In this paper, we empirically demonstrate that natural images can be reconstructed with high fidelity from compressed representations using a simple first-order norm-plus-linear autoregressive (FINOLA) process—without relying on explicit positional information. Through systematic analysis, we observe that the learned coefficient matrices ($\mathbf{A}$ and $\mathbf{B}$) in FINOLA are typically invertible, and their product, $\mathbf{AB}^{-1}$, is diagonalizable across training runs. This structure enables a striking interpretation: FINOLA’s latent dynamics resemble a system of one-way wave equations evolving in a compressed latent space. Under this framework, each image corresponds to a unique solution of these equations. This offers a new perspective on image invariance, suggesting that the underlying structure of images may be governed by simple, invariant dynamic laws. Our findings shed light on a novel avenue for understanding and modeling visual data through the lens of latent-space dynamics and wave propagation.

Understanding infographic charts with design-driven visual elements (e.g., pictograms, icons) requires both visual recognition and reasoning, posing challenges for multimodal large language models (MLLMs). However, existing visual question answering benchmarks fall short in evaluating these capabilities of MLLMs due to the lack of paired plain charts and visual-element-based questions. To bridge this gap, we introduce InfoChartQA, a benchmark for evaluating MLLMs on infographic chart understanding. It includes 5,642 pairs of infographic and plain charts, each sharing the same underlying data but differing in visual presentations. We further design visual-element-based questions to capture their unique visual designs and communicative intent. Evaluation of 20 MLLMs reveals a substantial performance decline on infographic charts, particularly for visual-element-based questions related to metaphors. The paired infographic and plain charts enable fine-grained error analysis and ablation studies, which highlight new opportunities for advancing MLLMs in infographic chart understanding. We release InfoChartQA at https://github.com/CoolDawnAnt/InfoChartQA.

Abstract Large multimodal language models (MLLMs) such as GPT-4V and GPT-4o have achieved remarkable advancements in understanding and generating multimodal content, showcasing superior quality and capabilities across diverse tasks. However, their deployment faces significant challenges, including slow inference, high computational cost, and impracticality for on-device applications. In contrast, the emergence of small MLLMs, exemplified by the LLava-series models and Phi-3-Vision, offers promising alternatives with faster inference, reduced deployment costs, and the ability to handle domain-specific scenarios. Despite their growing presence, the capability boundaries between large and small MLLMs remain underexplored. In this work, we conduct a systematic and comprehensive evaluation to benchmark both small and large MLLMs, spanning general capabilities such as object recognition, temporal reasoning, and multimodal comprehension, as well as real-world applications in domains like industry and automotive. Our evaluation reveals that small MLLMs can achieve comparable performance to large models in specific scenarios but lag significantly in complex tasks requiring deeper reasoning or nuanced understanding. Furthermore, we identify common failure cases in both small and large MLLMs, highlighting domains where even state-of-the-art models struggle. We hope our findings will guide the research community in pushing the quality boundaries of MLLMs, advancing their usability and effectiveness across diverse applications.

Matrix reordering permutes the rows and columns of a matrix to reveal meaningful visual patterns, such as blocks that represent clusters. A comprehensive collection of matrices, along with a scoring method for measuring the quality of visual patterns in these matrices, contributes to building a benchmark. This benchmark is essential for selecting or designing suitable reordering algorithms for revealing specific patterns. In this paper, we build a matrix-reordering benchmark, ReorderBench, with the goal of evaluating and improving matrix-reordering techniques. This is achieved by generating a large set of representative and diverse matrices and scoring these matrices with a convolution- and entropy-based method. Our benchmark contains 2,835,000 binary matrices and 5,670,000 continuous matrices, each generated to exhibit one of four visual patterns: block, off-diagonal block, star, or band, along with 450 real-world matrices featuring hybrid visual patterns. We demonstrate the usefulness of ReorderBench through three main applications in matrix reordering: 1) evaluating different reordering algorithms, 2) creating a unified scoring model to measure the visual patterns in any matrix, and 3) developing a deep learning model for matrix reordering.

