Search

Visual instruction tuning has made considerable strides in enhancing the capabilities of Large Multimodal Models (LMMs). However, existing open LMMs largely focus on single-image tasks, their applications to multi-image scenarios remains less explored. Additionally, prior LMM research separately tackles different scenarios, leaving it impossible to generalize cross scenarios with new emerging capabilities. To this end, we introduce LLaVA-NeXT-Interleave, which simultaneously tackles Multi-image, Multi-frame (video), Multi-view (3D), and Multi-patch (single-image) scenarios in LMMs. To enable these capabilities, we regard the interleaved data format as a general template and compile the M4-Instruct dataset with 1,177.6k samples, spanning 4 primary domains with 14 tasks and 41 datasets. We also curate the LLaVA-Interleave Bench to comprehensively evaluate the multi-image performance of LMMs. Through extensive experiments, LLaVA-NeXT-Interleave achieves leading results in multi-image, video, and 3D benchmarks, while maintaining the performance of single-image tasks. Besides, our model also exhibits several emerging capabilities, e.g., transferring tasks across different settings and modalities. Code is available at https://github.com/LLaVA-VL/LLaVA-NeXT, Comment: Project Page: https://llava-vl.github.io/blog/2024-06-16-llava-next-interleave/

As the most critical component in the communication process, channels have a great impact on the communication quality of network. With the continuous expansion of network scale, the limited channel resources lead to the limitation of communication network scale. Therefore, achieving reasonable channel assignment and utilization becomes an extremely challenging problem. In order to solve this issue effectively, the channel assignment problem in communication networks can be transformed into a graph labeling problem, utilizing graphs to simulate the communication networks. In this paper, the topologies of mesh networks and stars are studied by constructing Cartesian product, and the lower bound and exact value of the optimal radio label of the Cartesian product of square mesh network and star $G=P(m,m)\Box K_{1,n}$ are obtained, where $m\geq 2$.

Zero-shot learning (ZSL) aims to explore the semantic-visual interactions to discover comprehensive knowledge transferred from seen categories to classify unseen categories. Recently, prompt engineering has emerged in ZSL, demonstrating impressive potential as it enables the zero-shot transfer of diverse visual concepts to downstream tasks. However, these methods are still not well generalized to broad unseen domains. A key reason is that the fixed adaption of learnable prompts on seen domains makes it tend to over-emphasize the primary visual features observed during training. In this work, we propose a \textbf{P}rompt-to-\textbf{P}rompt generation methodology (\textbf{P2P}), which addresses this issue by further embracing the instruction-following technique to distill instructive visual prompts for comprehensive transferable knowledge discovery. The core of P2P is to mine semantic-related instruction from prompt-conditioned visual features and text instruction on modal-sharing semantic concepts and then inversely rectify the visual representations with the guidance of the learned instruction prompts. This enforces the compensation for missing visual details to primary contexts and further eliminates the cross-modal disparity, endowing unseen domain generalization. Through extensive experimental results, we demonstrate the efficacy of P2P in achieving superior performance over state-of-the-art methods.

Although recent generative image compression methods have demonstrated impressive potential in optimizing the rate-distortion-perception trade-off, they still face the critical challenge of flexible rate adaption to diverse compression necessities and scenarios. To overcome this challenge, this paper proposes a Controllable Generative Image Compression framework, Control-GIC, the first capable of fine-grained bitrate adaption across a broad spectrum while ensuring high-fidelity and generality compression. We base Control-GIC on a VQGAN framework representing an image as a sequence of variable-length codes (i.e. VQ-indices), which can be losslessly compressed and exhibits a direct positive correlation with the bitrates. Therefore, drawing inspiration from the classical coding principle, we naturally correlate the information density of local image patches with their granular representations, to achieve dynamic adjustment of the code quantity following different granularity decisions. This implies we can flexibly determine a proper allocation of granularity for the patches to acquire desirable compression rates. We further develop a probabilistic conditional decoder that can trace back to historic encoded multi-granularity representations according to transmitted codes, and then reconstruct hierarchical granular features in the formalization of conditional probability, enabling more informative aggregation to improve reconstruction realism. Our experiments show that Control-GIC allows highly flexible and controllable bitrate adaption and even once compression on an entire dataset to fulfill constrained bitrate conditions. Experimental results demonstrate its superior performance over recent state-of-the-art methods.

