Search

With the advancements in denoising diffusion probabilistic models (DDPMs), image inpainting has significantly evolved from merely filling information based on nearby regions to generating content conditioned on various prompts such as text, exemplar images, and sketches. However, existing methods, such as model fine-tuning and simple concatenation of latent vectors, often result in generation failures due to overfitting and inconsistency between the inpainted region and the background. In this paper, we argue that the current large diffusion models are sufficiently powerful to generate realistic images without further tuning. Hence, we introduce PILOT (in\textbf{P}ainting v\textbf{I}a \textbf{L}atent \textbf{O}p\textbf{T}imization), an optimization approach grounded on a novel \textit{semantic centralization} and \textit{background preservation loss}. Our method searches latent spaces capable of generating inpainted regions that exhibit high fidelity to user-provided prompts while maintaining coherence with the background. Furthermore, we propose a strategy to balance optimization expense and image quality, significantly enhancing generation efficiency. Our method seamlessly integrates with any pre-trained model, including ControlNet and DreamBooth, making it suitable for deployment in multi-modal editing tools. Our qualitative and quantitative evaluations demonstrate that PILOT outperforms existing approaches by generating more coherent, diverse, and faithful inpainted regions in response to provided prompts.

Multi-label image classification datasets are often partially labeled where many labels are missing, posing a significant challenge to training accurate deep classifiers. However, the powerful Mixup sample-mixing data augmentation cannot be well utilized to address this challenge, as it cannot perform linear interpolation on the unknown labels to construct augmented samples. In this paper, we propose LogicMix, a Mixup variant designed for such partially labeled datasets. LogicMix mixes the sample labels by logical OR so that the unknown labels can be correctly mixed by utilizing OR's logical equivalences, including the domination and identity laws. Unlike Mixup, which mixes exactly two samples, LogicMix can mix multiple ($\geq2$) partially labeled samples, constructing visually more confused augmented samples to regularize training. LogicMix is more general and effective than other compared Mixup variants in the experiments on various partially labeled dataset scenarios. Moreover, it is plug-and-play and only requires minimal computation, hence it can be easily inserted into existing frameworks to collaborate with other methods to improve model performance with a negligible impact on training time, as demonstrated through extensive experiments. In particular, through the collaboration of LogicMix, RandAugment, Curriculum Labeling, and Category-wise Fine-Tuning, we attain state-of-the-art performance on MS-COCO, VG-200, and Pascal VOC 2007 benchmarking datasets. The remarkable generality, effectiveness, collaboration, and simplicity suggest that LogicMix promises to be a popular and vital data augmentation method.

Model-Driven Engineering (MDE) is a technique that aims to boost productivity in software development and ensure the safety of critical systems. Central to MDE is the refinement of high-level requirement models into executable code. Given that requirement models form the foundation of the entire development process, ensuring their correctness is crucial. RM2PT is a widely used MDE platform that employs the REModel language for requirement modeling. REModel contains contract sections and other sections including a UML sequence diagram. This paper contributes a coroutine-based type system that represents pre- and post-conditions in the contract sections in a requirement model as the receiving and yielding parts of coroutines, respectively. The type system is capable of composing coroutine types, so that users can view functions as a whole system and check their collective behavior. By doing so, our type system ensures that the contracts defined in it are executed as outlined in the accompanied sequence diagram. We assessed our approach using four case studies provided by RM2PT, validating the accuracy of the models.

In the realm of point cloud scene understanding, particularly in indoor scenes, objects are arranged following human habits, resulting in objects of certain semantics being closely positioned and displaying notable inter-object correlations. This can create a tendency for neural networks to exploit these strong dependencies, bypassing the individual object patterns. To address this challenge, we introduce a novel self-supervised learning (SSL) strategy. Our approach leverages both object patterns and contextual cues to produce robust features. It begins with the formulation of an object-exchanging strategy, where pairs of objects with comparable sizes are exchanged across different scenes, effectively disentangling the strong contextual dependencies. Subsequently, we introduce a context-aware feature learning strategy, which encodes object patterns without relying on their specific context by aggregating object features across various scenes. Our extensive experiments demonstrate the superiority of our method over existing SSL techniques, further showing its better robustness to environmental changes. Moreover, we showcase the applicability of our approach by transferring pre-trained models to diverse point cloud datasets.