Recent studies have indicated that foundation models, such as BERT and GPT, excel at adapting to various downstream tasks. This adaptability has made them a dominant force in building artificial intelligence (AI) systems. Moreover, a new research paradigm has emerged as visualization techniques are incorporated into these models. This study divides these intersections into two research areas: visualization for foundation model (VIS4FM) and foundation model for visualization (FM4VIS). In terms of VIS4FM, we explore the primary role of visualizations in understanding, refining, and evaluating these intricate foundation models. VIS4FM addresses the pressing need for transparency, explainability, fairness, and robustness. Conversely, in terms of FM4VIS, we highlight how foundation models can be used to advance the visualization field itself. The intersection of foundation models with visualizations is promising but also introduces a set of challenges. By highlighting these challenges and promising opportunities, this study aims to provide a starting point for the continued exploration of this research avenue.

Combining multiple datasets enables performance boost on many computer vision tasks. But similar trend has not been witnessed in object detection when combining multiple datasets due to two inconsistencies among detection datasets: taxonomy difference and domain gap. In this paper, we address these challenges by a new design (named Detection Hub) that is dataset-aware and category-aligned. It not only mitigates the dataset inconsistency but also provides coherent guidance for the detector to learn across multiple datasets. In particular, the dataset-aware design is achieved by learning a dataset embedding that is used to adapt object queries as well as convolutional kernels in detection heads. The categories across datasets are semantically aligned into a unified space by replacing one-hot category representations with word embedding and leveraging the semantic coherence of language embedding. Detection Hub fulfills the benefits of large data on object detection. Experiments demonstrate that joint training on multiple datasets achieves significant performance gains over training on each dataset alone. Detection Hub further achieves SoTA performance on UODB benchmark with wide variety of datasets.

Benefiting from masked visual modeling, self-supervised video representation learning has achieved remarkable progress. However, existing methods focus on learning representations from scratch through reconstructing low-level features like raw pixel values. In this paper, we propose masked video distillation (MVD), a simple yet effective two-stage masked feature modeling framework for video representation learning: firstly we pretrain an image (or video) model by recovering low-level features of masked patches, then we use the resulting features as targets for masked feature modeling. For the choice of teacher models, we observe that students taught by video teachers perform better on temporally-heavy video tasks, while image teachers transfer stronger spatial representations for spatially-heavy video tasks. Visualization analysis also indicates different teachers produce different learned patterns for students. To leverage the advantage of different teachers, we design a spatial-temporal co-teaching method for MVD. Specifically, we distill student models from both video teachers and image teachers by masked feature modeling. Extensive experimental results demonstrate that video transformers pre-trained with spatial-temporal co-teaching outperform models distilled with a single teacher on a multitude of video datasets. Our MVD with vanilla ViT achieves state-of-the-art performance compared with previous methods on several challenging video downstream tasks. For example, with the ViT-Large model, our MVD achieves 86.4% and 76.7% Top-1 accuracy on Kinetics-400 and Something-Something-v2, outperforming VideoMAE by 1.2% and 2.4% respectively. When a larger ViT-Huge model is adopted, MVD achieves the state-of-the-art performance with 77.3% Top-1 accuracy on Something-Something-v2. Code will be available at https://github.com/ruiwang2021/mvd.

