Searcharxiv⌕ Search

arXiv subjects

Shu Liu

Publications and source records attributed to Shu Liu.

At least 109 records · Page 6Linked to original sources
arXiv2024

GridMask Data Augmentation

We propose a novel data augmentation method `GridMask' in this paper. It utilizes information removal to achieve state-of-the-art results in a variety of computer vision tasks. We analyze the requirement of information dropping. Then we show limitation of existing information dropping algorithms and propose our structured method, which is simple and yet very effective. It is based on the deletion of regions of the input image. Our extensive experiments show that our method outperforms the latest AutoAugment, which is way more computationally expensive due to the use of reinforcement learning to find the best policies. On the ImageNet dataset for recognition, COCO2017 object detection, and on Cityscapes dataset for semantic segmentation, our method all notably improves performance over baselines. The extensive experiments manifest the effectiveness and generality of the new method.

cs.CV↗
arXiv2024

LISA++: An Improved Baseline for Reasoning Segmentation with Large Language Model

While LISA effectively bridges the gap between segmentation and large language models to enable reasoning segmentation, it poses certain limitations: unable to distinguish different instances of the target region, and constrained by the pre-defined textual response formats. In this work, we introduce LISA++, an update to the existing LISA model, focusing on improving core functionalities while keeping the base architecture intact. The main enhancements in LISA++ include: \textbf{1) Enhanced Segmentation}: The instance segmentation ability has been added, providing a more detailed scene analysis along with the existing multi-region semantic segmentation. \textbf{2) More Natural Conversation}: Improved capability for multi-turn dialogue, with the ability to incorporate segmentation results directly into text responses, i.e., Segmentation in Dialogue (SiD). These improvements are achieved by curating the existing samples of generic segmentation datasets, aimed specifically at enhancing the segmentation and conversational skills without structural change and additional data sources. Comparative analysis with the original LISA model shows significant advancements in these areas, positioning LISA++ as a notable upgrade in visual understanding and interaction. LISA++'s adaptability and improved features highlight the versatility of the mask-as-embedding paradigm proposed by LISA, and the potential as a foundational model for diverse applications.

cs.CV↗
arXiv2024

MOODv2: Masked Image Modeling for Out-of-Distribution Detection

The crux of effective out-of-distribution (OOD) detection lies in acquiring a robust in-distribution (ID) representation, distinct from OOD samples. While previous methods predominantly leaned on recognition-based techniques for this purpose, they often resulted in shortcut learning, lacking comprehensive representations. In our study, we conducted a comprehensive analysis, exploring distinct pretraining tasks and employing various OOD score functions. The results highlight that the feature representations pre-trained through reconstruction yield a notable enhancement and narrow the performance gap among various score functions. This suggests that even simple score functions can rival complex ones when leveraging reconstruction-based pretext tasks. Reconstruction-based pretext tasks adapt well to various score functions. As such, it holds promising potential for further expansion. Our OOD detection framework, MOODv2, employs the masked image modeling pretext task. Without bells and whistles, MOODv2 impressively enhances 14.30% AUROC to 95.68% on ImageNet and achieves 99.98% on CIFAR-10.

cs.CV↗
arXiv2023

BAL: Balancing Diversity and Novelty for Active Learning

The objective of Active Learning is to strategically label a subset of the dataset to maximize performance within a predetermined labeling budget. In this study, we harness features acquired through self-supervised learning. We introduce a straightforward yet potent metric, Cluster Distance Difference, to identify diverse data. Subsequently, we introduce a novel framework, Balancing Active Learning (BAL), which constructs adaptive sub-pools to balance diverse and uncertain data. Our approach outperforms all established active learning methods on widely recognized benchmarks by 1.20%. Moreover, we assess the efficacy of our proposed framework under extended settings, encompassing both larger and smaller labeling budgets. Experimental results demonstrate that, when labeling 80% of the samples, the performance of the current SOTA method declines by 0.74%, whereas our proposed BAL achieves performance comparable to the full dataset. Codes are available at https://github.com/JulietLJY/BAL.

