Abstract Interpersonal communication in data science can yield sought‐after insights, but presentation environments are often not conducive for live analysis, forcing the process to move offline. Through a formative survey with 16 participants, we identified both technical (e.g., complexity of tools) and psychological (e.g., pressure of programming during presentation) factors constraining live data analysis. To enable live analysis, we present SlideSAVR, a data‐driven presentation assistant that leverages sketching and voice inputs in live discussion to support collaborative data analysis during presentations. Powered by an agentic framework that flexibly defines augmentation rules, updates slide content dynamically to match the live context, and automates backend computations, SlideSAVR enables fluid audience‐presenter interaction and reduces the need for offline reanalysis and follow‐up communication. We demonstrate SlideSAVR's ability to support a range of tasks through nine representative use cases. We further evaluate the system's accuracy and computation time across different settings, showing that SlideSAVR can reliably perform diverse tasks when provided with both sketch and voice inputs.

Federated fine-tuning (FFT) of large language models (LLMs) has recently emerged as a promising solution to enable domain-specific adaptation while preserving data privacy. Despite its benefits, FFT on resource-constrained clients relies on the high computational and memory demands of full-model fine-tuning, which limits the potential advancement. This paper presents FedSODA, a resource-efficient FFT framework that enables clients to adapt LLMs without accessing or storing the full model. Specifically, we first propose a similarity group pruning (SGP) module, which prunes redundant layers from the full LLM while retaining the most critical layers to preserve the model performance. Moreover, we introduce an orchestrated distillation alignment (ODA) module to reduce gradient divergence between the sub-LLM and the full LLM during FFT. Through the use of the QLoRA, clients only need to deploy quantized sub-LLMs and fine-tune lightweight adapters, significantly reducing local resource requirements. We conduct extensive experiments on three open-source LLMs across a variety of downstream tasks. The experimental results demonstrate that FedSODA reduces communication overhead by an average of 70.6%, decreases storage usage by 75.6%, and improves task accuracy by 3.1%, making it highly suitable for practical FFT applications under resource constraints.

Referential gestures, or as termed in linguistics, deixis, are an essential part of communication around data visualizations. Despite their importance, such gestures are often overlooked when documenting data analysis meetings. Transcripts, for instance, fail to capture gestures, and video recordings may not adequately capture or emphasize them. We introduce a novel method for documenting collaborative data meetings that treats deixis as a first-class citizen. Our proposed framework captures cursor-based gestural data along with audio and converts them into interactive documents. The framework leverages a large language model to identify word correspondences with gestures. These identified references are used to create context-based annotations in the resulting interactive document. We assess the effectiveness of our proposed method through a user study, finding that participants preferred our automated interactive documentation over recordings, transcripts, and manual note-taking. Furthermore, we derive a preliminary taxonomy of cursor-based deictic gestures from participant actions during the study. This taxonomy offers further opportunities for better utilizing cursor-based deixis in collaborative data analysis scenarios.

Existing deep learning-based speech enhancement methods only adopt clean speech as positive samples to guide the training of speech enhancement networks while negative samples, i.e., noisy speech, are unexploited. In this paper, we adopt contrastive regularization (CR) built upon contrastive learning to exploit both the information of noisy and clean speech as negative and positive samples, respectively. Particularly, CR minimizes the distance between clean and enhanced speech and maximizes the distance between noisy and enhanced speech in the representation space of the self-supervised learning model. However, the contrastive samples are non-consensual, as the negatives are usually represented distantly from the clean speech, leaving the solution space still under-constricted. To tackle this issue, we provide the negative samples assembled from (1) the noisy speech, and (2) the corresponding enhanced speech without using CR, and we customize a curriculum learning strategy to define the importance of these negative samples to balance the learning difficulty caused by different similarities between the embeddings of the positive and negative samples. Experiments show that our proposal improves SE performance effectively without introducing additional computation/parameters.