In this paper, we propose a novel cross-modal distillation method, called TinyCLIP, for large-scale language-image pre-trained models. The method introduces two core techniques: affinity mimicking and weight inheritance. Affinity mimicking explores the interaction between modalities during distillation, enabling student models to mimic teachers’ behavior of learning cross-modal feature alignment in a visual-linguistic affinity space. Weight inheritance transmits the pre-trained weights from the teacher models to their student counterparts to improve distillation efficiency. Moreover, we extend the method into a multi-stage progressive distillation to mitigate the loss of informative weights during extreme compression. Comprehensive experiments demonstrate the efficacy of TinyCLIP, showing that it can reduce the size of the pre-trained CLIP ViT-B/32 by 50%, while maintaining comparable zero-shot performance. While aiming for comparable performance, distillation with weight inheritance can speed up the training by 1.4 - 7.8× compared to training from scratch. Moreover, our TinyCLIP ViT-8M/16, trained on YFCC-15M, achieves an impressive zero-shot top-1 accuracy of 41.1% on ImageNet, surpassing the original CLIP ViT-B/16 by 3.5% while utilizing only 8.9% parameters. Finally, we demonstrate the good transferability of TinyCLIP in various downstream tasks. Code and models will be open-sourced at aka.ms/tinyclip.

Long-tailed object detection (LTOD) aims to handle the extreme data imbalance in real-world datasets, where many tail classes have scarce instances. One popular strategy is to explore extra data with image-level labels, yet it produces limited results due to (1) semantic ambiguity---an image-level label only captures a salient part of the image, ignoring the remaining rich semantics within the image; and (2) location sensitivity---the label highly depends on the locations and crops of the original image, which may change after data transformations like random cropping. To remedy this, we propose RichSem, a simple but effective method, which is robust to learn rich semantics from coarse locations without the need of accurate bounding boxes. RichSem leverages rich semantics from images, which are then served as additional ``soft supervision'' for training detectors. Specifically, we add a semantic branch to our detector to learn these soft semantics and enhance feature representations for long-tailed object detection. The semantic branch is only used for training and is removed during inference. RichSem achieves consistent improvements on both overall and rare-category of LVIS under different backbones and detectors. Our method achieves state-of-the-art performance without requiring complex training and testing procedures. Moreover, we show the effectiveness of our method on other long-tailed datasets with additional experiments.

Recent advances in artificial intelligence largely benefit from better neural network architectures. These architectures are a product of a costly process of trial-and-error. To ease this process, we develop ArchExplorer, a visual analysis method for understanding a neural architecture space and summarizing design principles. The key idea behind our method is to make the architecture space explainable by exploiting structural distances between architectures. We formulate the pairwise distance calculation as solving an all-pairs shortest path problem. To improve efficiency, we decompose this problem into a set of single-source shortest path problems. The time complexity is reduced from O(kn2N) to O(knN). Architectures are hierarchically clustered according to the distances between them. A circle-packing-based architecture visualization has been developed to convey both the global relationships between clusters and local neighborhoods of the architectures in each cluster. Two case studies and a post-analysis are presented to demonstrate the effectiveness of ArchExplorer in summarizing design principles and selecting better-performing architectures.

We present Mobile-Former, a parallel design of MobileNet and transformer with a two-way bridge in between. This structure leverages the advantages of MobileNet at local processing and transformer at global interaction. And the bridge enables bidirectional fusion of local and global features. Different from recent works on vision transformer, the transformer in Mobile-Former contains very few tokens (e.g. 6 or fewer tokens) that are randomly initialized to learn global priors, resulting in low computational cost. Combining with the proposed light-weight cross attention to model the bridge, Mobile-Former is not only computationally efficient, but also has more representation power. It outperforms MobileNetV3 at low FLOP regime from 25M to 500M FLOPs on ImageNet classification. For instance, Mobile-Former achieves 77.9% top-1 accuracy at 294M FLOPs, gaining 1.3% over MobileNetV3 but saving 17% of computations. When transferring to object detection, Mobile-Former outperforms MobileNetV3 by 8.6 AP in RetinaNet framework. Furthermore, we build an efficient end-to-end detector by replacing backbone, encoder and decoder in DETR with Mobile-Former, which outperforms DETR by 1.3 AP but saves 52% of computational cost and 36% of parameters. Code will be released at https://github.com/aaboys/mobileformer.

