论文检索

Jianzhou Wang, Yirui Wu, Lixin Yuan, Wenxiao Zhang, Jun Liu, Junyang Chen, Huan Wang, Wenhai Wang

Incremental Object Detection(IOD) targets at progressively extending capability of object detectors to recognize new classes. However, representation confusion between old and new classes leads to catastrophic forgetting. To alleviate this problem, we propose DiffKA, with intrinsic knowledge generated and aggregated by forward and backward diffusion, gradually establishing rigid class boundary. With incremental streaming data, forward diffusion spreads information to generate potential inter-class associations among new- and old-class prototypes within a hierarchical tree, named as Intrinsic Correlation Tree(ICTree), to store intrinsic knowledge. Afterwards, backward diffusion refines and aggregates the generated knowledge in ICTree, explicitly establishing rigid class boundary to mitigate representation confusion. To keep semantic consistency with extreme IOD settings, we reorganize semantic relevance of old- and new-class prototypes in paradigms to adaptively and effectively update DiffKA. Experiments on MS COCO dataset show DiffKA achieves state-of-the-art performance on IOD tasks with significant advantages.

Sergio M. Vanegas Arias, Lasse Lensu, Fredy Ruiz Palacios

Multiscale signals represent a formidable modelling challenge in Machine Learning as the ubiquitous Mean Squared Error loss function neglects signal behaviour at smaller values. Several scale-equalizing error metrics have been devised to tackle this problem, amongst which the Mean Absolute Percentage Error (MAPE) remains the most widely used due to its simplicity and interpretability. However, by its very definition, MAPE introduces three major issues: asymptotic behaviour at zero-target values, asymptotic gradient behaviour at zero error, and accuracy loss for large signal scales. We address these limitations by proposing the Symmetric Mean Arctangent Squared Percentage Error (SMASPE), which builds up from the Mean Arctangent Absolute Percentage Error (MAAPE) and leverages a mathematically smoother definition along with user-provided signal bounds to extend its functionality. The numerical properties of SMASPE are explored, and its performance is tested in two real-life cases for deterministic and stochastic optimization. The experiments show a clear advantage of the proposed loss function, with an improvement of up to 42% with respect to MAAPE in terms of Mean Absolute Error for deep learning models when appropriate bounds are selected.

Tao Tong, Yujie Mo, Yucheng Xie, Songyue Cai, Xiaoshuang Shi, Xiaofeng Zhu

Deep neural networks can easily lead to the over-fitting issue due to the influence of noisy labels. However, previous label correction methods for dealing with noisy labels often need expensive computation cost to achieve effectiveness and ignore the generalization ability of the model. To address these issues, in this paper, we propose a new meta-based self-correction method to achieve accurate filtering of noisy labels and to enhance the generalization ability of the label correction model. Specifically, we first investigate a new gradient score method to filter noisy labels with less computation cost, and then theoretically design a new generalization regularizer into the meta-learner and the base learner, for correcting noisy labels as well as achieving the generalization ability. Experimental results on real datasets verify the effectiveness of our proposed method in terms of different classification tasks.

Zilin Tian, Yunfei Long, Liguo Zhang, Jiahong Zhao

Audio adversarial examples impose acoustically imperceptible perturbations to clean audio examples, fooling classification models into producing incorrect results. Transferability is a critical property of audio adversarial examples, making black-box attacks applicable in practice and attracting increasing interest. Despite recent studies achieving transferability across models within the same domain, they consistently fail to achieve transferability across different domains. Given that time-domain and frequency-domain models are the two predominant approaches in audio classification, we observe that adversarial examples generated for one domain demonstrate significantly constrained transferability to the other. To address this limitation, we propose an Adaptive Inter-domain Ensemble (AIE) attack, which integrates transferable adversarial information from both domains and dynamically optimizes their contributions through adaptive weighting, improving the cross-domain transferability of audio adversarial examples. Extensive evaluations on diverse datasets consistently demonstrate that AIE outperforms existing methods, establishing its effectiveness in enhancing adversarial transferability across domains.

