Search

Vision Transformers (ViTs) have achieved remarkable success in various computer vision tasks. However, ViTs have a huge computational cost due to their inherent reliance on multi-head self-attention (MHSA), prompting efforts to accelerate ViTs for practical applications. To this end, recent works aim to reduce the number of tokens, mainly focusing on how to effectively prune or merge them. Nevertheless, since ViT tokens are generated from non-overlapping grid patches, they usually do not convey sufficient semantics, making it incompatible with efficient ViTs. To address this, we propose ImagePiece, a novel re-tokenization strategy for Vision Transformers. Following the MaxMatch strategy of NLP tokenization, ImagePiece groups semantically insufficient yet locally coherent tokens until they convey meaning. This simple retokenization is highly compatible with previous token reduction methods, being able to drastically narrow down relevant tokens, enhancing the inference speed of DeiT-S by 54% (nearly 1.5$\times$ faster) while achieving a 0.39% improvement in ImageNet classification accuracy. For hyper-speed inference scenarios (with 251% acceleration), our approach surpasses other baselines by an accuracy over 8%., Comment: Accepted to AAAI 2025

While hydroxyl radicals ($\cdot$OH) play an important role as potent oxidizing agents in various plasma applications, their high reactivity confines them to a thin layer at the plasma-liquid interface, posing challenges in comprehending the intricate generation and transport processes. Similarly, solvated electrons ($\mathrm{e_{aq}}$), highly reactive reducing agents, are expected to exhibit distribution beneath the liquid surface and interact with $\cdot$OH in the thin layer. In this study, we have determined the penetration depth and concentration of ($\mathrm{e_{aq}}$) at the interface between an atmospheric argon plasma plume and an electrolyte anode via a lock-in amplification absorbance measurement. With bias voltages from 1000 to 2500 V, the penetration depth remains approximately 10 nm, and the peak concentration near the surface reaches 1 mM. Diffusion is the primary mechanism for $\cdot$OH generation in the electrolyte, with most $\cdot$OH reacting with ($\mathrm{e_{aq}}$) at the interface, thus influencing the ($\mathrm{e_{aq}}$) distribution. In contrast, the electrolyte cathode significantly boosts $\cdot$OH generation, leading to rapid recombination into hydrogen peroxide.

Crystal symmetry plays an important role in the Hall effects. Unconventional spin Hall effect (USHE), characterized by Dresselhaus and out-of-plane spins, has been observed in materials with low crystal symmetry. Recently, antisymmetric planar Hall effect (APHE) was discovered in rutile RuO$_2$ and IrO$_2$ (101) thin films, which also exhibit low crystal symmetry. However, the correlation between USHE and APHE remains largely unexplored. In this study, we report the observation of both USHE and APHE in IrO$_2$ (111) films, using spin-torque ferromagnetic resonance (ST-FMR) and harmonic Hall measurements, respectively. Notably, the unconventional spin torque efficiency from Dresselhaus spin was more than double that of a previous report. Additionally, the temperature dependence of APHE suggests that it arises from the Lorentz force, constrained by crystal symmetry. Symmetry analysis confirms the coexistence of USHE and APHE in IrO$_2$ (111), paving the way for a deeper understanding of Hall effects and related physical phenomena.

Spin-orbit torque (SOT) can be used to efficiently manipulate the magnetic state of magnetic materials, which is an essential element for memory and logic applications. Due to symmetry constraints, only in-plane spins can be injected into the ferromagnet from the underlying SOT layer for conventional SOT materials such as heavy metals and topological materials. Through the use of materials with low symmetries, or other symmetry breaking approaches, unconventional spin currents with out-of-plane polarization has been demonstrated and enabled field-free deterministic switching of perpendicular magnetization. Despite this progress, the SOT efficiency of these materials has typically remained low. Here, we report a giant SOT efficiency of 0.85 in sputtered Ni4W/CoFeB heterostructure at room temperature, as evaluated by second harmonic Hall measurements. In addition, due to the low crystal symmetry of Ni4W, unconventional out-of-plane and Dresselhaus-like spin components were observed. Macro-spin simulation suggests our spin Hall tensor to provide about an order of magnitude improvement in the magnetization switching efficiency, thus broadening the path towards energy efficient spintronic devices using low-symmetry materials.

