Search

Autoregressive transformers have revolutionized generative models in language processing and shown substantial promise in image and video generation. However, these models face significant challenges when extended to 3D generation tasks due to their reliance on next-token prediction to learn token sequences, which is incompatible with the unordered nature of 3D data. Instead of imposing an artificial order on 3D data, in this paper, we introduce G3PT, a scalable coarse-to-fine 3D generative model utilizing a cross-scale querying transformer. The key is to map point-based 3D data into discrete tokens with different levels of detail, naturally establishing a sequential relationship between different levels suitable for autoregressive modeling. Additionally, the cross-scale querying transformer connects tokens globally across different levels of detail without requiring an ordered sequence. Benefiting from this approach, G3PT features a versatile 3D generation pipeline that effortlessly supports diverse conditional structures, enabling the generation of 3D shapes from various types of conditions. Extensive experiments demonstrate that G3PT achieves superior generation quality and generalization ability compared to previous 3D generation methods. Most importantly, for the first time in 3D generation, scaling up G3PT reveals distinct power-law scaling behaviors., Comment: 9 pages, 5 figures

Solving Algebra Problems with Geometry Diagrams (APGDs) is still a challenging problem because diagram processing is not studied as intensively as language processing. To work against this challenge, this paper proposes a hologram reasoning scheme and develops a high-performance method for solving APGDs by using this scheme. To reach this goal, it first defines a hologram, being a kind of graph, and proposes a hologram generator to convert a given APGD into a hologram, which represents the entire information of APGD and the relations for solving the problem can be acquired from it by a uniform way. Then HGR, a hologram reasoning method employs a pool of prepared graph models to derive algebraic equations, which is consistent with the geometric theorems. This method is able to be updated by adding new graph models into the pool. Lastly, it employs deep reinforcement learning to enhance the efficiency of model selection from the pool. The entire HGR not only ensures high solution accuracy with fewer reasoning steps but also significantly enhances the interpretability of the solution process by providing descriptions of all reasoning steps. Experimental results demonstrate the effectiveness of HGR in improving both accuracy and interpretability in solving APGDs.

We present a real-time LiDAR-Inertial-Camera SLAM system with 3D Gaussian Splatting as the mapping backend. Leveraging robust pose estimates from our LiDAR-Inertial-Camera odometry, Coco-LIC, an incremental photo-realistic mapping system is proposed in this paper. We initialize 3D Gaussians from colorized LiDAR points and optimize them using differentiable rendering powered by 3D Gaussian Splatting. Meticulously designed strategies are employed to incrementally expand the Gaussian map and adaptively control its density, ensuring high-quality mapping with real-time capability. Experiments conducted in diverse scenarios demonstrate the superior performance of our method compared to existing radiance-field-based SLAM systems., Comment: Submitted to IROS 2024

Neural radiance fields~(NeRF) have recently been applied to render large-scale scenes. However, their limited model capacity typically results in blurred rendering results. Existing large-scale NeRFs primarily address this limitation by partitioning the scene into blocks, which are subsequently handled by separate sub-NeRFs. These sub-NeRFs, trained from scratch and processed independently, lead to inconsistencies in geometry and appearance across the scene. Consequently, the rendering quality fails to exhibit significant improvement despite the expansion of model capacity. In this work, we present global-guided focal neural radiance field (GF-NeRF) that achieves high-fidelity rendering of large-scale scenes. Our proposed GF-NeRF utilizes a two-stage (Global and Focal) architecture and a global-guided training strategy. The global stage obtains a continuous representation of the entire scene while the focal stage decomposes the scene into multiple blocks and further processes them with distinct sub-encoders. Leveraging this two-stage architecture, sub-encoders only need fine-tuning based on the global encoder, thus reducing training complexity in the focal stage while maintaining scene-wide consistency. Spatial information and error information from the global stage also benefit the sub-encoders to focus on crucial areas and effectively capture more details of large-scale scenes. Notably, our approach does not rely on any prior knowledge about the target scene, attributing GF-NeRF adaptable to various large-scale scene types, including street-view and aerial-view scenes. We demonstrate that our method achieves high-fidelity, natural rendering results on various types of large-scale datasets. Our project page: https://shaomq2187.github.io/GF-NeRF/, Comment: WACV 2025

SemEval-2024 Task 8 introduces the challenge of identifying machine-generated texts from diverse Large Language Models (LLMs) in various languages and domains. The task comprises three subtasks: binary classification in monolingual and multilingual (Subtask A), multi-class classification (Subtask B), and mixed text detection (Subtask C). This paper focuses on Subtask A & B. Each subtask is supported by three datasets for training, development, and testing. To tackle this task, two methods: 1) using traditional machine learning (ML) with natural language preprocessing (NLP) for feature extraction, and 2) fine-tuning LLMs for text classification. The results show that transformer models, particularly LoRA-RoBERTa, exceed traditional ML methods in effectiveness, with majority voting being particularly effective in multilingual contexts for identifying machine-generated texts.