Martin Tappler, Ignacio D. Lopez-Miguel, Sebastian Tschiatschek, Ezio Bartocci

We consider the challenging problem of using domain knowledge to improve deep reinforcement learning policies. To this end, we propose LEGIBLE, a novel approach, following a multi-step process, which starts by mining rules from a deep RL policy, constituting a partially symbolic representation. These rules describe which decisions the RL policy makes and which it avoids making. In the second step, we generalize the mined rules using domain knowledge expressed as metamorphic relations. We adapt these relations from software testing to RL to specify expected changes of actions in response to changes in observations. The third step is evaluating generalized rules to determine which generalizations improve performance when enforced. These improvements show weaknesses in the policy, where it has not learned the general rules and thus can be improved by rule guidance. LEGIBLE supported by metamorphic relations provides a principled way of expressing and enforcing domain knowledge about RL environments. We show the efficacy of our approach by demonstrating that it effectively finds weaknesses, accompanied by explanations of these weaknesses, in eleven RL environments and by showcasing that guiding policy execution with rules improves performance w.r.t. gained reward.

Lue Tao, Tian-Zuo Wang, Yuan Jiang, Zhi-Hua Zhou

Machine learning has advanced in predictive tasks, but practitioners often need to proactively avoid undesired outcomes rather than just predicting them. To this end, a framework called rehearsal has been introduced, which tackles the avoiding undesired future (AUF) problem by modeling how variables influence each other and searching for a decision that leads to desired results. In this paper, we propose a novel rehearsal approach for addressing the AUF problem by making a sequence of decisions, where each decision is dynamically informed by the latest observations via retrospective inference. Theoretically, we show that sequential decisions in our approach tend to achieve a higher success rate in avoiding undesired outcomes by more reliably inferring the outcome of actions compared with existing solutions. Perhaps surprisingly, our approach remains advantageous even under imprecise modeling of relations between variables, and we provide a sufficient condition under which the advantage holds. Finally, experimental results confirm the practical effectiveness of the proposed approach in both simulated and real-world tasks.

Huidong Tang, Chen Li, Sayaka Kamei, Yoshihiro Yamanishi, Yasuhiko Morimoto

Deep generative models, such as generative adversarial networks (GANs), have been employed for de~novo molecular generation in drug discovery. Most prior studies have utilized reinforcement learning (RL) algorithms, particularly Monte Carlo tree search (MCTS), to handle the discrete nature of molecular representations in GANs. However, due to the inherent instability in training GANs and RL models, along with the high computational cost associated with MCTS sampling, MCTS RL-based GANs struggle to scale to large chemical databases. To tackle these challenges, this study introduces a novel GAN based on actor-critic RL with instant and global rewards, called InstGAN, to generate molecules at the token-level with multi-property optimization. Furthermore, maximized information entropy is leveraged to alleviate the mode collapse. The experimental results demonstrate that InstGAN outperforms other baselines, achieves comparable performance to state-of-the-art models, and efficiently generates molecules with multi-property optimization. The code is available at: https://github.com/tang777777/InstGAN.

Hao Tang, Shengfeng He, Jing Qin

Few-shot learning (FSL) addresses the challenge of classifying novel classes with limited training samples. While some methods leverage semantic knowledge from smaller-scale models to mitigate data scarcity, these approaches often introduce noise and bias due to the data’s inherent simplicity. In this paper, we propose a novel framework, Synergistic Knowledge Transfer (SynTrans), which effectively transfers diverse and complementary knowledge from large multimodal models to empower the off-the-shelf few-shot learner. Specifically, SynTrans employs CLIP as a robust teacher and uses a few-shot vision encoder as a weak student, distilling semantic-aligned visual knowledge via an unsupervised proxy task. Subsequently, a training-free synergistic knowledge mining module facilitates collaboration among large multimodal models to extract high-quality semantic knowledge. Building upon this, a visual-semantic bridging module enables bi-directional knowledge transfer between visual and semantic spaces, transforming explicit visual and implicit semantic knowledge into category-specific classifier weights. Finally, SynTrans introduces a visual weight generator and a semantic weight reconstructor to adaptively construct optimal multimodal FSL classifiers. Experimental results on four FSL datasets demonstrate that SynTrans, even when paired with a simple few-shot vision encoder, significantly outperforms current state-of-the-art methods.

Yuxia Sun, Aoxiang Sun, Siyi Pan, Zhixiao Fu, Jingcai Guo

Personalized federated learning (PFL) tailors models to clients' unique data distributions while preserving privacy. However, existing aggregation-weight-based PFL methods often struggle with heterogeneous data, facing challenges in accuracy, computational efficiency, and communication overhead. We propose FedAPA, a novel PFL method featuring a server-side, gradient-based adaptive aggregation strategy to generate personalized models, by updating aggregation weights based on gradients of client-parameter changes with respect to the aggregation weights in a centralized manner. FedAPA guarantees theoretical convergence and achieves superior accuracy and computational efficiency compared to 10 PFL competitors across three datasets, with competitive communication overhead. The code and full proofs are available at: https://github.com/Yuxia-Sun/FL_FedAPA.