Transformers have achieved great success in pluralistic image inpainting recently. However, we find existing transformer based solutions regard each pixel as a token, thus suffer from information loss issue from two aspects: 1) They downsample the input image into much lower resolutions for efficiency consideration, incurring information loss and extra misalignment for the boundaries of masked regions. 2) They quantize 2563RGB pixels to a small number (such as 512) of quantized pixels. The indices of quantized pixels are used as tokens for the inputs and prediction targets of transformer. Although an extra CNN network is used to upsample and refine the low-resolution results, it is difficult to retrieve the lost information back. To keep input information as much as possible, we propose a new transformer based framework “PUT”. Specifically, to avoid input downsampling while maintaining the computation efficiency, we design a patch-based auto-encoder P-VQVAE, where the encoder converts the masked image into non-overlapped patch tokens and the decoder recovers the masked regions from the inpainted tokens while keeping the unmasked regions unchanged. To eliminate the information loss caused by quantization, an Un-Quantized Transformer (UQ-Transformer) is applied, which directly takes the features from P-VQVAE encoder as input without quantization and regards the quantized tokens only as prediction targets. Extensive experiments show that PUT greatly outperforms state-of-the-art methods on image fidelity, especially for large masked regions and complex large-scale datasets.

This paper studies the BERT pretraining of video transformers. It is a straightforward but worth-studying extension given the recent success from BERT pretraining of image transformers. We introduce BEVT which decouples video representation learning into spatial representation learning and temporal dynamics learning. In particular, BEVT first performs masked image modeling on image data, and then conducts masked image modeling jointly with masked video modeling on video data. This design is motivated by two observations: 1) transformers learned on image datasets provide decent spatial priors that can ease the learning of video transformers, which are often times computationally-intensive if trained from scratch; 2) discriminative clues, i.e., spatial and temporal information, needed to make correct predictions vary among different videos due to large intra-class and inter-class variations. We conduct extensive experiments on three challenging video benchmarks where BEVT achieves very promising results. On Kinetics 400, for which recognition mostly relies on discriminative spatial representations, BEVT achieves comparable results to strong supervised baselines. On Something-Something-V2 and Diving 48, which contain videos relying on temporal dynamics, BEVT outperforms by clear margins all alternative baselines and achieves state-of-the-art performance with a 71.4% and 87.2% Top-1 accuracy respectively. Code is available at https://github.com/xyzforever/BEVT.

Vision Transformer (ViT) models have recently drawn much attention in computer vision due to their high model capability. However, ViT models suffer from huge number of parameters, restricting their applicability on devices with limited memory. To alleviate this problem, we propose MiniViT, a new compression framework, which achieves parameter reduction in vision transformers while retaining the same performance. The central idea of MiniViT is to multiplex the weights of consecutive transformer blocks. More specifically, we make the weights shared across layers, while imposing a transformation on the weights to increase diversity. Weight distillation over self-attention is also applied to transfer knowledge from large-scale ViT models to weight-multiplexed compact models. Comprehensive experiments demonstrate the efficacy of MiniViT, showing that it can reduce the size of the pre-trained Swin-B transformer by 48%, while achieving an increase of 1.0% in Top-1 accuracy on ImageNet. Moreover, using a single-layer of parameters, MiniViT is able to compress DeiT-B by 9.7 times from 86M to 9M parameters, without seriously compromising the performance. Finally, we verify the transferability of MiniViT by reporting its performance on downstream benchmarks. Code and models are available at here.

The complex nature of combining localization and classification in object detection has resulted in the flourished development of methods. Previous works tried to improve the performance in various object detection heads but failed to present a unified view. In this paper, we present a novel dynamic head framework to unify object detection heads with attentions. By coherently combining multiple self-attention mechanisms between feature levels for scale-awareness, among spatial locations for spatial-awareness, and within output channels for task-awareness, the proposed approach significantly improves the representation ability of object detection heads without any computational overhead. Further experiments demonstrate that the effectiveness and efficiency of the proposed dynamic head on the COCO benchmark. With a standard ResNeXt-101-DCN backbone, we largely improve the performance over popular object detectors and achieve a new state-of-the-art at 54.0 AP. The code will be released at https://github.com/microsoft/DynamicHead.