We examine the response of the Fermi arc in the context of quasi-particle interference (QPI) with regard to a localized surface impurity on various three-dimensional Weyl semimetals (WSMs). Our study also reveals the variation of the local density of states (LDOS), obtained by Fourier transforming the QPI profile, on the two-dimensional surface. We use the $T$-matrix formalism to numerically (analytically and numerically) capture the details of the momentum space scattering in QPI (real space decay in LDOS), considering relevant tight-binding lattice and/or low-energy continuum models modeling a range of different WSMs. In particular, we consider multi-WSM (mWSM), hosting multiple Fermi arcs between two opposite chirality Weyl nodes (WNs), where we find a universal $1/r$-decay ($r$ measuring the radial distance from the impurity core) of the impurity-induced LDOS, irrespective of the topological charge. Interestingly, the inter-Fermi arc scattering is only present for triple WSMs, where we find an additional $1/r^3$-decay as compared to double and single WSMs. The untilted single (double) [triple] WSM shows a straight-line (leaf-like) [oval-shaped] QPI profile. The above QPI profiles are canted for hybrid WSMs where type-I and type-II Weyl nodes coexist, however, hybrid single WSM demonstrates strong non-uniformity, unlike the hybrid double and triple WSMs. We also show that the chirality and the positions of the Weyl nodes imprint marked signatures in the QPI profile. This allows us to distinguish between different WSMs, including the time-reversal-broken WSMs from the time-reversal-invariant WSM, even though both of the WSMs can host two pairs of Weyl nodes. Our study can thus shed light on experimentally obtainable complex QPI profiles and help differentiate different WSMs and their surface band structures., Comment: 19 pages, 6 figures

We place a molecular Bose-Einstein condensate in a 1D shaken lattice with a Floquet-engineered dispersion, and observe the dynamics in both position and momentum space. At the initial condition of zero momentum, our engineered dispersion is inverted, and therefore unstable. We observe that the condensate is destabilized by the lattice shaking as expected, but rather than decaying incoherently or producing jets, as in other unstable condensates, under our conditions the condensate bifurcates into two portions in momentum space, with each portion subsequently following semi-classical trajectories that suffer minimal spreading in momentum space as they evolve. We can model the evolution with a Gross-Pitaevskii equation, which suggests the initial bifurcation is facilitate by a nearly linear "inverted V"-shaped dispersion at the zone center, while the lack of spreading in momentum space is facilitated by interactions, as in a soliton. We propose that this relatively clean bifurcation in momentum space has applications for counter-diabatic preparation of exotic ground states in many-body quantum simulation schemes.

Unmanned Aerial Vehicles (UAVs), specifically drones equipped with remote sensing object detection technology, have rapidly gained a broad spectrum of applications and emerged as one of the primary research focuses in the field of computer vision. Although UAV remote sensing systems have the ability to detect various objects, small-scale objects can be challenging to detect reliably due to factors such as object size, image degradation, and real-time limitations. To tackle these issues, a real-time object detection algorithm (YOLO-Drone) is proposed and applied to two new UAV platforms as well as a specific light source (silicon-based golden LED). YOLO-Drone presents several novelties: 1) including a new backbone Darknet59; 2) a new complex feature aggregation module MSPP-FPN that incorporated one spatial pyramid pooling and three atrous spatial pyramid pooling modules; 3) and the use of Generalized Intersection over Union (GIoU) as the loss function. To evaluate performance, two benchmark datasets, UAVDT and VisDrone, along with one homemade dataset acquired at night under silicon-based golden LEDs, are utilized. The experimental results show that, in both UAVDT and VisDrone, the proposed YOLO-Drone outperforms state-of-the-art (SOTA) object detection methods by improving the mAP of 10.13% and 8.59%, respectively. With regards to UAVDT, the YOLO-Drone exhibits both high real-time inference speed of 53 FPS and a maximum mAP of 34.04%. Notably, YOLO-Drone achieves high performance under the silicon-based golden LEDs, with a mAP of up to 87.71%, surpassing the performance of YOLO series under ordinary light sources. To conclude, the proposed YOLO-Drone is a highly effective solution for object detection in UAV applications, particularly for night detection tasks where silicon-based golden light LED technology exhibits significant superiority., Comment: Some contributing authors are not signed

