AAAI 2024 论文列表

Point-based interactive image segmentation can ease the burden of mask annotation in applications such as semantic segmentation and image editing. However, fully extracting the target mask with limited user inputs remains challenging. We introduce a novel method, Variance-Insensitive and Target-Preserving Mask Refinement to enhance segmentation quality with fewer user inputs. Regarding the last segmentation result as the initial mask, an iterative refinement process is commonly employed to continually enhance the initial mask. Nevertheless, conventional techniques suffer from sensitivity to the variance in the initial mask. To circumvent this problem, our proposed method incorporates a mask matching algorithm for ensuring consistent inferences from different types of initial masks. We also introduce a target-aware zooming algorithm to preserve object information during downsampling, balancing efficiency and accuracy. Experiments on GrabCut, Berkeley, SBD, and DAVIS datasets demonstrate our method's state-of-the-art performance in interactive image segmentation.

Explaining the decision process of machine learning algorithms is nowadays crucial for both model’s performance enhancement and human comprehension. This can be achieved by assessing the variable importance of single variables, even for high-capacity non-linear methods, e.g. Deep Neural Networks (DNNs). While only removal-based approaches, such as Permutation Importance (PI), can bring statistical validity, they return misleading results when variables are correlated. Conditional Permutation Importance (CPI) bypasses PI’s limitations in such cases. However, in high-dimensional settings, where high correlations between the variables cancel their conditional importance, the use of CPI as well as other methods leads to unreliable results, besides prohibitive computation costs. Grouping variables statistically via clustering or some prior knowledge gains some power back and leads to better interpretations. In this work, we introduce BCPI (Block-Based Conditional Permutation Importance), a new generic framework for variable importance computation with statistical guarantees handling both single and group cases. Furthermore, as handling groups with high cardinality (such as a set of observations of a given modality) are both time-consuming and resource-intensive, we also introduce a new stacking approach extending the DNN architecture with sub-linear layers adapted to the group structure. We show that the ensuing approach extended with stacking controls the type-I error even with highly-correlated groups and shows top accuracy across benchmarks. Furthermore, we perform a real-world data analysis in a large-scale medical dataset where we aim to show the consistency between our results and the literature for a biomarker prediction.

Deep learning methods on graph data have achieved remarkable efficacy across a variety of real-world applications, such as social network analysis and transaction risk detection. Nevertheless, recent studies have illuminated a concerning fact: even the most expressive Graph Neural Networks (GNNs) are vulnerable to graph adversarial attacks. While several methods have been proposed to enhance the robustness of GNN models against adversarial attacks, few have focused on a simple yet realistic approach: valuing the adversarial risks and focused safeguards at the node level. This empowers defenders to allocate heightened security level to vulnerable nodes, while lower to robust nodes. With this new perspective, we propose a novel graph defense strategy RisKeeper, such that the adversarial risk can be directly kept in the input graph. We start at valuing the adversarial risk, by introducing a cost-aware projected gradient descent attack that takes into account both cost avoidance and compliance with costs budgets. Subsequently, we present a learnable approach to ascertain the ideal security level for each individual node by solving a bi-level optimization problem. Through extensive experiments on four real-world datasets, we demonstrate that our method achieves superior performance surpassing state-of-the-art methods. Our in-depth case studies provide further insights into vulnerable and robust structural patterns, serving as inspiration for practitioners to exercise heightened vigilance.

Human values are crucial to human decision-making. Value pluralism is the view that multiple correct values may be held in tension with one another (e.g., when considering lying to a friend to protect their feelings, how does one balance honesty with friendship?). As statistical learners, AI systems fit to averages by default, washing out these potentially irreducible value conflicts. To improve AI systems to better reflect value pluralism, the first-order challenge is to explore the extent to which AI systems can model pluralistic human values, rights, and duties as well as their interaction. We introduce ValuePrism, a large-scale dataset of 218k values, rights, and duties connected to 31k human-written situations. ValuePrism’s contextualized values are generated by GPT-4 and deemed high-quality by human annotators 91% of the time. We conduct a large-scale study with annotators across diverse social and demographic backgrounds to try to understand whose values are represented. With ValuePrism, we build Value Kaleidoscope (or Kaleido), an open, light-weight, and structured language-based multi-task model that generates, explains, and assesses the relevance and valence (i.e., support or oppose) of human values, rights, and duties within a specific context. Humans prefer the sets of values output by our system over the teacher GPT- 4, finding them more accurate and with broader coverage. In addition, we demonstrate that Kaleido can help explain variability in human decision-making by outputting contrasting values. Finally, we show that Kaleido’s representations transfer to other philosophical frameworks and datasets, confirming the benefit of an explicit, modular, and interpretable approach to value pluralism. We hope that our work will serve as a step to making more explicit the implicit values behind human decision-making and to steering AI systems to make decisions that are more in accordance with them.

