Search

We present Polynomial Attention Drop-in Replacement (PADRe), a novel and unifying framework designed to replace the conventional self-attention mechanism in transformer models. Notably, several recent alternative attention mechanisms, including Hyena, Mamba, SimA, Conv2Former, and Castling-ViT, can be viewed as specific instances of our PADRe framework. PADRe leverages polynomial functions and draws upon established results from approximation theory, enhancing computational efficiency without compromising accuracy. PADRe's key components include multiplicative nonlinearities, which we implement using straightforward, hardware-friendly operations such as Hadamard products, incurring only linear computational and memory costs. PADRe further avoids the need for using complex functions such as Softmax, yet it maintains comparable or superior accuracy compared to traditional self-attention. We assess the effectiveness of PADRe as a drop-in replacement for self-attention across diverse computer vision tasks. These tasks include image classification, image-based 2D object detection, and 3D point cloud object detection. Empirical results demonstrate that PADRe runs significantly faster than the conventional self-attention (11x ~ 43x faster on server GPU and mobile NPU) while maintaining similar accuracy when substituting self-attention in the transformer models.

M-type stars are the ones that flare most frequently, but how big their maximum flare energy can reach is still unknown. We present 163 flares from 162 individual M2 through L1-type stars that triggered the GWAC, with flare energies ranging from $10^{32.2}$ to $10^{36.4}$ erg . The flare amplitudes range from $\triangle G = 0.84$ to $\sim 10$ mag. Flare energy increases with stellar surface temperature ($T_{\rm eff}$) but both $\triangle G$ and equivalent duration $\log_{10}(ED)$ seem to be independent of $T_{\rm eff}$. Combining periods detected from light curves of TESS and K2, spectra from LAMOST, SDSS and the 2.16 m Telescope, and the Gaia DR3 data, we found that these GWAC flare stars are young. For the stars that have spectra, we found that these stars are in or very near to the saturation region, and $\log_{10}(L_{\rm H\alpha}/L_{\rm bol})$ is lower for M7-L1 stars than for M2-M6 stars. We also studied the relation between GWAC flare bolometric energy $E_{\rm bol}$ and stellar hemispherical area $S$, and found that $\log_{10}E_{\rm bol}$ (in erg) increases with increasing $S$ (in cm$^2$), and the maximum flare energy $\log_{10}E_{\rm bol, max} \geqslant \log_{10}S + 14.25$. For M7-L1 stars, there seem to be other factors limiting their maximum flare energies in addition to stellar hemispherical area., Comment: 18 pages, 11 figures, 4 tables

In this paper, we propose a novel token selective attention approach, ToSA, which can identify tokens that need to be attended as well as those that can skip a transformer layer. More specifically, a token selector parses the current attention maps and predicts the attention maps for the next layer, which are then used to select the important tokens that should participate in the attention operation. The remaining tokens simply bypass the next layer and are concatenated with the attended ones to re-form a complete set of tokens. In this way, we reduce the quadratic computation and memory costs as fewer tokens participate in self-attention while maintaining the features for all the image patches throughout the network, which allows it to be used for dense prediction tasks. Our experiments show that by applying ToSA, we can significantly reduce computation costs while maintaining accuracy on the ImageNet classification benchmark. Furthermore, we evaluate on the dense prediction task of monocular depth estimation on NYU Depth V2, and show that we can achieve similar depth prediction accuracy using a considerably lighter backbone with ToSA., Comment: Accepted at CVPRW 2024

Optical flow estimation is crucial to a variety of vision tasks. Despite substantial recent advancements, achieving real-time on-device optical flow estimation remains a complex challenge. First, an optical flow model must be sufficiently lightweight to meet computation and memory constraints to ensure real-time performance on devices. Second, the necessity for real-time on-device operation imposes constraints that weaken the model's capacity to adequately handle ambiguities in flow estimation, thereby intensifying the difficulty of preserving flow accuracy. This paper introduces two synergistic techniques, Self-Cleaning Iteration (SCI) and Regression Focal Loss (RFL), designed to enhance the capabilities of optical flow models, with a focus on addressing optical flow regression ambiguities. These techniques prove particularly effective in mitigating error propagation, a prevalent issue in optical flow models that employ iterative refinement. Notably, these techniques add negligible to zero overhead in model parameters and inference latency, thereby preserving real-time on-device efficiency. The effectiveness of our proposed SCI and RFL techniques, collectively referred to as SciFlow for brevity, is demonstrated across two distinct lightweight optical flow model architectures in our experiments. Remarkably, SciFlow enables substantial reduction in error metrics (EPE and Fl-all) over the baseline models by up to 6.3% and 10.5% for in-domain scenarios and by up to 6.2% and 13.5% for cross-domain scenarios on the Sintel and KITTI 2015 datasets, respectively., Comment: CVPRW 2024

