instance precise location within an image. Current advanced methods propose the pixel-wise binary classification to tackle the congested prediction, in which the pixel-level thresholds binarize the prediction confidence of being the pedestrian head. Since the crowd scenes suffer from extremely varying contents, counts and scales, the confidence-threshold learner is fragile and under-generalized encountering domain knowledge shift. Moreover, at the most time, the target domain is agnostic in training. Hence, it is imperative to exploit how to enhance the generalization of confidence-threshold locator to the latent target domain. In this paper, we propose a Dynamic Proxy Domain (DPD) method to generalize the learner under domain shift. Concretely, based on the theoretical analysis to the generalization error risk upper bound on the latent target domain to a binary classifier, we propose to introduce a generated proxy domain to facilitate generalization. Then, based on the theory, we design a DPD algorithm which is composed by a training paradigm and proxy domain generator to enhance the domain generalization of the confidence-threshold learner. Besides, we conduct our method on five kinds of domain shift scenarios, demonstrating the effectiveness on generalizing the crowd localization. Our code will be available at https://github.com/zhangda1018/DPD.

cs.CV↗

Degenerate orbital effect in a three orbital periodic Anderson model

The competition between the Ruderman-Kittel-Kasuya-Yosida effect and Kondo effect is a central subject of periodic Anderson model. By using the density matrix embedding theory, we study a three orbital periodic Anderson model, in which the effects of degenerate conduction orbitals via the local magnetic moments, number of electrons, and spin-spin correlation functions, are investigated. From the phase diagram at half filling, we find there exists two different anti-ferromagnetic phases and one paramagnetic phase. To explore the difference of the two anti-ferromagnetic phases, the topology of Fermi surface and the connection with standard periodic Anderson model are considered. The spin-spin correlation functions yield insight into the competition between Ruderman-Kittel-Kasuya-Yosida interaction and Kondo interaction. We further find there is a "scaling transformation" connects different ratio of the hybridization strength.

cond-mat.str-el↗