One of the key challenges in Sequential Recommendation (SR) is how to extract and represent user preferences. Traditional SR methods rely on the next item as the supervision signal to guide preference extraction and representation. We propose a novel
learning strategy, named preference editing. The idea is to force the SR model to discriminate the common and unique preferences in different sequences of interactions between users and the recommender system. By doing so, the SR model is able to learn how to identify common and unique user preferences, and thereby do better user preference extraction and representation. We propose a transformer based SR model, named MrTransformer (Multi-preference Transformer), that concatenates some special tokens in front of the sequence to represent multiple user preferences and makes sure they capture different aspects through a preference coverage mechanism. Then, we devise a preference editing-based self-supervised learning mechanism for training MrTransformer which contains two main operations: preference separation and preference recombination. The former separates the common and unique user preferences for a given pair of sequences. The latter swaps the common preferences to obtain recombined user preferences for each sequence. Based on the preference separation and preference recombination operations, we define two types of SSL loss that require that the recombined preferences are similar to the original ones, and the common preferences are close to each other. We carry out extensive experiments on two benchmark datasets. MrTransformer with preference editing significantly outperforms state-of-the-art SR methods in terms of Recall, MRR and NDCG. We find that long sequences whose user preferences are harder to extract and represent benefit most from preference editing.
This paper introduces HPNet, a novel deep-learning approach for segmenting a 3D shape represented as a point cloud into primitive patches. The key to deep primitive segmentation is learning a feature representation that can separate points of differe
nt primitives. Unlike utilizing a single feature representation, HPNet leverages hybrid representations that combine one learned semantic descriptor, two spectral descriptors derived from predicted geometric parameters, as well as an adjacency matrix that encodes sharp edges. Moreover, instead of merely concatenating the descriptors, HPNet optimally combines hybrid representations by learning combination weights. This weighting module builds on the entropy of input features. The output primitive segmentation is obtained from a mean-shift clustering module. Experimental results on benchmark datasets ANSI and ABCParts show that HPNet leads to significant performance gains from baseline approaches.
Quantum materials with layered kagome structures have drawn considerable attention due to their unique lattice geometry, which gives rise to flat bands co-existing with Dirac-like dispersions. The interplay between strong Coulomb correlations and non
trivial band topology in these systems results in various exotic phenomena. Recently, vanadium-based materials with layered kagome structures are discovered to be topological metals, which exhibit charge density wave (CDW) properties, significant anomalous Hall effect, and unusual superconductivity at low temperatures. Here we exploit high-resolution angle-resolved photoemission spectroscopy to investigate the electronic structure evolution induced by the CDW transition in a vanadium-based kagome material RbV3Sb5. A remarkable band renormalization in the CDW state is observed, which is consistent with first principles calculations based on an inverse star-of-David superstructure. The CDW phase transition gives rise to a partial energy gap opening at the Fermi level, a shift in the band dispersion, and most importantly, the emergence of new van Hove singularities associated with large density of states, which are absent in the normal phase and may be related to superconductivity observed at lower temperatures. Our work would shed light on the microscopic mechanisms for the formation of the CDW and superconducting states in these topological kagome metals.
There is considerable interest in the pH-dependent, switchable, biocatalytic properties of cerium oxide (CeO2) nanoparticles (CeNPs) in biomedicine, where these materials exhibit beneficial antioxidant activity against reactive oxygen species (ROS) a
t basic physiological pH but cytotoxic prooxidant activity in acidic cancer cell pH microenvironment. While the general characteristics of the role of oxygen vacancies are known, the mechanism of their action at the atomic scale under different pH conditions has yet to be elucidated. The present work applies density functional theory (DFT) calculations to interpret, at the atomic scale, the pH-induced behavior of the stable {111} surface of CeO2 containing oxygen vacancies. Analysis of the surface-adsorbed media species reveals the critical role of pH on the interaction between ROS and the defective CeO2 {111} surface. Under basic conditions, the superoxide dismutase (SOD) and catalase (CAT) biomimetic reactions can be performed cyclically, scavenging and decomposing ROS to harmless products, making CeO2 an excellent antioxidant. However, under acidic conditions, the CAT biomimetic reaction is hindered owing to the limited reversibility of Ce3+ and Ce4+ and formation and annihilation of oxygen vacancies. A Fenton biomimetic reaction is predicted to occur simultaneously with the SOD and CAT biomimetic reactions, resulting in the formation of hydroxyl radicals, making CeO2 a cytotoxic prooxidant.
There is considerable interest in the pH-dependent switchable biocatalytic properties of cerium oxide nanoparticles (CeNPs) in biomedicine, where these materials exhibit beneficial antioxidant activity against reactive oxygen species at neutral and b
asic physiological pH but cytotoxic prooxidant activity at acidic pathological pH. Oxygen vacancies play a key role in such biocatalytic activities. While the general characteristics of the role of oxygen vacancies are known, the mechanism of their action at the atomic scale under different pH conditions has yet to be elucidated. The present work applies density functional theory (DFT) calculations to interpret the pH-induced behavior of the stable {111} surface of CeO2 at the atomic scale. Analysis of the surface-adsorbed media species reveals the critical role of pH on the reversibility of the Ce3+ and Ce4+ redox equilibria and the formation and annihilation of the oxygen vacancies. Under acidic conditions, this reversible switching is hindered owing to incomplete volumetric filling of the oxygen vacancies by the oxygen in the water molecules, hence effective retention of the oxygen vacancies, and consequent inhibition of redox biomimetic reactions. Under neutral and basic conditions, the capacity for this reversible switching is preserved due to complete filling of the oxygen vacancies by the OH ions owing to their ready size accommodation, thereby retaining the capacity for performing redox biomimetic reactions cyclically.
