Document retrieval, designed to recall query-relevant documents from expansive collections, is essential for information-seeking tasks, such as web search and open-domain question-answering. Advances in representation learning and pretrained language models (PLMs) have driven a paradigm shift from traditional sparse retrieval methods to more effective dense retrieval approaches, forging enhanced semantic connections between queries and documents and establishing new performance benchmarks. However, reliance on extensive annotated document-query pairs limits their competitiveness in low-resource scenarios. Recent research efforts employing the few-shot capabilities of large language models (LLMs) and prompt engineering for synthetic data generation have emerged as a promising solution. Nonetheless, these approaches are hindered by the generation of lower-quality data within the conventional dense retrieval training process. To this end, in this paper, we introduce iGFT, a framework aimed at enhancing low-resource dense retrieval by integrating a three-phase process --- Generation, Filtering, and Tuning --- coupled with an iterative optimization strategy. Specifically, we first employ supervised fine-tuning on limited ground truth data, enabling an LLM to function as the generator capable of producing potential queries from given documents. Subsequently, we present a multi-stage filtering module to minimize noise in the generated data while retaining samples poised to significantly improve the dense retrieval model's performance in the follow-up fine-tuning process. Furthermore, we design a novel iterative optimization strategy that dynamically optimizes the query generator for producing more informative queries, thereby enhancing the efficacy of the entire framework. Finally, extensive experiments conducted on a series of publicly available retrieval benchmark datasets have demonstrated the effectiveness of the proposed iGFT.
论文检索
输入标题、作者或关键词，从 8,216 篇学术成果中精准定位
In light of rising drug-related concerns and the increasing role of social media, sales and discussions of illicit drugs have become commonplace online. Social media platforms hosting user-generated content must therefore perform content moderation, which is a difficult task due to the vast amount of jargon used in drug discussions. Previous works on drug jargon detection were limited to extracting a list of terms, but these approaches have fundamental problems in practical application. First, they are trivially evaded using word substitutions. Second, they cannot distinguish whether euphemistic terms (pot, crack) are being used as drugs or as their benign meanings. We argue that drug content moderation should be done using contexts, rather than relying on a banlist. However, manually annotated datasets for training such a task are not only expensive but also prone to becoming obsolete. We present JEDIS, a framework for detecting illicit drug jargon terms by analyzing their contexts. JEDIS utilizes a novel approach that combines distant supervision and delexicalization, which allows JEDIS to be trained without human-labeled data while being robust to new terms and euphemisms. Experiments on two manually annotated datasets show JEDIS significantly outperforms state-of-the-art word-based baselines in terms of F1-score and detection coverage in drug jargon detection. We also conduct qualitative analysis that demonstrates JEDIS is robust against pitfalls faced by existing approaches.
Out-of-distribution (OOD) detection has emerged as a pivotal approach for enhancing the reliability of machine learning models, considering the potential for test data to be sampled from classes disparate from in-distribution (ID) data employed during model training. Detecting those OOD data is typically realized as a distance measurement problem, where those deviating far away from the training distribution in the learned feature space are considered OOD samples. Advanced works have shown great success in learning with prototypes for feature-based OOD detection methods, where each ID class is represented with single or multiple prototypes. However, modeling with a finite number of prototypes would fail to maximally capture intra-class variations. In view of this, this paper extends the existing prototype-based learning paradigm to an infinite setting. This motivates us to design two feasible formulations for the Distributional Prototype Learning (DPL) objective, where, to avoid intractable computation and exploding parameters caused by the infinity nature, our key idea is to model an infinite number of discrete prototypes of each ID class with a class-wise continuous distribution. We theoretically analyze both alternatives, identifying the more stable-converging version of the learning objective. We show that, by sampling prototypes from a mixture of class-conditioned Gaussian distributions, the objective can be efficiently computed in a closed form without resorting to the computationally expensive Monte-Carlo approximation of the involved expectation terms. Extensive evaluations across mainstream OOD detection benchmarks empirically manifest that our proposed DPL has established a new state-of-the-art in various OOD settings.
