论文检索

Yuichi Kamata, Quoc Hoan Tran, Yasuhiro Endo, Hirotaka Oshima

The Transformer model, renowned for its powerful attention mechanism, has achieved state-of-the-art performance in various artificial intelligence tasks but faces challenges with quantum data. With a growing focus on leveraging quantum machine learning for quantum data, particularly in quantum chemistry, we propose the Molecular Quantum Transformer (MQT) for modeling interactions in molecular quantum systems. By utilizing quantum circuits to implement the attention mechanism on the molecular configurations, MQT can efficiently calculate ground-state energies for all configurations. Numerical demonstrations show that in calculating ground-state energies for H2, LiH, BeH2, and H4, MQT outperforms the classical Transformer, highlighting the promise of quantum effects in Transformer structures. Furthermore, its pretraining capability on diverse molecular data facilitates the efficient learning of new molecules, extending its applicability to complex molecular systems with minimal additional effort. Our method offers an alternative to existing quantum algorithms for estimating ground-state energies, opening new avenues in quantum chemistry and materials science.

Rohan Juneja, Shivam Aggarwal, Safeen Huda, Tulika Mitra, Li-Shiuan Peh

Quantization is critical for efficiently deploying large language models (LLMs). Yet conventional methods remain hardware-agnostic, limited to bit-width constraints, and do not account for intrinsic circuit characteristics such as the timing behaviors and energy profiles of Multiply-Accumulate (MAC) units. This disconnect from circuit-level behavior limits the ability to exploit available timing margins and energy-saving opportunities, reducing the overall efficiency of deployment on modern accelerators. To address these limitations, we propose HALO, a versatile framework for Hardware-Aware Post-Training Quantization (PTQ). Unlike traditional methods, HALO explicitly incorporates detailed hardware characteristics, including critical-path timing and power consumption, into its quantization approach. HALO strategically selects weights with low critical-path-delays enabling higher operational frequencies and dynamic frequency scaling without disrupting the architecture's dataflow. Remarkably, HALO achieves these improvements with only a few dynamic voltage and frequency scaling (DVFS) adjustments, ensuring simplicity and practicality in deployment. Additionally, by reducing switching activity within the MAC units, HALO effectively lowers energy consumption. Evaluations on accelerators such as Tensor Processing Units (TPUs) and Graphics Processing Units (GPUs) demonstrate that HALO significantly enhances inference efficiency, achieving average performance improvements of 270% and energy savings of 51% over baseline quantization methods, all with minimal impact on accuracy.

Linglin Jing, Yuting Gao, Zhigang Wang, Wang Lan, Yiwen Tang, Weiyun Wang, Wenhai Wang, Qingpei Guo

Recent advancements have shown that the Mixture of Experts (MoE) approach significantly enhances the capacity of large language models (LLMs) and improves performance on downstream tasks. Building on these promising results, multi-modal large language models (MLLMs) have increasingly adopted MoE techniques. However, existing multi-modal MoE tuning methods typically face two key challenges: expert uniformity and router rigidity. Expert uniformity occurs because MoE experts are often initialized by simply replicating the FFN parameters from LLMs, leading to homogenized expert functions and weakening the intended diversification of the MoE architecture. Meanwhile, router rigidity stems from the prevalent use of static linear routers for expert selection, which fail to distinguish between visual and textual tokens, resulting in similar expert distributions for image and text. To address these limitations, we propose EvoMoE, an innovative MoE tuning framework. EvoMoE introduces a meticulously designed expert initialization strategy that progressively evolves multiple robust experts from a single trainable expert, a process termed expert evolution that specifically targets severe expert homogenization. Furthermore, we introduce the Dynamic Token-aware Router (DTR), a novel routing mechanism that allocates input tokens to appropriate experts based on their modality and intrinsic token values. This dynamic routing is facilitated by hypernetworks, which dynamically generate routing weights tailored for each individual token. Extensive experiments demonstrate that EvoMoE significantly outperforms other sparse MLLMs across a variety of multi-modal benchmarks, including MME, MMBench, TextVQA, and POPE. Our results highlight the effectiveness of EvoMoE in enhancing the performance of MLLMs by addressing the critical issues of expert uniformity and router rigidity.