Optical phase-change materials are highly promising for emerging applications such as tunable metasurfaces, reconfigurable photonic circuits, and non-von Neumann computing. However, these materials typically require both high melting temperatures and fast quenching rates to reversibly switch between their crystalline and amorphous phases, a significant challenge for large-scale integration. Here, we present an experimental technique which leverages the thermo-optic effect in GST to enable both spatial and temporal thermal measurements of two common electro-thermal microheater designs currently used by the phase-change community. Our approach shows excellent agreement between experimental results and numerical simulations and provides a non-invasive method for rapid characterization of electrically programmable phase-change devices.

Meta-learning is used to efficiently enable the automatic selection of machine learning models by combining data and prior knowledge. Since the traditional meta-learning technique lacks explainability, as well as shortcomings in terms of transparency and fairness, achieving explainability for meta-learning is crucial. This paper proposes FIND, an interpretable meta-learning framework that not only can explain the recommendation results of meta-learning algorithm selection, but also provide a more complete and accurate explanation of the recommendation algorithm's performance on specific datasets combined with business scenarios. The validity and correctness of this framework have been demonstrated by extensive experiments.

The quantum Hall effect in three-dimensional Weyl semimetal (WSM) receives significant attention for the emergence of the Fermi loop where the underlying two-dimensional Hall conductivity, namely, sheet Hall conductivity, shows quantized plateaus. Considering the tilted lattice models for multi Weyl semimetals (mWSMs), we systematically study the Landau levels (LLs) and magneto-Hall conductivity in the presence of parallel and perpendicular (with respect to the Weyl node's separation) magnetic field, i.e., $\mathbf{ B}\parallel z$ and $\mathbf{B}\parallel x$, to explore the impact of tilting and non-linearity in the dispersion. We make use of two (single) node low-energy models to qualitatively explain the emergence of mid-gap chiral (linear crossing of chiral) LLs on the lattice for $\mathbf{ B}\parallel z$ ($\mathbf{ B}\parallel x$). Remarkably, we find that the sheet Hall conductivity becomes quantized for $\mathbf{ B}\parallel z$ even when two Weyl nodes project onto a single Fermi point in two opposite surfaces, forming a Fermi loop with $k_z$ as the good quantum number. On the other hand, the Fermi loop, connecting two distinct Fermi points in two opposite surfaces, with $k_x$ being the good quantum number, causes the quantization in sheet Hall conductivity for $\mathbf{ B}\parallel x$. The quantization is almost lost (perfectly remained) in the type-II phase for $\mathbf{ B}\parallel x$ ($\mathbf{ B}\parallel z$). Interestingly, the jump profiles between the adjacent quantized plateaus change with the topological charge for both of the above cases. The momentum-integrated three-dimensional Hall conductivity is not quantized; however, it bears the signature of chiral LLs as resulting in the linear dependence on $\mu$ for small $\mu$. The linear zone (its slope) reduces (increases) as the tilt (topological charge) of the underlying WSM increases., Comment: 19 pages and 9 figures

Memristors have emerged as a promising technology for efficient neuromorphic architectures owing to their ability to act as programmable synapses, combining processing and memory into a single device. Although they are most commonly used for static encoding of synaptic weights, recent work has begun to investigate the use of their dynamical properties, such as Short Term Plasticity (STP), to integrate events over time in event-based architectures. However, we are still far from completely understanding the range of possible behaviors and how they might be exploited in neuromorphic computation. This work focuses on a newly developed Li$_\textbf{x}$WO$_\textbf{3}$-based three-terminal memristor that exhibits tunable STP and a conductance response modeled by a double exponential decay. We derive a stochastic model of the device from experimental data and investigate how device stochasticity, STP, and the double exponential decay affect accuracy in a hierarchy of time-surfaces (HOTS) architecture. We found that the device's stochasticity does not affect accuracy, that STP can reduce the effect of salt and pepper noise in signals from event-based sensors, and that the double exponential decay improves accuracy by integrating temporal information over multiple time scales. Our approach can be generalized to study other memristive devices to build a better understanding of how control over temporal dynamics can enable neuromorphic engineers to fine-tune devices and architectures to fit their problems at hand.