cs.LG↗
arXiv2023

Spatial distribution of NH2D in massive star-forming regions

To understand the relation between NH$_2$D and its physical environment, we mapped ortho-NH$_2$D $1_{11}^s-1_{01}^a$ at 85.9 GHz toward 24 Galactic late-stage massive star-forming regions with Institut de Radioastronomie Millim$ é$trique (IRAM) 30-m telescope. Ortho-NH$_2$D $1_{11}^s-1_{01}^a$ was detected in 18 of 24 sources. Comparing with the distribution of H$^{13}$CN 1-0 as a dense gas tracer and radio recombination line H42$α$, ortho-NH$_2$D $1_{11}^s-1_{01}^a$ present complex and diverse spatial distribution in these targets. 11 of the 18 targets, present a different distribution between ortho-NH$_2$D $1_{11}^s-1_{01}^a$ and H$^{13}$CN 1-0, while no significant difference between these two lines can be found in the other 7 sources, mainly due to limited spatial resolution and sensitivity. Moreover, with H42$α$ tracing massive young stellar objects, ortho-NH$_2$D $1_{11}^s-1_{01}^a$ seems to show a relatively weak emission near the massive young stellar objects.

astro-ph.GA↗
arXiv2023

HyperS2V: A Framework for Structural Representation of Nodes in Hyper Networks

In contrast to regular (simple) networks, hyper networks possess the ability to depict more complex relationships among nodes and store extensive information. Such networks are commonly found in real-world applications, such as in social interactions. Learning embedded representations for nodes involves a process that translates network structures into more simplified spaces, thereby enabling the application of machine learning approaches designed for vector data to be extended to network data. Nevertheless, there remains a need to delve into methods for learning embedded representations that prioritize structural aspects. This research introduces HyperS2V, a node embedding approach that centers on the structural similarity within hyper networks. Initially, we establish the concept of hyper-degrees to capture the structural properties of nodes within hyper networks. Subsequently, a novel function is formulated to measure the structural similarity between different hyper-degree values. Lastly, we generate structural embeddings utilizing a multi-scale random walk framework. Moreover, a series of experiments, both intrinsic and extrinsic, are performed on both toy and real networks. The results underscore the superior performance of HyperS2V in terms of both interpretability and applicability to downstream tasks.

cs.SI↗
arXiv2023

Nonlinear Codes with Low Redundancy

Determining the largest size, or equivalently finding the lowest redundancy, of q-ary codes for given length and minimum distance is one of the central and fundamental problems in coding theory. Inspired by the construction of Varshamov-Tenengolts (VT for short) codes via check-sums, we provide an explicit construction of nonlinear codes with lower redundancy than linear codes under the same length and minimum distance. Similar to the VT codes, our construction works well for small distance (or even constant distance). Furthermore, we design quasi-linear time decoding algorithms for both erasure and adversary errors.

cs.IT↗
arXiv2023

Enhancing the machine vision performance with multi-spectral light sources

This study mainly focuses on the performance of different multi-spectral light sources on different object colors in machine vision and tries to enhance machine vision with multi-spectral light sources. Using different color pencils as samples, by recognizing the collected images with two classical neural networks, AlexNet and VGG19, the performance was investigated under 35 different multi-spectral light sources. The results show that for both models there are always some non-pure white light sources, whose accuracy is better than pure white light, which suggests the potential of multi-spectral light sources to further enhance the effectiveness of machine vision. The comparison of both models is also performed, and surprised to find that the overall performance of VGG19 is lower than that of AlexNet, which shows that the importance of the choice of multi-spectral light sources and models.

eess.IV↗
arXiv2023

Opacities of dense gas tracers in galactic massive star-forming regions

Optical depths of dense molecular gas are commonly used in Galactic and extragalactic studies to constrain the dense gas mass of the clouds or galaxies. The optical depths are often obtained based on spatially unresolved data, especially in galaxies, which may affect the reliability of such measurements. We examine such effects in spatially resolved Galactic massive star-forming regions. Using the 10-m SMT telescope, we mapped HCN and H13CN 3-2, HCO+, and H13CO+ 3-2 towards 51 Galactic massive star-forming regions, 30 of which resulted in robust determination of spatially resolved optical depths. Conspicuous spatial variations of optical depths have been detected within each source. We first obtained opacities for each position and calculated an optical-thick line intensity-weighted average, then averaged all the spectra and derived a single opacity for each region. The two were found to agree extremely well, with a linear least square correlation coefficient of 0.997 for the whole sample.