The pre-trained vision-language model, exemplified by CLIP, advances zero-shot semantic segmentation by aligning visual features with class embeddings through a transformer decoder to generate semantic masks. Despite its effectiveness, prevailing methods within this paradigm encounter challenges, including overfitting on seen classes and small fragmentation in masks. To mitigate these issues, we propose a Language-Driven Visual Consensus (LDVC) approach, fostering improved alignment of semantic and visual information.Specifically, we leverage class embeddings as anchors due to their discrete and abstract nature, steering vision features toward class embeddings. Moreover, to circumvent noisy alignments from the vision part due to its redundant nature, we introduce route attention into self-attention for finding visual consensus, thereby enhancing semantic consistency within the same object. Equipped with a vision-language prompting strategy, our approach significantly boosts the generalization capacity of segmentation models for unseen classes. Experimental results underscore the effectiveness of our approach, showcasing mIoU gains of 4.5 on the PASCAL VOC 2012 and 3.6 on the COCO-Stuff 164k for unseen classes compared with the state-of-the-art methods.

Large-scale image datasets are often partially labeled, where only a few categories' labels are known for each image. Assigning pseudo-labels to unknown labels to gain additional training signals has become prevalent for training deep classification models. However, some pseudo-labels are inevitably incorrect, leading to a notable decline in the model classification performance. In this paper, we propose a novel method called Category-wise Fine-Tuning (CFT), aiming to reduce model inaccuracies caused by the wrong pseudo-labels. In particular, CFT employs known labels without pseudo-labels to fine-tune the logistic regressions of trained models individually to calibrate each category's model predictions. Genetic Algorithm, seldom used for training deep models, is also utilized in CFT to maximize the classification performance directly. CFT is applied to well-trained models, unlike most existing methods that train models from scratch. Hence, CFT is general and compatible with models trained with different methods and schemes, as demonstrated through extensive experiments. CFT requires only a few seconds for each category for calibration with consumer-grade GPUs. We achieve state-of-the-art results on three benchmarking datasets, including the CheXpert chest X-ray competition dataset (ensemble mAUC 93.33%, single model 91.82%), partially labeled MS-COCO (average mAP 83.69%), and Open Image V3 (mAP 85.31%), outperforming the previous bests by 0.28%, 2.21%, 2.50%, and 0.91%, respectively. The single model on CheXpert has been officially evaluated by the competition server, endorsing the correctness of the result. The outstanding results and generalizability indicate that CFT could be substantial and prevalent for classification model development. Code is available at: https://github.com/maxium0526/category-wise-fine-tuning.

Despite the potential benefits of data augmentation for mitigating the data insufficiency, traditional augmentation methods primarily rely on the prior intra-domain knowledge. On the other hand, advanced generative adversarial networks (GANs) generate inter-domain samples with limited variety. These previous methods make limited contributions to describing the decision boundaries for binary classification. In this paper, we propose a distance guided GAN (DisGAN) which controls the variation degrees of generated samples in the hyperplane space. Specifically, we instantiate the idea of DisGAN by combining two ways. The first way is vertical distance GAN (VerDisGAN) where the inter-domain generation is conditioned on the vertical distances. The second way is horizontal distance GAN (HorDisGAN) where the intra-domain generation is conditioned on the horizontal distances. Furthermore, VerDisGAN can produce the class-specific regions by mapping the source images to the hyperplane. Experimental results show that DisGAN consistently outperforms the GAN-based augmentation methods with explainable binary classification. The proposed method can apply to different classification architectures and has potential to extend to multi-class classification., Comment: 12 pages, 8 figures. This work has been submitted to the IEEE TNNLS for possible publication. Personal use of this material is permitted. Permission from IEEE must be obtained for all other uses, in any current or future media

