Accurate IP geolocation plays a critical role in a wide range of location-aware applications, from cybersecurity to content delivery. While recent advances in deep learning have led to substantial improvements in geolocation accuracy, existing methods often fail to generalize under out-of-distribution (OOD) scenarios caused by distribution shifts. To address this challenge, we propose a novel framework-Graph Invariant Learning (GIL)-for IP geolocation, referred to as GILGeo. Our approach is designed to identify invariant structural patterns in IP graphs across diverse environments, thereby enhancing model generalizability. By dynamically recombining invariant and spurious features, GILGeo simulates a variety of environmental conditions during training. This promotes the learning of domain-invariant representations and leads to significantly improved performance in unseen OOD settings. Extensive experiments on three real-world datasets show that GILGeo outperforms state-of-the-art baselines, establishing a new benchmark for IP geolocation under distributional shift. Our anonymized code and datasets are publicly available at: https://github.com/xiaohanwang01/GILGeo.

Recent developments have demonstrated the efficacy of enhancing the reliability and reasoning ability of Large Language Models (LLMs) through the integration of Knowledge Graphs (KGs). However, previous works have not fully considered the challenges present in expert domains, including the prevalence of professional terms with abbreviations and ambiguities, knowledge confusion stemming from the mixture of KGs with other retrieval methods, and others, which hinder their deployment in expert question-answering (QA) systems. In this work, we propose a QA framework for expert fields by Synergizing large Language models And knowledge Graphs (SLAG). SLAG is designed to improve the accuracy of KG retrieval and to intelligently coordinate KG retrievers with other retrieval methods through bidirectional enhancement between LLMs and KGs: (1) An LLM-Enhanced KG retriever is proposed to resolve entity ambiguities in queries and to accurately extract pertinent subgraphs. (2) A KG-Enhanced LLM reasoner is designed to distill subgraphs and to flexibly generate answers or explicit queries for cooperative retrievers. We compared SLAG with other retrieval techniques on multiple datasets across diverse expert domains, tasks, and languages. We observed that our approach achieved a relative improvement of 13% over state-of-the-art methods on public benchmarks in terms of the F1 score. Moreover, we successfully deployed SLAG in an industrial financial QA system and outperformed the latest online version. Finally, we provide an open source version of SLAG with code and datasets, which is available at https://anonymous.4open.science/r/SLAG-5611.

In long-context LLM serving, the prefill stage often dominates time-to-first-token and computational cost. Although Prefix Cache in vLLM/PagedAttention has been widely used to reuse identical prompt prefixes, repeated content in practical applications frequently appears as non-prefix, cross-request, cross-turn, and cross-agent segments, which makes conventional cache mechanisms insufficient. This paper presents SparseX, a segment-level KV Cache sharing method for common serving scenarios. SparseX uses contiguous token segments as reuse units and exploits Sparse-Q indices that naturally arise in KV Cache reuse workloads to estimate the key tokens that require correction. Based on this estimate, SparseX performs Sparse-KV Recomputation within a single forward pass, thereby restoring cross-segment contextual interactions under complex interleaved reuse patterns while avoiding additional models or separate preprocessing stages for token selection. SparseX further implements a full+sparse hybrid attention mode based on a layer-specific threshold: early layers retain full attention to obtain a more stable token-importance signal, and later layers switch to sparse recomputation to improve reuse quality on complex long-context tasks. We implement SparseX-vLLM on top of vLLM, integrating segment-level cache lookup, PagedAttention management, RoPE alignment, Sparse-Q token selection, and FlashAttention backends into a unified execution path. SparseX is model-agnostic, training-free, and compatible with Prefix Cache, and it provides unified support for common online serving scenarios including multi-round chat, retrieval-augmented generation (RAG), and agent workflows.