Compound graphs are networks in which vertices can be grouped into larger subsets, with these subsets capable of further grouping, resulting in a nesting that can be many levels deep. In several applications, including biological workflows, chemical equations, and computational data flow analysis, these graphs often exhibit a tree-like nesting structure, where sibling clusters are disjoint. Common compound graph layouts prioritize the lowest level of the grouping, down to the individual ungrouped vertices, which can make the higher level grouped structures more difficult to discern, especially in deeply nested networks. Leveraging the additional structure of the tree-like nesting, we contribute an overview+detail layout for this class of compound graphs that preserves the saliency of the higher level network structure when groups are expanded to show internal nested structure. Our layout draws inner structures adjacent to their parents, using a modified tree layout to place substructures. We describe our algorithm and then present case studies demonstrating the layout's utility to a domain expert working on data flow analysis. Finally, we discuss network parameters and analysis situations in which our layout is well suited.

Abstract Person re‐identification represents a pivotal sub‐problem in image retrieval, boasting broad application prospects in fields such as intelligent security and video surveillance. However, most existing person re‐identification methods predominantly focus solely on visual features pertaining to the person targets, thereby disregarding some supporting information closely related to the scene context. In the context of athlete re‐identification during sports event scenes, the athlete bib number is fully considered, an important clue that can provide different athletes' identities, and the traditional visual features of the person and high‐level semantic information of the bib number text are fused. A multi‐source information mutual gain mechanism is designed to improve the accuracy of the person re‐identification task. In the existing only publicly available marathon bib number dataset RBNR, the recognition accuracy of this method is significantly superior to that of the existing person re‐identification method. In addition, this paper constructs and publishes an athlete re‐identification dataset (HNNU‐ReID8000) for mainstream sports events, and the mean average precision (mAP) value of this method reaches 96.1% on this dataset, significantly ahead of existing state‐of‐the‐art person re‐identification methods. The code and the HNNU‐ReID8000 dataset will be released at https://github.com/yanbin‐zhu/zyb_person‐reid .

As global communication demands continue to rapidly expand, traditional terrestrial networks are facing significant challenges in terms of coverage, capacity, and reliability. To overcome these limitations, large-scale low-earth orbit (LEO) satellite networks have emerged as a promising solution, offering ubiquitous and seamless connectivity worldwide. However, this solution brings its own set of challenges, including limited satellite resources, diverse quality of service (QoS) requirements for random service function chain (SFC) requests, the complexity of managing large-scale networks, and the unpredictability of network changes. To tackle these challenges, this paper presents a novel adaptive online SFC deployment algorithm based on deep reinforcement learning. The proposed algorithm effectively handles real-time network changes and diverse service requirements while minimizing resource usage and enhancing QoS, leveraging the sharing of virtual network functions (VNFs) among multiple SFCs on satellite nodes and effectively balancing computing occupancy across satellites. To reduce complexity, we employ subnet segmentation to diminish the dimensionality of the state space. Simulation results validate the effectiveness of the proposed algorithm in significantly reducing resource occupancy and end-to-end delay, even in scenarios involving a large number of requests.

We present SizePairs, a new technique to create stable and balanced treemap layouts that visualize values changing over time in hierarchical data. To achieve an overall high-quality result across all time steps in terms of stability and aspect ratio, SizePairs employs a new hierarchical size-based pairing algorithm that recursively pairs two nodes that complement their size changes over time and have similar sizes. SizePairs maximizes the visual quality and stability by optimizing the splitting orientation of each internal node and flipping leaf nodes, if necessary. We also present a comprehensive comparison of SizePairs against the state-of-the-art treemaps developed for visualizing time-dependent data. SizePairs outperforms existing techniques in both visual quality and stability, while being faster than the local moves technique.

We present Target Netgrams as a visualization technique for radial layouts of graphs. Inspired by manually created target sociograms, we propose an annulus-constrained stress model that aims to position nodes onto the annuli between adjacent circles for indicating their radial hierarchy, while maintaining the network structure (clusters and neighborhoods) and improving readability as much as possible. This is achieved by having more space on the annuli than traditional layout techniques. By adapting stress majorization to this model, the layout is computed as a constrained least square optimization problem. Additional constraints (e.g., parent-child preservation, attribute-based clusters and structure-aware radii) are provided for exploring nodes, edges, and levels of interest. We demonstrate the effectiveness of our method through a comprehensive evaluation, a user study, and a case study.