Zhihao Sui, Liang Hu, Jian Cao, Dora D. Liu, Usman Naseem, Zhongyuan Lai, Qi Zhang

Machine Unlearning (MU) technology facilitates the removal of the influence of specific data instances from trained models on request. Despite rapid advancements in MU technology, its vulnerabilities are still underexplored, posing potential risks of privacy breaches through leaks of ostensibly unlearned information. Current limited research on MU attacks requires access to original models containing privacy data, which violates the critical privacy-preserving objective of MU. To address this gap, we initiate the innovative study on recalling the forgotten class memberships from unlearned models (ULMs) without requiring access to the original one. Specifically, we implement a Membership Recall Attack (MRA) framework with a teacher-student knowledge distillation architecture, where ULMs serve as noisy labelers to transfer knowledge to student models. Then, it is translated into a Learning with Noisy Labels (LNL) problem for inferring correct labels of the forgetting instances. Extensive experiments on state-of-the-art MU methods with multiple real datasets demonstrate that the proposed MRA strategy exhibits high efficacy in recovering class memberships of unlearned instances. As a result, our study and evaluation have established a benchmark for future research on MU vulnerabilities.

Zunhai Su, Hanyu Wei, Zhe Chen, Wang Shen, Linge Li, Huangqi Yu, Kehong Yuan

Key-Value (KV) cache facilitates efficient large language models (LLMs) inference by avoiding recomputation of past KVs. As the batch size and context length increase, the oversized KV caches become a significant memory bottleneck, highlighting the need for efficient compression. Existing KV quantization rely on fine-grained quantization or the retention of a significant portion of high bit-widths caches, both of which compromise compression ratio and often fail to maintain robustness at extremely low average bit-widths. In this work, we explore the potential of rotation technique for 2-bit KV quantization and propose RotateKV, which achieves accurate and robust performance through the following innovations: (i) Outlier-Aware Rotation, which utilizes channel-reordering to adapt the rotations to varying channel-wise outlier distributions without sacrificing the computational efficiency of the fast Walsh-Hadamard transform (FWHT); (ii) Pre-RoPE Grouped-Head Rotation, which mitigates the impact of rotary position embedding (RoPE) on proposed outlier-aware rotation and further smooths outliers across heads; (iii) Attention-Sink-Aware Quantization, which leverages the massive activations to precisely identify and protect attention sinks. RotateKV achieves less than 0.3 perplexity (PPL) degradation with 2-bit quantization on WikiText-2 using LLaMA-2-13B, maintains strong CoT reasoning and long-context capabilities, with less than 1.7% degradation on GSM8K, outperforming existing methods even at lower average bit-widths. RotateKV also showcases a 3.97× reduction in peak memory usage, supports 5.75× larger batch sizes, and achieves a 2.32× speedup in decoding stage.

Lingyun Song, Xiaofan Sun, Xinbiao Gan, Yudai Pan, Xiaolin Han, Jie Ma, Jun Liu, Xuequn Shang

Considerable attention has been paid to predicting student performance on exercises. The performance of prior studies is determined by the quality of the trait features of students and exercises. Nevertheless, most of the prior study primarily examines simple pairwise interactions in learning trait features, like those between students and exercises or exercises and concepts, while disregarding the complex higher-order interactions that typically exist among these components, which in turn hinders the prediction results. In this paper, we using an innovative Multi-Channel Graph Contrastive Learning (MCGCL) framework that integrates various high-order interactions for predicting student performance. MCGCL characterizes graph structures reflecting various high-order relationships among students, exercises, and concepts through multiple channels, thereby enhancing the trait features of both students and exercises. Moreover, graph contrastive learning is employed to enhance the representation of trait features acquired from high-order graph structures in diverse views. Extensive experiments on real-world datasets show that MCGCL achieves state-of-the-art results on the task of predicting student performance. The code is available at https://github.com/sunlitsong/MCGCL.