This work undertakes studies to evaluate Interpretability Methods for Time Series Deep Learning. Sensitivity analysis assesses how input changes affect the output, constituting a key component of interpretation. Among the post-hoc interpretation methods such as back-propagation, perturbation, and approximation, my work will investigate perturbation-based sensitivity Analysis methods on modern Transformer models to benchmark their performances. Specifically, my work intends to answer three research questions: 1) Do different sensitivity analysis methods yield comparable outputs and attribute importance rankings? 2) Using the same sensitivity analysis method, do different Deep Learning models impact the output of the sensitivity analysis? 3) How well do the results from sensitivity analysis methods align with the ground truth?

The recent contrastive language-image pre-training (CLIP) model has shown great success in a wide range of image-level tasks, revealing remarkable ability for learning powerful visual representations with rich semantics. An open and worthwhile problem is efficiently adapting such a strong model to the video domain and designing a robust video anomaly detector. In this work, we propose VadCLIP, a new paradigm for weakly supervised video anomaly detection (WSVAD) by leveraging the frozen CLIP model directly without any pre-training and fine-tuning process. Unlike current works that directly feed extracted features into the weakly supervised classifier for frame-level binary classification, VadCLIP makes full use of fine-grained associations between vision and language on the strength of CLIP and involves dual branch. One branch simply utilizes visual features for coarse-grained binary classification, while the other fully leverages the fine-grained language-image alignment. With the benefit of dual branch, VadCLIP achieves both coarse-grained and fine-grained video anomaly detection by transferring pre-trained knowledge from CLIP to WSVAD task. We conduct extensive experiments on two commonly-used benchmarks, demonstrating that VadCLIP achieves the best performance on both coarse-grained and fine-grained WSVAD, surpassing the state-of-the-art methods by a large margin. Specifically, VadCLIP achieves 84.51% AP and 88.02% AUC on XD-Violence and UCF-Crime, respectively. Code and features are released at https://github.com/nwpu-zxr/VadCLIP.

In content-based video retrieval (CBVR), dealing with large-scale collections, efficiency is as important as accuracy; thus, several video-level feature-based studies have actively been conducted. Nevertheless, owing to the severe difficulty of embedding a lengthy and untrimmed video into a single feature, these studies have been insufficient for accurate retrieval compared to frame-level feature-based studies. In this paper, we show that appropriate suppression of irrelevant frames can provide insight into the current obstacles of the video-level approaches. Furthermore, we propose a Video-to-Video Suppression network (VVS) as a solution. VVS is an end-to-end framework that consists of an easy distractor elimination stage to identify which frames to remove and a suppression weight generation stage to determine the extent to suppress the remaining frames. This structure is intended to effectively describe an untrimmed video with varying content and meaningless information. Its efficacy is proved via extensive experiments, and we show that our approach is not only state-of-the-art in video-level approaches but also has a fast inference time despite possessing retrieval capabilities close to those of frame-level approaches. Code is available at https://github.com/sejong-rcv/VVS

Correspondence pruning aims to find correct matches (inliers) from an initial set of putative correspondences, which is a fundamental task for many applications. The process of finding is challenging, given the varying inlier ratios between scenes/image pairs due to significant visual differences. However, the performance of the existing methods is usually limited by the problem of lacking visual cues (e.g., texture, illumination, structure) of scenes. In this paper, we propose a Visual-Spatial Fusion Transformer (VSFormer) to identify inliers and recover camera poses accurately. Firstly, we obtain highly abstract visual cues of a scene with the cross attention between local features of two-view images. Then, we model these visual cues and correspondences by a joint visual-spatial fusion module, simultaneously embedding visual cues into correspondences for pruning. Additionally, to mine the consistency of correspondences, we also design a novel module that combines the KNN-based graph and the transformer, effectively capturing both local and global contexts. Extensive experiments have demonstrated that the proposed VSFormer outperforms state-of-the-art methods on outdoor and indoor benchmarks. Our code is provided at the following repository: https://github.com/sugar-fly/VSFormer.

