Instance normalization (IN) is widely used in non-stationary multivariate time series forecasting to reduce distribution shifts and highlight common patterns across samples. However, IN can over-smooth instance-specific structural information that is essential for modeling temporal and cross-channel heterogeneity. While prior methods further suppress distribution discrepancies or attempt to recover temporal specific dependencies, they often ignore a central tension: how to adaptively model common and instance-specific dependency based on each instance's non-stationary structures. To address this dilemma, we propose SeesawNet, a unified architecture that dynamically balances common and instance-specific dependency modeling in both temporal and channel dimensions. At its core is Adaptive Stationary–Nonstationary Attention (ASNA), which captures common dependencies from normalized sequences and specific dependencies from raw sequences, and adaptively fuses them according to instance-level non-stationarity. Built upon ASNA, SeesawNet alternates dedicated temporal and channel relationship modeling to jointly capture long-range and cross-variable dependencies. Extensive experiments on multiple real-world benchmarks demonstrate that SeesawNet consistently outperforms state-of-the-art methods.
论文检索
输入标题、作者或关键词，从 990 篇学术成果中精准定位
Federated learning (FL) enables collaborative model training without centralizing raw data, but privacy regulations such as the right to be forgotten require FL systems to remove the influence of previously used training data upon request. Retraining a federated model from scratch is prohibitively expensive, motivating federated unlearning (FU). However, existing FU methods suffer from high unlearning overhead, utility degradation caused by entangled knowledge, and unintended relearning during post-unlearning recovery. In this paper, we propose FedCARE, a unified and low overhead FU framework that enables conflict-aware unlearning and relearning-resistant recovery. FedCARE leverages gradient ascent for efficient forgetting when target data are locally available and employs data free model inversion to construct class level proxies of shared knowledge. Based on these insights, FedCARE integrates a pseudo-sample generator, conflict-aware projected gradient ascent for utility preserving unlearning, and a recovery strategy that suppresses rollback toward the pre-unlearning model. FedCARE supports client, instance, and class level unlearning with modest overhead. Extensive experiments on multiple datasets and model architectures under both IID and non-IID settings show that FedCARE achieves effective forgetting, improved utility retention, and reduced relearning risk compared to state of the art FU baselines. The source code can be found at https://github.com/thechosenchu/FedCARE.
Automatic radiology report generation (RRG) aims to simulate the workflow of radiologists, assisting them in clinical diagnosis. However, existing methods often fall short in utilizing all information relevant to the examination, as is typically done in clinical practice. Although some works attempt to incorporate multi-view images and historical data, these additional inputs may sometimes lead to avoidable diagnostic errors on the contrary. To address these challenges, we introduce a decision-making stage after report generation for the first time and propose a Posterior Diagnostic Decision framework (PDD-RRG) to integrate potentially conflicting diagnoses. Specifically, we create various subsets of input data and utilize an existing RRG model to generate reports from different perspectives. Then the Bayesian posterior probability and the learned thresholds for each clinical observation are calculated to obtain an aggregated diagnostic conclusion, which is subsequently used to refine the generated report. Experiments on MIMIC-CXR demonstrate that our proposed PDD-RRG can effectively enhance the clinical efficacy of existing RRG models without any retraining.
The learnware paradigm supports model reuse by pairing each submitted model with a specification, a lightweight representation used by the learnware dock system to identify, match, and reuse models without accessing raw data. While specifications are essential for learnware identification, they are also data-dependent public artifacts and it is not clear whether they reveal private information. Recently, the Parameter Vector (PAVE) specification has been proposed and shown to be effective for learnwares, yet its privacy properties remain largely unexplored. In this paper, we provide the first theoretical privacy analysis for PAVE. Specifically, we first formalize two specification-induced risks in the learnware paradigm: the disclosure risk of the released specification and the amplification risk that the specification may strengthen attacks against the released model. Second, we characterize when compact PAVE releases admit intrinsic differential privacy (DP): under natural structural conditions of learnware docks, the compact PAVE specification satisfies an (ε, δ)-DP guarantee without explicit additive noise through a Gaussian-sketch view of stable parameter variations, and for regimes outside these conditions, we further provide DP-Stabilized-PAVE as a certified differentially private variant. Third, we show that the resulting DP guarantees control both disclosure risk and amplification risk, and we analyze the induced privacy--utility trade-off to guide effective learnware identification while preserving privacy.
