Multi-modal dataset distillation (DD) condenses large datasets into compact ones that retain task efficacy by capturing correspondence patterns, i.e., shared semantics between paired modalities. However, such patterns rely on cross-modal similarity and cannot be faithfully captured by intra-modal similarity of current unimodal strategies. As a result, current multi-modal DD methods tend to over-concentrate, redundantly encoding similar correspondence patterns and thus limiting generalizability. To this end, we propose a novel multi-modal DD framework to systematically Promote Correspondence coverage, i.e., ProCo. Initially, we develop a correspondence consistency metric based on cross-modal retrieval distributions to cluster correspondence patterns. These clusters capture the underlying correspondence distribution, enabling ProCo to initialize distilled data with representative patterns while regularizing optimization to promote correspondence representativeness and diversity. Moreover, we employ conditional neural fields for efficient distilled data parameterization, enhancing fine-grained pattern capture while allowing more distilled data under a fixed budget to boost correspondence coverage. Extensive experiments verify that our ProCo achieves superior and elastic budget-efficacy trade-offs, surpassing prior methods by over 15% with 10x distillation budget reduction, highlighting its real-world practicality.
论文检索
输入标题、作者或关键词，从 100,903 篇学术成果中精准定位
Urban air pollution is a major health crisis causing millions of premature deaths annually, underscoring the urgent need for accurate and scalable monitoring of air quality (AQ). While low-cost sensors (LCS) offer a scalable alternative to expensive reference-grade stations, their readings are affected by drift, calibration errors, and environmental interference. To address these challenges, we introduce Veli (Reference free Variational Estimation via Latent Inference), an unsupervised Bayesian model that leverages variational inference to correct LCS readings without requiring co-location with reference stations, eliminating a major deployment barrier. Specifically, Veli constructs a disentangled representation of the LCS readings, effectively separating the true pollutant reading from the sensor noise. To build our model and address the lack of standardized benchmarks in AQ monitoring, we also introduce the Air Quality Sensor Data Repository (AQ-SDR). AQ-SDR is the largest AQ sensor benchmark to date, with readings from 23,737 LCS and reference stations across multiple regions. Veli demonstrates strong generalization across both in-distribution and out-of-distribution settings, effectively handling sensor drift and erratic sensor behavior. Appendices are available in the extended version.
MemoryART: Enhancing LLMs via Multi-Memory Models with Adaptive Resonance Theory for Healthcare Agents
PDF ↗Though promising in healthcare consultation applications, large language models (LLMs) face critical limitations in retaining and utilizing long-term memory across multi-turn interactions. In particular, existing memory enhancing paradigms are constrained by limited context windows and embedding-based retrieval, often failing to maintain task relevance and still suffering from memory prototype collapse in multi-turn healthcare consultation. To address these challenges, we propose a cognitively-inspired memory framework named MemoryART, which is grounded in Adaptive Resonance Theory (ART)—a cognitive and learning theory of how humans and animals adapt to dynamic environments. MemoryART employs three memory modules—working memory, episodic memory, and semantic memory to support task-aware memory organization and dynamic retrieval. Specifically, episodic memory provides the storage of specific experiences along with contextual clues, which is crucial for managing patient-specific information and perfect for multi-turn healthcare consultation interactions. Building upon this concept, MemoryART leverages multi-channel competitive learning and resonance matching to enable efficient and interpretable episodic memory encoding, alleviating issues of prototype collapse and noisy memory associations. For evaluation, we construct a long-term medical dialogue benchmark called MediLongChat using a LLM-based generation pipeline. The resulting dataset features realistic, multi-disease chat histories, each exceeding 100K tokens across 20–30 dialogues, simulating real-world healthcare interaction patterns. Our experimental results show that MemoryART outperforms mainstream approaches in memory-intensive tasks, achieving SOTA results and significantly reducing token consumption across five popular LLMs, confirming its effectiveness and efficiency in providing scalable, reliable memory for LLMs in healthcare.