Night photography often struggles with challenges like low light and blurring, stemming from dark environments and prolonged exposures. Current methods either disregard priors and directly fitting end-to-end networks, leading to inconsistent illumination, or rely on unreliable handcrafted priors to constrain the network, thereby bringing the greater error to the final result. We believe in the strength of data-driven high-quality priors and strive to offer a reliable and consistent prior, circumventing the restrictions of manual priors. In this paper, we propose Clearer Night Image Restoration with Vector-Quantized Codebook (VQCNIR) to achieve remarkable and consistent restoration outcomes on real-world and synthetic benchmarks. To ensure the faithful restoration of details and illumination, we propose the incorporation of two essential modules: the Adaptive Illumination Enhancement Module (AIEM) and the Deformable Bi-directional Cross-Attention (DBCA) module. The AIEM leverages the inter-channel correlation of features to dynamically maintain illumination consistency between degraded features and high-quality codebook features. Meanwhile, the DBCA module effectively integrates texture and structural information through bi-directional cross-attention and deformable convolution, resulting in enhanced fine-grained detail and structural fidelity across parallel decoders. Extensive experiments validate the remarkable benefits of VQCNIR in enhancing image quality under low-light conditions, showcasing its state-of-the-art performance on both synthetic and real-world datasets. The code is available at https://github.com/AlexZou14/VQCNIR.

Visual Question Answering (VQA) is a fundamental task in computer vision and natural language process fields. Although the “pre-training & finetuning” learning paradigm significantly improves the VQA performance, the adversarial robustness of such a learning paradigm has not been explored. In this paper, we delve into a new problem: using a pre-trained multimodal source model to create adversarial image-text pairs and then transferring them to attack the target VQA models. Correspondingly, we propose a novel VQATTACK model, which can iteratively generate both im- age and text perturbations with the designed modules: the large language model (LLM)-enhanced image attack and the cross-modal joint attack module. At each iteration, the LLM-enhanced image attack module first optimizes the latent representation-based loss to generate feature-level image perturbations. Then it incorporates an LLM to further enhance the image perturbations by optimizing the designed masked answer anti-recovery loss. The cross-modal joint attack module will be triggered at a specific iteration, which updates the image and text perturbations sequentially. Notably, the text perturbation updates are based on both the learned gradients in the word embedding space and word synonym-based substitution. Experimental results on two VQA datasets with five validated models demonstrate the effectiveness of the proposed VQATTACK in the transferable attack setting, compared with state-of-the-art baselines. This work reveals a significant blind spot in the “pre-training & fine-tuning” paradigm on VQA tasks. The source code can be found in the link https://github.com/ericyinyzy/VQAttack.

Few-shot font generation is challenging, as it needs to capture the fine-grained stroke styles from a limited set of reference glyphs, and then transfer to other characters, which are expected to have similar styles. However, due to the diversity and complexity of Chinese font styles, the synthesized glyphs of existing methods usually exhibit visible artifacts, such as missing details and distorted strokes. In this paper, we propose a VQGAN-based framework (i.e., VQ-Font) to enhance glyph fidelity through token prior refinement and structure-aware enhancement. Specifically, we pre-train a VQGAN to encapsulate font token prior within a code-book. Subsequently, VQ-Font refines the synthesized glyphs with the codebook to eliminate the domain gap between synthesized and real-world strokes. Furthermore, our VQ-Font leverages the inherent design of Chinese characters, where structure components such as radicals and character components are combined in specific arrangements, to recalibrate fine-grained styles based on references. This process improves the matching and fusion of styles at the structure level. Both modules collaborate to enhance the fidelity of the generated fonts. Experiments on a collected font dataset show that our VQ-Font outperforms the competing methods both quantitatively and qualitatively, especially in generating challenging styles. Our code is available at https://github.com/Yaomingshuai/VQ-Font.