We demonstrate that two-dimensional Kramers-Weyl fermions can be engineered in spin-orbit coupled twisted bilayers, where the chiral structure of these moir\'e systems breaks all mirror symmetries, confining Kramers-Weyl fermions to high-symmetry points in the Brillouin zone under time reversal symmetry. Our theoretical analysis reveals a symmetry-enforced Weyl-like interlayer moir\'e coupling that universally ensures an ideal radial spin-texture at arbitrary twist angles, under $C_{nz}$ symmetry with n>2. First principles density functional calculation confirm the realization of these fermions in twisted $\alpha$-In$_2$Se$_3$ bilayers, where flat bands and out-of-plane ferroelectric polarization in each layer guarantee two-dimensional Kramers-Weyl physics with perfectly ideal radial spin textures.

We introduce EXAONE 3.0 instruction-tuned language model, the first open model in the family of Large Language Models (LLMs) developed by LG AI Research. Among different model sizes, we publicly release the 7.8B instruction-tuned model to promote open research and innovations. Through extensive evaluations across a wide range of public and in-house benchmarks, EXAONE 3.0 demonstrates highly competitive real-world performance with instruction-following capability against other state-of-the-art open models of similar size. Our comparative analysis shows that EXAONE 3.0 excels particularly in Korean, while achieving compelling performance across general tasks and complex reasoning. With its strong real-world effectiveness and bilingual proficiency, we hope that EXAONE keeps contributing to advancements in Expert AI. Our EXAONE 3.0 instruction-tuned model is available at https://huggingface.co/LGAI-EXAONE/EXAONE-3.0-7.8B-Instruct

Standard single-image super-resolution creates paired training data from high-resolution images through fixed downsampling kernels. However, real-world super-resolution (RWSR) faces unknown degradations in the low-resolution inputs, all the while lacking paired training data. Existing methods approach this problem by learning blind general models through complex synthetic augmentations on training inputs; they sacrifice the performance on specific degradation for broader generalization to many possible ones. We address the unsupervised RWSR for a targeted real-world degradation. We study from a distillation perspective and introduce a novel pairwise distance distillation framework. Through our framework, a model specialized in synthetic degradation adapts to target real-world degradations by distilling intra- and inter-model distances across the specialized model and an auxiliary generalized model. Experiments on diverse datasets demonstrate that our method significantly enhances fidelity and perceptual quality, surpassing state-of-the-art approaches in RWSR. The source code is available at https://github.com/Yuehan717/PDD., Comment: ECCV 2024

Advances in material fabrication techniques and growth methods have opened up a new chapter for twistronics, in the form of twisted freestanding three-dimensional material membranes. Through first-principles calculations based on density functional theory, we investigate the crystal and electronic structures of twisted bilayer BaTiO$_3$. Our findings reveal that large stacking fault energy leads to chiral in-plane vortex pattern that was recently observed in experiments. Moreover, we also found non-zero out-of-plane local dipole moments, indicating that the strong interlayer interaction might offer promising strategy to stabilize ferroelectric order in the two-dimensional limit. Remarkably, the vortex pattern in the twisted BaTiO$_3$ bilayer support localized electronic states with quasi-flat bands, associated with the interlayer hybridization of oxygen $p_z$ orbitals. We found that the associated band width reaches a minimum at $\sim$19$^{\circ}$ twisting, configuring the largest magic angle in moir\'e systems reported so far. Further, the moir\'e vortex pattern bears a striking resemblance to two interpenetrating Lieb lattices and corresponding tight-binding model provides a comprehensive description of the evolution the moir\'e bands with twist angle and reveals the topological nature of these states., Comment: 8 pages, 5 figures

Translating major language resources to build minor language resources becomes a widely-used approach. Particularly in translating complex data points composed of multiple components, it is common to translate each component separately. However, we argue that this practice often overlooks the interrelation between components within the same data point. To address this limitation, we propose a novel MT pipeline that considers the intra-data relation in implementing MT for training data. In our MT pipeline, all the components in a data point are concatenated to form a single translation sequence and subsequently reconstructed to the data components after translation. We introduce a Catalyst Statement (CS) to enhance the intra-data relation, and Indicator Token (IT) to assist the decomposition of a translated sequence into its respective data components. Through our approach, we have achieved a considerable improvement in translation quality itself, along with its effectiveness as training data. Compared with the conventional approach that translates each data component separately, our method yields better training data that enhances the performance of the trained model by 2.690 points for the web page ranking (WPR) task, and 0.845 for the question generation (QG) task in the XGLUE benchmark., Comment: Accepted to EMNLP2024 findings