This paper introduces Grounding DINO 1.5, a suite of advanced open-set object detection models developed by IDEA Research, which aims to advance the "Edge" of open-set object detection. The suite encompasses two models: Grounding DINO 1.5 Pro, a high-performance model designed for stronger generalization capability across a wide range of scenarios, and Grounding DINO 1.5 Edge, an efficient model optimized for faster speed demanded in many applications requiring edge deployment. The Grounding DINO 1.5 Pro model advances its predecessor by scaling up the model architecture, integrating an enhanced vision backbone, and expanding the training dataset to over 20 million images with grounding annotations, thereby achieving a richer semantic understanding. The Grounding DINO 1.5 Edge model, while designed for efficiency with reduced feature scales, maintains robust detection capabilities by being trained on the same comprehensive dataset. Empirical results demonstrate the effectiveness of Grounding DINO 1.5, with the Grounding DINO 1.5 Pro model attaining a 54.3 AP on the COCO detection benchmark and a 55.7 AP on the LVIS-minival zero-shot transfer benchmark, setting new records for open-set object detection. Furthermore, the Grounding DINO 1.5 Edge model, when optimized with TensorRT, achieves a speed of 75.2 FPS while attaining a zero-shot performance of 36.2 AP on the LVIS-minival benchmark, making it more suitable for edge computing scenarios. Model examples and demos with API will be released at https://github.com/IDEA-Research/Grounding-DINO-1.5-API, Comment: homepage: https://deepdataspace.com/home

Fragile topology, akin to twisted bilayer graphene and the exotic phases therein, is a notable topological class with intriguing properties. However, due to its unique nature and the lack of bulk-edge correspondence, the experimental signature of fragile topology has been under debated since its birth. Here, we demonstrate experimentally that fragile topological phases with filling anomaly can be probed via screw dislocations, despite that they do not support gapless edge states. Using a designer hexagonal phononic crystal with a fragile topological band gap, we find that 1D gapless bound modes can emerge at a screw dislocation due to the bulk fragile topology. We then establish a connection between our system and the twisted boundary condition via the gauge invariance principle and illustrate that such an emergent phenomenon is an intrinsic property of fragile topological phases with filling anomaly. We observe experimentally the 1D topological bound states using the pump-probe measurements of their dispersion and wavefunctions, which unveils a novel bulk-defect correspondence of fragile topology and a powerful tool for probing fragile topological phases and materials., Comment: Submitted to Science Bulletin

This paper provides a comprehensive review of the NTIRE 2024 challenge, focusing on efficient single-image super-resolution (ESR) solutions and their outcomes. The task of this challenge is to super-resolve an input image with a magnification factor of x4 based on pairs of low and corresponding high-resolution images. The primary objective is to develop networks that optimize various aspects such as runtime, parameters, and FLOPs, while still maintaining a peak signal-to-noise ratio (PSNR) of approximately 26.90 dB on the DIV2K_LSDIR_valid dataset and 26.99 dB on the DIV2K_LSDIR_test dataset. In addition, this challenge has 4 tracks including the main track (overall performance), sub-track 1 (runtime), sub-track 2 (FLOPs), and sub-track 3 (parameters). In the main track, all three metrics (ie runtime, FLOPs, and parameter count) were considered. The ranking of the main track is calculated based on a weighted sum-up of the scores of all other sub-tracks. In sub-track 1, the practical runtime performance of the submissions was evaluated, and the corresponding score was used to determine the ranking. In sub-track 2, the number of FLOPs was considered. The score calculated based on the corresponding FLOPs was used to determine the ranking. In sub-track 3, the number of parameters was considered. The score calculated based on the corresponding parameters was used to determine the ranking. RLFN is set as the baseline for efficiency measurement. The challenge had 262 registered participants, and 34 teams made valid submissions. They gauge the state-of-the-art in efficient single-image super-resolution. To facilitate the reproducibility of the challenge and enable other researchers to build upon these findings, the code and the pre-trained model of validated solutions are made publicly available at https://github.com/Amazingren/NTIRE2024_ESR/., Comment: The report paper of NTIRE2024 Efficient Super-resolution, accepted by CVPRW2024