In the field of vanishing point detection, previous works commonly relied on extracting and clustering straight lines or classifying candidate points as vanishing points. This paper proposes a novel end-to-end framework, called VPDETR (Vanishing Point DEtection TRansformer), that views vanishing point detection as a set prediction problem, applicable to both Manhattan and non-Manhattan world datasets. By using the positional embedding of anchor points as queries in Transformer decoders and dynamically updating them layer by layer, our method is able to directly input images and output their vanishing points without the need for explicit straight line extraction and candidate points sampling. Additionally, we introduce an orthogonal loss and a cross-prediction loss to improve accuracy on the Manhattan world datasets. Experimental results demonstrate that VPDETR achieves competitive performance compared to state-of-the-art methods, without requiring post-processing.

Large-scale pre-trained models have achieved remarkable success in various computer vision tasks. A standard approach to leverage these models is to fine-tune all model parameters for downstream tasks, which poses challenges in terms of computational and storage costs. Recently, inspired by Natural Language Processing (NLP), parameter-efficient transfer learning has been successfully applied to vision tasks. However, most existing techniques primarily focus on single-task adaptation, and despite limited research on multi-task adaptation, these methods often exhibit suboptimal training/inference efficiency. In this paper, we first propose an once-for-all Vision Multi-Task Adapter (VMT-Adapter), which strikes approximately O(1) training and inference efficiency w.r.t task number. Concretely, VMT-Adapter shares the knowledge from multiple tasks to enhance cross-task interaction while preserves task-specific knowledge via independent knowledge extraction modules. Notably, since task-specific modules require few parameters, VMT-Adapter can handle an arbitrary number of tasks with a negligible increase of trainable parameters. We also propose VMT-Adapter-Lite, which further reduces the trainable parameters by learning shared parameters between down- and up-projections. Extensive experiments on four dense scene understanding tasks demonstrate the superiority of VMT-Adapter(-Lite), achieving a 3.96% (1.34%) relative improvement compared to single-task full fine-tuning, while utilizing merely ～1% (0.36%) trainable parameters of the pre-trained model.

Outdoor Vision-and-Language Navigation (VLN) requires an agent to navigate through realistic 3D outdoor environments based on natural language instructions. The performance of existing VLN methods is limited by insufficient diversity in navigation environments and limited training data. To address these issues, we propose VLN-Video, which utilizes the diverse outdoor environments present in driving videos in multiple cities in the U.S. augmented with automatically generated navigation instructions and actions to improve outdoor VLN performance. VLN-Video combines the best of intuitive classical approaches and modern deep learning techniques, using template infilling to generate grounded non-repetitive navigation instructions, combined with an image rotation similarity based navigation action predictor to obtain VLN style data from driving videos for pretraining deep learning VLN models. We pre-train the model on the Touchdown dataset and our video-augmented dataset created from driving videos with three proxy tasks: Masked Language Modeling, Instruction and Trajectory Matching, and Next Action Prediction, so as to learn temporally-aware and visually-aligned instruction representations. The learned instruction representation is adapted to the state-of-the-art navigation agent when fine-tuning on the Touchdown dataset. Empirical results demonstrate that VLN-Video significantly outperforms previous state-of-the-art models by 2.1% in task completion rate, achieving a new state-of-the-art on the Touchdown dataset.

Vision and language foundation models (VLMs) have showcased impressive capabilities in 2D scene understanding. However, their latent potential in elevating the understanding of 3D autonomous driving scenes remains untapped. In this paper, we propose VLM2Scene, which exploits the potential of VLMs to enhance 3D self-supervised representation learning through our proposed image-text-LiDAR contrastive learning strategy. Specifically, in the realm of autonomous driving scenes, the inherent sparsity of LiDAR point clouds poses a notable challenge for point-level contrastive learning methods. This method often grapples with limitations tied to a restricted receptive field and the presence of noisy points. To tackle this challenge, our approach emphasizes region-level learning, leveraging regional masks without semantics derived from the vision foundation model. This approach capitalizes on valuable contextual information to enhance the learning of point cloud representations. First, we introduce Region Caption Prompts to generate fine-grained language descriptions for the corresponding regions, utilizing the language foundation model. These region prompts then facilitate the establishment of positive and negative text-point pairs within the contrastive loss framework. Second, we propose a Region Semantic Concordance Regularization, which involves a semantic-filtered region learning and a region semantic assignment strategy. The former aims to filter the false negative samples based on the semantic distance, and the latter mitigates potential inaccuracies in pixel semantics, thereby enhancing overall semantic consistency. Extensive experiments on representative autonomous driving datasets demonstrate that our self-supervised method significantly outperforms other counterparts. Codes are available at https://github.com/gbliao/VLM2Scene.

