Search

Capturing high dynamic range (HDR) scenes is one of the most important issues in camera design. Majority of cameras use exposure fusion technique, which fuses images captured by different exposure levels, to increase dynamic range. However, this approach can only handle images with limited exposure difference, normally 3-4 stops. When applying to very high dynamic scenes where a large exposure difference is required, this approach often fails due to incorrect alignment or inconsistent lighting between inputs, or tone mapping artifacts. In this work, we propose UltraFusion, the first exposure fusion technique that can merge input with 9 stops differences. The key idea is that we model the exposure fusion as a guided inpainting problem, where the under-exposed image is used as a guidance to fill the missing information of over-exposed highlight in the over-exposed region. Using under-exposed image as a soft guidance, instead of a hard constrain, our model is robust to potential alignment issue or lighting variations. Moreover, utilizing the image prior of the generative model, our model also generates natural tone mapping, even for very high-dynamic range scene. Our approach outperforms HDR-Transformer on latest HDR benchmarks. Moreover, to test its performance in ultra high dynamic range scene, we capture a new real-world exposure fusion benchmark, UltraFusion Dataset, with exposure difference up to 9 stops, and experiments show that \model~can generate beautiful and high-quality fusion results under various scenarios. An online demo is provided at https://openimaginglab.github.io/UltraFusion/.

With the rapid advancement of Multi-modal Large Language Models (MLLMs), MLLM-based Image Quality Assessment (IQA) methods have shown promising performance in linguistic quality description. However, current methods still fall short in accurately scoring image quality. In this work, we aim to leverage MLLMs to regress accurate quality scores. A key challenge is that the quality score is inherently continuous, typically modeled as a Gaussian distribution, whereas MLLMs generate discrete token outputs. This mismatch necessitates score discretization. Previous approaches discretize the mean score into a one-hot label, resulting in information loss and failing to capture inter-image relationships. We propose a distribution-based approach that discretizes the score distribution into a soft label. This method preserves the characteristics of the score distribution, achieving high accuracy and maintaining inter-image relationships. Moreover, to address dataset variation, where different IQA datasets exhibit various distributions, we introduce a fidelity loss based on Thurstone's model. This loss captures intra-dataset relationships, facilitating co-training across multiple IQA datasets. With these designs, we develop the distribution-based Depicted image Quality Assessment model for Score regression (DeQA-Score). Experiments across multiple benchmarks show that DeQA-Score stably outperforms baselines in score regression. Also, DeQA-Score can predict the score distribution that closely aligns with human annotations. Codes and model weights have been released in https://depictqa.github.io/deqa-score/.

Modern 3D generation methods can rapidly create shapes from sparse or single views, but their outputs often lack geometric detail due to computational constraints. We present DetailGen3D, a generative approach specifically designed to enhance these generated 3D shapes. Our key insight is to model the coarse-to-fine transformation directly through data-dependent flows in latent space, avoiding the computational overhead of large-scale 3D generative models. We introduce a token matching strategy that ensures accurate spatial correspondence during refinement, enabling local detail synthesis while preserving global structure. By carefully designing our training data to match the characteristics of synthesized coarse shapes, our method can effectively enhance shapes produced by various 3D generation and reconstruction approaches, from single-view to sparse multi-view inputs. Extensive experiments demonstrate that DetailGen3D achieves high-fidelity geometric detail synthesis while maintaining efficiency in training., Comment: We need to refine it. The experiment and description is not satisfied

Reinforcement Learning (RL) empowers agents to acquire various skills by learning from reward signals. Unfortunately, designing high-quality instance-level rewards often demands significant effort. An emerging alternative, RL with delayed reward, focuses on learning from rewards presented periodically, which can be obtained from human evaluators assessing the agent's performance over sequences of behaviors. However, traditional methods in this domain assume the existence of underlying Markovian rewards and that the observed delayed reward is simply the sum of instance-level rewards, both of which often do not align well with real-world scenarios. In this paper, we introduce the problem of RL from Composite Delayed Reward (RLCoDe), which generalizes traditional RL from delayed rewards by eliminating the strong assumption. We suggest that the delayed reward may arise from a more complex structure reflecting the overall contribution of the sequence. To address this problem, we present a framework for modeling composite delayed rewards, using a weighted sum of non-Markovian components to capture the different contributions of individual steps. Building on this framework, we propose Composite Delayed Reward Transformer (CoDeTr), which incorporates a specialized in-sequence attention mechanism to effectively model these contributions. We conduct experiments on challenging locomotion tasks where the agent receives delayed rewards computed from composite functions of observable step rewards. The experimental results indicate that CoDeTr consistently outperforms baseline methods across evaluated metrics. Additionally, we demonstrate that it effectively identifies the most significant time steps within the sequence and accurately predicts rewards that closely reflect the environment feedback.

