Fine-tuning large language models (LLMs) is difficult due to their huge model size. Recent Fourier domain-based methods show potential for reducing fine-tuning costs. We propose a block circulant matrix-based fine-tuning method with a stable training heuristic to leverage the properties of circulant matrices and one-dimensional Fourier transforms to reduce storage and computation costs. Experiments show that our method uses 14× less number of parameters than VeRA, 16× smaller than LoRA and 32× less FLOPs than FourierFT, while maintaining close or better task performance. Our approach presents a promising way in frequency domain to fine-tune large models on downstream tasks.
论文检索
输入标题、作者或关键词，从 1,280 篇学术成果中精准定位
已筛选
IJCAI 2025
1,280篇论文
第 37 / 64 页
FedDLAD: A Federated Learning Dual-Layer Anomaly Detection Framework for Enhancing Resilience Against Backdoor Attacks
PDF ↗In Federated Learning (FL), the decentralized nature of client training introduces vulnerabilities, notably backdoor attacks. Prevailing anomaly detection approaches typically perform binary classification, dividing clients into trusted and untrusted groups. However, these methods face two critical challenges: the insider threat, where malicious clients concealed within the trusted group compromise the global model, and the benign exclusion, where legitimate contributions from benign clients are mistakenly classified as untrusted and disregarded. These issues weaken both the robustness and fairness of FL systems, exposing inherent defense vulnerabilities. To address these challenges, we propose FedDLAD, a Federated Learning Dual-Layer Anomaly Detection framework designed to enhance resilience against backdoor attacks. The framework leverages the Connectivity-Based Outlier Factor (COF) module to perform a robust initial classification of clients by analyzing structural data connectivity. The Interquartile Range (IQR) module further reinforces this by mitigating the insider threat through the removal of residual malicious influences within the trusted group. Furthermore, the Pardon module dynamically reintegrates misclassified benign clients from the untrusted group, thereby preserving their valuable contributions and addressing the benign exclusion. We conduct extensive evaluations of FedDLAD against state-of-the-art defenses on real-world datasets, demonstrating its superior ability to reduce backdoor attack success rates while maintaining robust model performance. Code is available at: https://github.com/dingbinb/FedDLAD.
An important class of neuro-symbolic (NeSy) methods relies on knowledge compilation (KC) techniques to transform logical constraints into a differentiable exact arithmetic circuit (AC) that represents all models of a logical formula. However, given the complexity of KC, compiling such exact circuits can be infeasible. Previous works in such cases proposed to compile a circuit for a subset of models. In this work, we will show that gradients calculated on a subset of models can be very far from true gradients. We propose a new framework that calculates gradients based on compiling logical constraints partially in not only a lower-bound circuit but also an upper-bound circuit. We prove that from this pair of ACs, gradients that are within a bounded distance from true gradients can be calculated. Our experiments show that adding the upper-bound AC also helps the learning process in practice, allowing for similar or better generalisation than working solely with fully compiled ACs, even with less than 150 seconds of partial compilation.
Neural architecture search-based multi-modal classification (NAS-MMC) aims to automatically find optimal network structures for improving the multi-modal classification performance. However, most current NAS-MMC methods are quite time-consuming during the training process. In this paper, we propose a knowledge sharing-based neural architecture search (KS-NAS) method for multi-modal classification. The KS-NAS optimizes the search process by introducing a dynamically updated knowledge base to reduce the consumption of computational resource. Specifically, during the deep evolutionary search, individuals in the initial population acquire initial parameters from a knowledge base, and then undergo training and optimization until convergence is reached, avoiding the need for training from scratch. The knowledge base is dynamically updated by aggregating the parameters of high-quality individuals trained within the population, thus progressively improving the quality of the knowledge base. As the population evolves, the knowledge base continues to optimize, ensuring that subsequent individuals can obtain higher-quality initialization parameters, which significantly accelerates the training speed of the population. Experimental results show that the KS-NAS method achieves state-of-the-art results in terms of classification performance and training efficiency across multiple popular multi-modal tasks.
