Under the shift toward Industry 4.0, mass-customized manufacturing systems have introduced complex scheduling problems, such as the Flexible Job Shop Scheduling Problem (FJSSP). Recent Deep Reinforcement Learning (DRL)-based heuristics have shown promise, yet existing methods often suffer from two key limitations: they typically rely on single-policy optimization, which limits exploration, and on imprecise reward functions, which fail to accurately reflect decision quality. To address these challenges simultaneously, we propose PGMPO (Preference-Guided Multi-Policy Optimization), a novel learning framework consisting of (1) a simple but effective multi-policy modeling approach that allows a single network to represent multiple decision-makers, and (2) a preference-driven model optimization method that effectively guides policies to learn diverse and specialized problem-solving strategies without the need for explicit reward functions. Experimental results demonstrate that PGMPO substantially boosts the performance of existing neural solvers across several benchmarks.
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
Empowering Precise Embodied Agents with Executable Analytic Concepts as Semantic-Physical Blueprints
A core challenge for embodied agents is the ``semantic-to-physical gap"—the difficulty of mapping symbolic reasoning to precise execution. While Vision-Language Models (VLMs) enhance agent task planning, they often fail in problem classes requiring accurate alignment between functional geometry and physical constraints, such as articulated object manipulation or precision assembly. To address these challenges, we propose GRACE, an agent framework that adopts Executable Analytic Concepts (EAC) as a core knowledge annotation paradigm for object understanding. Under this paradigm, the agent does not merely perceive objects as unstructured visual data; instead, it interprets and annotates them as ``Semantic-Physical Blueprints", which provide a structured cognitive representation that encodes geometric primitives, mechanical affordances, and manipulation constraints. Central to our framework is a policy scaffolding mechanism, which allows the agent to dynamically ground VLM-based insights into these instantiated blueprints. This enables the agent to model and resolve complex decision-making tasks involving parameterized object descriptions and constrained motion planning. Experimental results demonstrate that agents utilizing this paradigm achieve high robustness in complex manipulation tasks. Furthermore, we demonstrate that GRACE supports automated concept discovery, offering a scalable and modular solution for high-precision embodied intelligence.
Trajectory representation learning transforms trajectory data into low-dimensional embeddings for downstream analytics. Although trajectory data inherently contains rich spatiotemporal information that remains to be more deeply explored, recent approaches have increasingly favored integrating external multimodal information rather than deeply mining internal trajectory patterns, resulting in the underutilization of the significant potential inherent in spatiotemporal features. To address these issues, we propose SimTRL, a novel unimodal framework specifically designed for road network-based trajectory data. Specifically, we design a bi-directional mamba-traj encoder that leverages the linear computational efficiency of State Space Models to capture global non-causal topological dependencies, overcoming the limitations of traditional causal models. Furthermore, we introduce a context-aware time encoder mechanism to explicitly model dynamic time embeddings. Extensive experiments conducted on two real-world datasets demonstrate that our model achieves average performance gains of 5.19%, 13.91%, 7.2%, and 2.69% across four tasks, respectively. Compared to multimodal models, our approach achieves an average improvement of 9.06% across all evaluation metrics, with a pre-training speedup of 11.2×. The source code of our model is available at https://github.com/sxgu1/SimTRL.
While class-agnostic object counting has recently evolved from image-exemplar to language-guided paradigms, existing methods are limited by text polysemy and the lack of prompts in audio-sensing scenarios. To overcome these challenges, we introduce a sound-centric counting paradigm, enabling models to "listen and count" using the vivid and discriminative signatures of sound cues. We propose S2ICount, which achieves seamless, fine-grained audio-visual alignment for sound-guided counting. At its core, SoundMamba leverages linear-complexity long-range semantic modelling to align cross-modal features, while the Sound Calibration Module designs a circular scanning mechanism to progressively reconcile modality discrepancies using multi-scale cues from a Stable Diffusion backbone. To enhance saliency awareness, a Multi-tier Hybrid Optimization strategy enforces consistency across magnitude, semantics, and distribution. In addition, we introduce SoundCount, the first large-scale sound-oriented counting dataset, comprising over 2,000 scenes and 1,650 audio sources across 66 object categories. Extensive experiments demonstrate that S2ICount achieves state-of-the-art performance, validating the effectiveness of sound as a robust guidance modality for counting. The dataset, demo and implementation code are made publicly accessible at https://github.com/ZSTU-CV-Lab/S2ICount.