astro-ph.GA↗
arXiv2023

A Lower Bound on the List-Decodability of Insdel Codes

For codes equipped with metrics such as Hamming metric, symbol pair metric or cover metric, the Johnson bound guarantees list-decodability of such codes. That is, the Johnson bound provides a lower bound on the list-decoding radius of a code in terms of its relative minimum distance $δ$, list size $L$ and the alphabet size $q.$ For study of list-decodability of codes with insertion and deletion errors (we call such codes insdel codes), it is natural to ask the open problem whether there is also a Johnson-type bound. The problem was first investigated by Wachter-Zeh and the result was amended by Hayashi and Yasunaga where a lower bound on the list-decodability for insdel codes was derived. The main purpose of this paper is to move a step further towards solving the above open problem. In this work, we provide a new lower bound for the list-decodability of an insdel code. As a consequence, we show that unlike the Johnson bound for codes under other metrics that is tight, the bound on list-decodability of insdel codes given by Hayashi and Yasunaga is not tight. Our main idea is to show that if an insdel code with a given Levenshtein distance $d$ is not list-decodable with list size $L$, then the list decoding radius is lower bounded by a bound involving $L$ and $d$. In other words, if the list decoding radius is less than this lower bound, the code must be list-decodable with list size $L$. At the end of the paper we use such bound to provide an insdel-list-decodability bound for various well-known codes, which has not been extensively studied before.

cs.IT↗
arXiv2023

Ref-NeuS: Ambiguity-Reduced Neural Implicit Surface Learning for Multi-View Reconstruction with Reflection

Neural implicit surface learning has shown significant progress in multi-view 3D reconstruction, where an object is represented by multilayer perceptrons that provide continuous implicit surface representation and view-dependent radiance. However, current methods often fail to accurately reconstruct reflective surfaces, leading to severe ambiguity. To overcome this issue, we propose Ref-NeuS, which aims to reduce ambiguity by attenuating the effect of reflective surfaces. Specifically, we utilize an anomaly detector to estimate an explicit reflection score with the guidance of multi-view context to localize reflective surfaces. Afterward, we design a reflection-aware photometric loss that adaptively reduces ambiguity by modeling rendered color as a Gaussian distribution, with the reflection score representing the variance. We show that together with a reflection direction-dependent radiance, our model achieves high-quality surface reconstruction on reflective surfaces and outperforms the state-of-the-arts by a large margin. Besides, our model is also comparable on general surfaces.

cs.CV↗
arXiv2023

Not All Steps are Created Equal: Selective Diffusion Distillation for Image Manipulation

Conditional diffusion models have demonstrated impressive performance in image manipulation tasks. The general pipeline involves adding noise to the image and then denoising it. However, this method faces a trade-off problem: adding too much noise affects the fidelity of the image while adding too little affects its editability. This largely limits their practical applicability. In this paper, we propose a novel framework, Selective Diffusion Distillation (SDD), that ensures both the fidelity and editability of images. Instead of directly editing images with a diffusion model, we train a feedforward image manipulation network under the guidance of the diffusion model. Besides, we propose an effective indicator to select the semantic-related timestep to obtain the correct semantic guidance from the diffusion model. This approach successfully avoids the dilemma caused by the diffusion process. Our extensive experiments demonstrate the advantages of our framework. Code is released at https://github.com/AndysonYs/Selective-Diffusion-Distillation.

cs.CV↗
arXiv2023

Parameterized Wasserstein Hamiltonian Flow

In this work, we propose a numerical method to compute the Wasserstein Hamiltonian flow (WHF), which is a Hamiltonian system on the probability density manifold. Many well-known PDE systems can be reformulated as WHFs. We use parameterized function as push-forward map to characterize the solution of WHF, and convert the PDE to a finite-dimensional ODE system, which is a Hamiltonian system in the phase space of the parameter manifold. We establish error analysis results for the continuous time approximation scheme in Wasserstein metric. For the numerical implementation, we use neural networks as push-forward maps. We apply an effective symplectic scheme to solve the derived Hamiltonian ODE system so that the method preserves some important quantities such as total energy. The computation is done by fully deterministic symplectic integrator without any neural network training. Thus, our method does not involve direct optimization over network parameters and hence can avoid the error introduced by stochastic gradient descent (SGD) methods, which is usually hard to quantify and measure. The proposed algorithm is a sampling-based approach that scales well to higher dimensional problems. In addition, the method also provides an alternative connection between the Lagrangian and Eulerian perspectives of the original WHF through the parameterized ODE dynamics.

math.NA↗
arXiv2023

Codes with Biochemical Constraints and Single Error Correction for DNA-Based Data Storage