Reinforcement Learning (RL) has made significant strides in various domains, and policy gradient methods like Proximal Policy Optimization (PPO) have gained popularity due to their balance in performance, training stability, and computational efficiency. These methods directly optimize policies through gradient-based updates. However, developing effective control policies for environments with complex and non-linear dynamics remains a challenge. High variance in gradient estimates and non-convex optimization landscapes often lead to unstable learning trajectories. Koopman Operator Theory has emerged as a powerful framework for studying non-linear systems through an infinite-dimensional linear operator that acts on a higher-dimensional space of measurement functions. In contrast with their non-linear counterparts, linear systems are simpler, more predictable, and easier to analyze. In this paper, we present Koopman-Inspired Proximal Policy Optimization (KIPPO), which learns an approximately linear latent-space representation of the underlying system’s dynamics while retaining essential features for effective policy learning. This is achieved through a Koopman-approximation auxiliary network that can be added to the baseline policy optimization algorithms without altering the architecture of the core policy or value function. Extensive experimental results demonstrate consistent improvements over the PPO baseline with 6–60% increased performance while reducing variability by up to 91% when evaluated on various continuous control tasks.
A classifier is considered interpretable if each of its decisions has an explanation which is small enough to be easily understood by a human user. A DNF can be seen as a binary classifier kappa over boolean domains. The size of an explanation of a positive decision taken by a DNF kappa is bounded by the size of the terms in kappa, since we can explain a positive decision by giving a term of kappa that evaluates to true. Since both positive and negative decisions must be explained, we consider that interpretable DNFs are those kappa for which both kappa and its complement can be expressed as DNFs composed of terms of bounded size. In this paper, we investigate the family of k-DNFs whose complements can also be expressed as k-DNFs. We compare two such families, namely depth-k decision trees and nested k-DNFs, a novel family of models. Experimental evidence indicates that nested k-DNFs are an interesting alternative to decision trees in terms of interpretability and accuracy.
SPoRt - Safe Policy Ratio: Certified Training and Deployment of Task Policies in Model-Free RL
PDF ↗To apply reinforcement learning to safety-critical applications, we ought to provide safety guarantees during both policy training and deployment. In this work we present novel theoretical results that provide a bound on the probability of violating a safety property for a new task-specific policy in a model-free, episodic setup: the bound, based on a 'maximum policy ratio' that is computed with respect to a 'safe' base policy, can also be more generally applied to temporally-extended properties (beyond safety) and to robust control problems. We thus present SPoRt, which also provides a data-driven approach for obtaining such a bound for the base policy, based on scenario theory, and which includes Projected PPO, a new projection-based approach for training the task-specific policy while maintaining a user-specified bound on property violation. Hence, SPoRt enables the user to trade off safety guarantees in exchange for task-specific performance. Accordingly, we present experimental results demonstrating this trade-off, as well as a comparison of the theoretical bound to posterior bounds based on empirical violation rates.
Many important scientific problems involve multivariate optimization coupled with slow and laborious experimental measurements. These high-dimensional searches can be defined by complex, non-convex optimization landscapes that resemble needle-in-a-haystack surfaces, leading to entrapment in local minima. Contextualizing optimizers with human domain knowledge is a powerful approach to guide searches to localized fruitful regions. However, this approach is susceptible to human confirmation bias. It is also challenging for domain experts to keep track of the rapidly expanding scientific literature. Here, we propose the use of Large Language Models (LLMs) for contextualizing Bayesian optimization (BO) via a hybrid optimization framework that intelligently and economically blends stochastic inference with domain knowledge-based insights from the LLM, which is used to suggest new, better-performing areas of the search space for exploration. Our method fosters user engagement by offering real-time commentary on the optimization progress, explaining the reasoning behind the search strategies. We validate the effectiveness of our approach on synthetic benchmarks with up to 15 variables and demonstrate the ability of LLMs to reason in four real-world experimental tasks where context-aware suggestions boost optimization performance substantially.
Entropy Regularisation is a widely adopted technique that enhances policy optimisation performance and stability. Maximum entropy reinforcement learning (MaxEnt RL) regularises policy evaluation by augmenting the objective with an entropy term, showing theoretical benefits in policy optimisation. However, its practical application in straightforward direct policy gradient settings remains surprisingly underexplored. We hypothesise that this is due to the difficulty of managing the entropy reward in practice. This paper proposes Entropy Advantage Policy Optimisation (EAPO), a simple method that facilitates MaxEnt RL implementation by separately estimating task and entropy objectives. Our empirical evaluations demonstrate that extending Proximal Policy Optimisation (PPO) and Trust Region Policy Optimisation (TRPO) within the MaxEnt framework improves optimisation performance, generalisation, and exploration in various environments. Moreover, our method provides a stable and performant MaxEnt RL algorithm for discrete action spaces.