Driven by privacy regulations, federated unlearning (FU) aims to remove the influence of specific clients or samples from a trained federated model, approximating the behavior of retraining from scratch without the target data. However, existing FU methods are largely reactive: retraining-based solutions are accurate but prohibitively expensive, while parameter-based approaches are more efficient yet may cause irreversible knowledge damage and catastrophic forgetting. We introduce PFU-SKLA, a proactive FU framework that endows models with built-in forgettability. During pretraining, we perform Orthogonality-guided Representation Disentanglement (ORD) to learn a robust, disentangled feature space that reduces cross-client/class interference. Additionally, we use Dynamic Sensitivity-based Key Layer Identification (DS-KLI) to identify sensitive layers that hold target knowledge. During unlearning, we propose a sparse, lightweight adaptation strategy that precisely erases target knowledge by inserting sparse adapters into the identified critical layers while preserving non-target knowledge by freezing the backbone. Extensive experiments across multiple datasets and three standard unlearning settings show that PFU-SKLA consistently approaches retraining-from-scratch performance, while substantially reducing communication and computation costs compared to state-of-the-art (SOTA) FU methods.
Accurately predicting the popularity of information cascades facilitates the development of social media network applications. During the cascade propagation process, user-generated comments continually evolve, thereby substantially affecting overall information popularity. However, existing methods primarily focus on learning structural–temporal cascade dynamics while neglecting the modeling of user-generated comment evolution, leading to suboptimal predictions. In this paper, we propose a novel framework, CasUGC, that aligns Cascade dynamics with User-Generated Comment evolution for popularity prediction. Specifically, we develop a dual-granularity alignment strategy that bridges the representation gap between comment semantics and structural–temporal dynamics at both the user and cascade levels. Building on this alignment, we further design a dynamic cascade feature generation module to produce temporally enhanced cascade embeddings. To further address the inherent uncertainties in both alignment and prediction, we introduce two diffusion-based components: a cascade-level diffusion network that learns a latent distribution mapping between semantic and structural–temporal representations, and a second diffusion process that captures the stochasticity of propagation dynamics for robust prediction. Extensive experiments conducted on four real-world social media datasets demonstrate the superiority of CasUGC over state-of-the-art methods.
Unmanned aerial vehicle swarms in pursuit-evasion requires encirclement efficiency while maintaining safety constraints, facing a critical safety-efficiency trade-off. Existing safe multi-agent reinforcement learning (MARL) methods often yield either unsafe task policies or conservative policies. This is challenging since forward feasible safety under dense interactions requires online safety preview. To address this issue, we propose Safety-Constrained Online Preview Enforcement (SCOPE), a MARL-based algorithm that balances encirclement efficiency and safety by short-horizon preview and safety enforcement. SCOPE learns an online safety-preview dynamics model that rolls out future trajectories to inform encirclement decisions checking. By proposing preview-fused actor-critic, SCOPE uses short-horizon previews for efficient encirclement and less unsafe behavior. Hierarchical safety enforcement performs safety look-ahead and online action correction to maintain forward feasible safety. Experiments show that SCOPE better balances encirclement efficiency and safety than safe MARL baselines, maintaining average encirclement time while reducing the average agent cost by 65.6%. Code could be found at https://github.com/98177qdn/SCOPE.
Parameter-efficient fine-tuning (PEFT) has emerged as an critical technique for adapting large-scale foundation models across natural language processing and computer vision. While existing methods such as low-rank adaptations achieve parameter efficiency via low-rank weight updates, they are limited in their ability to preserve the geometric structure of pretrained representations. We introduce Low-rank Compositional Orthogonal fine-tuning (LoCO), a novel PEFT method that constructs orthogonal transformations through low-rank skew-symmetric matrices and compositional rotation chains. We propose an approximation scheme that enables fully parallel computation of compositional rotations, making the approach practical for high-dimensional feature spaces. Our method maintains low computational complexity while maintaining orthogonality with controlled approximation error. We validate LoCO across diverse domains, including diffusion transformer fine-tuning, vision transformer adaptation, and language model adaptation. Our approach demonstrates superior or competitive performance compared to both existing orthogonal and non-orthogonal baselines.