In Deep Neural Networks (DNNs), manipulating gradients is central to various algorithms, including data subset selection and instance attribution. For better tractability, practitioners often resort to using only the gradients of the last layer as a heuristic, instead of the full gradient across all model parameters, which we show is detrimental due to the Support Vector Effect (SVE). We introduce SVE, a max-margin-like behavior in the last layer(s) of DNNs and employ it to thoroughly scrutinize prevalent data selection and attribution methods relying on last layer gradients. Our investigation exposes limitations in these techniques and not only provides explanations for previously observed pitfalls, like lack of diversity and temporal performance degradation, but also offers fresh insights, including the vulnerability of existing methods to basic poisoning attacks and the potential for competitive performance using much simpler alternatives. Based on insights from SVE, we craft new methods RandE and PAE for data subset selection and instance attribution, respectively, which often outperform the purported state-of-the-art at a fraction of the cost, emphasizing the practical advantages of more efficient and less complex approaches.
Spatio-temporal forecasting is a critical component of various smart city applications, such as transportation optimization, energy management, and socio-economic analysis. Recently, several automated spatio-temporal forecasting methods have been proposed to automatically search the optimal neural network architecture for capturing complex spatio-temporal dependencies. However, the existing automated approaches suffer from expensive neural architecture search overhead, which hinders their practical use and the further exploration of diverse spatio-temporal operators in a finer granularity. In this paper, we propose AutoSTF, a decoupled automatic neural architecture search framework for cost-effective automated spatio-temporal forecasting. From the efficiency perspective, we first decouple the mixed search space into temporal space and spatial space and respectively devise representation compression and parameter-sharing schemes to mitigate the parameter explosion. The decoupled spatio-temporal search not only expedites the model optimization process but also leaves new room for more effective spatio-temporal dependency modeling. From the effectiveness perspective, we propose a multi-patch transfer module to jointly capture multi-granularity temporal dependencies and extend the spatial search space to enable finer-grained layer-wise spatial dependency search. Extensive experiments on eight datasets demonstrate the superiority of AutoSTF in terms of both accuracy and efficiency. Specifically, our proposed method achieves up to 13.48x speed-up compared to state-of-the-art automatic spatio-temporal forecasting methods while maintaining the best forecasting accuracy. The source code and data are available at https://github.com/usail-hkust/AutoSTF.
With the rise of e-commerce and short videos, online recommender systems that can capture users' interests and update new items in real-time play an increasingly important role. In both online and offline recommendation systems, the cold-start problem caused by interaction sparsity has been impacting the effectiveness of recommendations for cold-start items. Many cold-start scheme based on fine-tuning or knowledge transferring shows excellent performance on offline recommendation. Yet, these schemes are infeasible for online recommendation on streaming data pipelines due to different training method, computational overhead and time constraints. Inspired by the above questions, we propose a model-agnostic recommendation algorithm called Popularity-Aware Meta-learning (PAM), to address the item cold-start problem under streaming data settings. PAM divides the incoming data into different meta-learning tasks by predefined item popularity thresholds. The model can distinguish and reweight behavior-related and content-related features in each task based on their different roles in different popularity levels, thus adapting to recommendations for cold-start samples. These task-fixing design significantly reduces additional computation and storage costs compared to offline methods. Furthermore, PAM also introduced data augmentation and an additional self-supervised loss specifically designed for low-popularity tasks, leveraging insights from high-popularity samples. This approach effectively mitigates the issue of inadequate supervision due to the scarcity of cold-start samples. Experimental results across multiple public datasets demonstrate the superiority of our approach over other baseline methods in addressing cold-start challenges in online streaming data scenarios.
This paper investigates the problem of learning mesh-based physical simulations, which is a crucial task with applications in fluid mechanics and aerodynamics. Recent works typically utilize graph neural networks (GNNs) to produce next-time states on irregular meshes by modeling interacting dynamics, and then adopt iterative rollouts for the whole trajectories. However, these methods cannot achieve satisfactory performance in long-term predictions due to the failure of capturing long-term dependency and potential error accumulations. To tackle this, we introduce a new future-to-present learning perspective, and further develop a simple yet effective approach named Foresight And Interpolation (FAIR) for long-term mesh-based simulations. The main idea of our FAIR is to first learn a graph ODE model for coarse long-term predictions and then refine short-term predictions via interpolation. Specifically, FAIR employs a continuous graph ODE model that incorporates past states into the evolution of interacting node representations, which is capable of learning coarse long-term trajectories under a multi-task learning framework. Then, we leverage a channel aggregation strategy to summarize the trajectories for refined short-term predictions, which can be illustrated using an interpolation process. Through pyramid-like alternative propagation between the foresight step and refinement step, our proposed framework FAIR can generate accurate long-term trajectories, achieving a significant error reduction compared with the best baseline on four benchmark datasets. Extensive ablation studies and visualization further validate the superiority of our proposed FAIR.