We consider the uniqueness of the following positive solutions of $m$-Laplacian equation: \begin{equation} \left\{ \begin{aligned} -\Delta _m u&=\lambda u^{m-1}+u^{p-1} \quad \text{in} \quad \Omega\\ u&=0 \quad \text{on} \quad \partial \Omega \end{aligned} \right. \qquad(0.1)\end{equation} where $m>1$ is a constant. When $p\rightarrow m$, the uniqueness of positive solutions of $(0.1)$ is shown which is based on the essential uniqueness of first eigenfunction for $m$-Laplacian equation. Futhermore, we also prove the uniqueness results when $(0.1)$ is a perturbation of Laplacian equation., Comment: arXiv admin note: substantial text overlap with arXiv:2201.03394 by other authors

Photonics offers a transformative approach to artificial intelligence (AI) and neuromorphic computing by providing low latency, high bandwidth, and energy-efficient computations. Here, we introduce a photonic tensor core processor enabled by time-multiplexed inputs and charge-integrated outputs. This fully integrated processor, comprising only two thin-film lithium niobate (TFLN) modulators, a III-V laser, and a charge-integration photoreceiver, can implement an entire layer of a neural network. It can execute 65 billion operations per second (GOPS) per neuron, including simultaneous weight updates-a hitherto unachieved speed. Our processor stands out from conventional photonic processors, which have static weights set during training, as it supports fast "hardware-in-the-loop" training, and can dynamically adjust the inputs (fan-in) and outputs (fan-out) within a layer, thereby enhancing its versatility. Our processor can perform large-scale dot-product operations with vector dimensions up to 131,072. Furthermore, it successfully classifies (supervised learning) and clusters (unsupervised learning) 112*112-pixel images after "hardware-in-the-loop" training. To handle "hardware-in-the-loop" training for clustering AI tasks, we provide a solution for multiplications involving two negative numbers based on our processor., Comment: 19 pages, 6 figures

Although current data augmentation methods are successful to alleviate the data insufficiency, conventional augmentation are primarily intra-domain while advanced generative adversarial networks (GANs) generate images remaining uncertain, particularly in small-scale datasets. In this paper, we propose a parameterized GAN (ParaGAN) that effectively controls the changes of synthetic samples among domains and highlights the attention regions for downstream classification. Specifically, ParaGAN incorporates projection distance parameters in cyclic projection and projects the source images to the decision boundary to obtain the class-difference maps. Our experiments show that ParaGAN can consistently outperform the existing augmentation methods with explainable classification on two small-scale medical datasets., Comment: 5 pages, 4 figures. This work has been submitted to the IEEE ICASSP for possible publication. Copyright may be transferred without notice, after which this version may no longer be accessible

Coroutine, as a powerful programming construct, is widely used in asynchronous applications to replace thread-based programming or the callback hell. Using coroutines makes code more readable and maintainable, for its ability to transfer control while keeping the literal scope. However, reasoning about coroutine behavior can be challenging without proper typing. We propose a type notation and calculus for composing asymmetric, first-class, stackless coroutines. Given the types of a list of coroutines, we can compute a composed type matching the collective behavior of the coroutines, so that the input and output can be type-checked by a type system. Our coroutine types can model the data received by or yielded from a coroutine, which be of coroutine types as well. On top of our type calculus, we discuss its soundness and evaluation issues, then provide four application scenarios of our coroutine types. Not only can our types be used in modern programming languages, such as Python, but also model program behaviors in OCaml and even Prolog.