In this paper, we investigate Open-Vocabulary 3D Instance Segmentation (OV-3DIS) with free-form language instructions. Earlier works that rely on only annotated base categories for training suffer from limited generalization to unseen novel categories. Recent works mitigate poor generalizability to novel categories by generating class-agnostic masks or projecting generalized masks from 2D to 3D, but disregard semantic or geometry information, leading to sub-optimal performance. Instead, generating generalizable but semantic-related masks directly from 3D point clouds would result in superior outcomes. In this paper, we introduce Segment any 3D Object with LanguagE (SOLE), which is a semantic and geometric-aware visual-language learning framework with strong generalizability by generating semantic-related masks directly from 3D point clouds. Specifically, we propose a multimodal fusion network to incorporate multimodal semantics in both backbone and decoder. In addition, to align the 3D segmentation model with various language instructions and enhance the mask quality, we introduce three types of multimodal associations as supervision. Our SOLE outperforms previous methods by a large margin on ScanNetv2, ScanNet200, and Replica benchmarks, and the results are even close to the fully-supervised counterpart despite the absence of class annotations in the training. Furthermore, extensive qualitative results demonstrate the versatility of our SOLE to language instructions., Comment: Project Page: https://cvrp-sole.github.io

The piezoelectricity of group IV monochalcogenides (MXs, with M = Ge, Sn and X = S, Se) has attracted much attention due to their substantially higher piezoelectric coefficients compared to other 2D materials. However, with increasing layer number, their piezoelectricity rapidly disappears due to the antiferroelectric stacking order, severely limiting their practical applications. Using first-principles calculations, we investigated the piezoelectricity of MXs with the ferroelectric AA stacking configuration, which has recently been stabilized in experiments. We found that AA-stacked MXs have a ferroelectric ground state with the smallest lattice constant among other stacking configurations, resulting in a giant piezoelectric coefficient, which is the first demonstration of a strategy where the piezoelectric coefficients can increase with the number of layers. This can be attributed to a strong negative correlation between the lattice constant along the armchair direction and the piezoelectric coefficient, and spontaneous compressive strain stabilized in ferroelectric AA stacking configuration., Comment: 16 pages and 5 figures for main manuscript. 9pages and 3 figures for SI

A charge density wave (CDW) represents an exotic state in which electrons are arranged in a long range ordered pattern in low-dimensional materials. Although our understanding of the fundamental character of CDW has been enriched after extensive studies, its relationship with functional phenomena remains relatively limited. Here, we show an unprecedented demonstration of a tunable charge-spin interconversion (CSI) in graphene/1T-TaS$_2$ van der Waals heterostructures by manipulating the distinct CDW phases in 1T-TaS$_2$. Whereas CSI from spins polarized in all three directions are observed in the heterostructure when the CDW phase does not show commensurability, the output of one of the components disappears and the other two are enhanced when the CDW phase becomes commensurate. The experimental observation is supported by first-principles calculations, which evidence that chiral CDW multidomains are at the origin of the switching of CSI. Our results uncover a new approach for on-demand CSI in low-dimensional systems, paving the way for advanced spin-orbitronic devices., Comment: Funding information: Marie Sklodowska-Curie Actions, H2020-MSCA-ITN-2020; Project Acronym SPEAR; Grant Agreement No. 955671

For object detection task with noisy labels, it is important to consider not only categorization noise, as in image classification, but also localization noise, missing annotations, and bogus bounding boxes. However, previous studies have only addressed certain types of noise (e.g., localization or categorization). In this paper, we propose Universal-Noise Annotation (UNA), a more practical setting that encompasses all types of noise that can occur in object detection, and analyze how UNA affects the performance of the detector. We analyzed the development direction of previous works of detection algorithms and examined the factors that impact the robustness of detection model learning method. We open-source the code for injecting UNA into the dataset and all the training log and weight are also shared., Comment: appendix and code : https://github.com/Ryoo72/UNA