The scarcity of ground-truth labels poses one major challenge in developing optical flow estimation models that are both generalizable and robust. While current methods rely on data augmentation, they have yet to fully exploit the rich information available in labeled video sequences. We propose OCAI, a method that supports robust frame interpolation by generating intermediate video frames alongside optical flows in between. Utilizing a forward warping approach, OCAI employs occlusion awareness to resolve ambiguities in pixel values and fills in missing values by leveraging the forward-backward consistency of optical flows. Additionally, we introduce a teacher-student style semi-supervised learning method on top of the interpolated frames. Using a pair of unlabeled frames and the teacher model's predicted optical flow, we generate interpolated frames and flows to train a student model. The teacher's weights are maintained using Exponential Moving Averaging of the student. Our evaluations demonstrate perceptually superior interpolation quality and enhanced optical flow accuracy on established benchmarks such as Sintel and KITTI., Comment: CVPR 2024

In this paper, we propose a novel video depth estimation approach, FutureDepth, which enables the model to implicitly leverage multi-frame and motion cues to improve depth estimation by making it learn to predict the future at training. More specifically, we propose a future prediction network, F-Net, which takes the features of multiple consecutive frames and is trained to predict multi-frame features one time step ahead iteratively. In this way, F-Net learns the underlying motion and correspondence information, and we incorporate its features into the depth decoding process. Additionally, to enrich the learning of multiframe correspondence cues, we further leverage a reconstruction network, R-Net, which is trained via adaptively masked auto-encoding of multiframe feature volumes. At inference time, both F-Net and R-Net are used to produce queries to work with the depth decoder, as well as a final refinement network. Through extensive experiments on several benchmarks, i.e., NYUDv2, KITTI, DDAD, and Sintel, which cover indoor, driving, and open-domain scenarios, we show that FutureDepth significantly improves upon baseline models, outperforms existing video depth estimation methods, and sets new state-of-the-art (SOTA) accuracy. Furthermore, FutureDepth is more efficient than existing SOTA video depth estimation models and has similar latencies when comparing to monocular models

In this paper, we introduce a novel approach that harnesses both 2D and 3D attentions to enable highly accurate depth completion without requiring iterative spatial propagations. Specifically, we first enhance a baseline convolutional depth completion model by applying attention to 2D features in the bottleneck and skip connections. This effectively improves the performance of this simple network and sets it on par with the latest, complex transformer-based models. Leveraging the initial depths and features from this network, we uplift the 2D features to form a 3D point cloud and construct a 3D point transformer to process it, allowing the model to explicitly learn and exploit 3D geometric features. In addition, we propose normalization techniques to process the point cloud, which improves learning and leads to better accuracy than directly using point transformers off the shelf. Furthermore, we incorporate global attention on downsampled point cloud features, which enables long-range context while still being computationally feasible. We evaluate our method, DeCoTR, on established depth completion benchmarks, including NYU Depth V2 and KITTI, showcasing that it sets new state-of-the-art performance. We further conduct zero-shot evaluations on ScanNet and DDAD benchmarks and demonstrate that DeCoTR has superior generalizability compared to existing approaches., Comment: Accepted at CVPR 2024

This paper presents Neural Mesh Fusion (NMF), an efficient approach for joint optimization of polygon mesh from multi-view image observations and unsupervised 3D planar-surface parsing of the scene. In contrast to implicit neural representations, NMF directly learns to deform surface triangle mesh and generate an embedding for unsupervised 3D planar segmentation through gradient-based optimization directly on the surface mesh. The conducted experiments show that NMF obtains competitive results compared to state-of-the-art multi-view planar reconstruction, while not requiring any ground-truth 3D or planar supervision. Moreover, NMF is significantly more computationally efficient compared to implicit neural rendering-based scene reconstruction approaches.