Zikun Jin, Yuhua Qian, Xinyan Liang, Jiaqian Zhang, Jinpeng Yuan, Shen Hu, Haijun Geng, Honghong Cheng

Robust signal enhancement under non-stationary and low SNR conditions remains challenging, as methods based on the short-time Fourier transform (STFT) with fixed resolution struggle to represent complex and time–frequency structures. While leveraging the fractional domain as an auxiliary view offers flexibility in modeling time-frequency structures, existing methods typically adopt fixed transform orders and overlook alignment between views, hindering effective integration of complementary representations and leaving frequency domain misalignment unresolved. Therefore, we propose FracFusion, a novel framework that integrates a learnable short-time fractional Fourier Transform (STFrFT) module to generate dynamic auxiliary views, combined with two stage alignment-aware fusion modules: Pearson Channel Fusion for correlation-guided consistency and Efficient Align Fusion for fine-grained, frequency aligned interaction. Experiments on speech and electromagnetic (EM) datasets show that FracFusion consistently outperforms state-of-the-art baselines across diverse noise levels and signal types, demonstrating robust adaptability across domains.

Zewen Jin, Shen Fu, Chengjie Tang, Youhui Bai, Shengnan Wang, Jiaan Zhu, Chizheng Fang, Ping Gong, Cheng Li

To accelerate Mixture-of-Experts (MoE) inference, the hybrid parallelism paradigm is first applying pipeline parallelism (PP) to vertically divide the model into stages, with each stage further divided horizontally using tensor or expert parallelism. On the algorithm side, dynamic Top-K routing reduces computation by activating fewer experts per token on average. In this paper, we explore the application of dynamic Top-K routing to PP-enabled MoE inference, aiming to fully unleash their combined potential. We identify key performance bottlenecks arising from Top-K value variation across layers, which conflicts with PP's typically uniform stage partitioning, as well as opportunities to optimize memory usage through their integration. To address these challenges, we present SMIDT, an efficient MoE inference framework tailored for dynamic Top-K routing. SMIDT features: (1) an adaptive, module-level uneven partitioning strategy to balance computation across PP stages, (2) a memory-aware expert replication scheme (DPMoE) that reduces communication overhead, and (3) a lightweight search algorithm combining binary search and dynamic programming to generate efficient parallelism plans. We implement SMIDT on SGLang, a state-of-the-art LLM inference framework, evaluate it on 32 A40 GPUs and 16 A100 GPUs, and compare with manually tuned parallelism strategies. Experimental results show that, when co-locating prefill and decoding phases, SMIDT achieves 1.20–3.13x throughput improvements for prefill-only tasks and 1.05–1.89x for prefill-decoding tasks. When disaggregating prefill and decoding tasks, SMIDT improves average and P99 time-to-first-token (TTFT) by 1.10–1.17x and 1.21–1.26x, respectively.

Yue Jin, Giovanni Montana

Offline multi-agent reinforcement learning (MARL) is severely hampered by the challenge of evaluating out-of-distribution (OOD) joint actions. Our core finding is that when the behavior policy is factorized—a common scenario where agents act fully or partially independently during data collection—a strategy of partial action replacement (PAR) can significantly mitigate this challenge. PAR updates a single or part of agents' actions while the others remain fixed to the behavioral data, reducing distribution shift compared to full joint-action updates. Based on this insight, we develop Soft-Partial Conservative Q-Learning (SPaCQL), using PAR to mitigate OOD issue and dynamically weighting different PAR strategies based on the uncertainty of value estimation. We provide a rigorous theoretical foundation for this approach, proving that under factorized behavior policies, the induced distribution shift scales linearly with the number of deviating agents rather than exponentially with the joint-action space. This yields a provably tighter value error bound for this important class of offline MARL problems. Our theoretical results also indicate that SPaCQL adaptively addresses distribution shift using uncertainty-informed weights. Our empirical results demonstrate SPaCQL enables more effective policy learning, and manifest its remarkable superiority over baseline algorithms when the offline dataset exhibits the independence structure.