In DNA-based data storage, DNA codes with biochemical constraints and error correction are designed to protect data reliability. Single-stranded DNA sequences with secondary structure avoidance (SSA) help to avoid undesirable secondary structures which may cause chemical inactivity. Homopolymer run-length limit and GC-balanced limit also help to reduce the error probability of DNA sequences during synthesizing and sequencing. In this letter, based on a recent work \cite{bib7}, we construct DNA codes free of secondary structures of stem length $\geq m$ and have homopolymer run-length $\leq\ell$ for odd $m\leq11$ and $\ell\geq3$ with rate $1+\log_2ρ_m-3/(2^{\ell-1}+\ell+1)$, where $ρ_m$ is in Table \ref{tm}. In particular, when $m=3$, $\ell=4$, its rate tends to 1.3206 bits/nt, beating a previous work by Benerjee {\it et al.}. We also construct DNA codes with all of the above three constraints as well as single error correction. At last, codes with GC-locally balanced constraint are presented.

cs.IT↗
arXiv2023

Hierarchical Dense Correlation Distillation for Few-Shot Segmentation-Extended Abstract

Few-shot semantic segmentation (FSS) aims to form class-agnostic models segmenting unseen classes with only a handful of annotations. Previous methods limited to the semantic feature and prototype representation suffer from coarse segmentation granularity and train-set overfitting. In this work, we design Hierarchically Decoupled Matching Network (HDMNet) mining pixel-level support correlation based on the transformer architecture. The self-attention modules are used to assist in establishing hierarchical dense features, as a means to accomplish the cascade matching between query and support features. Moreover, we propose a matching module to reduce train-set overfitting and introduce correlation distillation leveraging semantic correspondence from coarse resolution to boost fine-grained segmentation. Our method performs decently in experiments. We achieve 50.0% mIoU on COCO dataset one-shot setting and 56.0% on five-shot segmentation, respectively. The code will be available on the project website. We hope our work can benefit broader industrial applications where novel classes with limited annotations are required to be decently identified.

cs.CV↗
arXiv2023

Explicit Construction of q-ary 2-deletion Correcting Codes with Low Redundancy

We consider the problem of efficient construction of q-ary 2-deletion correcting codes with low redundancy. We show that our construction requires less redundancy than any existing efficiently encodable q-ary 2-deletion correcting codes. Precisely speaking, we present an explicit construction of a q-ary 2-deletion correcting code with redundancy 5 log(n)+10log(log(n)) + 3 log(q)+O(1). Using a minor modification to the original construction, we obtain an efficiently encodable q-ary 2-deletion code that is efficiently list-decodable. Similarly, we show that our construction of list-decodable code requires a smaller redundancy compared to any existing list-decodable codes. To obtain our sketches, we transform a q-ary codeword to a binary string which can then be used as an input to the underlying base binary sketch. This is then complemented with additional q-ary sketches that the original q-ary codeword is required to satisfy. In other words, we build our codes via a binary 2-deletion code as a black-box. Finally we utilize the binary 2-deletion code proposed by Guruswami and Hastad to our construction to obtain the main result of this paper.

cs.IT↗
arXiv2023

Generalized Parametric Contrastive Learning

In this paper, we propose the Generalized Parametric Contrastive Learning (GPaCo/PaCo) which works well on both imbalanced and balanced data. Based on theoretical analysis, we observe that supervised contrastive loss tends to bias high-frequency classes and thus increases the difficulty of imbalanced learning. We introduce a set of parametric class-wise learnable centers to rebalance from an optimization perspective. Further, we analyze our GPaCo/PaCo loss under a balanced setting. Our analysis demonstrates that GPaCo/PaCo can adaptively enhance the intensity of pushing samples of the same class close as more samples are pulled together with their corresponding centers and benefit hard example learning. Experiments on long-tailed benchmarks manifest the new state-of-the-art for long-tailed recognition. On full ImageNet, models from CNNs to vision transformers trained with GPaCo loss show better generalization performance and stronger robustness compared with MAE models. Moreover, GPaCo can be applied to the semantic segmentation task and obvious improvements are observed on the 4 most popular benchmarks. Our code is available at https://github.com/dvlab-research/Parametric-Contrastive-Learning.

cs.CV↗
arXiv2023

A first-order computational algorithm for reaction-diffusion type equations via primal-dual hybrid gradient method

We propose an easy-to-implement iterative method for resolving the implicit (or semi-implicit) schemes arising in solving reaction-diffusion (RD) type equations. We formulate the nonlinear time implicit scheme as a min-max saddle point problem and then apply the primal-dual hybrid gradient (PDHG) method. Suitable precondition matrices are applied to the PDHG method to accelerate the convergence of algorithms under different circumstances. Furthermore, our method is applicable to various discrete numerical schemes with high flexibility. From various numerical examples tested in this paper, the proposed method converges properly and can efficiently produce numerical solutions with sufficient accuracy.

math.NA↗