This paper aims at addressing the problem of substantial performance degradation at extremely low computational cost (e.g. 5M FLOPs on ImageNet classification). We found that two factors, sparse connectivity and dynamic activation function, are effective to improve the accuracy. The former avoids the significant reduction of network width, while the latter mitigates the detriment of reduction in network depth. Technically, we propose micro-factorized convolution, which factorizes a convolution matrix into low rank matrices, to integrate sparse connectivity into convolution. We also present a new dynamic activation function, named Dynamic Shift Max, to improve the non-linearity via maxing out multiple dynamic fusions between an input feature map and its circular channel shift. Building upon these two new operators, we arrive at a family of networks, named MicroNet, that achieves significant performance gains over the state of the art in the low FLOP regime. For instance, under the constraint of 12M FLOPs, MicroNet achieves 59.4% top-1 accuracy on ImageNet classification, outperforming MobileNetV3 by 9.6%. Source code is at https://github.com/liyunsheng13/micronet.

We present in this paper a new architecture, named Convolutional vision Transformer (CvT), that improves Vision Transformer (ViT) in performance and efficiency by introducing convolutions into ViT to yield the best of both de-signs. This is accomplished through two primary modifications: a hierarchy of Transformers containing a new convolutional token embedding, and a convolutional Transformer block leveraging a convolutional projection. These changes introduce desirable properties of convolutional neural networks (CNNs) to the ViT architecture (i.e. shift, scale, and distortion invariance) while maintaining the merits of Transformers (i.e. dynamic attention, global context, and better generalization). We validate CvT by conducting extensive experiments, showing that this approach achieves state-of-the-art performance over other Vision Transformers and ResNets on ImageNet-1k, with fewer parameters and lower FLOPs. In addition, performance gains are maintained when pretrained on larger datasets (e.g. ImageNet-22k) and fine-tuned to downstream tasks. Pretrained on ImageNet-22k, our CvT-W24 obtains a top-1 accuracy of 87.7% on the ImageNet-1k val set. Finally, our results show that the positional encoding, a crucial component in existing Vision Transformers, can be safely re-moved in our model, simplifying the design for higher resolution vision tasks. Code will be released at https://github.com/microsoft/CvT.

Neural Architecture Search (NAS) often trains and evaluates a large number of architectures. Recent predictor-based NAS approaches attempt to alleviate such heavy computation costs with two key steps: sampling some architecture-performance pairs and fitting a proxy accuracy predictor. Given limited samples, these predictors, however, are far from accurate to locate top architectures due to the difficulty of fitting the huge search space. This paper reflects on a simple yet crucial question: if our final goal is to find the best architecture, do we really need to model the whole space well?. We propose a paradigm shift from fitting the whole architecture space using one strong predictor, to progressively fitting a search path towards the high-performance sub-space through a set of weaker predictors. As a key property of the weak predictors, their probabilities of sampling better architectures keep increasing. Hence we only sample a few well-performed architectures guided by the previously learned predictor and estimate a new better weak predictor. This embarrassingly easy framework, dubbed WeakNAS, produces coarse-to-fine iteration to gradually refine the ranking of sampling space. Extensive experiments demonstrate that WeakNAS costs fewer samples to find top-performance architectures on NAS-Bench-101 and NAS-Bench-201. Compared to state-of-the-art (SOTA) predictor-based NAS methods, WeakNAS outperforms all with notable margins, e.g., requiring at least 7.5x less samples to find global optimal on NAS-Bench-101. WeakNAS can also absorb their ideas to boost performance more. Further, WeakNAS strikes the new SOTA result of 81.3% in the ImageNet MobileNet Search Space. The code is available at: https://github.com/VITA-Group/WeakNAS.