We study the allocation of m indivisible chores among n agents with additive disutilities under the fairness notion of envy-freeness up to one chore (EF1) and the efficiency notion of Pareto-optimality (PO). Although the existence of an allocation satisfying both EF1 and PO was recently established using a highly non-constructive fixed-point argument, an effective algorithm for computing such an allocation remains elusive, prompting the study of meaningful relaxations of these desiderata. Prior work introduced a natural relaxation through the concept of epistemic fairness: an allocation is said to be epistemic EF1 (EEF1), if for every agent i, it is possible to re-allocate the bundles of agents other than i such that i becomes EF1. In this work, we present a pseudo-polynomial time algorithm for computing an allocation of chores that is both EEF1 and PO. This gives an efficient polynomial-time algorithm for most practical settings where disutility values are integral and polynomially bounded in m and n. Our result employs the competitive equilibrium framework and relies on several technical insights that both utilize the distinct structure of epistemic EF1 and address the challenges it introduces.
LLMs' overconfidence, particularly when hallucinating, poses a significant challenge for the deployment of the models in safety-critical settings and makes a reliable estimation of uncertainty necessary. Existing approaches for uncertainty quantification typically prioritize lexical or probabilistic measures; however, these techniques often ignore the semantic variance of different responses with similar meaning. In this paper, we propose Adaptive Conformal Semantic Entropy (ACSE), a method for estimating prompt-level uncertainty by adaptively measuring semantic dispersion in LLMs outputs. Our uncertainty scoring function is based on clustering semantic entropy of multiple diverse responses to the same prompt. The function adaptively adjusts the uncertainty score based on semantic features of each cluster. To ensure statistical reliability of our score, we use conformal calibration to apply a decision rule to accept/abstain the prompts, providing a finite-sample, distribution-free guarantee such that the error rate among the accepted responses remains bounded by a user-specified tolerance. Our extensive experimental evaluations using different LLMs and datasets, demonstrate that our approach consistently outperforms state-of-the-art uncertainty quantification baselines using discriminative performance, acceptance rate, conformal guarantees, and probabilistic calibration indicators. As a highlight, for TriviaQA dataset, AUROC of our approach is 0.88 compared to 0.65 produced by the token entropy approach.
Qualitative preferences specified in CP-theory define a partial order over outcomes based on induced dominance relations. We investigate the problem of measuring divergence between such induced orders. We introduce two formally equivalent distance metrics, each leveraging a distinct computational strategy based on different types of normalization. Preliminary empirical results support the soundness of the proposed methodologies and offer comparative insights into their computational properties. To the best of our knowledge, this is the first study to compute distance measures between qualitative preferences that jointly accounts for conditional dependencies and relative importance statements.
Spatio-temporal graph networks form the foundation of modern traffic prediction, yet their deployment is fundamentally challenged by the pervasive reality of distribution shifts. While out-of-distribution (OOD) learning holds promise for robustness, existing methods rely on static graph structures, failing to capture the inherent topological dynamics of real-world traffic systems and thus limiting long-term deployment reliability. To bridge this gap, we propose DynaSTar, a Dynamic Spatio-Temporal Graph Invariant Learning model designed for reliable out-of-time (OOT) traffic prediction under evolving topologies. Our model employs a dynamic probabilistic graph structure, which is continuously refined through momentum-based updates and differentiable sparse sampling to model evolving inter-node dependencies. Besides, it utilizes node-level environment construction and modulated prediction to extract representations invariant to heterogeneous neighborhood fluctuations, enabling robust generalization. Comprehensive experiments on large-scale, long-term real-world datasets demonstrate that by effectively tracking evolving topological shifts, DynaSTar consistently outperforms state-of-the-art baselines across various OOT scenarios.
Current remote sensing (RS) perception systems suffer from task heterogeneity, necessitating distinct architectures for classification, localization, and reasoning. While vision--language models (VLMs) offer a route toward unification, their computational cost can hinder deployment. In this work, we propose RS-Florence, a compact unified model that addresses these tasks through a Prompt-Driven Sequence-to-Sequence framework. Unlike traditional approaches that segregate semantic understanding and geometric localization, RS-Florence maps images and task-specific prompts into a unified sequence of natural language and discrete geometric tokens. This formulation helps bridge high-level semantics and low-level pixel perception. Experiments across 4 task families and 8 benchmarks show that our 0.23B model remains competitive with task-specific specialists. These results also suggest that multitask joint training improves performance on several benchmarks over single-task fine-tuning, indicating that a shared prompt-driven interface can serve both language and geometry tasks.
Evaluating retrieval-augmented generation (RAG) systems in legal domains is challenging due to the nuanced nature of legal reasoning and scarcity of domain-specific benchmarks. We investigate semantic similarity and legal correctness in Brazilian legal question answering, developing a synthetic dataset of 3,012 evaluation instances from the Brazilian Civil Procedure Code spanning seven query types, validated through human expert assessment. Our evaluation framework combines BERTScore with domain-adapted LLM-as-Judge (GPT-4o-mini), validated against expert legal assessment. Analysis reveals 50.8% disagreement between semantic similarity and legal correctness, with correlation varying substantially by query type (ρ = 0.464 to ρ = 0.732). Explicit article citation emerges as the strongest quality predictor (Cohen’s d = 1.099), with cited responses achieving 145% higher legal correctness scores. Concrete examples demonstrate bidirectional divergence: responses may achieve high semantic similarity yet cite incorrect articles, or provide legally perfect answers with minimal lexical overlap. Findings demonstrate that semantic metrics alone are insufficient for legal RAG evaluation, with disagreement patterns structured by query type.