Process Reward Models (PRMs) supervise intermediate reasoning steps in large language models (LLMs), but existing PRMs are mainly trained on general-domain data and struggle with the structured, symbolic, and fact-sensitive nature of financial reasoning. Financial tasks require not only correct final answers but also verifiable intermediate steps grounded in domain knowledge. In this paper, we propose Fin-PRM, a domain-specialized, trajectory-aware PRM for financial reasoning that jointly models step-level correctness and trajectory-level coherence, producing binary supervision signals for both local and global reasoning quality. To support reliable supervision, we construct a high-quality financial reasoning dataset of 3K trajectories, where step- and trajectory-level labels are automatically derived from multi-source reward signals, including Monte Carlo rollouts, LLM-based evaluation, and explicit financial knowledge verification. Fin-PRM defines a unified ranking score that integrates step- and trajectory-level rewards, enabling consistent use across multiple settings. We evaluate Fin-PRM in three scenarios: (1) offline trajectory selection for supervised fine-tuning, (2) reward-guided Best-of-N inference for test-time scaling, and (3) process-aware reward shaping for reinforcement learning. Experiments on financial reasoning benchmarks, including CFLUE and FinQA, show that Fin-PRM consistently outperforms general-purpose PRMs and strong baselines. Our project resources will be available at https://github.com/aliyun/qwen-dianjin.
End-to-end autonomous driving systems have demonstrated advantages over traditional modular systems. Despite this progress, these end-to-end systems still struggle to be deployed in real-world driving environments, as they inevitably encounter undertrained scenarios in which autonomous vehicles may take unsafe actions. Reinforcement Learning (RL) provides a theoretical framework for addressing this challenge by enabling autonomous vehicles to self-improve: continuously collecting additional scenarios and learning from them. However, training autonomous vehicles with RL is not straightforward in the real world. Collecting real-world driving data involves costly interactions with the environment, and significant human intervention is required both to prevent autonomous vehicles from entering unsafe states and to reset them for subsequent episodes. In this paper, we introduce a novel real-world RL algorithm that allows autonomous vehicles to collect informative scenarios and learn from them with minimal human intervention. Our algorithm considers the learning progress of autonomous vehicles to identify informative scenarios and abort episodes before they enter unsafe states. To evaluate our algorithm, we introduce challenging urban driving tasks that require autonomous vehicles to reset themselves to initial states. The experimental results show that our real-world RL algorithm outperforms baselines with much less human intervention.
Fair dynamic graph embedding is crucial for real-world systems, such as recommendation and social networks. Prior studies impose a single-axis fairness formulation, treating attribute and structural bias as separable artifacts. This overlooks their coupling relationship, under which debiasing along one axis can induce cross-axis amplification. This coupling further introduces opposing gradient constraints under joint optimization, leading to optimization conflicts. Furthermore, the evolution of dynamic graphs causes shifts in bias distribution, leading to unstable optimization and exacerbating these conflicts. To address these issues, we propose FairTCD, a Fair Dual-Teacher Temporal Contrastive Distillation framework. FairTCD employs two adversarial fairness teachers to decouple attribute and structural fairness representations. To reconcile dynamic conflicts between two fairness objectives, we introduce a temporal contrastive distillation to induce consistency between attribute and structural fairness representations across time while retaining temporal semantics. A unified student model distills complementary knowledge from both teachers to achieve twofold fairness. Experiments on three real-world benchmarks demonstrate that FairTCD preserves performance while improving twofold fairness metrics by at least 4.53%.
In this work, we propose the first constant-approximation algorithms, LinAst and LinAtg, which simultaneously achieve optimal query complexity O(n) and adaptive complexity O(log n) for non-monotone submodular maximization under a cardinality constraint k over a ground set of size n. Specifically, compared with existing algorithms that attain the best known adaptive complexity of O(log n), our approach preserves this adaptivity while reducing the query complexity from O(n log k) to O(n) and improving the approximation ratio from 0.172-epsilon to 0.193-epsilon. Our algorithms are built upon LinAdapt, which achieves a constant approximation ratio with O(log n) adaptive rounds and linear query complexity by requiring only O(1) candidate guesses of the optimal value. We further introduce the BoostAdapt algorithm, which improves the approximation guarantee to 0.25-\epsilon with O(log n log k) adaptive complexity and O(n log k) query complexity, based on a novel staggered greedy threshold framework that alternately constructs two disjoint solution sets over O(log k) sequential rounds. Extensive experiments on standard benchmark datasets demonstrate that our algorithms consistently outperform state-of-the-art methods in terms of solution quality, query complexity, and running time.
