Yuhang Niu, Hongyuan Xu, Ciyi Liu, Bofan Wei, Jiaqi Ye, Yanlong Wen, Xiaojie Yuan. Proceedings of the 64th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers). 2026.

Oblivious RAM (ORAM) hides memory access patterns but imposes high bandwidth and latency overheads. Prior work (e.g., Hitchhiker) introducedpath diversionto serve multiple logical requests in one traversal, yet diversion rarely triggers because it requires requests to lie on the same ancestor–descendant chain.We present Path Weaver, an ORAM that turns diversion into the common case viarandom-path scheduling. Instead of a rigid root-to-leaf path, each traversal selects one uniformly random bucket per level; the controller then “stitches” per-level picks so different outstanding requests can be satisfied at different levels of thesametraversal. ATail-Wideningscheme fetches a few additional buckets at the dense bottom levels, substantially increasing match opportunities with constant overhead. To keep hardware small, Path Weaver uses a compact position map and regenerates per-level choices on the fly with a pipelined pseudorandom function (PRF) overlapped with DRAM access; we micro-profile these stages to quantify their cost. The design targets random-access main memory (DRAM) and requires no changes to the software stack.Path Weaver preserves the Path-style invariant of one legitimate bucket per level, thereby retaining indistinguishability and the same stash-overflow guarantees as Path ORAM. Evaluated in gem5 on SPEC CPU 2017, Path Weaver processes 2.32× more requests per traversal than Hitchhiker, reduces single-core ORAM latency by 2.1×, and accelerates four-core workloads by 3.3×. By making path diversion a core architectural feature, Path Weaver brings ORAM significantly closer to practical, ’deployonce” secure memory.

Multivariate time series forecasting has become increasingly crucial in fields such as energy and transportation. Recent research has focused on local lag relationships across variates, yielding impressive results. However, these methods typically require pre-calculating lag indicators and steps between variates based on historical data. This reliance on pre-calculated lag steps neglects the potential variability in lag steps over the historical and predicted time series. In this paper, we propose LagTS, a novel method that adaptively models lag relationships across variates in multivariate time series data and effectively leverages these relationships to improve prediction accuracy. Specifically, the proposed method extracts the lag relationships using a dedicated module. Moreover, it employs a lag relation loss to facilitate the adaptive modeling of the lag relationships. Extensive experimental results on four publicly available datasets and an industry dataset demonstrate the effectiveness of the proposed method.

Accurate forecasting of resource consumption, such as gas, is essential for efficient energy management, cost reduction, and sustainability. Time series forecasting (TSF) techniques like recurrent neural networks (RNNs), convolutional networks (TCNs), and Transformers have been employed to model complex temporal variations but face challenges in capturing long-term dependencies or inter-series relationships. In this paper, we propose a Hierarchical Contrastive Learning approach for Time Series (HCLTS) to improve gas consumption prediction for industrial and commercial users. HCLTS employs a prototype-based approach to construct positive and negative samples, leveraging industry labels and hierarchical sampling. We further adopt a hierarchical contrastive training objective to model underlying consumption patterns across industries. Extensive experiments on a real-world gas consumption dataset demonstrate that HCLTS achieves superior performance over existing methods, highlighting its potential for practical applications in diverse industrial contexts.

Taxonomy completion aims to integrate new concepts into existing taxonomies by determining their appropriate hypernym and hyponym. While semantic and structural information are crucial for this task, existing approaches often struggle to balance these aspects effectively. In this paper, we propose COMI, an efficient taxonomy completion framework that leverages large language models (LLMs) to capture both semantic and structural information in a unified manner. COMI compresses node semantics into token representations, enabling LLMs to efficiently process the input structure composed of these tokens. To enhance the model's understanding of the structure, a further fine-tuning process using contrastive learning with mixup data augmentation is applied, where mixup generates diverse and challenging negative samples. Through these innovations, COMI improves the integration of semantic and structural information, leading to more accurate taxonomy completion. The experimental results on three real-world datasets demonstrate that COMI achieves state-of-the-art performance while showing up to 284x faster inference compared to the previous best method. Our code and compressed tokens are available at https://github.com/cyclexu/COMI.