Reducing noise in quantum systems is a major challenge towards the application of quantum technologies. Here, we propose and demonstrate a scheme to reduce noise using a quantum autoencoder with rigorous performance guarantees. The quantum autoencoder learns to compresses noisy quantum states into a latent subspace and removes noise via projective measurements. We find various noise models where we can perfectly reconstruct the original state even for high noise levels. We apply the autoencoder to cool thermal states to the ground state and reduce the cost of magic state distillation by several orders of magnitude. Our autoencoder can be implemented using only unitary transformations without ancillas, making it immediately compatible with the state of the art. We experimentally demonstrate our methods to reduce noise in a photonic integrated circuit. Our results can be directly applied to make quantum technologies more robust to noise.

In the modern financial industry system, the structure of products has become more and more complex, and the bottleneck constraint of classical computing power has already restricted the development of the financial industry. Here, we present a photonic chip that implements the unary approach to European option pricing, in combination with the quantum amplitude estimation algorithm, to achieve a quadratic speedup compared to classical Monte Carlo methods. The circuit consists of three modules: a module loading the distribution of asset prices, a module computing the expected payoff, and a module performing the quantum amplitude estimation algorithm to introduce speed-ups. In the distribution module, a generative adversarial network is embedded for efficient learning and loading of asset distributions, which precisely capture the market trends. This work is a step forward in the development of specialized photonic processors for applications in finance, with the potential to improve the efficiency and quality of financial services., Comment: 11 pages, 7 figures

Stellar white-light flares are believed to play an essential role on the physical and chemical properties of the atmosphere of the surrounding exoplanets. Here we report an optical monitoring campaign on the nearby flaring system EI~Cnc carried out by the Ground-based Wide Angle Cameras (GWAC) and its dedicated follow-up telescope. A superflare, coming from the brighter component EI~CncA, was detected and observed, in which four components are required to properly model the complex decay light curve. The lower limit of flare energy in the $R-$band is estimated to be $3.3\times10^{32}$ ergs. 27 flares are additionally detected from the GWAC archive data with a total duration of 290 hours. The inferred cumulative flare frequency distribution follows a quite shallow power-law function with a slope of $\beta=-0.50\pm 0.03$ over the energy range between $10^{30}$ and $10^{33}$ erg, which reinforces the trend that stars cooler than M4 show enhanced superflare activity. The flares identified in EI~Cnc enable us to extend the $\tau-E$ relationship previously established in the white-light superflares of solar-type stars down to an energy as low as $\sim10^{30}$erg (i.e., by three orders): $\tau\propto E^{0.42\pm0.02}$, which suggests a common flare mechanism for stars with a type from M to solar-like, and implies an invariant of $B^{1/3}\upsilon_{\rm A}$ in the white-light flares., Comment: 14 pages, 8 figures

We propose MAMo, a novel memory and attention frame-work for monocular video depth estimation. MAMo can augment and improve any single-image depth estimation networks into video depth estimation models, enabling them to take advantage of the temporal information to predict more accurate depth. In MAMo, we augment model with memory which aids the depth prediction as the model streams through the video. Specifically, the memory stores learned visual and displacement tokens of the previous time instances. This allows the depth network to cross-reference relevant features from the past when predicting depth on the current frame. We introduce a novel scheme to continuously update the memory, optimizing it to keep tokens that correspond with both the past and the present visual information. We adopt attention-based approach to process memory features where we first learn the spatio-temporal relation among the resultant visual and displacement memory tokens using self-attention module. Further, the output features of self-attention are aggregated with the current visual features through cross-attention. The cross-attended features are finally given to a decoder to predict depth on the current frame. Through extensive experiments on several benchmarks, including KITTI, NYU-Depth V2, and DDAD, we show that MAMo consistently improves monocular depth estimation networks and sets new state-of-the-art (SOTA) accuracy. Notably, our MAMo video depth estimation provides higher accuracy with lower latency, when omparing to SOTA cost-volume-based video depth models., Comment: Accepted at ICCV 2023