Toxicity classification in textual content remains a significant problem. Data with labels from a single annotator fall short of capturing the diversity of human perspectives. Therefore, there is a growing need to incorporate crowdsourced annotations for training an effective toxicity classifier. Additionally, the standard approach to training a classifier using empirical risk minimization (ERM) may fail to address the potential shifts between the training set and testing set due to exploiting spurious correlations. This work introduces a novel bi-level optimization framework that integrates crowdsourced annotations with the soft-labeling technique and optimizes the soft-label weights by Group Distributionally Robust Optimization (GroupDRO) to enhance the robustness against out-of-distribution (OOD) risk. We theoretically prove the convergence of our bi-level optimization algorithm. Experimental results demonstrate that our approach outperforms existing baseline methods in terms of both average and worst-group accuracy, confirming its effectiveness in leveraging crowdsourced annotations to achieve more effective and robust toxicity classification., Comment: 38th Conference on Neural Information Processing Systems (NeurIPS 2024)

Lensless cameras offer significant advantages in size, weight, and cost compared to traditional lens-based systems. Without a focusing lens, lensless cameras rely on computational algorithms to recover the scenes from multiplexed measurements. However, current algorithms struggle with inaccurate forward imaging models and insufficient priors to reconstruct high-quality images. To overcome these limitations, we introduce a novel two-stage approach for consistent and photorealistic lensless image reconstruction. The first stage of our approach ensures data consistency by focusing on accurately reconstructing the low-frequency content with a spatially varying deconvolution method that adjusts to changes in the Point Spread Function (PSF) across the camera's field of view. The second stage enhances photorealism by incorporating a generative prior from pre-trained diffusion models. By conditioning on the low-frequency content retrieved in the first stage, the diffusion model effectively reconstructs the high-frequency details that are typically lost in the lensless imaging process, while also maintaining image fidelity. Our method achieves a superior balance between data fidelity and visual quality compared to existing methods, as demonstrated with two popular lensless systems, PhlatCam and DiffuserCam. Project website: https://phocolens.github.io/., Comment: NeurIPS 2024 Spotlight

Lensless cameras, innovatively replacing traditional lenses for ultra-thin, flat optics, encode light directly onto sensors, producing images that are not immediately recognizable. This compact, lightweight, and cost-effective imaging solution offers inherent privacy advantages, making it attractive for privacy-sensitive applications like face verification. Typical lensless face verification adopts a two-stage process of reconstruction followed by verification, incurring privacy risks from reconstructed faces and high computational costs. This paper presents an end-to-end optimization approach for privacy-preserving face verification directly on encoded lensless captures, ensuring that the entire software pipeline remains encoded with no visible faces as intermediate results. To achieve this, we propose several techniques to address unique challenges from the lensless setup which precludes traditional face detection and alignment. Specifically, we propose a face center alignment scheme, an augmentation curriculum to build robustness against variations, and a knowledge distillation method to smooth optimization and enhance performance. Evaluations under both simulation and real environment demonstrate our method outperforms two-stage lensless verification while enhancing privacy and efficiency. Project website: \url{lenslessface.github.io}., Comment: under review