Abstract Low-resource taxonomy completion aims to automatically insert new concepts into the existing taxonomy, in which only a few in-domain training samples are available. Recent studies have achieved considerable progress by incorporating prior knowledge from pre-trained language models (PLMs). However, these studies tend to overly rely on such knowledge and neglect the shareable knowledge across different taxonomies. In this paper, we propose TaxoPro, a plug-in LoRA-based cross-domain method, that captures shareable knowledge from the high- resource taxonomy to improve PLM-based low-resource taxonomy completion techniques. To prevent negative interference between domain-specific and domain-shared knowledge, TaxoPro decomposes cross- domain knowledge into domain-shared and domain-specific components, storing them using low-rank matrices (LoRA). Additionally, TaxoPro employs two auxiliary losses to regulate the flow of shareable knowledge. Experimental results demonstrate that TaxoPro improves PLM-based techniques, achieving state-of-the-art performance in completing low-resource taxonomies. Code is available at https://github.com/cyclexu/TaxoPro.

Automatic taxonomy completion aims to attach the emerging concept to an appropriate pair of hypernym and hyponym in the existing taxonomy.Existing methods suffer from the overfitting to leaf-only problem caused by imbalanced leaf and non-leaf samples when training the newly initialized classification head.Besides, they only leverage subtasks, namely attaching the concept to its hypernym or hyponym, as auxiliary supervision for representation learning yet neglect the effects of subtask results on the final prediction.To address the aforementioned limitations, we propose TacoPrompt, a Collaborative Multi-Task Prompt Learning Method for Self-Supervised Taxonomy Completion.First, we perform triplet semantic matching using the prompt learning paradigm to effectively learn non-leaf attachment ability from imbalanced training samples.Second, we design the result context to relate the final prediction to the subtask results by a contextual approach, enhancing prompt-based multi-task learning.Third, we leverage a two-stage retrieval and re-ranking approach to improve the inference efficiency.Experimental results on three datasets show that TacoPrompt achieves state-of-the-art taxonomy completion performance.Codes are available at https://github.com/cyclexu/TacoPrompt.

ICD coding is designed to assign the disease codes to electronic health records (EHRs) upon discharge, which is crucial for billing and clinical statistics. In an attempt to improve the effectiveness and efficiency of manual coding, many methods have been proposed to automatically predict ICD codes from clinical notes. However, most previous works ignore the decisive information contained in structured medical data in EHRs, which is hard to be captured from the noisy clinical notes. In this paper, we propose a Tree-enhanced Multimodal Attention Network (TreeMAN) to fuse tabular features and textual features into multimodal representations by enhancing the text representations with tree-based features via the attention mechanism. Tree-based features are constructed according to decision trees learned from structured multimodal medical data, which capture the decisive information about ICD coding. We can apply the same multi-label classifier from previous text models to the multimodal representations to predict ICD codes. Experiments on two MIMIC datasets show that our method outperforms prior state-of-the-art ICD coding approaches. The code is available at https://github.com/liu-zichen/TreeMAN.

Taxonomies are hierarchical classifications widely exploited to facilitate downstream natural language processing tasks. The taxonomy expansion task aims to incorporate emergent concepts into the existing taxonomies. Prior works focus on modeling the local substructure of taxonomies but neglect the global structure. In this paper, we propose TaxoPrompt, a framework that learns the global structure by prompt tuning with taxonomic context. Prompt tuning leverages a template to formulate downstream tasks into masked language model form for better distributed semantic knowledge use. To further infuse global structure knowledge into language models, we enhance the prompt template by exploiting the taxonomic context constructed by a variant of the random walk algorithm. Experiments on seven public benchmarks show that our proposed TaxoPrompt is effective and efficient in automatically expanding taxonomies and achieves state-of-the-art performance.