As one of the basic tasks of computer vision, semantic segmentation is widely used in many fields, e.g., medical images parsing, scene parsing, autonomous driving, etc. In the current mainstream approaches, downsampling or patching operation is required to ensure that the GPU memory is not overloaded for dealing with the high-resolution input images. However, the corresponding cost is the lack of details in the final segmentation map. In this work, we proposed RFNet, a refinement network which resolves the lack of detailed information in coarse predictions by fusing the coarse predictions and the fine predictions gained by fine input image patches. There are three key characteristics: (i) designing a spatial information extraction module which can efficiently process information in coarse and fine feature maps at spatial level. (ii) proposing an auxiliary-fusion information branch calculated from the prediction maps, which contribute to refine predictions. (iii) designing a boundary auxiliary loss function in the training process, which makes the model pay more attention to those pixels belonging to the boundary of objects. We show the superiority of the proposed RFNet on the Cityscapes dataset, the experimental results illustrate that ours RFNet performance outperforms other state-of-the-art approaches with low computation consumption. The codes will be available at: https://github.com/zhu-gl-ux/RFNet.

Due to technological limitations and budget constraints, spatiotemporal fusion is considered a promising way to deal with the tradeoff between the temporal and spatial resolutions of remote sensing images. Furthermore, the generative adversarial network (GAN) has shown its capability in a variety of applications. This article presents a remote sensing image spatiotemporal fusion method using a GAN (STFGAN), which adopts a two-stage framework with an end-to-end image fusion GAN (IFGAN) for each stage. The IFGAN contains a generator and a discriminator in competition with each other under the guidance of the optimization function. Considering the huge spatial resolution gap between the high-spatial, low-temporal (HSLT) resolution Landsat imagery and the corresponding low-spatial, high-temporal (LSHT) resolution MODIS imagery, a feature-level fusion strategy is adopted. Specifically, for the generator, we first super-resolve the MODIS images while also extracting the high-frequency features of the Landsat images. Finally, we integrate the features from the MODIS and Landsat images. STFGAN is able to learn an end-to-end mapping between the Landsat-MODIS image pairs and predicts the Landsat-like image for a prediction date by considering all the bands. STFGAN significantly improves the accuracy of phenological change and land-cover-type change prediction with the help of residual blocks and two prior Landsat-MODIS image pairs. To examine the performance of the proposed STFGAN method, experiments were conducted on three representative Landsat-MODIS data sets. The results clearly illustrate the effectiveness of the proposed method.

Metabolic disorders play an important role in the development of many common diseases, including obesity, diabetes and coronary heart disease. Identifying key metabolites associated with disease can help us understand the mechanism of disease better and improve clinical diagnosis. Predicting diseases-related metabolites through computational approaches can provide potential biomarkers for further biological experiments. Text annotations on metabolites in existing databases provide prior information, which could provide more information about metabolites. However, current approaches haven't taken this information into consideration. In this work, we proposed a probability matrix factorization method which combined deep textual features to predict metabolite-disease associations. The deep neural network combining convolutional neural network and gated recurrent unit network is used to extract the corresponding features from text annotations of metabolites and diseases. Then, associations between metabolites and diseases are predicted through the matrix factorization based on these textual features. The main contributions in the work is that our model shows that adding textual features could help to improve the prediction of metabolite-disease associations. Case studies have indicated our model got predictive ability for diseases-related metabolites.

Cosparse analysis model has shown its superior performance in image reconstruction. However, this analysis frame has not been exploited yet for hyperspectral image restoration task. An analysis operator learning method called GOAL (GeOmetric Analysis operator Learning) is applied for hyperspectral image. Considering the correlation of the hyperspectral bands, the hyperspectral images were cropped into cube cells to get training samples. To avoid the window effect by image patch strategy, the analysis sparse filter method which provides global support from local information of the image was adopted. The denoising experiments and inpainting experiments are implemented on the hyperspectral images. The results are compared with the state of the art method, which shows our method is robust and efficient.