Relay-assisted free-space optical (FSO) communication systems are exploited as a means to mitigate the limiting effects of the turbulence induced atmospheric scintillation. However, conventional ground relays are stationary, and their optimal placeme
nt is not always feasible. Due to their mobility and flexibility, unmanned aerial vehicles (UAVs) provide new opportunities for FSO relaying systems. In this paper, a hovering UAV-based serial FSO decode-and-forward relaying system is investigated. In the channel modelling for such a system, four types of impairments (i.e., atmospheric loss, atmospheric turbulence, pointing error, and link interruption due to angle-of-arrival fluctuation) are considered. Based on the proposed channel model, a tractable expression for the probability density function of the total channel gain is obtained. Closed-form expressions of the link outage probability and end-to-end outage probability are derived. Asymptotic outage performance bounds for each link and the overall system are also presented to reveal insights into the impacts of different impairments. To improve system performance, we optimize the beam width, field-of-view and UAVs locations. Numerical results show that the derived theoretical expressions are accurate to evaluate the outage performance of the system. Moreover, the proposed optimization schemes are efficient and can improve performance significantly.
We propose a new type of spin-valley locking (SVL), named $textit{C}$-paired SVL, in antiferromagnetic systems, which directly connects the spin/valley space with the real space, and hence enables both static and dynamical controls of spin and valley
to realize a multifunctional antiferromagnetic material. The new emergent quantum degree of freedom in the $textit{C}$-paired SVL is comprised of spin-polarized valleys related by a crystal symmetry instead of the time-reversal symmetry. Thus, both spin and valley can be accessed by simply breaking the corresponding crystal symmetry. Typically, one can use a strain field to induce a large net valley polarization/magnetization and use a charge current to generate a large noncollinear spin current. We predict the realization of the $textit{C}$-paired SVL in monolayer V$_2$Se$_2$O, which indeed exhibits giant piezomagnetism and can generate a large transverse spin current. Our findings provide unprecedented opportunities to integrate various controls of spin and valley with nonvolatile information storage in a single material, which is highly desirable for versatile fundamental research and device applications.
Despite the widespread utilization of deep neural networks (DNNs) for speech emotion recognition (SER), they are severely restricted due to the paucity of labeled data for training. Recently, segment-based approaches for SER have been evolving, which
train backbone networks on shorter segments instead of whole utterances, and thus naturally augments training examples without additional resources. However, one core challenge remains for segment-based approaches: most emotional corpora do not provide ground-truth labels at the segment level. To supervisely train a segment-based emotion model on such datasets, the most common way assigns each segment the corresponding utterances emotion label. However, this practice typically introduces noisy (incorrect) labels as emotional information is not uniformly distributed across the whole utterance. On the other hand, DNNs have been shown to easily over-fit a dataset when being trained with noisy labels. To this end, this work proposes a simple and effective deep self-learning (DSL) framework, which comprises a procedure to progressively correct segment-level labels in an iterative learning manner. The DSL method produces dynamically-generated and soft emotion labels, leading to significant performance improvements. Experiments on three well-known emotional corpora demonstrate noticeable gains using the proposed method.
In high-energy leptonic collisions, such as at a multi-TeV muon collider, the collinear splittings of the electroweak (EW) gauge bosons and leptons are the dominant phenomena, and the scattering processes should thus be formulated in terms of the EW
parton distribution functions (EW PDFs). We complete this formalism in the Standard Model to include the QCD sector and evaluate the quark and gluon PDFs inside a lepton at the double-log accuracy. The splittings of the photon and subsequently the quarks and gluons control the quark/gluon PDFs below the EW scale. The massive gauge bosons lead to substantial contributions at high scales. The jet production cross section can reach the order of a few nb (50 pb) in $e^+e^-$ ($mu^+mu^-$) collisions, at the TeV c.m. energies with a moderate acceptance cut, that governs the overall event shape up to about $p_T^j sim 60$ GeV.
Recent years have witnessed significant progress in 3D hand mesh recovery. Nevertheless, because of the intrinsic 2D-to-3D ambiguity, recovering camera-space 3D information from a single RGB image remains challenging. To tackle this problem, we divid
e camera-space mesh recovery into two sub-tasks, i.e., root-relative mesh recovery and root recovery. First, joint landmarks and silhouette are extracted from a single input image to provide 2D cues for the 3D tasks. In the root-relative mesh recovery task, we exploit semantic relations among joints to generate a 3D mesh from the extracted 2D cues. Such generated 3D mesh coordinates are expressed relative to a root position, i.e., wrist of the hand. In the root recovery task, the root position is registered to the camera space by aligning the generated 3D mesh back to 2D cues, thereby completing cameraspace 3D mesh recovery. Our pipeline is novel in that (1) it explicitly makes use of known semantic relations among joints and (2) it exploits 1D projections of the silhouette and mesh to achieve robust registration. Extensive experiments on popular datasets such as FreiHAND, RHD, and Human3.6M demonstrate that our approach achieves stateof-the-art performance on both root-relative mesh recovery and root recovery. Our code is publicly available at https://github.com/SeanChenxy/HandMesh.