GRAPE: Heterogeneous Graph Representation Learning for Genetic Perturbation with Coding and Non-Coding Biotype
PDF ↗Predicting genetic perturbations enables the identification of potentially crucial genes prior to wet-lab experiments, significantly improving overall experimental efficiency. Since genes are the foundation of cellular life, building gene regulatory networks (GRN) is essential to understand and predict the effects of genetic perturbations. However, current methods fail to fully leverage gene-related information, and solely rely on simple evaluation metrics to construct coarse-grained GRN. More importantly, they ignore functional differences between biotypes, limiting the ability to capture potential gene interactions. In this work, we leverage pre-trained large language model and DNA sequence model to extract features from gene descriptions and DNA sequence data, respectively, which serve as the initialization for gene representations. Additionally, we introduce gene biotype information for the first time in genetic perturbation, simulating the distinct roles of genes with different biotypes in regulating cellular processes, while capturing implicit gene relationships through graph structure learning (GSL). We propose GRAPE, a heterogeneous graph neural network (HGNN) that leverages gene representations initialized with features from descriptions and sequences, models the distinct roles of genes with different biotypes, and dynamically refines the GRN through GSL. The results on publicly available datasets show that our method achieves state-of-the-art performance. The code for reproducing the results can be seen at the link: https://github.com/ChangxiChi/GRAPE.
Tabular data is widely adopted in various machine learning tasks. Current tabular data learning mainly focuses on closed environments, while in real-world applications, open environments are often encountered, where distribution shifts and feature decrements occur, leading to severe performance degradation. Previous studies have primarily focused on addressing distribution shifts, while feature decrements, a unique challenge in tabular data learning, have received relatively little attention. In this paper, we present the first comprehensive study on the problem of Fully Test-Time Adaptation for Feature Decrement in Tabular Data. Through empirical analysis, we identify the suboptimality of existing missing-feature imputation methods and the limited applicability of missing-feature adaptation approaches. To address these challenges, we propose a novel method, LLM-IMPUTE, which leverages Large Language Models (LLMs) to impute missing features without relying on training data. Furthermore, we introduce Augmented-Training LLM (ATLLM), a method designed to enhance the robustness of feature decrements by simulating feature-decrement scenarios during the training phase to address tasks that can not be imputed by LLM-IMPUTE. Extensive experimental results demonstrate that our proposal significantly improves both performance and robustness in missing feature imputation and adaptation scenarios.
Automated processor design, which can significantly reduce human efforts and accelerate design cycles, has received considerable attention. While recent advancements have automatically designed single-cycle processors that execute one instruction per cycle, their performance cannot compete with modern superscalar processors that execute multiple instructions per cycle. Previous methods fail on superscalar processor design because they cannot address inter-instruction data dependencies, leading to inefficient sequential instruction execution. This paper proposes a novel approach to automatically designing superscalar processors using a hardware-friendly model called the Stateful Binary Speculation Diagram (State-BSD). We observe that processor parallelism can be enhanced through on-the-fly inter-instruction dependent data predictors, reusing the processor's internal states to learn the data dependency. To meet the challenge of both hardware-resource limitation and design functional correctness, State-BSD consists of two components: 1) a lightweight state-selector trained by simulated annealing method to detect the most reusable processor states and store them in a small buffer; and 2) a highly precise state-speculator trained by BSD expansion method to predict the inter-instruction dependent data using the selected states. It is the first work to achieve the automated superscalar processor design, i.e. QiMeng-CPU-v2, which improves the performance by about 380x than the state-of-the-art automated design and is comparable to human-designed superscalar processors such as ARM Cortex A53.