Shengkai Jin, Tianyu Chen, Chonghan Gao, Jun Han

Large language models excel at semantic reasoning yet struggle with numerical tasks because tokenization disrupts geometric continuity. Traditional methods fragment numerically close values into inconsistent token sequences, severing the correspondence between numerical proximity and representational similarity, which is essential for numerical cognition. We introduce GeoNum, a geometrically coherent numerical embedding based on polar coordinate decomposition. By encoding integer magnitudes through classification and fractional components via trigonometric regression, GeoNum constructs a continuous manifold where numerical distance is preserved geometrically. A three-stage framework progressively integrates GeoNum into pretrained language models via self-supervised pretraining, projection alignment, and efficient adaptation. Experimental results across diverse arithmetic benchmarks demonstrate consistent gains in high-precision accuracy and improved interpolation and extrapolation, underscoring the promising benefits of geometric continuity for numerical modeling in large language models.

Pengfei Jin, Peng Shu, Sifan Song, Sekeun Kim, Qing Xiao, Cheng Chen, Tianming Liu, Xiang Li, Quanzheng Li

Recent advances in parameter-efficient transfer learning have demonstrated the utility of composing LoRA adapters from libraries of pretrained modules. However, most existing approaches rely on simple retrieval heuristics or uniform averaging, which overlook the latent structure of task relationships in representation space. We propose a new framework for adapter reuse that moves beyond retrieval, formulating adapter composition as a geometry-aware sparse reconstruction problem. Specifically, we represent each task by a latent prototype vector derived from the base model’s encoder and aim to approximate the target task prototype as a sparse linear combination of retrieved reference prototypes, under an L1-regularized optimization objective. The resulting combination weights are then used to blend the corresponding LoRA adapters, yielding a composite adapter tailored to the target task. This formulation not only preserves the local geometric structure of the task representation manifold, but also promotes interpretability and efficient reuse by selecting a minimal set of relevant adapters. We demonstrate the effectiveness of our approach across multiple domains—including medical image segmentation, medical report generation and image synthesis. Our results highlight the benefit of coupling retrieval with latent geometry-aware optimization for improved zero-shot generalization.

Kun Jin, Tian Xie, Yang Liu, Xueru Zhang

In many real-world applications of machine learning—such as recommendations, hiring, and lending—deployed models influence the data they are trained on, leading to feedback loops between predictions and data distribution. The performative prediction (PP) framework captures this phenomenon by modeling the data distribution as a function of the deployed model. While prior work has focused on finding performative stable (PS) solutions for robustness, their societal impacts, particularly regarding fairness, remain underexplored. We show that PS solutions can lead to severe polarization and prediction performance disparities, and that conventional fairness interventions in previous works often fail under model-dependent distribution shifts due to failing the PS criteria. To address these challenges in PP, we introduce novel fairness mechanisms that provably ensure both stability and fairness, validated by theoretical analysis and empirical results.

Kaifeng Jin, Ignavier Ng, Kun Zhang, Biwei Huang

Recent advances in differentiable structure learning have framed the combinatorial problem of learning directed acyclic graphs as a continuous optimization problem. Various aspects, including data standardization, have been studied to identify factors that influence the empirical performance of these methods. In this work, we investigate critical limitations in differentiable structure learning methods, focusing on settings where the true structure can be identified up to Markov equivalence classes, particularly in the linear Gaussian case. While recent work highlighted potential non-convexity issues in this setting, we demonstrate and explain why the use of L1-penalized likelihood in such cases is fundamentally inconsistent, even if the global optimum of the optimization problem can be found. To resolve this limitation, we develop a hybrid differentiable structure learning method based on L0-penalized likelihood with hard acyclicity constraint, where the L0 penalty can be approximated by different techniques including Gumbel-Softmax. Specifically, we first estimate the underlying moral graph, and use it to restrict the search space of the optimization problem, which helps alleviate the non-convexity issue. Experimental results show that the proposed method enhances empirical performance both before and after data standardization, providing a more reliable path for future advancements in differentiable structure learning, especially for learning Markov equivalence classes.