Modern AI services must continually adapt to newly joined domains, yet delivering high-quality customized models is hampered by label sparsity, domain shifts, and tight budgets. We formulate this challenge as the learning system expansion problem and introduce HaT, an efficient heterogeneity-aware knowledge-transfer framework. HaT first selects a small set of high-quality source models with minimal overhead, and then fuses their imperfect predictions through a sample-wise attention mixer. Later, it adaptively distills the fused knowledge into target models via a knowledge dictionary. Extensive experiments on different tasks and modalities show that HaT outperforms state-of-the-art baselines by up to 16.5% accuracy, and saves 31.1% training time and up to 93.0% traffic.
This paper presents a novel generative framework for learning shared latent representations across multimodal data. Many advanced multimodal methods focus on capturing all combinations of modality-specific details across inputs, which can inadvertently obscure the high-level semantic concepts that are shared across modalities. Notably, Multimodal VAEs with low-dimensional latent variables are designed to capture shared representations, enabling various tasks such as joint multimodal synthesis and cross-modal inference. However, multimodal VAEs often struggle to design expressive joint variational posteriors and suffer from low-quality synthesis. In this work, ShaLa addresses these challenges by integrating a novel architectural inference model and a second-stage expressive diffusion prior, which not only facilitates effective inference of shared latent representation but also significantly improves the quality of downstream multimodal synthesis. We validate ShaLa extensively across multiple benchmarks, demonstrating superior coherence and synthesis quality compared to state-of-the-art multimodal VAEs. Furthermore, ShaLa scales to many more modalities while prior multimodal VAEs have fallen short in capturing the increasing complexity of the shared latent space.
Fair clustering has attracted increased attention in recent years. In this work, we study the individually fair clustering problem in Euclidean space. While single-swap local search methods have achieved near-linear running time and constant approximation guarantees, their performance often depends on the aspect ratio of the dataset (the ratio between the diameter and the minimum interpoint distance of the dataset). How to apply multi-swap local search while obtaining linear running time with better approximation ratio is still a challenging task. To address this, we introduce a collaborative initialization framework for that integrates greedy with sampling techniques. This framework eliminates the dependence on the aspect ratio and produces a constant-factor bicriteria approximation in linear time. In contrast to the current state-of-the-art near-linear time algorithm, which requires a restrictive assumption about the relationship between optimal centers and cluster centroids, we propose a multi-swap local search algorithm that provides an improved approximation guarantee. Our method runs in linear time with high probability and does not rely on the aforementioned assumption. We validate our theoretical results through extensive experiments on both real-world and synthetic datasets, including large-scale benchmarks with up to 100 million points. Our empirical evaluation demonstrates superior performance in terms of clustering quality and computational efficiency, along with scalability under varying parameter settings.
In this paper we address the problem of discovering causal relationships from observational event sequence data. Existing methods typically assume that events are instantaneous point events, however in many real-world settings, events have duration. For example, in healthcare, a patient's symptoms may persist over a time interval and influence clinical actions while ongoing. To address this, we introduce a causal model for interval-based event sequences that captures rich causal structures, including interactions between events and causal mechanisms that depend on whether other events are ongoing. We prove that our model is identifiable in the limit and present a practical causal discovery algorithm, Niagara, grounded in the algorithmic Markov condition. To select among candidate models, we employ a minimum description length (MDL) criterion, enabling robust inference even with limited data. We validate our approach on synthetic and real data and demonstrate its utility on a real-world medical case study, where it uncovers meaningful causal relationships from noisy, interval-based event data.
CATAL: Causally Disentangled Task Representation Learning for Offline Meta-Reinforcement Learning
PDF ↗Context-based Offline Meta Reinforcement Learning (COMRL) has shown promising results in improving the cross-task generalization ability of meta-policies. However, current methods often lead to entangled task representations, in which each latent dimension is influenced by multiple causal factors that govern variations in environment dynamics and reward mechanisms. This entanglement can degrade generalization performance, particularly when multiple causal factors vary simultaneously across tasks. To address this limitation, we propose CAusally disentangled TAsk representation Learning (CATAL) method for COMRL that aims to improve the generalization ability of the meta-policy, where each latent dimension in the task representations aligns to a single causal factor.Theoretically, we show that under mild conditions, the task representations learned by CATAL are causally disentangled. Empirically, extensive results on multi-task MuJoCo benchmarks show that CATAL consistently outperforms existing COMRL baselines in both in-distribution and out-of-distribution generalization.