The difference between the spin alignments of $K^\ast$ and those of $\phi$ at the low collision energies is a puzzle raised by the recent experiments. Unlike $\phi$ meson, $K^\ast$, carrying a unit strange charge, should react to strange chemical potential $\mu_S$. In this paper, we shall first convince you that $\mu_S$ is not small in a brayon-rich medium for keeping strange neutrality, and then derive the spin alignment induced by the gradient of $\mu_S$, and hence of baryon chemical potential $\mu_B$, using linear response theory, with the transport coefficients expressed, without any approximation, in terms of the $K^\ast$'s in-medium spectral properties by employing Ward-Takahashi identity. It turns out that such an effect applies mainly to the particles whose longitudinal and transverse modes diverge, and induces only the local spin alignment in a static medium. The magnitudes of these coefficients will be further estimated under the quasi-particle approximation., Comment: 10 pages, 3 figures

Low earth orbit (LEO) satellite systems play an important role in next generation communication networks due to their ability to provide extensive global coverage with guaranteed communications in remote areas and isolated areas where base stations cannot be cost-efficiently deployed. With the pervasive adoption of LEO satellite systems, especially in the LEO Internet-of-Things (IoT) scenarios, their spectrum resource management requirements have become more complex as a result of massive service requests and high bandwidth demand from terrestrial terminals. For instance, when leasing the spectrum to terrestrial users and controlling the uplink transmit power, satellites collect user data for machine learning purposes, which usually are sensitive information such as location, budget and quality of service (QoS) requirement. To facilitate model training in LEO IoT while preserving the privacy of data, blockchain-driven federated learning (FL) is widely used by leveraging on a fully decentralized architecture. In this paper, we propose a hybrid spectrum pricing and power control framework for LEO IoT by combining blockchain technology and FL. We first design a local deep reinforcement learning algorithm for LEO satellite systems to learn a revenue-maximizing pricing and power control scheme. Then the agents collaborate to form a FL system. We also propose a reputation-based blockchain which is used in the global model aggregation phase of FL. Based on the reputation mechanism, a node is selected for each global training round to perform model aggregation and block generation, which can further enhance the decentralization of the network and guarantee the trust. Simulation tests are conducted to evaluate the performances of the proposed scheme. Our results show the efficiency of finding the maximum revenue scheme for LEO satellite systems while preserving the privacy of each agent.

We present T-Rex2, a highly practical model for open-set object detection. Previous open-set object detection methods relying on text prompts effectively encapsulate the abstract concept of common objects, but struggle with rare or complex object representation due to data scarcity and descriptive limitations. Conversely, visual prompts excel in depicting novel objects through concrete visual examples, but fall short in conveying the abstract concept of objects as effectively as text prompts. Recognizing the complementary strengths and weaknesses of both text and visual prompts, we introduce T-Rex2 that synergizes both prompts within a single model through contrastive learning. T-Rex2 accepts inputs in diverse formats, including text prompts, visual prompts, and the combination of both, so that it can handle different scenarios by switching between the two prompt modalities. Comprehensive experiments demonstrate that T-Rex2 exhibits remarkable zero-shot object detection capabilities across a wide spectrum of scenarios. We show that text prompts and visual prompts can benefit from each other within the synergy, which is essential to cover massive and complicated real-world scenarios and pave the way towards generic object detection. Model API is now available at \url{https://github.com/IDEA-Research/T-Rex}., Comment: Technical Report

In this paper, we propose a simple and strong framework for Tracking Any Point with TRansformers (TAPTR). Based on the observation that point tracking bears a great resemblance to object detection and tracking, we borrow designs from DETR-like algorithms to address the task of TAP. In the proposed framework, in each video frame, each tracking point is represented as a point query, which consists of a positional part and a content part. As in DETR, each query (its position and content feature) is naturally updated layer by layer. Its visibility is predicted by its updated content feature. Queries belonging to the same tracking point can exchange information through self-attention along the temporal dimension. As all such operations are well-designed in DETR-like algorithms, the model is conceptually very simple. We also adopt some useful designs such as cost volume from optical flow models and develop simple designs to provide long temporal information while mitigating the feature drifting issue. Our framework demonstrates strong performance with state-of-the-art performance on various TAP datasets with faster inference speed.