Cheng Jin, Zhenyu Xiao, Yuantao Gu

Diffusion models deliver state-of-the-art generative performance across diverse modalities but remain computationally expensive due to their inherently iterative sampling process. Existing training-free acceleration methods typically improve numerical solvers for the reverse-time ODE, yet their effectiveness is fundamentally constrained by the inefficiency of the underlying sampling trajectories. We propose A-FloPS (Adaptive Flow Path Sampler), a principled, training-free framework that reparameterizes the sampling trajectory of any pre-trained diffusion model into a flow-matching form and augments it with an adaptive velocity decomposition. The reparameterization analytically maps diffusion scores to flow-compatible velocities, yielding integration-friendly trajectories without retraining. The adaptive mechanism further factorizes the velocity field into a linear drift term and a residual component whose temporal variation is actively suppressed, restoring the accuracy benefits of high-order integration even in extremely low-NFE regimes. Extensive experiments on conditional image generation and text-to-image synthesis show that A-FloPS consistently outperforms state-of-the-art training-free samplers in both sample quality and efficiency. Notably, with as few as 5 function evaluations, A-FloPS achieves substantially lower FID and generates sharper, more coherent images. The adaptive mechanism also improves native flow-based generative models, underscoring its generality. These results position A-FloPS as a versatile and effective solution for high-quality, low-latency generative modeling.

Ying Jiao, Rodrigo Castellano Ontiveros, Luc De Raedt, Marco Gori, Francesco Giannini, Michelangelo Diligenti, Giuseppe Marra

Neurosymbolic (NeSy) AI combines neural architectures and symbolic reasoning to improve accuracy, interpretability, and generalization. While logic inference on top of subsymbolic modules has been shown to effectively guarantee these properties, this often comes at the cost of reduced scalability, which can severely limit the usability of NeSy models. This paper introduces DeepProofLog (DPrL), a novel NeSy system based on stochastic logic programs, which addresses the scalability limitations of previous methods. DPrL parameterizes all derivation steps with neural networks, allowing efficient neural guidance over the proving system. Additionally, we establish a formal mapping between the resolution process of our deep stochastic logic programs and Markov Decision Processes, enabling the application of dynamic programming and reinforcement learning techniques for efficient inference and learning. This theoretical connection improves scalability for complex proof spaces and large knowledge bases. Our experiments on standard NeSy benchmarks and knowledge graph reasoning tasks demonstrate that DPrL outperforms existing state-of-the-art NeSy systems, advancing scalability to larger and more complex settings than previously possible.

Jiajun Jiao, Haowei Zhu, Puyuan Yang, Jianghui Wang, Ji Liu, Ziqiong Liu, Dong Li, Yuejian Fang, Jun-Hai Yong, Bin Wang 等

Diffusion models have achieved remarkable success in image and video generation. However, their inherently multiple step inference process imposes substantial computational overhead, hindering real-world deployment. Accelerating diffusion models is therefore essential, yet determining how to combine multiple model acceleration techniques remains a significant challenge. To address this issue, we introduce a framework driven by large language models (LLMs) for automated acceleration code generation and evaluation. First, we present DiffBench, a comprehensive benchmark that implements a three stage automated evaluation pipeline across diverse diffusion architectures, optimization combinations and deployment scenarios. Second, we propose DiffAgent, an agent that generates optimal acceleration strategies and codes for arbitrary diffusion models. DiffAgent employs a closed-loop workflow in which a planning component and a debugging component iteratively refine the output of a code generation component, while a genetic algorithm extracts performance feedback from the execution environment to guide subsequent code refinements. We provide a detailed explanation of the DiffBench construction and the design principles underlying DiffAgent. Extensive experiments show that DiffBench offers a thorough evaluation of generated codes and that DiffAgent significantly outperforms existing LLMs in producing effective diffusion acceleration strategies.