Gradient Boosting Decision Trees (GBDTs) are widely used in industry and academia for their high accuracy and efficiency, particularly on structured data. However, the subject of watermarking GBDT models remains underexplored, especially compared to neural networks. In this work, we present the first robust watermarking framework tailored to GBDT models, utilizing in-place fine-tuning to embed imperceptible and resilient watermarks. We propose four embedding strategies, each designed to minimize impact on model accuracy while ensuring watermark robustness. Through experiments across diverse datasets, we demonstrate that our methods achieve high watermark embedding rates, low accuracy degradation, and strong resistance to post-deployment fine-tuning.
Continual test-time adaptation (CTTA) enables online model adjustment under dynamic distribution shifts in real-world environments. However, most existing CTTA frameworks adopt fixed model architectures, lacking the structural flexibility required for deployment across heterogeneous edge devices with varying computational capacities. To address this, we propose an elastic framework for edge CTTA that performs resource-aware dynamic model search based on a pre-trained binary Supernet. This enables architectural flexibility by generating personalized models tailored to the resource constraints of different edge devices. Considering the evolving distribution of unlabeled data on edge devices during deployment, we introduce a pluggable lightweight fine-tuning mechanism. By inserting low-rank adapters into the frozen binary backbone, the model enables continual self-supervised adaptation with minimal computational overhead. In addition, we propose a structure-aware knowledge reflux mechanism that transfers the adaptation experience from fine-tuned edge models back into the Supernet. By distilling knowledge into structurally aligned Supernet paths, future architecture search is improved without requiring retraining. Experiments on multiple benchmarks validate that our method achieves state-of-the-art performance while significantly reducing resource consumption, with re-searched models after knowledge reflux showing further improvements.
Clustering approaches that utilize convex loss functions have recently attracted growing interest in the formation of compact data clusters. Although classical methods like kmeans and its wide family of variants are still widely used, all of them require the number of clusters (k) to be supplied as input, and many are notably sensitive to initialization. Convex clustering provides a more stable alternative by formulating the clustering task as a convex optimization problem, ensuring a unique global solution. However, it faces challenges in handling high dimensional data, especially in the presence of noise and outliers. Additionally, strong fusion regularization, controlled by the tuning parameter, can hinder effective cluster formation within a convex clustering framework. To overcome these challenges, we introduce a robust approach that integrates convex clustering with the Median of Means (MoM) estimator, thus developing an outlier resistant and efficient clustering framework that does not necessitate a prior knowledge of the number of clusters. By leveraging the robustness of MoM alongside the stability of convex clustering, our method enhances both performance and efficiency, especially on large scale datasets. Theoretical analysis demonstrates weak consistency under specific conditions, while experiments on synthetic and real world datasets validate the method’s superior performance compared to existing approaches.
T3Time: Tri-Modal Time Series Forecasting via Adaptive Multi-Head Alignment and Residual Fusion
PDF ↗Multivariate time series forecasting (MTSF) seeks to model temporal dynamics among variables to predict future trends. Transformer-based models and large language models (LLMs) have shown promise due to their ability to capture long-range dependencies and patterns. However, current methods often rely on rigid inductive biases, ignore intervariable interactions, or apply static fusion strategies that limit adaptability across forecast horizons. These limitations create bottlenecks in capturing nuanced, horizon-specific relationships in time-series data. To solve this problem, we propose T3Time, a novel trimodal framework consisting of time, spectral, and prompt branches, where the dedicated frequency encoding branch captures the periodic structures along with a gating mechanism that learns prioritization between temporal and spectral features based on the prediction horizon. We also proposed a mechanism which adaptively aggregates multiple cross-modal alignment heads by dynamically weighting the importance of each head based on the features. Extensive experiments on benchmark datasets demonstrate that our model consistently outperforms state-of-the-art baselines, achieving an average reduction of 3.28% in MSE and 2.29% in MAE. Furthermore, it shows strong generalization in few-shot learning settings: with 5% training data, we see a reduction in MSE and MAE by 4.13% and 1.91%, respectively; and with 10% data, by 3.62% and 1.98% on average. Code will be released upon publication.