A Hybrid Multi-Factor Network with Dynamic Sequence Modeling for Early Warning of Intraoperative Hypotension
PDF ↗Intraoperative hypotension (IOH) prediction using past physiological signals is crucial, as IOH may lead to inadequate organ perfusion and significantly elevate the risk of severe complications and mortality. However, current methods often rely on static modeling, overlooking the complex temporal dependencies and the inherently non-stationary nature of physiological signals. We propose a Hybrid Multi-Factor (HMF) network that formulates IOH prediction as a dynamic sequence forecasting task, explicitly capturing both temporal dependencies and physiological non-stationarity. We represent signal dynamics as multivariate time series and decompose them into trend and seasonal components, enabling separate modeling of long-term and periodic variations. Each component is encoded with a patch-based Transformer to balance computational efficiency and feature representation. To address distributional drift from evolving signals, we introduce a symmetric normalization mechanism. Experiments on both public and real-world clinical datasets show that HMF significantly outperforms competitive baselines. We hope HMF offers new insights into IOH prediction and ultimately promotes safer surgical care. Our code is available at https://github.com/Mingyue-Cheng/HMF.
Multi-objective multi-armed bandit (MOMAB) problems are crucial for complex decision-making scenarios where multiple conflicting objectives must be simultaneously optimized. However, most existing works are based on the linear assumption of the feedback rewards, which significantly constrains their applicability and efficacy in capturing the intricate dynamics of real-world environments. This paper explores a multi-objective neural bandit (MONB) framework, which integrates the universal approximators, neural networks, with the classical MOMABs. We adopt random scalarization to accommodate the special needs of a practitioner by setting an appropriate distribution on the regions of interest. Using the trade-off capabilities of upper confidence bound (UCB) and Thompson sampling (TS) strategies, we propose two novel algorithms, MONeural-UCB and MONeural-TS. Theoretical and empirical analysis demonstrate the superiority of our methods in multi-objective or multi-task bandit problems, which makes great improvement over the classical linear MOMABs.
DO-CoLM: Dynamic 3D Conformation Relationships Capture with Self-Adaptive Ordering Molecular Relational Modeling in Language Models
PDF ↗Molecular Relational Learning (MRL) aims to understand interactions between molecular pairs, playing a critical role in advancing biochemical research. Recently, Large Language Models (LLMs), with their extensive knowledge bases and advanced reasoning capabilities, have emerged as powerful tools for MRL. However, existing LLMs, which primarily rely on SMILES strings and molecular graphs, face two major challenges. They struggle to capture molecular stereochemistry and dynamics, as molecules possess multiple 3D conformations with varying reactivity and dynamic transformation relationships that are essential for accurately predicting molecular interactions but cannot be effectively represented by 1D SMILES or 2D molecular graphs. Additionally, these models do not consider the autoregressive nature of LLMs, overlooking the impact of input order on model performance. To address these issues, we propose DO-CoLM: a Dynamic relationship capture and self-adaptive Ordering 3D molecular Conformation LM for MRL. By introducing modules to dynamically model intra-molecular and inter-molecular conformational relationships and adaptively adjust the molecular modality input order, DO-CoLM achieves superior performance, as demonstrated by experimental results on 12 cross-domain datasets.
In Partial Multi-Label Learning (PML), each instance is associated with a candidate label set containing multiple relevant labels along with other false positive labels. Currently, most PML methods directly extract instance correlation from instance features while ignoring the candidate labels, which may contain more discriminative instance-related information. This paper argues that, with a well-designed model, more accurate instance correlation can be mined from the candidate labels to facilitate label disambiguation. To this end, we propose a novel PML method based on pseudo-label reconstruction (PML-PLR). Specifically, we first propose a novel orthogonal candidate label reconstruction method, which jointly optimizes with instance features to extract more consistent instance correlation. Then, we use instance correlation as reconstruction coefficient to reconstruct pseudo-labels. Subsequently, through local manifold learning, the reconstructed pseudo-labels are leveraged to propagate the consistency relationship between labels and instances, thereby improving the accuracy of pseudo-labels. Extensive experiments and analyses demonstrate that the proposed PML-PLR outperforms state-of-the-art methods.