Yuntao Shou, Jun Yao, Tao Meng, Wei Ai, Cen Chen, Keqin Li

Multimodal Emotion Recognition (MER) combines technologies from multiple fields (e.g., computer vision, natural language processing, and audio signal processing), aiming to infer an individual's emotional state by analyzing information from different sources (i.e., video, audio, and text). Compared with single modality, by fusing complementary semantic information from different modalities, the model can obtain more robust knowledge representation. However, the modality missing problem limits the performance of MERC in practical scenarios. Recent work has achieved impressive performance on modality completion using graph neural networks and diffusion models, respectively. This inspires us to combine these two dimensions in the completion network to obtain more powerful representation capabilities. However, we argue that directly running a full-rank score-based diffusion model on the entire graph adjacency matrix space may adversely affect the learning process of the diffusion model. This is because the model assumes a direct relationship between each pair of nodes and ignores local structural features and sparse connections between nodes, thereby significantly reducing the quality of the generated data. Based on the above ideas, we propose a novel Graph Spectral Diffusion Network (GSDNet), which utilizes a low-rank score-based diffusion model to map Gaussian noise to the graph spectral distribution space of missing modalities and recover the missing data according to its original distribution. Extensive experiments have demonstrated that GSDNet achieves state-of-the-art emotion recognition performance in various modality loss scenarios.

Ruxue Shi, Hengrui Gu, Hangting Ye, Yiwei Dai, Xu Shen, Xin Wang

Few-shot tabular learning, in which machine learning models are trained with a limited amount of labeled data, provides a cost-effective approach to addressing real-world challenges. The advent of Large Language Models (LLMs) has sparked interest in leveraging their pre-trained knowledge for few-shot tabular learning. Despite promising results, existing approaches either rely on test-time knowledge extraction, which introduces undesirable latency, or text-level knowledge, which leads to unreliable feature engineering. To overcome these limitations, we propose Latte, a training-time knowledge extraction framework that transfers the latent prior knowledge within LLMs to optimize a more generalized downstream model. Latte enables general knowledge-guided downstream tabular learning, facilitating the weighted fusion of information across different feature values while reducing the risk of overfitting to limited labeled data. Furthermore, Latte is compatible with existing unsupervised pre-training paradigms and effectively utilizes available unlabeled samples to overcome the performance limitations imposed by an extremely small labeled dataset. Extensive experiments on various few-shot tabular learning benchmarks demonstrate the superior performance of Latte, establishing it as a state-of-the-art approach in this domain. Our code is available at https://github.com/ruxueshi/Latte.git.

Jiandong Shi, Ming Li, Lu Bai, Feilong Cao, Ke Lu, Jiye Liang

Multimodal emotion recognition aims to identify emotions by integrating multimodal features derived from spoken utterances. However, existing work often neglects the calibration of conversational entities, focusing mainly on extracting potential intra- or cross-modal information. This leads to the underutilization of utterance information that is essential for accurately characterizing emotion. Additionally, the lack of effective modeling of conversational patterns limits the ability to capture emotional pathways across contexts, modalities and speakers, impacting the overall emotional understanding. In this study, we propose the modality-calibrated hypergraph fusion network (MATCH), which leverages multimodal fusion and hypergraph learning techniques to address these challenges. In particular, we introduce an entity calibration strategy that refines the representations of conversational entities both at the modality and context levels, allowing for deeper insights into emotion-related cues. Furthermore, we present an emotion-aligned hypergraph fusion method that incorporates a line graph to explore conversational patterns, facilitating flexible knowledge transfer across modalities through hyperedge-level and graph-level alignments. Experiments demonstrate that MATCH outperforms state-of-the-art approaches on two benchmark datasets.

Jiajun Shen, Yufei Jin, Kaibu Feng, Yi He, Xingquan Zhu

This paper presents HGEN that pioneers ensemble learning for heterogeneous graphs. We argue that the heterogeneity in node types, nodal features, and local neighborhood topology poses significant challenges for ensemble learning, particularly in accommodating diverse graph learners. Our HGEN framework ensembles multiple learners through a meta-path and transformation-based optimization pipeline to uplift classification accuracy. Specifically, HGEN uses meta-path combined with random dropping to create Allele Graph Neural Networks (GNNs), whereby the base graph learners are trained and aligned for later ensembling. To ensure effective ensemble learning, HGEN presents two key components:1) a residual-attention mechanism to calibrate allele GNNs of different meta-paths, thereby enforcing node embeddings to focus on more informative graphs to improve base learner accuracy, and 2) a correlation-regularization term to enlarge the disparity among embedding matrices generated from different meta-paths, thereby enriching base learner diversity. We analyze the convergence of HGEN and attest its higher regularization magnitude over simple voting. Experiments on five heterogeneous networks validate that HGEN consistently outperforms its state-of-the-art competitors by substantial margin. Codes are available at https://github.com/Chrisshen12/HGEN.