Bird's-eye-view (BEV) grid is a typical representation of the perception of road components, e.g., drivable area, in autonomous driving. Most existing approaches rely on cameras only to perform segmentation in BEV space, which is fundamentally constrained by the absence of reliable depth information. The latest works leverage both camera and LiDAR modalities but suboptimally fuse their features using simple, concatenation-based mechanisms. In this paper, we address these problems by enhancing the alignment of the unimodal features in order to aid feature fusion, as well as enhancing the alignment between the cameras' perspective view (PV) and BEV representations. We propose X-Align, a novel end-to-end cross-modal and cross-view learning framework for BEV segmentation consisting of the following components: (i) a novel Cross-Modal Feature Alignment (X-FA) loss, (ii) an attention-based Cross-Modal Feature Fusion (X-FF) module to align multi-modal BEV features implicitly, and (iii) an auxiliary PV segmentation branch with Cross-View Segmentation Alignment (X-SA) losses to improve the PV-to-BEV transformation. We evaluate our proposed method across two commonly used benchmark datasets, i.e., nuScenes and KITTI-360. Notably, X-Align significantly outperforms the state-of-the-art by 3 absolute mIoU points on nuScenes. We also provide extensive ablation studies to demonstrate the effectiveness of the individual components., Comment: Accepted for publication at Springer Machine Vision and Applications Journal. The Version of Record of this article is published in Machine Vision and Applications Journal, and is available online at https://doi.org/10.1007/s00138-023-01400-7. arXiv admin note: substantial text overlap with arXiv:2210.06778

We introduce OpenShape, a method for learning multi-modal joint representations of text, image, and point clouds. We adopt the commonly used multi-modal contrastive learning framework for representation alignment, but with a specific focus on scaling up 3D representations to enable open-world 3D shape understanding. To achieve this, we scale up training data by ensembling multiple 3D datasets and propose several strategies to automatically filter and enrich noisy text descriptions. We also explore and compare strategies for scaling 3D backbone networks and introduce a novel hard negative mining module for more efficient training. We evaluate OpenShape on zero-shot 3D classification benchmarks and demonstrate its superior capabilities for open-world recognition. Specifically, OpenShape achieves a zero-shot accuracy of 46.8% on the 1,156-category Objaverse-LVIS benchmark, compared to less than 10% for existing methods. OpenShape also achieves an accuracy of 85.3% on ModelNet40, outperforming previous zero-shot baseline methods by 20% and performing on par with some fully-supervised methods. Furthermore, we show that our learned embeddings encode a wide range of visual and semantic concepts (e.g., subcategories, color, shape, style) and facilitate fine-grained text-3D and image-3D interactions. Due to their alignment with CLIP embeddings, our learned shape representations can also be integrated with off-the-shelf CLIP-based models for various applications, such as point cloud captioning and point cloud-conditioned image generation., Comment: Project Website: https://colin97.github.io/OpenShape/

In this paper, we study the Lipschitz continuous dependence of conservative H\"older continuous weak solutions to a variational wave system derived from a model for nematic liquid crystals. Since the solution of this system generally forms finite time cusp singularity, the solution flow is not Lipschitz continuous under the Sobolev metric used in the existence and uniqueness theory. We establish a Finsler type optimal transport metric, and show the Lipschitz continuous dependence of solution on the initial data under this metric. This kind of Finsler type optimal transport metrics was first established in [A. Bressan and G. Chen, Arch. Ration. Mech. Anal. 226(3) (2017), 1303-1343] for the scalar variational wave equation. This equation can be used to describe the unit direction n of mean orientation of nematic liquid crystals, when n is restricted on a circle. The model considered in this paper describes the propagation of n without this restriction, i.e. n takes any value on the unite sphere. So we need to consider a wave system instead of a scalar equation., Comment: arXiv admin note: text overlap with arXiv:2007.15201

Inverse path tracing has recently been applied to joint material and lighting estimation, given geometry and multi-view HDR observations of an indoor scene. However, it has two major limitations: path tracing is expensive to compute, and ambiguities exist between reflection and emission. Our Factorized Inverse Path Tracing (FIPT) addresses these challenges by using a factored light transport formulation and finds emitters driven by rendering errors. Our algorithm enables accurate material and lighting optimization faster than previous work, and is more effective at resolving ambiguities. The exhaustive experiments on synthetic scenes show that our method (1) outperforms state-of-the-art indoor inverse rendering and relighting methods particularly in the presence of complex illumination effects; (2) speeds up inverse path tracing optimization to less than an hour. We further demonstrate robustness to noisy inputs through material and lighting estimates that allow plausible relighting in a real scene. The source code is available at: https://github.com/lwwu2/fipt, Comment: Updated experiment results; modified real-world sections