With the rapid advancement of Vision Language Models (VLMs), VLM-based Image Quality Assessment (IQA) seeks to describe image quality linguistically to align with human expression and capture the multifaceted nature of IQA tasks. However, current methods are still far from practical usage. First, prior works focus narrowly on specific sub-tasks or settings, which do not align with diverse real-world applications. Second, their performance is sub-optimal due to limitations in dataset coverage, scale, and quality. To overcome these challenges, we introduce Depicted image Quality Assessment in the Wild (DepictQA-Wild). Our method includes a multi-functional IQA task paradigm that encompasses both assessment and comparison tasks, brief and detailed responses, full-reference and non-reference scenarios. We introduce a ground-truth-informed dataset construction approach to enhance data quality, and scale up the dataset to 495K under the brief-detail joint framework. Consequently, we construct a comprehensive, large-scale, and high-quality dataset, named DQ-495K. We also retain image resolution during training to better handle resolution-related quality issues, and estimate a confidence score that is helpful to filter out low-quality responses. Experimental results demonstrate that DepictQA-Wild significantly outperforms traditional score-based methods, prior VLM-based IQA models, and proprietary GPT-4V in distortion identification, instant rating, and reasoning tasks. Our advantages are further confirmed by real-world applications including assessing the web-downloaded images and ranking model-processed images. Datasets and codes will be released in https://depictqa.github.io/depictqa-wild/.

The model inputs play a key role in the performance of the Bayesian optimization approach. In this paper, we investigate the influence of the inputs on the improved predictions of phenomenological nuclear charge radius formulas using an approach combining those original formulas and the Bayesian neural network (BNN). We find that there is no improvement in predictions after the abnormal odd-even staggering effect of 181,183,185Hg is injected into the BNN, while the original phenomenological formulas themselves possess rich physical information or rigid constraints. It indicates the abundance and intensity of physical inputs affect the performance of the Bayesian optimization approach as well as the robustness of the BNN. We further demonstrate that, by ensuring that the number of neurons in the hidden layer is larger than the number of NN inputs, adding hidden layers into the BNN can significantly improve the predictions of nuclear charge radii formulas within the Bayesian optimization approach., Comment: 10 pages,4 figures, 4 talbes

In this paper, we investigate an offline reinforcement learning (RL) problem where datasets are collected from two domains. In this scenario, having datasets with domain labels facilitates efficient policy training. However, in practice, the task of assigning domain labels can be resource-intensive or infeasible at a large scale, leading to a prevalence of domain-unlabeled data. To formalize this challenge, we introduce a novel offline RL problem setting named Positive-Unlabeled Offline RL (PUORL), which incorporates domain-unlabeled data. To address PUORL, we develop an offline RL algorithm utilizing positive-unlabeled learning to predict the domain labels of domain-unlabeled data, enabling the integration of this data into policy training. Our experiments show the effectiveness of our method in accurately identifying domains and learning policies that outperform baselines in the PUORL setting, highlighting its capability to leverage domain-unlabeled data effectively.

Action recognition, an essential component of computer vision, plays a pivotal role in multiple applications. Despite significant improvements brought by Convolutional Neural Networks (CNNs), these models suffer performance declines when trained with discontinuous video frames, which is a frequent scenario in real-world settings. This decline primarily results from the loss of temporal continuity, which is crucial for understanding the semantics of human actions. To overcome this issue, we introduce the 4A (Action Animation-based Augmentation Approach) pipeline, which employs a series of sophisticated techniques: starting with 2D human pose estimation from RGB videos, followed by Quaternion-based Graph Convolution Network for joint orientation and trajectory prediction, and Dynamic Skeletal Interpolation for creating smoother, diversified actions using game engine technology. This innovative approach generates realistic animations in varied game environments, viewed from multiple viewpoints. In this way, our method effectively bridges the domain gap between virtual and real-world data. In experimental evaluations, the 4A pipeline achieves comparable or even superior performance to traditional training approaches using real-world data, while requiring only 10% of the original data volume. Additionally, our approach demonstrates enhanced performance on In-the-wild videos, marking a significant advancement in the field of action recognition., Comment: Accepted by ECAI24, model and code are available at https://github.com/xingyu-song/4A