Deploying Vision Transformers (ViTs) in real-world multi-task learning remains challenging due to their massive computational costs and the difficulty of pruning shared backbones without harming task performance. Single-task pruning often causes destructive interference by discarding weights critical to other tasks, while existing multi-task pruning strategies remain costly and unscalable for billion-parameter models. We propose Multi-LoRA Guided Importance Consensus (LoGIC), a unified framework for efficient and robust multi-task ViT pruning. LoGIC follows a two-phase procedure: (i) task-consistent pruning of LoRA modules, guided by a task-adaptive gating mechanism that balances shared and task-specific contributions while enforcing structured sparsity for deployment; and (ii) cross-task consensus pruning of the frozen ViT backbone, which retains both universally shared and task-specialized capabilities, enabling aggressive sparsity without sacrificing accuracy. Across five diverse vision benchmarks, LoGIC achieves up to 50% structured sparsity while maintaining competitive accuracy and surpassing all baselines.
Open-set semi-supervised learning (OSSL) leverages unlabeled data containing both in-distribution (ID) and unknown out-of-distribution (OOD) samples, aiming simultaneously to improve closed-set accuracy and detect novel OOD instances. Existing methods either discard valuable information from uncertain samples or force-align every unlabeled sample into one or a few synthetic “catch-all” representations, resulting in geometric collapse and overconfidence on only seen OODs. To address the limitations, we introduce selective non-alignment, adding a novel “skip” operator into conventional pull and push operations of contrastive learning. Our framework, SkipAlign, selectively skips alignment (pulling) for low-confidence unlabeled samples, retaining only gentle repulsion against ID prototypes. This approach transforms uncertain samples into a pure repulsion signal, resulting in tighter ID clusters and naturally dispersed OOD features. Extensive experiments demonstrate that SkipAlign significantly outperforms state-of-the-art methods in detecting unseen OOD data without sacrificing ID classification accuracy.
Over-smoothing in Graph Neural Networks (GNNs) causes collapse in distinct node features, particularly on heterophilic graphs where adjacent nodes often have dissimilar labels. Although sheaf neural networks partially mitigate this problem, they typically rely on static or heavily parameterized sheaf structures that hinder generalization and scalability. Existing sheaf-based models either predefine restriction maps or introduce excessive complexity, yet fail to provide rigorous stability guarantees. In this paper, we introduce a novel scheme called SGPC (Sheaf GNNs with PAC-Bayes Calibration), a unified architecture that combines cellular-sheaf message passing with several mechanisms, including optimal transport-based lifting, variance-reduced diffusion, and PAC-Bayes spectral regularization for robust semi-supervised node classification. We establish performance bounds theoretically and demonstrate that end-to-end training in linear computational complexity can achieve the resulting bound-aware objective. Experiments on nine homophilic and heterophilic benchmarks show that SGPC outperforms state-of-the-art spectral and sheaf-based GNNs while providing certified confidence intervals on unseen nodes.
Tabular data synthesis is a key technique for protecting data privacy and addressing class imbalance, yet existing generative models struggle to capture the complex intrinsic structure of the data. To overcome this limitation, we propose TabGeoFlow, a novel geometric flow matching model for tabular data synthesis. The core innovation of TabGeoFlow is the injection of an explicit geometric inductive bias into the conditional flow matching framework. We decompose the learned vector field into local tangent and normal components of the data manifold. By dynamically suppressing the predicted normal component via a controlling loss function, we constrain the generative path to follow the data's intrinsic structure. Implemented with a shared backbone for parameter efficiency, TabGeoFlow achieves competitive or better fidelity and utility, while exhibiting near random black box MIA accuracy and DCR ≈ 50%, suggesting reduced memorization without sacrificing quality.