Zero-Shot Object Counting~(ZSOC) aims to count referred instances of arbitrary classes in a query image without human-annotated exemplars. To deal with ZSOC, preceding studies proposed a two-stage pipeline: discovering exemplars and counting. However, there remains a challenge of vulnerability to error propagation of the sequentially designed two-stage process. In this work, we propose an one-stage baseline, Visual-Language Baseline (VLBase), exploring the implicit association of the semantic-patch embeddings of CLIP. Subsequently, we extend the VLBase to Visual-language Counter (VLCounter) by incorporating three modules devised to tailor VLBase for object counting. First, we introduce Semantic-conditioned Prompt Tuning (SPT) within the image encoder to acquire target-highlighted representations. Second, Learnable Affine Transformation (LAT) is employed to translate the semantic-patch similarity map to be appropriate for the counting task. Lastly, we transfer the layer-wisely encoded features to the decoder through Segment-aware Skip Connection (SaSC) to keep the generalization capability for unseen classes. Through extensive experiments on FSC147, CARPK, and PUCPR+, we demonstrate the benefits of our end-to-end framework, VLCounter. Code is available at https://github.com/seunggu0305/VLCounter

We present VIXEN - a technique that succinctly summarizes in text the visual differences between a pair of images in order to highlight any content manipulation present. Our proposed network linearly maps image features in a pairwise manner, constructing a soft prompt for a pretrained large language model. We address the challenge of low volume of training data and lack of manipulation variety in existing image difference captioning (IDC) datasets by training on synthetically manipulated images from the recent InstructPix2Pix dataset generated via prompt-to-prompt editing framework. We augment this dataset with change summaries produced via GPT-3. We show that VIXEN produces state-of-the-art, comprehensible difference captions for diverse image contents and edit types, offering a potential mitigation against misinformation disseminated via manipulated image content. Code and data are available at http://github.com/alexblck/vixen

We address the medication recommendation problem, which aims to recommend effective medications for a patient's current visit by utilizing information (e.g., diagnoses and procedures) given at the patient's current and past visits. While there exist a number of recommender systems designed for this problem, we point out that they are challenged in accurately capturing the relation (spec., the degree of relevance) between the current and each of the past visits for the patient when obtaining her current health status, which is the basis for recommending medications. To address this limitation, we propose a novel medication recommendation framework, named VITA, based on the following two novel ideas: (1) relevant-Visit selectIon; (2) Target-aware Attention. Through extensive experiments using real-world datasets, we demonstrate the superiority of VITA (spec., up to 5.67% higher accuracy, in terms of Jaccard, than the best competitor) and the effectiveness of its two core ideas. The code is available at https://github.com/jhheo0123/VITA.

The integration of visual encoders and large language models (LLMs) has driven recent progress in multimodal large language models (MLLMs). However, the scarcity of high-quality instruction-tuning data for vision-language tasks remains a challenge. The current leading paradigm, such as LLaVA, relies on language-only GPT-4 to generate data, which requires pre-annotated image captions and detection bounding boxes, suffering from understanding image details. A practical solution to this problem would be to utilize the available multimodal large language models to generate instruction data for vision-language tasks. However, it's worth noting that the currently accessible MLLMs are not as powerful as their LLM counterparts, as they tend to produce inadequate responses and generate false information. As a solution for addressing the current issue, this paper proposes the Visual Instruction Generation and Correction (VIGC) framework that enables multimodal large language models to generate instruction-tuning data and progressively enhance its quality on-the-fly. Specifically, Visual Instruction Generation (VIG) guides the vision-language model to generate diverse instruction-tuning data. To ensure generation quality, Visual Instruction Correction (VIC) adopts an iterative update mechanism to correct any inaccuracies in data produced by VIG, effectively reducing the risk of hallucination. Leveraging the diverse, high-quality data generated by VIGC, we finetune mainstream models and validate data quality based on various evaluations. Experimental results demonstrate that VIGC not only compensates for the shortcomings of language-only data generation methods, but also effectively enhances the benchmark performance. The models, datasets, and code are available at https://opendatalab.github.io/VIGC