Hierarchical massive quadruple systems are ideal laboratories for examining the theories of star formation, dynamical evolution, and stellar evolution. The successive mergers of hierarchical quadruple systems might explain the mass gap between neutron stars and black holes. Looking for light curves of O-type binaries identified by LAMOST, we find a (2+2) quadruple system: TYC 3340-2437-1, located in the stellar bow-shock nebula (SBN). It has a probability of over 99.99\% being a quadruple system derived from the surface density of the vicinity stars. Its inner orbital periods are 3.390602(89) days and 2.4378(16) days, respectively, and the total mass is about (11.47 + 5.79) + (5.2 + 2.02) = 24.48 $M_{\odot}$. The line-of-sight inclinations of the inner binaries, B$_1$ and B$_2$, are 55.94 and 78.2 degrees, respectively, indicating that they are not co-planar. Based on observations spanning 34 months and the significance of the astrometric excess noise ($D>2$) in Gaia DR3 data, we guess that its outer orbital period might be a few years. If it were true, the quadruple system might form through the disk fragmentation mechanism with outer eccentric greater than zero. This eccentricity could be the cause of both the arc-like feature of the SBN and the noncoplanarity of the inner orbit. The outer orbital period and outer eccentric could be determined with the release of future epoch astrometric data of Gaia.

Diffusion models (DM) have achieved remarkable promise in image super-resolution (SR). However, most of them are tailored to solving non-blind inverse problems with fixed known degradation settings, limiting their adaptability to real-world applications that involve complex unknown degradations. In this work, we propose BlindDiff, a DM-based blind SR method to tackle the blind degradation settings in SISR. BlindDiff seamlessly integrates the MAP-based optimization into DMs, which constructs a joint distribution of the low-resolution (LR) observation, high-resolution (HR) data, and degradation kernels for the data and kernel priors, and solves the blind SR problem by unfolding MAP approach along with the reverse process. Unlike most DMs, BlindDiff firstly presents a modulated conditional transformer (MCFormer) that is pre-trained with noise and kernel constraints, further serving as a posterior sampler to provide both priors simultaneously. Then, we plug a simple yet effective kernel-aware gradient term between adjacent sampling iterations that guides the diffusion model to learn degradation consistency knowledge. This also enables to joint refine the degradation model as well as HR images by observing the previous denoised sample. With the MAP-based reverse diffusion process, we show that BlindDiff advocates alternate optimization for blur kernel estimation and HR image restoration in a mutual reinforcing manner. Experiments on both synthetic and real-world datasets show that BlindDiff achieves the state-of-the-art performance with significant model complexity reduction compared to recent DM-based methods. Code will be available at \url{https://github.com/lifengcs/BlindDiff}

Due to the high magnetic coupling strength between the Cr elements, the bulk phase CrO2 is one of several ferromagnetic oxides known to have the highest Curie temperature. When the dimensionality of the material is reduced from 3D to 2D, the 2D CrO2 system material is expected to maintain a high Curie temperature. In this work, we predict two new phases of CrO2 monolayer (D3d and D2h) by using first-principles calculations. We have found that the Curie temperature of 2D CrO2 is much lower than that of its bulk phase, but still remains as high as 191K, which is comparable to that of Fe2Cr2Ge6. In addition, 1L D3d-CrO2 is in the ferromagnetic state, while 1L D2h-CrO2 is in the antiferromagnetic state. Also, the different geometric structure affects its electrical properties: the 1L D3d-CrO2 is a half-metal while 1L D2h-CrO2 is a semiconductor. Our studies have shown that there is a wealth of electrical and magnetic properties in CrO2., Comment: 5 pages,4 figures

The polarization control of micro/nano lasers is an important topic in nanophotonics. Up to now, the simultaneous generation of two distinguishable orthogonally-polarized lasing modes from a single organic microlaser remains a critical challenge. Here, we demonstrate simultaneously orthogonally-polarized dual lasing from a microcavity filled with an organic single crystal exhibiting selective strong coupling. We show that the non-Hermiticity due to polarization-dependent losses leads to the formation of real and imaginary Fermi arcs with exceptional points. Simultaneous orthogonally-polarized lasing becomes possible thanks to the eigenstate mixing by the photonic spin-orbit coupling at the imaginary Fermi arcs. Our work provides a novel way to develop linearly-polarized lasers and paves the way for the future fundamental research in topological photonics, non-Hermitian optics, and other fields., Comment: arXiv admin note: text overlap with arXiv:2110.13456

In this paper, we generalize the Poincar\'e-Lyapunov method for systems with linear type centers to study nilpotent centers in switching polynomial systems and use it to investigate the bi-center problem of planar $Z_2$-equivariant cubic switching systems associated with two symmetric nilpotent singular points. With a properly designed perturbation, 6 explicit bi-center conditions for such polynomial systems are derived. Then, based on the $6$ center conditions, by using Bogdanov-Takens bifurcation theory with general perturbations, we prove that there exist at least $20$ small-amplitude limit cycles around the nilpotent bi-center for a class of $Z_2$-equivariant cubic switching systems. This is a new lower bound of cyclicity for such cubic polynomial systems, increased from $12$ to $20$.