The ubiquitous multi-camera setup on modern autonomous vehicles provides an opportunity to construct surround-view depth. Existing methods, however, either perform independent monocular depth estimations on each camera or rely on computationally heavy self attention mechanisms. In this paper, we propose a novel guided attention architecture, EGA-Depth, which can improve both the efficiency and accuracy of self-supervised multi-camera depth estimation. More specifically, for each camera, we use its perspective view as the query to cross-reference its neighboring views to derive informative features for this camera view. This allows the model to perform attention only across views with considerable overlaps and avoid the costly computations of standard self-attention. Given its efficiency, EGA-Depth enables us to exploit higher-resolution visual features, leading to improved accuracy. Furthermore, EGA-Depth can incorporate more frames from previous time steps as it scales linearly w.r.t. the number of views and frames. Extensive experiments on two challenging autonomous driving benchmarks nuScenes and DDAD demonstrate the efficacy of our proposed EGA-Depth and show that it achieves the new state-of-the-art in self-supervised multi-camera depth estimation., Comment: CVPR 2023 Workshop on Autonomous Driving

White-light superflares from ultra cool stars are thought to be resulted from magnetic reconnection, but the magnetic dynamics in a fully convective star is not clear yet. In this paper, we report a stellar superflare detected with the Ground Wide Angle Camera (GWAC), along with rapid follow-ups with the F60A, Xinglong 2.16m and LCOGT telescopes. The effective temperature of the counterpart is estimated to be $2200\pm50$K by the BT-Settl model, corresponding to a spectral type of L0. The $R-$band light curve can be modeled as a sum of three exponential decay components, where the impulsive component contributes a fraction of 23\% of the total energy, while the gradual and the shallower decay phases emit 42\% and 35\% of the total energy, respectively. The strong and variable Balmer narrow emission lines indicate the large amplitude flare is resulted from magnetic activity. The bolometric energy released is about $6.4\times10^{33}$ ergs, equivalent to an energy release in a duration of 143.7 hours at its quiescent level. The amplitude of $\Delta R=-8.6 $mag ( or $\Delta V=-11.2$ mag), placing it one of the highest amplitudes of any ultra cool star recorded with excellent temporal resolution. We argue that a stellar flare with such rapidly decaying and huge amplitude at distances greater than 1 kpc may be false positive in searching for counterparts of catastrophic events such as gravitational wave events or gamma-ray bursts, which are valuable in time-domain astronomy and should be given more attention., Comment: 9 pages, 5 figures, 1 table, MNRAS accepted

In this paper, we develop rotation-equivariant neural networks for 4D panoptic segmentation. 4D panoptic segmentation is a benchmark task for autonomous driving that requires recognizing semantic classes and object instances on the road based on LiDAR scans, as well as assigning temporally consistent IDs to instances across time. We observe that the driving scenario is symmetric to rotations on the ground plane. Therefore, rotation-equivariance could provide better generalization and more robust feature learning. Specifically, we review the object instance clustering strategies and restate the centerness-based approach and the offset-based approach as the prediction of invariant scalar fields and equivariant vector fields. Other sub-tasks are also unified from this perspective, and different invariant and equivariant layers are designed to facilitate their predictions. Through evaluation on the standard 4D panoptic segmentation benchmark of SemanticKITTI, we show that our equivariant models achieve higher accuracy with lower computational costs compared to their non-equivariant counterparts. Moreover, our method sets the new state-of-the-art performance and achieves 1st place on the SemanticKITTI 4D Panoptic Segmentation leaderboard., Comment: 13 pages. Accepted at ICCV 2023

We propose a novel data augmentation approach, DistractFlow, for training optical flow estimation models by introducing realistic distractions to the input frames. Based on a mixing ratio, we combine one of the frames in the pair with a distractor image depicting a similar domain, which allows for inducing visual perturbations congruent with natural objects and scenes. We refer to such pairs as distracted pairs. Our intuition is that using semantically meaningful distractors enables the model to learn related variations and attain robustness against challenging deviations, compared to conventional augmentation schemes focusing only on low-level aspects and modifications. More specifically, in addition to the supervised loss computed between the estimated flow for the original pair and its ground-truth flow, we include a second supervised loss defined between the distracted pair's flow and the original pair's ground-truth flow, weighted with the same mixing ratio. Furthermore, when unlabeled data is available, we extend our augmentation approach to self-supervised settings through pseudo-labeling and cross-consistency regularization. Given an original pair and its distracted version, we enforce the estimated flow on the distracted pair to agree with the flow of the original pair. Our approach allows increasing the number of available training pairs significantly without requiring additional annotations. It is agnostic to the model architecture and can be applied to training any optical flow estimation models. Our extensive evaluations on multiple benchmarks, including Sintel, KITTI, and SlowFlow, show that DistractFlow improves existing models consistently, outperforming the latest state of the art., Comment: CVPR 2023