Time series forecasting aims to predict future values using historical observations, which is crucial for many practical applications with complex temporal dynamics. Recent frequency-domain forecasting methods have utilized spectral representations to model periodicity, but they usually rely on an implicit assumption: periodic components and aperiodic residuals are orthogonal in the frequency domain. However, in practice, this assumption often does not hold due to issues such as spectral leakage, superposition of multiple periods, and the gradual periodic changes. This paper re-examines frequency-domain modeling from the perspective of spectral non-identifiability and proposes a Spectral Memory Learner with Dual Retrieval (SMLDR) for time series forecasting. Instead of modeling periodicity in the frequency dimension, SMLDR represents periodic components as reusable spectral prototypes stored in learnable memory. The dual retrieval mechanism extracts historical periodic components and predicts them in the frequency domain, while a lightweight time-domain branch is used to model aperiodic dynamics to predict future residuals.Extensive experiments on multiple time series forecasting benchmarks show that SMLDR consistently achieves excellent performance, and the learnable memory can capture complex periodicity.
Reward models (RMs) play a pivotal role in aligning large language models (LLMs) with human preferences. Due to the difficulty of obtaining high-quality human preference annotations, distilling preferences from generative LLMs has emerged as a standard practice. However, existing approaches predominantly treat teacher models as simple binary annotators, failing to fully exploit the rich knowledge and capabilities for RM distillation. To address this, we propose RM-Distiller, a framework designed to systematically exploit the multifaceted capabilities of teacher LLMs: (1) Refinement capability, which synthesizes highly correlated response pairs to create fine-grained and contrastive signals. (2) Scoring capability, which guides the RM in capturing precise preference strength via a margin-aware optimization objective. (3) Generation capability, which incorporates the teacher's generative distribution to regularize the RM to preserve its fundamental linguistic knowledge. Extensive experiments demonstrate that RM-Distiller significantly outperforms traditional distillation methods both on RM benchmarks and reinforcement learning-based alignment, proving that exploiting multifaceted teacher capabilities is critical for effective reward modeling. To the best of our knowledge, this is the first systematic research on RM distillation from generative LLMs.
Safety-aware shared autonomy aims to enable an autonomous agent to collaborate with a human operator, completing tasks under safety constraints while maximally preserving human intent. However, existing methods often underperform in long-horizon tasks that require balancing task performance under safety constraints with the degree of intervention—particularly when accurately predicting future unsafe events is critical. This limitation largely stems from their reliance on the current state alone to decide when to intervene and how to modify actions. We propose World Model Assisted Safety Planning (WASP), a model-predictive shared autonomy framework for explicit safety constraint satisfaction. We first formulate a safety-aware world model, and leverage its online predictive reasoning to decide when to intervene by jointly assessing prospective safety violations and degradation in task performance, thereby intervening only when necessary while enforcing safety constraints at decision time. Once intervention is triggered, WASP plans a short-horizon residual correction using world model rollouts, filters out unsafe candidates, and executes the least-deviating correction among the remaining high-return options in a receding-horizon loop. Experiments across diverse vision-based safety-critical domains show that WASP substantially reduces safety violations while preserving task performance and reducing unnecessary interventions over prior shared autonomy baselines.
Current mainstream research in multivariate time series (MTS) forecasting often assumes that all data are static. However, real-world MTS data typically arrives continuously in a streaming manner, which we refer to as streaming MTS. The statistical characteristics and spatiotemporal graph topology of these data evolve over time, presenting two key challenges: the model's ability to adapt to new data distributions and the enhancement of cross-domain generalization capabilities. In this paper, we propose a streaming MTS prediction framework. We begin by designing a lightweight spatiotemporal causal learning model that captures generalizable causal spatiotemporal features from a decoupling perspective. Next, we introduce a framework to enhance the model's streaming learning capability, leveraging the adaptability of continual learning while strengthening cross-domain representation abilities. Specifically, we reformulate continual learning as a multi-task learning problem and present a multi-task optimization algorithm that identifies a set of Pareto-optimal solutions to address the inherent stability-plasticity dilemma in continual learning. Finally, we propose a topology-aware feature propagation strategy that disseminates well-trained node embedding features to unseen graph structures, thereby improving the model's cross-domain generalization. Results on real-world datasets demonstrate that our model achieves superior forecasting performance while substantially improving computational efficiency and memory usage.