Temporal processing is vital for extracting meaningful information from time-varying signals. Recent advancements in Spiking Neural Networks (SNNs) have shown immense promise in efficiently processing these signals. However, progress in this field has been impeded by the lack of effective and standardized benchmarks, which complicates the consistent measurement of technological advancements and limits the practical applicability of SNNs. To bridge this gap, we introduce the Neuromorphic Sequential Arena (NSA), a comprehensive benchmark that offers an effective, versatile, and application-oriented evaluation framework for neuromorphic temporal processing. The NSA includes seven real-world temporal processing tasks from a diverse range of application scenarios, each capturing rich temporal dynamics across multiple timescales. Utilizing NSA, we conduct extensive comparisons of recently introduced spiking neuron models and neural architectures, presenting comprehensive baselines in terms of task performance, training speed, memory usage, and energy efficiency. Our findings emphasize an urgent need for efficient SNN designs that can consistently deliver high performance across tasks with varying temporal complexities while maintaining low computational costs. NSA enables systematic tracking of advancements in neuromorphic algorithm research and paves the way for developing effective and efficient neuromorphic temporal processing systems.
Multimodal Learning with visual and tabular modalities has become more and more popular nowadays, especially in the healthcare area. Due to the adaptation of new equipment or new factors being introduced, the tabular modality keeps changing. However, the standard process of training multimodal AI models requires tables to have fixed columns in training and inference; thus, it is not suitable for handling dynamically changed tables. Therefore, new methods are needed for efficiently handling such tables in multimodal learning. In this paper, we introduce a new task, multimodal inference with incremental tabular attributes, which aims to enable trained multimodal models to leverage incremental attributes in tabular modality during the inference stage efficiently. We implement a specialized encoder to disentangle the latent representation of incremental tabular attributes inside itself and with the old attributes to reduce information redundancy and further align the incremental attributes with the visual modality with consistency loss to improve information richness. Experimental results across five public datasets show that our method effectively utilizes incremental tabular attributes, achieving state-of-the-art performance in general scenarios. Beyond the inference, we also find that our method achieved better performance in fully supervised settings, evoking a new training style for multimodal learning with tables.
Learning to Extrapolate and Adjust: Two-Stage Meta-Learning for Concept Drift in Online Time Series Forecasting
PDF ↗The inherent non-stationarity of time series in practical applications poses significant challenges for accurate forecasting. This paper tackles the concept drift problem where the underlying distribution or environment of time series changes. To better describe the characteristics and effectively model concept drifts, we first classify them into macro-drift (stable, long-term changes) and micro-drift (sudden, short-term fluctuations). Next, we propose a unified meta-learning framework called LEAF (Learning to Extrapolate and Adjust for Forecasting), where an extrapolation module is first introduced to track and extrapolate the prediction model in latent space considering macro-drift, and then an adjustment module incorporates meta-learnable surrogate loss to capture sample-specific micro-drift patterns. LEAF’s dual-stage approach effectively addresses diverse concept drifts and is model-agnostic which can be compatible with any deep prediction model. We further provide theoretical analysis to justify why the proposed framework can handle macro-drift and micro-drift. To facilitate further research in this field, we release three electric load time series datasets collected from real-world scenarios, exhibiting diverse and typical concept drifts. Extensive experiments on multiple datasets demonstrate the effectiveness of LEAF.
Real-life deployment of federated Learning (FL) often faces non-IID data, which leads to poor accuracy and slow convergence. Personalized FL (pFL) tackles these issues by tailoring local models to individual data sources and using weighted aggregation methods for client-specific learning. However, existing pFL methods often fail to provide each local model with global knowledge on demand while maintaining low computational overhead. Additionally, local models tend to over-personalize their data during the training process, potentially dropping previously acquired global information. We propose FLAYER, a novel layer-wise learning method for pFL that optimizes local model personalization performance. FLAYER considers the different roles and learning abilities of neural network layers of individual local models. It incorporates global information for each local model as needed to initialize the local model cost-effectively. It then dynamically adjusts learning rates for each layer during local training, optimizing the personalized learning process for each local model while preserving global knowledge. Additionally, to enhance global representation in pFL, FLAYER selectively uploads parameters for global aggregation in a layer-wise manner. We evaluate FLAYER on four representative datasets in computer vision and natural language processing domains. Compared to eight state-of-the-art pFL methods, FLAYER improves the inference accuracy, on average, by 5.20% (up to 14.29%). Code is available at https://github.com/lancasterJie/FLAYER/.