Session-based recommendation (SR) gains increasing popularity because it helps greatly maintain users' privacy. Aside from its efficacy, explainability is also critical for developing a successful SR model, since it can improve the persuasiveness of the results, the users' satisfaction, and the debugging efficiency. However, the majority of current SR models are unexplainable and even those that claim to be interpretable cannot provide clear and convincing explanations of users' intentions and how they influence the models' decisions. To solve this problem, in this research, we propose a meta-path guided model which uses path instances to capture item dependencies, explicitly reveal the underlying motives, and illustrate the entire reasoning process. To begin with, our model explores meta-path guided instances and leverages the multi-head self-attention mechanism to disclose the hidden motivations beneath these path instances. To comprehensively model the user interest and interest shifting, we search paths in both adjacent and non-adjacent items. Then, we update item representations by incorporating the user-item interactions and meta-path-based context sequentially. Compared with recent strong baselines, our method is competent to the SOTA performance on three datasets and meanwhile provides sound and clear explanations.

The federated learning provides an effective solution to train collaborative models over a large scale of participated Industrial Internet of Things (IIoT) applications with the help of a global server, building an intelligent life. However, the federated learning is vulnerable to the backdoor attack from strong malicious participants. The backdoor attack is inconspicuous and may result in devastating consequences. To resist the attack on IIoT applications, we propose the federated backdoor filter defense that can identify backdoor inputs and restore the data to availability by theblur-label-flippingstrategy. We build multiple filters with eXplainable AI models on the server and send them to clients randomly, preventing advanced attackers from evading the defense. Our backdoor filters show significant backdoor recognition with the accuracy up to 99%. After the implementation of the blur-label-flipping strategy, victim's local model on suspicious backdoor samples can achieve the accuracy up to 88%.

As an essential form of knowledge representation, taxonomies are widely used in various downstream natural language processing tasks.However, with the continuously rising of new concepts, many existing taxonomies are unable to maintain coverage by manual expansion.In this paper, we propose TEMP, a self-supervised taxonomy expansion method, which predicts the position of new concepts by ranking the generated taxonomy-paths.For the first time, TEMP employs pre-trained contextual encoders in taxonomy construction and hypernym detection problems.Experiments prove that pre-trained contextual embeddings are able to capture hypernym-hyponym relations.To learn more detailed differences between taxonomy-paths, we train the model with dynamic margin loss by a novel dynamic margin function.Extensive evaluations exhibit that TEMP outperforms prior state-of-the-art taxonomy expansion approaches by 14.3% in accuracy and 15.8% in mean reciprocal rank on three public benchmarks.

Predicting future link states of a temporal network from past evolution is an important task from application perspective. Different from static networks analysis, temporal link prediction considers the trends of the links as time goes by, which is proved can achieve higher performance than that only exploring the interactions at a specific moment. The key challenge of this task is to obtain effective features of the evolving network. We propose a novel supervised method CTMF based on motif feature, which can characterize nearly complete local topological characteristic of node-pairs in the temporal network. We also attach cluster information as well as temporal transition factor to capture the influence of global structure and network transition on each motif over time. An unsupervised learning strategy is used on a concatenation of the cluster and temporal information based motif features (CTMF) from the continuous sets of snapshots to capture a better feature representation. We validate the effectiveness of the CTMF by utilizing it for link prediction on a number of temporal datasets, which shows our approach outperforms state-of-art methods using other features.

Time series classification presents a specific machine learning challenge due to the ordering of variables. Recent studies show that the simple nearest neighbor classifier with elastic distance measures is hard to beat and many researchers focus on alternative distance measures. Unlike nearest neighbor classifier try to find a training sample which has the minimum distance with test instance, we utilize a reconstruction strategy to determine the label of new time series in this paper. Concretely, for each test time series, we reconstruct it by using as few training samples as possible and then calculate the residuals between the test time series and the selected training samples of each class. The test time series is classified to the class with minimum residual. To get the required time series from the training set, we employ sparse restriction technique to discover the optimal combination of different training samples while fitting test time series. Meanwhile, to solve the scenarios where the time series dataset is linearly inseparable, we extend our method by the kernel trick. Extensive experimental results show that the proposed method can gain the significant improvement on commonly used time series datasets.