Solving partial differential equations (PDEs) by learning the solution operators has emerged as an attractive alternative to traditional numerical methods. However, implementing such architectures presents two main challenges: flexibility in handling irregular and arbitrary input and output formats and scalability to large discretizations. Most existing architectures are limited by their desired structure or infeasible to scale large inputs and outputs. To address these issues, we introduce an attention-based model called an inducing-point operator transformer (IPOT). Inspired by inducing points methods, IPOT is designed to handle any input function and output query while capturing global interactions in a computationally efficient way. By detaching the inputs/outputs discretizations from the processor with a smaller latent bottleneck, IPOT offers flexibility in processing arbitrary discretizations and scales linearly with the size of inputs/outputs. Our experimental results demonstrate that IPOT achieves strong performances with manageable computational complexity on an extensive range of PDE benchmarks and real-world weather forecasting scenarios, compared to state-of-the-art methods., Comment: AAAI 2024

We propose a spin Hall effect (SHE) enhancement mechanism due to Kramers-Weyl point (KWP) alignment in chiral topological semimetals with high Chern numbers (CNs). Through model Hamiltonian calculations, we identify enhancements in the intrinsic spin Hall conductivity (SHC) and the spin Hall angle (SHA). Such enhancements, attributed to a unique high CN KWP energetic alignment and a high degree of SOC-induced band nesting, strongly depend on orbital-orbital interactions. This represents a novel mechanism to enhance SHE, differing from the spin-orbit induced anticrossing mechanism in gapped systems. Guided by this principle, we corroborate our results by means of first-principles calculations and reveal multiple realistic materials with large intrinsic SHCs and even larger SHAs than Pt., Comment: 6 pages, 4 figures

Thru-hole epitaxy has recently been reported to be able to grow readily detachable domains crystallographically aligned with the underlying substrate over 2D mask material transferred onto a substrate. [Jang \textit{et al.}, \textit{Adv. Mater. Interfaces}, \textbf{2023} \textit{10}, 4 2201406] While the experimental demonstration of thru-hole epitaxy of GaN over multiple stacks of $h$-BN was evident, the detailed mechanism of how small holes in each stack of $h$-BN survived as thru-holes during multiple stacking of $h$-BN was not intuitively clear. Here, we use Monte Carlo simulations to investigate the conditions under which holes in each stack of 2D mask layers can survive as thru-holes during multiple stacking. If holes are highly anisotropic in shape by connecting smaller holes in a particular direction, thru-holes can be maintained with a high survival rate per stack, establishing more epitaxial connectedness. Our work verifies and supports that thru-hole epitaxy is attributed to the epitaxial connectedness established by thru-holes surviving even through multiple stacks., Comment: 16 pages, 6 figures

We investigate the twist angle dependence of spin-orbit coupling (SOC) proximity effects and charge-to-spin conversion (CSC) in graphene/WSe$_2$ heterostructures from first principles. The CSC is shown to strongly depend on the twist angle, with both the spin Hall and standard Rashba-Edelstein efficiencies optimized at or near 30{\deg} twisting. Symmetry breaking due to twisting also gives rise to an unconventional Rashba-Edelstein effect, with electrically generated non-equilibrium spin densities possessing spins collinear to the applied electric field. We further discuss how the carrier doping concentration and band broadening control the crossover between the Fermi-sea and -surface spin response, which reconciles the seemingly disparate experimental observations of different CSC phenomena., Comment: 5 pages, 4 figures, Supplementary Information (13 pages, 7 figures)

Human-Object Interaction detection is a holistic visual recognition task that entails object detection as well as interaction classification. Previous works of HOI detection has been addressed by the various compositions of subset predictions, e.g., Image -> HO -> I, Image -> HI -> O. Recently, transformer based architecture for HOI has emerged, which directly predicts the HOI triplets in an end-to-end fashion (Image -> HOI). Motivated by various inference paths for HOI detection, we propose cross-path consistency learning (CPC), which is a novel end-to-end learning strategy to improve HOI detection for transformers by leveraging augmented decoding paths. CPC learning enforces all the possible predictions from permuted inference sequences to be consistent. This simple scheme makes the model learn consistent representations, thereby improving generalization without increasing model capacity. Our experiments demonstrate the effectiveness of our method, and we achieved significant improvement on V-COCO and HICO-DET compared to the baseline models. Our code is available at https://github.com/mlvlab/CPChoi., Comment: CVPR2022 accepted