Visual analytics for machine learning has recently evolved as one of the most exciting areas in the field of visualization. To better identify which research topics are promising and to learn how to apply relevant techniques in visual analytics, we systematically review 259 papers published in the last ten years together with representative works before 2010. We build a taxonomy, which includes three first-level categories: techniques before model building, techniques during modeling building, and techniques after model building. Each category is further characterized by representative analysis tasks, and each task is exemplified by a set of recent influential works. We also discuss and highlight research challenges and promising potential future research opportunities useful for visual analytics researchers.

Adversarial examples, generated by adding small but intentionally imperceptible perturbations to normal examples, can mislead deep neural networks (DNNs) to make incorrect predictions. Although much work has been done on both adversarial attack and defense, a fine-grained understanding of adversarial examples is still lacking. To address this issue, we present a visual analysis method to explain why adversarial examples are misclassified. The key is to compare and analyze the datapaths of both the adversarial and normal examples. A datapath is a group of critical neurons along with their connections. We formulate the datapath extraction as a subset selection problem and solve it by constructing and training a neural network. A multi-level visualization consisting of a network-level visualization of data flows, a layer-level visualization of feature maps, and a neuron-level visualization of learned features, has been designed to help investigate how datapaths of adversarial and normal examples diverge and merge in the prediction process. A quantitative evaluation and a case study were conducted to demonstrate the promise of our method to explain the misclassification of adversarial examples.

Light-weight convolutional neural networks (CNNs) suffer performance degradation as their low computational budgets constrain both the depth (number of convolution layers) and the width (number of channels) of CNNs, resulting in limited representation capability. To address this issue, we present Dynamic Convolution, a new design that increases model complexity without increasing the network depth or width. Instead of using a single convolution kernel per layer, dynamic convolution aggregates multiple parallel convolution kernels dynamically based upon their attentions, which are input dependent. Assembling multiple kernels is not only computationally efficient due to the small kernel size, but also has more representation power since these kernels are aggregated in a non-linear way via attention. By simply using dynamic convolution for the state-of-the-art architecture MobileNetV3-Small, the top-1 accuracy of ImageNet classification is boosted by 2.9% with only 4% additional FLOPs and 2.9 AP gain is achieved on COCO keypoint detection.

Recently, extensive research attention has been drawn to the program executing security of embedded processor since increasing code tamper attacks, as well as transient faults severely affect the safety of embedded systems. The security monitoring and fault recovery technique is one of the effective methods to ensure the security and performance of embedded devices. In this article, an architecture for the security monitoring and fault recovery is proposed for run-time program execution, which builds a Monitoring Cache (M-Cache) and then checks the integrity based on reference data. Especially, the proposed architecture will build the checkpoint once the M-Cache is missed and also take the rollback operation after the unsuccessful of integrity check. In addition, three tampered positions (e.g., instruction register within the pipeline, instruction within the cache, and code within memory) have been elaborately focused on to guarantee the normal running of the embedded system. Eventually, by adopting the open RISC processor for algorithm implementation and verification, the proposal has been proven to be promising for the detection of a fault or tampered program, as well as the fast recovery of running environment and code.