SMOTE (Synthetic Minority Oversampling Technique) is the established geometric approach to random oversampling to balance classes in the imbalanced learning problem, followed by many extensions. Its idea is to introduce synthetic data points of the minor class, with each new point being the convex combination of an existing data point and one of its k-nearest neighbors. In this paper, by viewing SMOTE as sampling from the edges of a geometric neighborhood graph and borrowing tools from the topological data analysis, we propose a novel technique, Simplicial SMOTE, that samples from the simplices of a geometric neighborhood simplicial complex. A new synthetic point is defined by the barycentric coordinates w.r.t. a simplex spanned by an arbitrary number of data points being sufficiently close rather than a pair. Such a replacement of the geometric data model results in better coverage of the underlying data distribution compared to existing geometric sampling methods and allows the generation of synthetic points of the minority class closer to the majority class on the decision boundary. We experimentally demonstrate that our Simplicial SMOTE outperforms several popular geometric sampling methods, including the original SMOTE. Moreover, we show that simplicial sampling can be easily integrated into existing SMOTE extensions. We generalize and evaluate simplicial extensions of the classic Borderline SMOTE, Safe-level SMOTE, and ADASYN algorithms, all of which outperform their graph-based counterparts.
This paper addresses the Poisson πps sampling problem, a topic of significant academic interest in various domains and with practical data mining applications, such as influence maximization. The problem includes a set S of n elements, where each element v is assigned a weight w(v) reflecting its importance. The goal is to generate a random subset X of S, where each element v ∈ S is included in X independently with probability c ⋅ w(v)over ∑v ∈ S w(v), where 0 < c≤ 1 is a constant. The subsets must be independent across different queries. While the Poisson πps sampling problem can be reduced to the well-studied subset sampling problem, updates in Poisson πps sampling, such as adding a new element or removing an element, would cause the probabilities of all n elements to change in the corresponding subset sampling problem, making this approach impractical for dynamic scenarios. To address this, we propose a dynamic index specifically tailored for the Poisson πps sampling problem, supporting optimal expected O (1) query time and O (1) index update time, with an optimal O (n) space cost. Our solution involves recursively partitioning the set by weights and ultimately using table lookup. The core of our solution lies in addressing the challenges posed by weight explosion and correlations between elements. Empirical evaluations demonstrate that our approach achieves significant speedups in update time while maintaining consistently competitive query time compared to the subset-sampling-based methods.
When modifying existing policies in high-risk settings, it is often necessary to ensure with high certainty that the newly proposed policy improves upon a baseline, such as the status quo. In this work, we consider the problem of safe policy improvement, where one only adopts a new policy if it is deemed to be better than the specified baseline with at least a pre-specified probability. We focus on threshold policies, a ubiquitous class of policies with applications in economics, healthcare, and digital advertising. Existing methods rely on potentially underpowered safety checks and limit the opportunities for finding safe improvements, so too often they must revert to the baseline to maintain safety. We overcome these issues by leveraging the most powerful safety test in the asymptotic regime and allowing for multiple candidates to be tested for improvement over the baseline. We show that in adversarial settings, our approach controls the rate of adopting a policy worse than the baseline to the pre-specified error level, even in moderate sample sizes. We present CSPI and CSPI-MT, two novel algorithms for selecting cutoff(s) to maximize the policy improvement from baseline. We demonstrate through both synthetic and external datasets that our approaches improve both the detection rates of safe policies and the realized improvement, particularly under stringent safety requirements and low signal-to-noise conditions.
The rise of link recommendation systems in online social networks has sparked significant research interest in strategically adding links to enhance social influence. This paper delves into the influence maximization with augmentation (IMA) problem that aims to add k edges connecting seed nodes and ordinary nodes to boost the influence propagation of the given seed set. IMA is a monotone submodular maximization problem so that the greedy algorithm provides a (1-1/e-ε)-approximate solution, where ε is an error term caused by the intractable nature of influence spread computation. Previous work often utilizes an unbiased estimator that relies on the chosen edges for influence estimation, resulting in non-submodular estimate with respect to edge selection. To ensure the overall error being bounded by ε, such an estimator requires Θ(ε/k) multiplicative error for each estimation, incurring prohibitive overhead. Meanwhile, some other work approximates IMA via conventional influence maximization (IM) on an augmented graph by adding a new node for every edge candidate, leading to heavy extra sampling due to a significant increase in graph size. To address these challenges, we design a novel unbiased estimator on the original graph that is independent of the chosen edges by leveraging the tractability of one-hop influence computation. We show that the estimate via our estimator is submodular so that it enables the estimate of all k edges in a whole with a bounded estimation error of Θ(ε), saving O(k2) time compared to the chosen-edge-dependent estimator while retaining the same graph size. Moreover, we propose several techniques based on the properties of our estimator to further speed up the greedy selection. Putting it together, we develop a scalable algorithm for the IMA problem, namely ScaLIM. Finally, extensive experiments are conducted to validate the effectiveness and efficiency of our proposed approach, e.g., ScaLIM is faster than baselines by nearly two orders of magnitude.