The mass and decay of the $s\bar{s}$ member of the $1^{3}F_{4}$ meson nonet are investigated in the framework of the Regge phenomenology and the $^{3}P_{0}$ model. We propose, based on the results, that the assignment of the $s\bar{s}$ member of the $1^{3}F_{4}$ meson nonet will require additional testing in the future. Our results also provide information for future studies of the $1^{3}F_{4}$ meson nonet., Comment: 11 pages

Community detection becomes an important problem with the booming of social networks. The Medoid-Shift algorithm preserves the benefits of Mean-Shift and can be applied to problems based on distance matrix, such as community detection. One drawback of the Medoid-Shift algorithm is that there may be no data points within the neighborhood region defined by a distance parameter. To deal with the community detection problem better, a new algorithm called Revised Medoid-Shift (RMS) in this work is thus proposed. During the process of finding the next medoid, the RMS algorithm is based on a neighborhood defined by KNN, while the original Medoid-Shift is based on a neighborhood defined by a distance parameter. Since the neighborhood defined by KNN is more stable than the one defined by the distance parameter in terms of the number of data points within the neighborhood, the RMS algorithm may converge more smoothly. In the RMS method, each of the data points is shifted towards a medoid within the neighborhood defined by KNN. After the iterative process of shifting, each of the data point converges into a cluster center, and the data points converging into the same center are grouped into the same cluster. The RMS algorithm is tested on two kinds of datasets including community datasets with known ground truth partition and community datasets without ground truth partition respectively. The experiment results show sthat the proposed RMS algorithm generally produces betster results than Medoid-Shift and some state-of-the-art together with most classic community detection algorithms on different kinds of community detection datasets.

This paper presents a new annotation method called Sparse Annotation (SA) for crowd counting, which reduces human labeling efforts by sparsely labeling individuals in an image. We argue that sparse labeling can reduce the redundancy of full annotation and capture more diverse information from distant individuals that is not fully captured by Partial Annotation methods. Besides, we propose a point-based Progressive Point Matching network (PPM) to better explore the crowd from the whole image with sparse annotation, which includes a Proposal Matching Network (PMN) and a Performance Restoration Network (PRN). The PMN generates pseudo-point samples using a basic point classifier, while the PRN refines the point classifier with the pseudo points to maximize performance. Our experimental results show that PPM outperforms previous semi-supervised crowd counting methods with the same amount of annotation by a large margin and achieves competitive performance with state-of-the-art fully-supervised methods.

Dense Self-Supervised Learning (SSL) methods address the limitations of using image-level feature representations when handling images with multiple objects. Although the dense features extracted by employing segmentation maps and bounding boxes allow networks to perform SSL for each object, we show that they suffer from coupling and positional bias, which arise from the receptive field increasing with layer depth and zero-padding. We address this by introducing three data augmentation strategies, and leveraging them in (i) a decoupling module that aims to robustify the network to variations in the object's surroundings, and (ii) a de-positioning module that encourages the network to discard positional object information. We demonstrate the benefits of our method on COCO and on a new challenging benchmark, OpenImage-MINI, for object classification, semantic segmentation, and object detection. Our extensive experiments evidence the better generalization of our method compared to the SOTA dense SSL methods

Self-supervised learning (SSL) has the potential to benefit many applications, particularly those where manually annotating data is cumbersome. One such situation is the semantic segmentation of point clouds. In this context, existing methods employ contrastive learning strategies and define positive pairs by performing various augmentation of point clusters in a single frame. As such, these methods do not exploit the temporal nature of LiDAR data. In this paper, we introduce an SSL strategy that leverages positive pairs in both the spatial and temporal domain. To this end, we design (i) a point-to-cluster learning strategy that aggregates spatial information to distinguish objects; and (ii) a cluster-to-cluster learning strategy based on unsupervised object tracking that exploits temporal correspondences. We demonstrate the benefits of our approach via extensive experiments performed by self-supervised training on two large-scale LiDAR datasets and transferring the resulting models to other point cloud segmentation benchmarks. Our results evidence that our method outperforms the state-of-the-art point cloud SSL methods., Comment: CVPR accepted

In this work, the mass spectrum of the first radial excitation axial vector meson nonet is considered in the framework of the nonrelativistic constituent quark model and Regge phenomenology. After that, we investigate the strong decay characteristics of these states within the $^{3}P_{0}$ model. The results are compared to values from other phenomenological models, and they may be beneficial in the prospective search for $2^{3}P_{1}$ meson nonets., Comment: 11 pages