MOTIVATION: The interactions between protein and nucleic acids play a key role in various biological processes. Accurate recognition of the residues that bind nucleic acids can facilitate the study of uncharacterized protein-nucleic acids interactions. The accuracy of existing nucleic acids-binding residues prediction methods is relatively low. RESULTS: In this work, we introduce NucBind, a novel method for the prediction of nucleic acids-binding residues. NucBind combines the predictions from a support vector machine-based ab-initio method SVMnuc and a template-based method COACH-D. SVMnuc was trained with features from three complementary sequence profiles. COACH-D predicts the binding residues based on homologous templates identified from a nucleic acids-binding library. The proposed methods were assessed and compared with other peering methods on three benchmark datasets. Experimental results show that NucBind consistently outperforms other state-of-the-art methods. Though with higher accuracy, similar to many other ab-initio methods, cross prediction between DNA and RNA-binding residues was also observed in SVMnuc and NucBind. We attribute the success of NucBind to two folds. The first is the utilization of improved features extracted from three complementary sequence profiles in SVMnuc. The second is the combination of two complementary methods: the ab-initio method SVMnuc and the template-based method COACH-D. AVAILABILITY AND IMPLEMENTATION: http://yanglab.nankai.edu.cn/NucBind. SUPPLEMENTARY INFORMATION: Supplementary data are available at Bioinformatics online.

We propose a novel semi -supervised fully convolutional network for Polarimetric synthetic aperture radar (PoISAR) terrain classification. First, by designing a fully convolutional structure, we can perform pixel-based classification tasks. Then, by applying semi -supervised generative adversarial networks (GANs), we utilize both labeled and unlabeled samples and aim to obtain higher classification accuracy. Through a mini-max two-player game, GAN has better performance than other “single-player” classifiers. Finally, we combine the fully convolutional structure with the semi-supervised GAN. Our fully convolutional semi-supervised GAN (FC-SGAN) has excellent spatial feature learning ability and can perform end-to-end pixel-based classification tasks. Experimental results show that compared with existing works, the proposed method has better performances. Even when the training set gets smaller, our method keeps high accuracy.

This paper sets up the real-time user information search system under the home environment by integrating the robot technology, intelligent wearable technology, speech recognition technology, speech synthesis technology, facial expression recognition technology, and cloud computing technology, and builds the user emotion perception means based on big data clouds, to make more convenient human-machine interaction and more accurate and reasonable feedback. RoCoSense system built in this paper can greatly enrich the intelligent system interactive means by the real-time perception of user emotion, and has a wide application prospect in the fields of consumption, entertainment, and health care.

Among the many types of deep models, deep generative models (DGMs) provide a solution to the important problem of unsupervised and semi-supervised learning. However, training DGMs requires more skill, experience, and know-how because their training is more complex than other types of deep models such as convolutional neural networks (CNNs). We develop a visual analytics approach for better understanding and diagnosing the training process of a DGM. To help experts understand the overall training process, we first extract a large amount of time series data that represents training dynamics (e.g., activation changes over time). A blue-noise polyline sampling scheme is then introduced to select time series samples, which can both preserve outliers and reduce visual clutter. To further investigate the root cause of a failed training process, we propose a credit assignment algorithm that indicates how other neurons contribute to the output of the neuron causing the training failure. Two case studies are conducted with machine learning experts to demonstrate how our approach helps understand and diagnose the training processes of DGMs. We also show how our approach can be directly used to analyze other types of deep models, such as CNNs.

Although several effective learning-from-crowd methods have been developed to infer correct labels from noisy crowdsourced labels, a method for post-processed expert validation is still needed. This paper introduces a semi-supervised learning algorithm that is capable of selecting the most informative instances and maximizing the influence of expert labels. Specifically, we have developed a complete uncertainty assessment to facilitate the selection of the most informative instances. The expert labels are then propagated to similar instances via regularized Bayesian inference. Experiments on both real-world and simulated datasets indicate that given a specific accuracy goal (e.g., 95%) our method reduces expert effort from 39% to 60% compared with the state-of-the-art method.

Deep convolutional neural networks (CNNs) have achieved breakthrough performance in many pattern recognition tasks such as image classification. However, the development of high-quality deep models typically relies on a substantial amount of trial-and-error, as there is still no clear understanding of when and why a deep model works. In this paper, we present a visual analytics approach for better understanding, diagnosing, and refining deep CNNs. We formulate a deep CNN as a directed acyclic graph. Based on this formulation, a hybrid visualization is developed to disclose the multiple facets of each neuron and the interactions between them. In particular, we introduce a hierarchical rectangle packing algorithm and a matrix reordering algorithm to show the derived features of a neuron cluster. We also propose a biclustering-based edge bundling method to reduce visual clutter caused by a large number of connections between neurons. We evaluated our method on a set of CNNs and the results are generally favorable.