In Reinforcement Learning (RL), it is commonly assumed that an immediate reward signal is generated for each action taken by the agent, helping the agent maximize cumulative rewards to obtain the optimal policy. However, in many real-world scenarios, designing immediate reward signals is difficult; instead, agents receive a single reward that is contingent upon a partial sequence or a complete trajectory. In this work, we define this challenging problem as RL from Bagged Reward (RLBR), where sequences of data are treated as bags with non-Markovian bagged rewards, leading to the formulation of Bagged Reward Markov Decision Processes (BRMDPs). Theoretically, we demonstrate that RLBR can be addressed by solving a standard MDP with properly redistributed bagged rewards allocated to each instance within a bag. Empirically, we find that reward redistribution becomes more challenging as the bag length increases, due to reduced informational granularity. Existing reward redistribution methods are insufficient to address these challenges. Therefore, we propose a novel reward redistribution method equipped with a bidirectional attention mechanism, enabling the accurate interpretation of contextual nuances and temporal dependencies within each bag. We experimentally demonstrate that our proposed method consistently outperforms existing approaches.

Stochastic Differential Equations (SDEs) serve as a powerful modeling tool in various scientific domains, including systems science, engineering, and ecological science. While the specific form of SDEs is typically known for a given problem, certain model parameters remain unknown. Efficiently inferring these unknown parameters based on observations of the state in discrete time series represents a vital practical subject. The challenge arises in nonlinear SDEs, where maximum likelihood estimation of parameters is generally unfeasible due to the absence of closed-form expressions for transition and stationary probability density functions of the states. In response to this limitation, we propose a novel two-step parameter inference mechanism. This approach involves a global-search phase followed by a local-refining procedure. The global-search phase is dedicated to identifying the domain of high-value likelihood functions, while the local-refining procedure is specifically designed to enhance the surrogate likelihood within this localized domain. Additionally, we present two simulation-based approximations for the transition density, aiming to efficiently or accurately approximate the likelihood function. Numerical examples illustrate the efficacy of our proposed methodology in achieving posterior parameter estimation.

This paper studies a distributed continuous-time aggregative optimization problem, which is a fundamental problem in the price-based energy management. The objective of the distributed aggregative optimization is to minimize the sum of local objective functions, which have a specific expression that relies on agents' own decisions and the aggregation of all agents' decisions. To solve the problem, a novel distributed continuous-time algorithm is proposed by combining gradient dynamics with a dynamic average consensus estimator in a two-time scale. The exponential convergence of the proposed algorithm is established under the assumption of a convex global cost function by virtue of the stability theory of singular perturbation systems. Motivated by practical applications, the implementation of the continuous-time algorithm with event-triggered communication is investigated. Simulations on the price-based energy management of distributed energy resources are given to illustrate the proposed method., Comment: 7 pages,7 figures

Learning multi-lingual sentence embeddings is a fundamental task in natural language processing. Recent trends in learning both mono-lingual and multi-lingual sentence embeddings are mainly based on contrastive learning (CL) among an anchor, one positive, and multiple negative instances. In this work, we argue that leveraging multiple positives should be considered for multi-lingual sentence embeddings because (1) positives in a diverse set of languages can benefit cross-lingual learning, and (2) transitive similarity across multiple positives can provide reliable structural information for learning. In order to investigate the impact of multiple positives in CL, we propose a novel approach, named MPCL, to effectively utilize multiple positive instances to improve the learning of multi-lingual sentence embeddings. Experimental results on various backbone models and downstream tasks demonstrate that MPCL leads to better retrieval, semantic similarity, and classification performances compared to conventional CL. We also observe that in unseen languages, sentence embedding models trained on multiple positives show better cross-lingual transfer performance than models trained on a single positive instance., Comment: Accepted to EACL 2024, main conference

Cloud-native databases have become the de-facto choice for mission-critical applications on the cloud due to the need for high availability, resource elasticity, and cost efficiency. Meanwhile, driven by the increasing connectivity between data generation and analysis, users prefer a single database to efficiently process both OLTP and OLAP workloads, which enhances data freshness and reduces the complexity of data synchronization and the overall business cost. In this paper, we summarize five crucial design goals for a cloud-native HTAP database based on our experience and customers' feedback, i.e., transparency, competitive OLAP performance, minimal perturbation on OLTP workloads, high data freshness, and excellent resource elasticity. As our solution to realize these goals, we present PolarDB-IMCI, a cloud-native HTAP database system designed and deployed at Alibaba Cloud. Our evaluation results show that PolarDB-IMCI is able to handle HTAP efficiently on both experimental and production workloads; notably, it speeds up analytical queries up to $\times149$ on TPC-H (100 $GB$). PolarDB-IMCI introduces low visibility delay and little performance perturbation on OLTP workloads (< 5%), and resource elasticity can be achieved by scaling out in tens of seconds., Comment: 14 pages, 16 figures, to be published in ACM SIGMOD 2023