Tao Jiang, Zichuan Lin, Lihe Li, Yi-Chen Li, Cong Guan, Lei Yuan, Zongzhang Zhang, Yang Yu, Deheng Ye

Large transformer models, trained on diverse datasets, have demonstrated impressive few-shot performance on previously unseen tasks without requiring parameter updates. This capability has also been explored in Reinforcement Learning (RL), where agents interact with the environment to retrieve context and maximize cumulative rewards, showcasing strong adaptability in complex settings. However, in cooperative Multi-Agent Reinforcement Learning (MARL), where agents must coordinate toward a shared goal, decentralized policy deployment can lead to mismatches in task alignment and reward assignment, limiting the efficiency of policy adaptation. To address this challenge, we introduce Multi-agent In-context Coordination via Decentralized Memory Retrieval (MAICC), a novel approach designed to enhance coordination by fast adaptation. Our method involves training a centralized embedding model to capture fine-grained trajectory representations, followed by decentralized models that approximate the centralized one to obtain team-level task information. Based on the learned embeddings, relevant trajectories are retrieved as context, which, combined with the agents' current sub-trajectories, inform decision-making. During decentralized execution, we introduce a novel memory mechanism that effectively balances test-time online data with offline memory. Based on the constructed memory, we propose a hybrid utility score that incorporates both individual- and team-level returns, ensuring credit assignment across agents. Extensive experiments on cooperative MARL benchmarks, including Level-Based Foraging (LBF) and SMAC (v1/v2), show that MAICC enables faster adaptation to unseen tasks compared to existing methods.

Tao Jiang, Yucheng Jiang, Xiwen Yao, Gong Cheng, Junwei Han

Post-Training Quantization enables efficient Vision Transformer (ViTs) deployment with a small calibration data, and its prevalent use of uniform quantization harnesses AI accelerator matrix cores for high-speed inference. However, the application of uniform quantization is fundamentally challenged by the extreme non-uniformity of activation distributions.Specifically, the power-law nature of post-Softmax attention scores and the significant inter-channel variance in post-GELU activations create a dilemma for conventional quantization, as it struggles to preserve critical high-magnitude values without sacrificing overall precision. To resolve this core conflict, we introduce UQ-ViT (Uniform Quantization for Vision Transformers), a novel uniform quantization framework designed to reconcile high precision with hardware efficiency. Central to UQ-ViT are two operators: Dynamic Elimination of Maximum (DeMax) and Normalization Quantization (NormQuant). DeMax is a quantization operator for post-Softmax attention scores that utilizes uniform quantization. It dynamically eliminates and preserves dominant values, effectively mitigating quantization loss from the extreme values in the power-law distribution. NormQuant utilizes a per-channel quantization strategy during quantization and reverts to a per-tensor format for dequantization, achieving both high accuracy and computational efficiency. Crucially, it is applicable to any linear layer, enabling effective quantization of post-GELU activations in ViTs. Through extensive experiments on various ViTs and vision tasks, including image classification, object detection, and instance segmentation, we demonstrate that our proposed approach outperforms existing methods, achieving superior accuracy while ensuring hardware friendliness.

Mingdie Jiang, Quanjiang Li, Tingjin Luo, Yiping Song, Chenping Hou

Feature dynamics have emerged as a critical topic about open-environment learning due to the instability of feature availability. While traditional feature evolution targets single-label tasks, multi-label learning is essential to accommodate the exploding annotation spaces. However, multi-label classification with incremental and decremental features is a crucial yet underexplored problem, which poses the challenge of preserving feature representations and label correlations from historical instances and simultaneously adapting to newly arriving streaming data. To address these issues, we propose a two-stage, one-pass learning approach termed MLID. It attempts to compress the informative content of vanished features into the domain of survived ones, facilitate the propagation of label dependencies via low-rank regularization of the classifier, and incorporate augmented features to construct an adaptive classification mechanism. Besides, we design optimization strategies for each stage and provide theoretical guarantees of convergence. Moreover, we establish the generalization error bound of MLID and demonstrate that the compactness of the trace norm and the reuse of models based on effective features can enhance the generalization performance. Finally, we extend it to multi-shot case and extensive experimental results validate the superiority of our MLID.