Are Graph Transformers Necessary? Efficient Long-Range Message Passing with Fractal Nodes in MPNNs
PDF ↗Graph Neural Networks (GNNs) have emerged as powerful tools for learning on graph-structured data, but often struggle to balance local and global information. While graph Transformers aim to address this by enabling long-range interactions, they often overlook the inherent locality and efficiency of Message Passing Neural Networks (MPNNs). We propose a new concept called 'fractal nodes', inspired by the fractal structure observed in real-world networks. Our approach is based on the intuition that graph partitioning naturally induces fractal structure, where subgraphs often reflect the connectivity patterns of the full graph. Fractal nodes are designed to coexist with the original nodes and adaptively aggregate subgraph-level feature representations, thereby enforcing feature similarity within each subgraph. We show that fractal nodes alleviate the over-squashing problem by providing direct shortcut connections that enable long-range propagation of subgraph-level representations. Experiment results show that our method improves the expressive power of MPNNs and achieves comparable or better performance to graph Transformers while maintaining the computational efficiency of MPNN by improving the long-range dependencies of MPNN.
Departures: Distributional Transport for Single-Cell Perturbation Prediction with Neural Schrödinger Bridges
PDF ↗Predicting single-cell perturbation outcomes directly advances gene function analysis and facilitates drug candidate selection, making it a key driver of both basic and translational biomedical research. However, a major bottleneck in this task is the unpaired nature of single-cell data, as the same cell cannot be observed both before and after perturbation due to the destructive nature of sequencing. Although some neural generative transport models attempt to tackle unpaired single-cell perturbation data, they either lack explicit conditioning or depend on prior spaces for indirect distribution alignment, limiting precise perturbation modeling. In this work, we approximate Schrödinger Bridge (SB), which defines stochastic dynamic mappings recovering the entropy-regularized optimal transport (OT), to directly align the distributions of control and perturbed single-cell populations across different perturbation conditions. Unlike prior SB approximations that rely on bidirectional modeling to infer optimal source-target sample coupling, we leverage Minibatch-OT based pairing to avoid such bidirectional inference and the associated ill-posedness of defining the reverse process. This pairing directly guides bridge learning, yielding a scalable approximation to the SB. We approximate two SB models, one modeling discrete gene activation states and the other continuous expression distributions. Joint training enables accurate perturbation modeling and captures single-cell heterogeneity. Experiments on public genetic and drug perturbation datasets show that our model effectively captures heterogeneous single-cell responses and achieves state-of-the-art performance.
We study active mitigation of selection bias in statistical learning. That is sequential maximization over a set A of the expectation of a reward function R(a,X) w.r.t. a r.v. X drawn from a target distribution PT possibly different from the (supposedly dominating) source distribution PS under which rewards are observed. The importance function dPT/dPS (x) with which the sequentially observed biased rewards should be ideally weighted being unknown in practice, auxiliary information is assumed to be available in the form of known moments of the target distribution PT for debiasing purposes. In the batch setting, this problem has already been studied and can be solved under certain conditions in two successive steps: 1) identify a weight function so as to approximate the moments 2) maximize the resulting (empirical version of the) weighted reward. In the active setting, if the problem boils down to identifying the best arm in a stochastic multi armed bandit (MAB) model, the presence of selection bias strongly affects the complexity of the sequential optimization problem and requires the development of a new algorithmic approach, as we show here. In a fixed confidence setting, we introduce a novel notion of complexity, which accounts for the balance between arm evaluation and (parametric) weight function estimation, establish lower bounds and propose an algorithm proved to be near optimal. Theoretical guarantees are backed up by numerical results.
Low-rank tensor decompositions (TDs) provide an effective framework for multiway data analysis. Traditional TD methods rely on predefined structural assumptions, such as CP or Tucker decompositions. From a probabilistic perspective, these methods effectively model the relationships between latent factors and the low-rank tensor using Dirac delta distributions. However, tensor low-rank decomposition is inherently non-unique, leading to a multimodal distribution over possible solutions. Critically, such prior knowledge is rarely available in practical scenarios, particularly regarding the optimal rank structure and contraction rules. To address this issue, we propose a score-based model that eliminates the need for predefined structural or distributional assumptions, enabling the learning of compatibility between tensors and latent factors. Specifically, a neural network is designed to learn the energy function, which is optimized via score matching to capture the gradient of the joint log-probability of tensor entries and latent factors. Our method allows for modeling structures and distributions beyond the Dirac delta assumption. Moreover, integrating the block coordinate descent (BCD) algorithm with the proposed smooth regularization enables the model to perform both tensor completion and denoising. Experimental results demonstrate significant performance improvements across various tensor types, including sparse and continuous-time tensors, as well as visual data.