High-cadence, multiwavelength observations have continuously revealed the diversity of tidal disruption events (TDEs), thus greatly advancing our knowledge and understanding of TDEs. In this work, we conducted an intensive optical-UV and X-ray follow-up campaign of TDE AT2023lli, and found a remarkable month-long bump in its UV/optical light curve nearly two months prior to maximum brightness. The bump represents the longest separation time from the main peak among known TDEs to date. The main UV/optical outburst declines as $t^{-4.10}$, making it one of the fastest decaying optically selected TDEs. Furthermore, we detected sporadic X-ray emission 30 days after the UV/optical peak, accompanied by a reduction in the period of inactivity. It is proposed that the UV/optical bump could be caused by the self-intersection of the stream debris, whereas the primary peak is generated by the reprocessed emission of the accretion process. In addition, our results suggest that episodic X-ray radiation during the initial phase of decline may be due to the patched obscurer surrounding the accretion disk, a phenomenon associated with the inhomogeneous reprocessing process. The double TDE scenario, in which two stars are disrupted in sequence, is also a possible explanation for producing the observed early bump and main peak. We anticipate that the multicolor light curves of TDEs, especially in the very early stages, and the underlying physics can be better understood in the near future with the assistance of dedicated surveys such as the deep high-cadence survey of the 2.5-meter Wide Field Survey Telescope (WFST)., Comment: 14 pages, 8 figures,accepted for publication by ApJL

While Reinforcement Learning from Human Feedback (RLHF) significantly enhances the generation quality of Large Language Models (LLMs), recent studies have raised concerns regarding the complexity and instability associated with the Proximal Policy Optimization (PPO) algorithm, proposing a series of order-based calibration methods as viable alternatives. This paper delves further into current order-based methods, examining their inefficiencies in utilizing reward values and addressing misalignment issues. Building upon these findings, we propose a novel \textbf{V}alue-based \textbf{C}ali\textbf{B}ration (VCB) method to better align LLMs with human preferences. Experimental results demonstrate that VCB surpasses existing alignment methods on AI assistant and summarization datasets, providing impressive generalizability, robustness, and stability in diverse settings., Comment: 19 pages, Under review

We introduce Grounded SAM, which uses Grounding DINO as an open-set object detector to combine with the segment anything model (SAM). This integration enables the detection and segmentation of any regions based on arbitrary text inputs and opens a door to connecting various vision models. As shown in Fig.1, a wide range of vision tasks can be achieved by using the versatile Grounded SAM pipeline. For example, an automatic annotation pipeline based solely on input images can be realized by incorporating models such as BLIP and Recognize Anything. Additionally, incorporating Stable-Diffusion allows for controllable image editing, while the integration of OSX facilitates promptable 3D human motion analysis. Grounded SAM also shows superior performance on open-vocabulary benchmarks, achieving 48.7 mean AP on SegInW (Segmentation in the wild) zero-shot benchmark with the combination of Grounding DINO-Base and SAM-Huge models.

Spin triplet exciton formation sets limits on technologies using organic semiconductors that are confined to singlet-triplet photophysics. In contrast, excitations in the spin doublet manifold in organic radical semiconductors can show efficient luminescence. Here we explore the dynamics of the spin allowed process of intermolecular energy transfer from triplet to doublet excitons. We employ a carbene-metal-amide (CMA-CF3) as a model triplet donor host, since following photoexcitation it undergoes extremely fast intersystem crossing to set up a population of triplet excitons within 4 ps. This enables a foundational study for tracking energy transfer from triplets to a model radical semiconductor, TTM-3PCz. Over 90% of all radical luminescence originates from the triplet channel in this system under photoexcitation. We find that intermolecular triplet-to-doublet energy transfer can occur directly and rapidly, with 12% of triplet excitons transferring already on sub-ns timescales. This enhanced triplet harvesting mechanism is utilised in efficient near-infrared organic light-emitting diodes, which can be extended to other opto-electronic and -spintronic technologies by radical-based spin control in molecular semiconductors.