Lifan Jiang, Mengying Zhu, Yangyang Wu, Xuan Liu, Xiaolin Zheng, Shenglin Ben

Few-shot graph learning remains a fundamental yet challenging problem, especially under heterophilic graph settings where connected nodes are likely to belong to different classes. In such scenarios, two key challenges arise: (1) unreliable or noisy graph structures that hinder effective message passing, and (2) semantic inconsistency: in heterophilic graphs, aggregating messages from neighbors of different classes entangles representations and introduces misleading semantics. These issues are further exacerbated by the limited labeled data inherent to few-shot learning, making it difficult to adaptively repair structure or disentangle semantics. To address these challenges, we propose DAPrompt, a Dual Alignment Prompt framework that jointly calibrates graph structure and semantic representations across the learning pipeline. In the pretraining stage, DAPrompt incorporates a graph structure learning module to denoise and repair the underlying topology, enhancing structural reliability. In the prompt tuning stage, we introduce two coordinated modules: a structure-aware prompt learner, which employs prompt tokens to repair unreliable graph structures and capture structure-level alignment, and a semantics-aligned prompt learner, which enhances the graph using target node semantics to mitigate representation noise caused by class-mismatched propagation. Extensive experiments on both node-level and graph-level few-shot benchmarks validate its effectiveness, achieving state-of-the-art performance and highlighting the value of structure-semantic dual alignment in heterophilic few-shot graph learning.

Ke Jiang, Wen Jiang, Xiaoyang Tan

The performance of Offline reinforcement learning is significantly impacted by the issue of state distributional shift, and out-of-distribution (OOD) state correction is a popular approach to address this problem. However, previous methods correct the agent's transition distributions in a supervised way, which significantly degrades the flexibility and robustness. In this paper, we propose a novel method named Density-Aware Safety Perception (DASP) for OOD state correction. Specifically, our method encourages the agent to prioritize actions that lead to outcomes with higher data density, thereby promoting its operation within or the return to in-distribution (safe) regions. To achieve this, we optimize the objective within a variational framework that concurrently considers both the potential outcomes of decision-making and their density, thus providing crucial contextual information for safe decision-making. Finally, we validate the effectiveness and feasibility of our proposed method through extensive experimental evaluations on the offline MuJoCo and AntMaze suites.

Jiecheng Jiang, Jiawei Tang, Jiahao Jiang, Hui Liu, Junhui Hou, Yuheng Jia

Label distribution learning (LDL) is a novel paradigm that describe the samples by label distribution of a sample. However, acquiring LDL dataset is costly and time-consuming, which leads to the birth of incomplete label distribution learning (IncomLDL). All the previous IncomLDL methods set the description degrees of "missing" labels in an instance to 0, but remains those of other labels unchanged. This setting is unrealistic because when certain labels are missing, the degrees of the remaining labels will increase accordingly. We fix this unrealistic setting in IncomLDL and raise a new problem: LDL with hidden labels (HidLDL), which aims to recover a complete label distribution from a real-world incomplete label distribution where certain labels in an instance are omitted during annotation. To solve this challenging problem, we discover the significance of proportional information of the observed labels and capture it by an innovative constraint to utilize it during the optimization process. We simultaneously use local feature similarity and the global low-rank structure to reveal the mysterious veil of hidden labels. Moreover, we **theoretically** give the recovery bound of our method, proving the feasibility of our method in learning from hidden labels. Extensive recovery and predictive experiments on various datasets prove the superiority of our method to state-of-the-art LDL and IncomLDL methods.

Haoning Jiang, Han Wu, Zhuoli Ouyang, Ziheng Wang, Tinghuan Chen, Junmin Jiang

This paper introduces the Functionality-Driven Multi-Agent Group Relative Policy Optimization (FD-MAGRPO) algorithm, which is designed to enhance exploration efficiency in reinforcement learning (RL) for analog integrated circuit sizing. Our proposed method integrates two key innovations: (1) a critic-free multi-agent optimization framework based on Group Relative Policy Optimization (GRPO), that eliminates the critic network and achieves stable and efficient policy updates; and (2) a functionality-driven grouping strategy, that enables agents to coordinate exploration by functional roles instead of circuit blocks, thereby improving credit assignment and cooperation. Experimental results on practical low-dropout regulator (LDO) circuits with 65–179 design parameters show that the proposed method achieves rapid convergence with only 800–3000 simulations, yielding a 4.8×–13.0× speedup over state-of-the-art methods. Mathematical analysis and empirical studies validate that the combination of critic-free optimization and functionality-based grouping leads to higher exploration efficiency and faster convergence. The proposed method enables the discovery of higher circuit performances that are inaccessible to conventional approaches, establishing FD-MAGRPO as a robust and efficient solution for complex analog-LDO sizing tasks.