Yimiao Shao, Wenzhong Li, Kang Xia, Kaijie Lin, Mingkai Lin, Sanglu Lu

Probabilistic time series forecasting has attracted an increasing attention in machine learning community for its potential applications in the fields of renewable energy, traffic management, healthcare, etc. Previous research mainly focused on extracting long-range dependencies for point-wise prediction, which fail to capture complex temporal patterns and statistical characteristics for probabilistic analysis. In this paper, we propose a novel pattern-mixture decomposition method that decomposes long-term series into quantile drift, divergence patterns, and Gaussian mixture components, which can effectively capture the intricate temporal patterns and stochastic characteristics in time series. Based on pattern-mixture decomposition, we propose a novel Transformer-based model called QuantileFormer for probabilistic time series forecasting. It takes the the comprehensive drift-divergence mixture patterns as features, and designs a variational inference based fusion Transformer architecture to generate quantile prediction results. Extensive experiments show that the proposed method consistently boosts the baseline methods by a large margin and achieves state-of-the-art performance on six real-world benchmarks.

Yihua Shao, Minxi Yan, Yang Liu, Siyu Chen, Wenjie Chen, Xinwei Long, Ziyang Yan, Lei Li, Chenyu Zhang, Nicu Sebe 等

Low-rank Adaptation (LoRA) has demonstrated remarkable capabilities for task specific fine-tuning. However, in scenarios that involve multiple tasks, training a separate LoRA model for each one results in considerable inefficiency in terms of storage and inference. Moreover, existing parameter generation methods fail to capture the correlations among these tasks, making multi-task LoRA parameter generation challenging. To address these limitations, we propose In-Context Meta LoRA (ICM-LoRA), a novel approach that efficiently achieves task-specific customization of large language models (LLMs). Specifically, we use training data from all tasks to train a tailored generator, Conditional Variational Autoencoder (CVAE). CVAE takes task descriptions as inputs and produces task-aware LoRA weights as outputs. These LoRA weights are then merged with LLMs to create task-specialized models without the need for additional fine-tuning. Furthermore, we utilize in-context meta-learning for knowledge enhancement and task mapping, to capture the relationship between tasks and parameter distributions. As a result, our method achieves more accurate LoRA parameter generation for diverse tasks using CVAE. ICM-LoRA enables more accurate LoRA parameter reconstruction than current parameter reconstruction methods and is useful for implementing task-specific enhancements of LoRA parameters. At the same time, our method occupies 283MB, only 1% storage compared with the original LoRA. The code is available at https://github.com/YihuaJerry/ICM-LoRA.

Jie-Jing Shao, Hong-Jie You, Guohao Cai, Quanyu Dai, Zhenhua Dong, Lan-Zhe Guo

Large Language Models (LLMs) have demonstrated remarkable capabilities in language understanding and commonsense reasoning, yet they often struggle with constraint satisfaction in planning problems. Previous studies relying on test-time improvement with self-evaluation fail to address this limitation effectively. In this work, we identify this critical gap and propose a novel neuro-symbolic framework, Reinforced Neuro-Symbolic Planning (\algo), that enhances LLM-powered planning by incorporating a symbolic verifier. The verifier provides explicit feedback on constraint satisfaction, enabling iterative refinement of the state evaluation. Specifically, we utilize the outcome feedback from each logical goal to update the process value along planning paths through a reinforcement value function maximization objective. We further employ T-norms to aggregate the satisfaction levels of multiple constraints, which provided more effective guidance for the test-time search. Our framework bridges the strengths of neural and symbolic methods, leveraging the generative power of LLMs while ensuring rigorous adherence to constraints through symbolic verification. Extensive experiments demonstrate that our approach significantly improves planning accuracy and constraint satisfaction across various domains, outperforming traditional self-evaluation methods. It highlights the potential of hybrid neuro-symbolic systems to address complex constrained planning tasks.

Andreas Roth, Thomas Liebig

Most graph neural networks (GNNs) utilize approximations of the general graph convolution derived in the graph Fourier domain. While GNNs are typically applied in the multi-input multi-output (MIMO) case, the approximations are performed in the single-input single-output (SISO) case. In this work, we first derive the MIMO graph convolution through the convolution theorem and approximate it directly in the MIMO case. We find the key MIMO-specific property of the graph convolution to be operating on multiple computational graphs, or equivalently, applying distinct feature transformations for each pair of nodes. As a localized approximation, we introduce localized MIMO graph convolutions (LMGCs), which generalize many linear message-passing neural networks. For almost every choice of edge weights, we prove that LMGCs with a single computational graph are injective on multisets, and the resulting representations are linearly independent when more than one computational graph is used. Our experimental results confirm that an LMGC can combine the benefits of various methods.