In real-world applications, node classification on graphs often faces the challenge of class imbalance, where majority classes dominate training, resulting in biased model performance. Traditional Graph Neural Networks (GNNs) often struggle in such scenarios, as they tend to overfit to majority classes while underrepresenting minority classes. Existing solutions, which either prioritize nodes based on class size or synthesize new nodes for minority classes, often fall short of effectively addressing this imbalance issue. This paper introduces a novel approach to class-imbalanced node classification by utilizing a balanced meta-set for importance measurement, where a training node is considered significant if it enhances model performance under an unbiased setting. Our method identifies important nodes that can counteract class imbalance and utilizes them for model training, allowing for fine-grained and dynamic node selection throughout the training process. We theoretically derive a formula to directly assess node importance, reducing computational overhead and providing an intuitive threshold for node selection. Guided by this metric, we develop a novel framework that filters valuable labeled, unlabeled, and synthetic nodes that enhance model performance in an unbiased context. A key advantage of this framework is its separation of the synthetic node generation process from the filtering process, ensuring compatibility with various node generation techniques. Furthermore, we introduce a strategy to construct a high-quality meta-set that closely approximates the overall feature distribution, ensuring robust representation of each class. We evaluate our framework, NodeImport, across multiple benchmark datasets using popular GNN architectures, demonstrating its superiority over state-of-the-art baselines. Our results highlight the flexibility and effectiveness of the framework in mitigating class imbalance, leading to improved node classification outcomes. The source code is available at https://github.com/NanChanNN/NodeImport.
The online information ecosystem enables influence campaigns of unprecedented scale and impact. We urgently need empirically grounded approaches to counter the growing threat of malicious campaigns, now amplified by generative AI. But, developing defenses in real-world settings is impractical. Social system simulations with agents modelled using Large Language Models (LLMs) are a promising alternative approach and a growing area of research. However, existing simulators lack features needed to capture the complex information-sharing dynamics of platform-based social networks. To bridge this gap, we present SandboxSocial, a new simulator that includes several key innovations, mainly: (1) a virtual social media platform (modelled as Mastodon and mirrored in an actual Mastodon server) that enables a realistic setting in which agents interact; (2) an adapter that uses real-world user data to create more grounded agents and social media content; and (3) multi-modal capabilities that enable our agents to interact using both text and images---just as humans do on social media. We make the simulator more useful to researchers by providing measurement and analysis tools that track simulation dynamics and compute evaluation metrics to compare experimental results.
While current time series research focuses on developing new models, crucial questions of selecting an optimal approach for training such models are underexplored. Tsururu, a Python library introduced in this paper, bridges SoTA research and industry by enabling flexible combinations of global and multivariate approaches and multi-step-ahead forecasting strategies. It also enables seamless integration with various forecasting models. Available at https://github.com/sb-ai-lab/tsururu.
We formulate a Coverage Path Planning (CPP) problem for a helicopter or a UAV tasked with mapping ground-level radiation while avoiding radiation that is too strong. We introduce a simulation environment that incorporates digital elevation models, altitude-dependent measurement footprints and realistic flight constraints, as well as state-of-the-art radiation scenario simulations, such as nuclear explosions, provided by the German Federal Office for Radiation Protection. We highlight the complexity of radiological survey missions and demonstrate the necessity for new CPP approaches that address these unique challenges. The code to our simulation environment can be found under https://github.com/JohannBlake/Aerial-Coverage-Path-Planning-in-Nuclear-Emergencies.