Foundation models possess strong capabilities in reasoning and memorizing across modalities. To further unleash the power of foundation models, we present FIND, a generalized interface for aligning foundation models' embeddings with unified image and dataset-level understanding spanning modality and granularity. As shown in the teaser figure, a lightweight transformer interface without tuning any foundation model weights is enough for segmentation, grounding, and retrieval in an interleaved manner. The proposed interface has the following favorable attributes: (1) Generalizable. It applies to various tasks spanning retrieval, segmentation, etc., under the same architecture and weights. (2) Interleavable. With the benefit of multi-task multi-modal training, the proposed interface creates an interleaved shared embedding space. (3) Extendable. The proposed interface is adaptive to new tasks, and new models. In light of the interleaved embedding space, we introduce FIND-Bench, which introduces new training and evaluation annotations to the COCO dataset for interleaved segmentation and retrieval. We are the first work aligning foundations models' embeddings for interleave understanding. Meanwhile, our approach achieves state-of-the-art performance on FIND-Bench and competitive performance on standard retrieval and segmentation settings., Comment: CODE: https://github.com/UX-Decoder/FIND

With the recent significant advancements in large multi-modal models (LMMs), the importance of their grounding capability in visual chat is increasingly recognized. Despite recent efforts to enable LMMs to support grounding, their capabilities for grounding and chat are usually separate, and their chat performance drops dramatically when asked to ground. The problem is the lack of a dataset for grounded visual chat (GVC). Existing grounding datasets only contain short captions. To address this issue, we have created GVC data that allows for the combination of grounding and chat capabilities. To better evaluate the GVC capabilities, we have introduced a benchmark called Grounding-Bench. Additionally, we have proposed a model design that can support GVC and various types of visual prompts by connecting segmentation models with language models. Experimental results demonstrate that our model outperforms other LMMs on Grounding-Bench. Furthermore, our model achieves competitive performance on classic grounding benchmarks like RefCOCO/+/g and Flickr30K Entities. Our code will be released at https://github.com/UX-Decoder/LLaVA-Grounding .

Malignant pleural mesothelioma (MPM) is the most common form of mesothelioma. To assess response to treatment, tumor measurements are acquired and evaluated based on a patient's longitudinal computed tomography (CT) scans. Tumor volume, however, is the more accurate metric for assessing tumor burden and response. Automated segmentation methods using deep learning can be employed to acquire volume, which otherwise is a tedious task performed manually. The deep learning-based tumor volume and contours can then be compared with a standard reference to assess the robustness of the automated segmentations. The purpose of this study was to evaluate the impact of probability map threshold on MPM tumor delineations generated using a convolutional neural network (CNN). Eighty-eight CT scans from 21 MPM patients were segmented by a VGG16/U-Net CNN. A radiologist modified the contours generated at a 0.5 probability threshold. Percent difference of tumor volume and overlap using the Dice Similarity Coefficient (DSC) were compared between the standard reference provided by the radiologist and CNN outputs for thresholds ranging from 0.001 to 0.9. CNN annotations consistently yielded smaller tumor volumes than radiologist contours. Reducing the probability threshold from 0.5 to 0.1 decreased the absolute percent volume difference, on average, from 43.96% to 24.18%. Median and mean DSC ranged from 0.58 to 0.60, with a peak at a threshold of 0.5; no distinct threshold was found for percent volume difference. No single output threshold in the CNN probability maps was optimal for both tumor volume and DSC. This work underscores the need to assess tumor volume and spatial overlap when evaluating CNN performance. While automated segmentations may yield comparable tumor volumes to that of the reference standard, the spatial region delineated by the CNN at a specific threshold is equally important., Comment: 10 pages, 7 figures, 2 tables

In-context prompting in large language models (LLMs) has become a prevalent approach to improve zero-shot capabilities, but this idea is less explored in the vision domain. Existing visual prompting methods focus on referring segmentation to segment the most relevant object, falling short of addressing many generic vision tasks like open-set segmentation and detection. In this paper, we introduce a universal visual in-context prompting framework for both tasks. In particular, we build on top of an encoder-decoder architecture, and develop a versatile prompt encoder to support a variety of prompts like strokes, boxes, and points. We further enhance it to take an arbitrary number of reference image segments as the context. Our extensive explorations show that the proposed visual in-context prompting elicits extraordinary referring and generic segmentation capabilities to refer and detect, yielding competitive performance to close-set in-domain datasets and showing promising results on many open-set segmentation datasets. By joint training on COCO and SA-1B, our model achieves $57.7$ PQ on COCO and $23.2$ PQ on ADE20K. Code will be available at https://github.com/UX-Decoder/DINOv., Comment: technical report