Satellite Image Time Series (SITS) representation learning is complex due to high spatiotemporal resolutions, irregular acquisition times, and intricate spatiotemporal interactions. These challenges result in specialized neural network architectures tailored for SITS analysis. The field has witnessed promising results achieved by pioneering researchers, but transferring the latest advances or established paradigms from Computer Vision (CV) to SITS is still highly challenging due to the existing suboptimal representation learning framework. In this paper, we develop a novel perspective of SITS processing as a direct set prediction problem, inspired by the recent trend in adopting query-based transformer decoders to streamline the object detection or image segmentation pipeline. We further propose to decompose the representation learning process of SITS into three explicit steps: collect-update-distribute, which is computationally efficient and suits for irregularly-sampled and asynchronous temporal satellite observations. Facilitated by the unique reformulation, our proposed temporal learning backbone of SITS, initially pre-trained on the resource efficient pixel-set format and then fine-tuned on the downstream dense prediction tasks, has attained new state-of-the-art (SOTA) results on the PASTIS benchmark dataset. Specifically, the clear separation between temporal and spatial components in the semantic/panoptic segmentation pipeline of SITS makes us leverage the latest advances in CV, such as the universal image segmentation architecture, resulting in a noticeable 2.5 points increase in mIoU and 8.8 points increase in PQ, respectively, compared to the best scores reported so far.

Many issues concerning the origin of high-temperature superconductivity (HTS) are still under debate. For example, how the magnetic ordering varies with doping and its relationship with the superconducting temperature; and why the maximal Tc always occurs near the quantum critical point. In this paper, taking hole-doped La2CuO4 as a classical example, we employ the first-principles band structure and total energy calculations and Monte Carlo simulations to explore how the symmetry-breaking magnetic ground state evolves with hole doping and the origin of a dome-shaped superconductivity region in the phase diagram. We demonstrate that the local antiferromagnetic ordering and doping play key roles in determining the electron-phonon coupling, thus Tc. Initially, the La2CuO4 possesses a checkerboard local antiferromagnetic ground state. As the hole doping increases, Tc increases with the increase of the density of states at the Fermi surface. But as the doping increases further, the strength of the antiferromagnetic interaction weakens. At the critical doping level, a magnetic phase transition occurs that reduces the local antiferromagnetism-assisted electron-phonon coupling, thus diminishing the Tc. The superconductivity disappears in the heavily overdoped region when the antiferromagnetic ordering disappears. These observations could account for why cuprates have a dome-shaped superconductivity region in the phase diagram. Our study, thus, contributes to a fundamental understanding of the correlation between doping, local magnetic ordering, and superconductivity of HTS., Comment: 14 pages, 3 figures in the main text; 11 pages, 7 figures in the supplementary materials

A resilient distributed algorithm is proposed to solve the distributed resource allocation problem of a first-order nonlinear multi-agent system who is subject to false data injection (FDI) attacks. An intelligent attacker injects false data into agents' actuators and sensors such that agents execute the algorithm according to the compromised control inputs and interactive information. The goal of the attacker is to make the multi-agent system to be unstable and to cause the deviance of agents' decisions from the optimal resource allocation. At first, we analyze the robustness of a distributed resource allocation algorithm under FDI attacks. Then, the unknown nonlinear term and the false data injected in agents are considered as extended states which can be estimated by extended state observers. The estimation was used in the feedback control to suppress the effect of the FDI attacks. A resilient distributed resource allocation algorithm based on the extended state observer is proposed to ensure that it can converge to the optimal allocation without requiring any information about the nature of the attacker. An example is given to illustrate the results.