Combining Code Generating Large Language Models and Self-Play to Iteratively Refine Strategies in Games
PDF ↗We propose a self-play approach to generating strategies for playing in multi-player games, where strategies are represented as computer code. We use large language models (LLMs) to generate pieces of code to play in the game, which we refer to as generated bots. We engage the LLM generated bots in competitions, designed to generate increasingly stronger strategies. We follow game theoretic principles in organizing these tournaments, and use a Policy Space Response Oracle (PSRO) approach. We start with an initial set of LLM generated bots, and continue in rounds for adding new bots into the population. Each round adds a bot to the population by asking the LLM to produce code for playing against a bot representing the Nash equilibrium mixture over the current population. Our analysis shows that even a few rounds are sufficient to produces strong bots for playing the game. Our demo shows the process for the game of Checkers. We allow users to select initial bots in the population, run the process, inspect how the bots evolve over time, and play against the generated bots.
Removing unwanted consequences from a knowledge base has been investigated in belief change under the name contraction and is called repair in ontology engineering. Simple repair and contraction approaches based on removing statements from the knowledge base (respectively called belief base contractions and classical repairs) have the disadvantage that they are syntax-dependent and may remove more consequences than necessary. Belief set contractions do not have these problems, but may result in belief sets that have no finite representation. Similarly, optimal repairs, which are syntax-independent and maximize the retained consequences, may not exist. Our KR 2024 paper leverage advances in characterizing and computing optimal repairs of ontologies based on the description logics EL to obtain contraction operations that combine the advantages of belief set and belief base contractions. It introduces this new approach in a very general setting, and proves a characterization theorem that relates the obtained contractions with well-known rationality postulates. Then, it describes a variety of interesting instances, not only in the standard repair/contraction setting where one wants to get rid of a consequence, but also in other settings such as variants of forgetting in propositional and description logic.
The proliferation of edge devices has generated an unprecedented volume of time series data across different domains, motivating a variety of well-customized methods. Recently, Large Language Models (LLMs) have emerged as a new paradigm for time series analytics by leveraging the shared sequential nature of textual data and time series. However, a fundamental cross-modality gap between time series and LLMs exists, as LLMs are pre-trained on textual corpora and are not inherently optimized for time series. Many recent proposals are designed to address this issue. In this survey, we provide an up-to-date overview of LLMs-based cross-modality modeling for time series analytics. We first introduce a taxonomy that classifies existing approaches into four groups based on the type of textual data employed for time series modeling. We then summarize key cross-modality strategies, e.g., alignment and fusion, and discuss their applications across a range of downstream tasks. Furthermore, we conduct experiments on multimodal datasets from different application domains to investigate effective combinations of textual data and cross-modality strategies for enhancing time series analytics. Finally, we suggest several promising directions for future research. This survey is designed for a range of professionals, researchers, and practitioners interested in LLM-based time series modeling.
Graph neural networks (GNNs) are powerful deep learning models for graph-structured data, demonstrating remarkable success across diverse domains. Recently, the database (DB) community has increasingly recognized the potentiality of GNNs, prompting a surge of researches focusing on improving database systems through GNN-based approaches. However, despite notable advances, There is a lack of a comprehensive review and understanding of how GNNs could improve DB systems. Therefore, this survey aims to bridge this gap by providing a structured and in-depth overview of GNNs for DB systems. Specifically, we propose a new taxonomy that classifies existing methods into two key categories: (1) Relational Databases, which includes tasks like performance prediction, query optimization, and Text-to-SQL, and (2) Graph Databases, addressing challenges like efficient graph query processing and graph similarity computation. We systematically review key methods in each category, highlighting their contributions and practical implications. Finally, we suggest promising avenues for integrating GNNs into Database systems.
A Comprehensive Survey on Physical Risk Control in the Era of Foundation Model-enabled Robotics
PDF ↗Recent Foundation Model-enabled robotics (FMRs) display greatly improved general-purpose skills, enabling more adaptable automation than conventional robotics. Their ability to handle diverse tasks thus creates new opportunities to replace human labor. However, unlike general foundation models, FMRs interact with the physical world, where their actions directly affect the safety of humans and surrounding objects, requiring careful deployment and control. Based on this proposition, our survey comprehensively summarizes robot control approaches to mitigate physical risks by covering all the lifespan of FMRs ranging from pre-deployment to post-accident stage. Specifically, we broadly divide the timeline into the following three phases: (1) pre-deployment phase, (2) pre-incident phase, and (3) post-incident phase. Throughout this survey, we find that there is much room to study (i) pre-incident risk mitigation strategies, (ii) research that assumes physical interaction with humans, and (iii) essential issues of foundation models themselves. We hope that this survey will be a milestone in providing a high-resolution analysis of the physical risks of FMRs and their control, contributing to the realization of a good human-robot relationship.