Over-smoothing (OSM) and over-squashing (OSQ) are two fundamental phenomena that limit the performance of Graph Neural Networks (GNNs), yet a unified spectral-spatial understanding of these phenomena remains underexplored. In this paper, we adopt polynomial spectral filters as an analytical tool to establish a unified spectral-spatial framework for graph convolution and systematically characterize the effect of the polynomial order k on information propagation in GNNs. Within this framework, we reveal an intrinsic trade-off induced by the polynomial order. Specifically, higher-order filters enhance spectral expressiveness and alleviate OSM caused by the dominant low-frequency components. However, they also expand the spatial receptive field, thereby intensifying information compression and increasing the risk of OSQ. Based on this analysis, we provide a principled guideline for selecting the polynomial order and propose a Quadratic Spectral Graph Convolution Network (QS-GCN) for graph classification. Experiments demonstrate the effectiveness and robustness of the proposed method.
Most existing Graph Neural Networks (GNNs) rely on the node-level message passing or attention mechanisms to propagate and extract useful information. Although recent advances attempt to move beyond purely the node-level propagation by constructing high-level representations, these approaches are often constrained by pre-computed substructures or unidirectional bottom-up aggregations. Consequently, high-level structural semantics cannot effectively feed back to guide node representation learning, limiting the collaborative optimization between fine-grained features and macroscopic structural semantics. To address these limitations, we propose a novel Adaptive Dual-level Collaborative GNN (ADC-GNN) associated with an adaptive dual-level collaborative mechanism. We commence by introducing a set of global, learnable latent prototypes as high-level semantic references, and then employ a relaxed Sinkhorn algorithm to establish differentiable, non-collapsing assignments between nodes and prototypes. Based on these assignments, the ADC-GNN constructs high-level representations and enables interactions among them. We show that the ADC-GNN can inject the learned high-level information back into the node level, forming a closed-loop, bidirectional optimization process. Experiments demonstrate the superior performance of the proposed ADC-GNN on graph classification.
Recent personalized federated learning research focuses on heterogeneous models across clients. However, existing methods often rely on external data, model decoupling, and partial learning, which makes them sensitive to settings. In contrast, we revisit hypernetworks and leverage their strong generalization ability to propose the first practical method for personalized federated learning. We first propose a model heterogeneous personalized federated learning framework based on hypernetworks, MH-pFedHN, which quantifies clients with different architectures using customized embedding vectors and then generates client-specific model parameters through a server-side hypernetwork. Besides the shared feature extractor, our hypernetwork consists of multiple heads, where clients with similar numbers of parameters are assigned the same number of customized embedding vectors and consequently share the same head. This design enables knowledge sharing across different architectures and reduces the computation of parameter generation. To further enhance the hypernetwork’s learning and generalization, we propose MH-pFedHNGD, which introduces a lightweight yet effective plug-in global model. Our framework requires no external data and does not disclose client model architectures, thereby effectively ensuring security and demonstrating great potential. Experiments across various models and tasks demonstrate that our approach outperforms standard baselines and exhibits strong generalization performance. Our code is available at https://github.com/DangDang1895/MH-pFL
Learning universal representations for time series is fundamental for diverse downstream tasks. However, current approaches largely rely on handcrafted data augmentations, which may distort intrinsic temporal dynamics and structural regularities. In addition, most static representation learning frameworks struggle to cope with the non-stationary nature of real-world time series. To address these issues, we propose Heterogeneous Hypergraph Structure-aware Contrastive Adaptive Network (H²SCAN), a novel augmentation-free framework that derives contrastive supervision directly from graph topology. Specifically, H²SCAN constructs a heterogeneous hypergraph with three node types to capture multi-scale temporal characteristics. Building upon this representation, a meta-adaptation network is introduced to dynamically reweight heterogeneous hyperedges, enabling the model to adapt to distribution shifts in real-time. Finally, a structure-aware contrastive learning objective is employed to align latent representation similarity with the intrinsic hypergraph topology. Experiments on multiple benchmarks and cloud Kafka cluster datasets demonstrate that H²SCAN outperforms existing methods by modeling high-order multi-domain dependencies and preserving the semantic integrity of time series data.