We introduce T-Rex, an interactive object counting model designed to first detect and then count any objects. We formulate object counting as an open-set object detection task with the integration of visual prompts. Users can specify the objects of interest by marking points or boxes on a reference image, and T-Rex then detects all objects with a similar pattern. Guided by the visual feedback from T-Rex, users can also interactively refine the counting results by prompting on missing or falsely-detected objects. T-Rex has achieved state-of-the-art performance on several class-agnostic counting benchmarks. To further exploit its potential, we established a new counting benchmark encompassing diverse scenarios and challenges. Both quantitative and qualitative results show that T-Rex possesses exceptional zero-shot counting capabilities. We also present various practical application scenarios for T-Rex, illustrating its potential in the realm of visual prompting., Comment: Technical report. Work in progress

Forecast combination integrates information from various sources by consolidating multiple forecast results from the target time series. Instead of the need to select a single optimal forecasting model, this paper introduces a deep learning ensemble forecasting model based on the Dirichlet process. Initially, the learning rate is sampled with three basis distributions as hyperparameters to convert the infinite mixture into a finite one. All checkpoints are collected to establish a deep learning sub-model pool, and weight adjustment and diversity strategies are developed during the combination process. The main advantage of this method is its ability to generate the required base learners through a single training process, utilizing the decaying strategy to tackle the challenge posed by the stochastic nature of gradient descent in determining the optimal learning rate. To ensure the method's generalizability and competitiveness, this paper conducts an empirical analysis using the weekly dataset from the M4 competition and explores sensitivity to the number of models to be combined. The results demonstrate that the ensemble model proposed offers substantial improvements in prediction accuracy and stability compared to a single benchmark model.

As the popularity of hierarchical point forecast reconciliation methods increases, there is a growing interest in probabilistic forecast reconciliation. Many studies have utilized machine learning or deep learning techniques to implement probabilistic forecasting reconciliation and have made notable progress. However, these methods treat the reconciliation step as a fixed and hard post-processing step, leading to a trade-off between accuracy and coherency. In this paper, we propose a new approach for probabilistic forecast reconciliation. Unlike existing approaches, our proposed approach fuses the prediction step and reconciliation step into a deep learning framework, making the reconciliation step more flexible and soft by introducing the Kullback-Leibler divergence regularization term into the loss function. The approach is evaluated using three hierarchical time series datasets, which shows the advantages of our approach over other probabilistic forecast reconciliation methods.

LLaVA-Plus is a general-purpose multimodal assistant that expands the capabilities of large multimodal models. It maintains a skill repository of pre-trained vision and vision-language models and can activate relevant tools based on users' inputs to fulfill real-world tasks. LLaVA-Plus is trained on multimodal instruction-following data to acquire the ability to use tools, covering visual understanding, generation, external knowledge retrieval, and compositions. Empirical results show that LLaVA-Plus outperforms LLaVA in existing capabilities and exhibits new ones. It is distinct in that the image query is directly grounded and actively engaged throughout the entire human-AI interaction sessions, significantly improving tool use performance and enabling new scenarios., Comment: 25 pages, 25M file size. Project Page: https://llava-vl.github.io/llava-plus/

The minimum network flow algorithm is widely used in multi-target tracking. However, the majority of the present methods concentrate exclusively on minimizing cost functions whose values may not indicate accurate solutions under occlusions. In this paper, by exploiting the properties of tracklets intersections and low-confidence detections, we develop a two-stage tracking pipeline with an intersection mask that can accurately locate inaccurate tracklets which are corrected in the second stage. Specifically, we employ the minimum network flow algorithm with high-confidence detections as input in the first stage to obtain the candidate tracklets that need correction. Then we leverage the intersection mask to accurately locate the inaccurate parts of candidate tracklets. The second stage utilizes low-confidence detections that may be attributed to occlusions for correcting inaccurate tracklets. This process constructs a graph of nodes in inaccurate tracklets and low-confidence nodes and uses it for the second round of minimum network flow calculation. We perform sufficient experiments on popular MOT benchmark datasets and achieve 78.4 MOTA on the test set of MOT16, 79.2 on MOT17, and 76.4 on MOT20, which shows that the proposed method is effective.