Interactive model analysis, the process of understanding, diagnosing, and refining a machine learning model with the help of interactive visualization, is very important for users to efficiently solve real-world artificial intelligence and data mining problems. Dramatic advances in big data analytics have led to a wide variety of interactive model analysis tasks. In this paper, we present a comprehensive analysis and interpretation of this rapidly developing area. Specifically, we classify the relevant work into three categories: understanding, diagnosis, and refinement. Each category is exemplified by recent influential work. Possible future research opportunities are also explored and discussed.

Although there has been a great deal of interest in analyzing customer opinions and breaking news in microblogs, progress has been hampered by the lack of an effective mechanism to discover and retrieve data of interest from microblogs. To address this problem, we have developed an uncertainty-aware visual analytics approach to retrieve salient posts, users, and hashtags. We extend an existing ranking technique to compute a multifaceted retrieval result: the mutual reinforcement rank of a graph node, the uncertainty of each rank, and the propagation of uncertainty among different graph nodes. To illustrate the three facets, we have also designed a composite visualization with three visual components: a graph visualization, an uncertainty glyph, and a flow map. The graph visualization with glyphs, the flow map, and the uncertainty analysis together enable analysts to effectively find the most uncertain results and interactively refine them. We have applied our approach to several Twitter datasets. Qualitative evaluation and two real-world case studies demonstrate the promise of our approach for retrieving high-quality microblog data.

A high-quality label layout is critical for effective information understanding and consumption. Existing labeling methods fail to help users quickly gain an overview of visualized data when the number of labels is large. Visual clutter is a major challenge preventing these methods from being applied to real-world applications. To address this, we propose a context-aware label layout that can measure and reduce visual clutter during the layout process. Our method formulates the clutter model using four factors: confusion, visual connection, distance, and intersection. Based on this clutter model, an effective clutter-aware labeling method has been developed that can generate clear and legible label layouts in different visualizations. We have applied our method to several types of visualizations and the results show promise, especially in support of an uncluttered and informative label layout.

It is important for many different applications such as government and business intelligence to analyze and explore the diffusion of public opinions on social media. However, the rapid propagation and great diversity of public opinions on social media pose great challenges to effective analysis of opinion diffusion. In this paper, we introduce a visual analysis system called OpinionFlow to empower analysts to detect opinion propagation patterns and glean insights. Inspired by the information diffusion model and the theory of selective exposure, we develop an opinion diffusion model to approximate opinion propagation among Twitter users. Accordingly, we design an opinion flow visualization that combines a Sankey graph with a tailored density map in one view to visually convey diffusion of opinions among many users. A stacked tree is used to allow analysts to select topics of interest at different levels. The stacked tree is synchronized with the opinion flow visualization to help users examine and compare diffusion patterns across topics. Experiments and case studies on Twitter data demonstrate the effectiveness and usability of OpinionFlow.

Storyline visualizations, which are useful in many applications, aim to illustrate the dynamic relationships between entities in a story. However, the growing complexity and scalability of stories pose great challenges for existing approaches. In this paper, we propose an efficient optimization approach to generating an aesthetically appealing storyline visualization, which effectively handles the hierarchical relationships between entities over time. The approach formulates the storyline layout as a novel hybrid optimization approach that combines discrete and continuous optimization. The discrete method generates an initial layout through the ordering and alignment of entities, and the continuous method optimizes the initial layout to produce the optimal one. The efficient approach makes real-time interactions (e.g., bundling and straightening) possible, thus enabling users to better understand and track how the story evolves. Experiments and case studies are conducted to demonstrate the effectiveness and usefulness of the optimization approach.