Referring image segmentation aims at localizing all pixels of the visual objects described by a natural language sentence. Previous works learn to straightforwardly align the sentence embedding and pixel-level embedding for highlighting the referred objects, but ignore the semantic consistency of pixels within the same object, leading to incomplete masks and localization errors in predictions. To tackle this problem, we propose CoupAlign, a simple yet effective multi-level visual-semantic alignment method, to couple sentence-mask alignment with word-pixel alignment to enforce object mask constraint for achieving more accurate localization and segmentation. Specifically, the Word-Pixel Alignment (WPA) module performs early fusion of linguistic and pixel-level features in intermediate layers of the vision and language encoders. Based on the word-pixel aligned embedding, a set of mask proposals are generated to hypothesize possible objects. Then in the Sentence-Mask Alignment (SMA) module, the masks are weighted by the sentence embedding to localize the referred object, and finally projected back to aggregate the pixels for the target. To further enhance the learning of the two alignment modules, an auxiliary loss is designed to contrast the foreground and background pixels. By hierarchically aligning pixels and masks with linguistic features, our CoupAlign captures the pixel coherence at both visual and semantic levels, thus generating more accurate predictions. Extensive experiments on popular datasets (e.g., RefCOCO and G-Ref) show that our method achieves consistent improvements over state-of-the-art methods, e.g., about 2% oIoU increase on the validation and testing set of RefCOCO. Especially, CoupAlign has remarkable ability in distinguishing the target from multiple objects of the same class., Comment: accept to NeurIPS 2022

Based on the mass relations from the Regge phenomenology, we revisit the mass spectrum of $1^{3}D_{1}$ meson nonet. The masses of kaon and $s\bar{s}$ member of $1^{3}D_{1}$ ($1^{3}D_{1}(s\bar{s})$) meson nonet are obtained and the results are compared with the values from different theoretical models. Moreover, the strong decay properties of the kaon and $1^{3}D_{1}(s\bar{s})$ are presented in the $^{3}P_{0}$ model. On the basis of results, we suggest the assignment of $1^{3}D_{1}$ meson nonet need further tested in the new experiment in the future. Our results also provide mass constraints for the study of these states.

In this work, we investigate the mass spectrum of $1^{1}D_{2}$ and $1^{3}D_{2}$ meson nonets in the framework of the meson mass matrix and Regge phenomenology. The results are compared with the values from different phenomenological models and may be useful for the assignment of the $1^{1}D_{2}$ and $1^{3}D_{2}$ meson nonets in the future., Comment: 13 pages, 6 figures

Mach cones are expected to form in the expanding quark-gluon plasma (QGP) when energetic quarks and gluons (called jets) traverse the hot medium at a velocity faster than the speed of sound in high-energy heavy-ion collisions. The shape of the Mach cone and the associated diffusion wake are sensitive to the initial jet production location and the jet propagation direction relative to the radial flow because of the distortion by the collective expansion of the QGP and large density gradient. The shape of jet-induced Mach cones and their distortions in heavy-ion collisions provide a unique and direct probe of the dynamical evolution and the equation of state of QGP. However, it is difficult to identify the Mach cone and the diffusion wake in current experimental measurements of final hadron distributions because they are averaged over all possible initial jet production locations and propagation directions. To overcome this difficulty, we develop a deep learning assisted jet tomography which uses the full information of the final hadrons from jets to localize the initial jet production positions. This method can help to constrain the initial regions of jet production in heavy-ion collisions and enable a differential study of Mach-cones with different jet path length and orientation relative to the radial flow of the QGP in heavy-ion collisions.

Based on the relations from the meson-meson mass mixing matrix, the mixing angles of isoscalar state have been re-evaluated via mass relations and latest experimental results. The results in the present work are compared with the values from different theoretical models, meanwhile, the quarkonia content of isoscalar state are presented. In order to check the validity of analysis, some predictions on the decays of the isoscalar state are presented. These predictions may be useful for the phenomenological analysis for meson nonet in future experiments., Comment: 8 pages