AI for materials science is a critical topic within AI for science, aiming to accelerate materials discovery and produce accurate property predictions. Bilayer 2D material stacking is essential for exploring new materials with novel functions and inherent phenomena, enabling the creation of new 2D bilayers for diverse real-world applications. Research on bilayer vdWs materials has made significant progress from experimental and computational perspectives. Various bilayer materials have been successfully synthesized experimentally and the increasing utilization of high-throughput computing technology has constructed several computational two-dimensional materials databases. However, the use of AI to model bilayer stacking and predict new properties remains underexplored, necessitating further research studies. In this work, we propose a novel multimodal learning approach to study the interfaces between dissimilar materials that jointly enable new or multiple functions, and to predict new properties arising from the vertical integration (stacking) of different functional material layers under given configurations. Comprehensive experiments demonstrate the effectiveness and efficiency of our approach compared to baseline methods. Our code is available at https://github.com/AnVuong123/bimat_ml.
Our work evolves around Fictitious Play, one of the first iterative methods that is known to converge to a Nash equilibrium in zero-sum games. In recent years, there has been a revived interest, due to applications in various machine learning problems, which has motivated a line of work on its convergence properties and on proposing new variants of the initial algorithm. Our paper is along this direction and introduces one new variant, which we refer to as Almost Greedy Fictitious Play. The proposed algorithm greedily attempts to find the optimal stepsize at each iteration but its search space is constrained and includes almost all the line between the cumulative mixed strategy and the current best response. Our main result is that the method achieves an instance dependent convergence rate of O(1/T) with respect to the duality gap. This matches the rate of Continuous Fictitious Play, and offers an alternative to discretization. We complement our theoretical findings with experiments that demonstrate the effectiveness of the method.
Structured claim decomposition is often proposed as a solution for verifying complex, multi-faceted claims, yet empirical results have been inconsistent. We argue that these inconsistencies stem from two overlooked bottlenecks: evidence alignment and sub-claim error profiles. To better understand these factors, we introduce a new dataset of real-world complex claims, featuring temporally bounded evidence and human-annotated sub-claim evidence spans. We evaluate decomposition under two evidence alignment setups: Sub-claim Aligned Evidence (SAE) and Repeated Claim-level Evidence (SRE). Our results reveal that decomposition brings significant performance improvement only when evidence is granular and strictly aligned. By contrast, standard setups that rely on repeated claim-level evidence (SRE) fail to improve and often degrade performance as shown across different datasets and domains (PHEMEPlus, MMM-Fact, COVID-Fact). Furthermore, we demonstrate that in the presence of noisy sub-claim labels, the nature of the error ends up determining downstream robustness. We find that conservative "abstention" significantly reduces error propagation compared to aggressive but incorrect predictions. These findings suggest that future claim decomposition frameworks must prioritize precise evidence synthesis and calibrate the label bias of sub-claim verification models.
We consider the problem of assessing a group of nodes in a network. Our focus is on vitality indices—a natural class of centrality measures that evaluate the importance of a node by examining the impact of its removal on the network. We conduct a comprehensive analysis of group vitality indices. Specifically, we show that every vitality index admits a unique extension to groups, which can be defined using a group variant of the Shapley value recently proposed in the literature. We also provide an axiomatization of the entire class, along with two specific group vitality indices that satisfy additional normalization conditions. Furthermore, we study the computational properties of all vitality indices, as well as Group Attachment Centrality.
Large Language Models (LLMs) achieve strong performance in many applications but remain limited in handling graph-structured data due to their reliance on textual context. Recent approaches integrate Graph Neural Networks (GNNs) to enhance structural modeling, yet they largely overlook fairness, leaving models vulnerable to bias amplification across graph and text modalities. To address this issue, we propose FairGEnt, a disentangled graph-enhanced large language model for fair graph learning. FairGEnt separates sensitive-related and sensitive-invariant factors in both graph and textual representations to mitigate bias while preserving task-relevant information, and further aligns the two modalities through a fairness-aware integration module. In addition, FairGEnt incorporates fair graph-enhanced instruction tuning to improve LLM understanding of complex graph structures. Experiments on multiple benchmark datasets demonstrate that FairGEnt consistently outperforms existing methods in both fairness and predictive performance.
Probabilities of causation (PoCs) are fundamental quantities for counterfactual analysis and personalized decision making. However, existing analytical results are largely confined to binary settings. This paper extends PoCs to multi-valued treatments and outcomes by deriving closed form bounds for a representative family of discrete PoCs within Structural Causal Models, using standard experimental and observational distributions. We introduce the notion of equivalence classes of PoCs, which reduces arbitrary discrete PoCs to this family, and establish a replaceability principle that transfers bounds across value permutations. For the resulting bounds, we prove soundness in all dimensions and empirically verify tightness in low dimensional cases via Balke's linear programming method; we further conjecture that this tightness extends to all dimensions. Simulations indicate that our closed form bounds consistently tighten recent recursive bounds while remaining simpler to compute. Finally, we illustrate the practical relevance of our results through toy examples.
Offline-to-online reinforcement learning harnesses the stability of offline pretraining and the flexibility of online fine-tuning. A key challenge lies in the non-stationary distribution shift between offline datasets and the evolving online policy. Common approaches often rely on static mixing ratios or heuristic-based replay strategies, which lack adaptability to different environments and varying training dynamics, resulting in suboptimal tradeoff between stability and asymptotic performance. In this work, we propose Reinforcement Learning with Optimized Adaptive Data-mixing (ROAD), a dynamic plug-and-play framework that automates the data replay process. We identify a fundamental objective misalignment in existing approaches. To tackle this, we formulate the data selection problem as a bi-level optimization process, interpreting the data mixing strategy as a meta-decision governing the policy performance (outer-level) during online fine-tuning, while the conventional Q-learning updates operate at the inner level. To make it tractable, we propose a practical algorithm using a multi-armed bandit mechanism. This is guided by a surrogate objective which simultaneously maintains offline priors and prevents value overestimation. Our empirical results demonstrate that this approach consistently outperforms existing data replay methods across various datasets, eliminating the need for manual, context-specific adjustments while achieving superior stability and asymptotic performance.
Temporal planning and scheduling often rely on constraint models to reason about activity durations. Simple Temporal Networks with Uncertainty (STNUs) address cases where some contingent durations are outside the control of the executing agent. The Multiple Interdependent STNU (MISTNU) model extends this setting to multi-agent systems by representing shared activities whose duration is decided by one agent and imposed on others. While controllability properties for MISTNUs have been defined, existing repair methods rely on a centralized SMT-based brute-force approach. This paper introduces a linear-constraint characterization of all negative cycles responsible for uncontrollability. Based on this formulation, we propose both a more efficient centralized linear-programming repair method and a distributed constraint reasoning approach that treats inconsistent cycles as inter-agent constraints. We experimentally compare several Distributed Constraint Optimization Problem (DCOP) solvers against the centralized baseline in terms of efficiency and repair quality.
Evaluation of tool-augmented Large Language Models (LLMs) has not advanced far beyond final answer accuracy, and neglects in-depth evaluation of reasoning ability despite it being a central claim of recent models. We aim to address this gap by developing a dependency graph-based evaluation to give insight into meta-level reasoning ability. We create a two-stage algorithm to construct dependency graphs from a series of LLM-generated tool calls: in the first stage creating nodes for tool calls, then in the second creating edges where tool call results are identified as being used as arguments in subsequent calls. The resulting graph facilitates the development of graph-theoretic metrics which allows deeper evaluation of the reasoning ability of LLMs over the task in question. We also use the dependency graph to assist with the generation of natural language explanations of the question-answering process. We create a baseline explanation generated from the list of tool calls, and a graph-augmented explanation which incorporates information about information provenance. A human evaluation study shows that our graph-augmented explanations are preferred across three key explainability criteria, although the extra detail contained impacts ease of understanding.
Recent reasoning-enhanced Large Language Models (LLMs) have achieved promising results in solving complex competitive programming problems. However, it remains unclear whether these reasoning abilities generalize to relevant tasks, like identifying algorithmically similar problems (ASPs). We introduce AlgoSimBench, a benchmark of 402 multiple-choice questions curated in an adversarial setting: each given reference problem is paired with one algorithmically similar problem and three distractors that are semantically close but algorithmically dissimilar. This design forces models to rely on algorithmic reasoning rather than superficial textual cues. Our evaluation shows that LLMs consistently struggle under this setting. To address this gap, we propose Attempted Solution Matching (ASM), which leverages LLM-generated solution attempts to assess similarity, yielding an average accuracy improvement of 9% across models. Beyond LLM evaluation, AlgoSimBench also probes code retrieval methods; when combined with BM25, ASM achieves an additional 11.8% gain over state-of-the-art embedding models. AlgoSimBench offers a challenging testbed that facilitates future studies on LLMs and retrieval methods.