Due to the high operational complexity and limited deployment scale of lossless RDMA networks, the community has been exploring efficient RDMA communication over lossy fabrics. State-of-the-art (SOTA) lossy RDMA solutions implement a simplified selective repeat mechanism in RDMA NICs (RNICs) to enhance loss recovery efficiency. However, these solutions still face performance challenges, such as unavoidable ECMP hash collisions and excessive retransmission timeouts (RTOs). In this paper, we enhance RDMA reliability with the goals of being independent of PFC, compatible with packet-level load balancing, free from RTO, and friendly to hardware offloading. To this end, we propose DCP, a transport architecture that co-designs both the switch and RNICs, fully meeting the design goals. At its core, DCP-Switch introduces a simple yet effective lossless control plane, which is leveraged by DCP-RNIC to enhance reliability support for high-speed lossy fabrics, primarily including header-only-based retransmission and bitmap-free packet tracking. We prototype DCP-Switch using P4 switch and DCP-RNIC using FPGA. Extensive experiments demonstrate that DCP achieves 1.6× and 2.1× performance improvements, compared to SOTA lossless and lossy RDMA solutions, respectively.

Anomaly Detection (AD) technology has received much attention recently, especially in industrial quality inspection applications. Most existing unsupervised AD methods assume that the training data contains only normal samples, which is difficult to satisfy in practice. When the training data are mixed with even a small number of defective samples, the AD methods, that use distillation learning, will be negatively affected, leading to significant performance drops. To tackle this issue, in this paper, we proposed an approach, namely AD^2 , to conduct anomaly detection during the training phase of an anomaly detection model. Specifically, we devise a Non-Major Feature Elimination (NMFE) module to eliminate the prominent anomaly-related discrepancy information and adopt an Anomaly Training Data Removal (ATDR) strategy to identify outliers in the training data, preventing abnormal information from affecting model training. During the inference phase, AD^2 does not introduce any extra computation overhead. Experiments demonstrate that AD^2 can successfully alleviate the performance deterioration caused by polluted training samples. On the MVTec LOCO dataset, when 10% of the training set is corrupted by anomalous samples, AD^2 can significantly improve the image-level AUROC from 0.793 to 0.865 compared to the ordinary AD method, without sacrificing any inference efficiency. AD^2 provides an effective solution for issues of data uncertainty in anomaly detection. The source code will be released.

Recent years have witnessed a plethora of learning-based solutions for congestion control (CC) that demonstrate better performance over traditional TCP schemes. However, they fail to provide consistently good convergence properties, including fairness, fast convergence and stability, due to the mismatch between their objective functions and these properties. Despite being intuitive, integrating these properties into existing learning-based CC is challenging, because: 1) their training environments are designed for the performance optimization of single flow but incapable of cooperative multi-flow optimization, and 2) there is no directly measurable metric to represent these properties into the training objective function. We present Astraea, a new learning-based congestion control that ensures fast convergence to fairness with stability. At the heart of Astraea is a multi-agent deep reinforcement learning framework that explicitly optimizes these convergence properties during the training process by enabling the learning of interactive policy between multiple competing flows, while maintaining high performance. We further build a faithful multi-flow environment that emulates the competing behaviors of concurrent flows, explicitly expressing convergence properties to enable their optimization during training. We have fully implemented Astraea and our comprehensive experiments show that Astraea can quickly converge to fairness point and exhibit better stability than its counterparts. For example, Astraea achieves near-optimal bandwidth sharing (i.e., fairness) when multiple flows compete for the same bottleneck, delivers up to 8.4 & times; faster convergence speed and 2.8 & times; smaller throughput deviation, while achieving comparable or even better performance over prior solutions.

The emergence of writable, cross-session persistent memory in LLM agents introduces a qualitatively different threat landscape from conventional input-centric security concerns, characterized by three properties: persistence, statefulness, and propagation. To systematically characterize this landscape, we propose a Memory Lifecycle Framework that organizes attacks, defenses, and their cross-phase dependencies along two axes: six lifecycle phases (Write, Store, Retrieve, Execute, Share Propagate, Forget Rollback) and four security objectives (Integrity, Confidentiality, Availability, Governance). This analysis in turn exposes the need for formal security guarantees at the system level, motivating Verifiable Memory Governance(VMG), a framework of five architectural primitives that specifies what verifiable mechanisms a long-term-memory system must provide to maintain auditable, recoverable control over its memory state. Our analysis indicates that robust Long-Term Memory (LTM) security cannot be retrofitted at retrieval or execution time alone, but must be anchored in storage-time provenance, versioning, and policy-aware retention from the outset.

Federated fine-tuning of Large Language Models (LLMs) is obstructed by a trilemma of challenges: protecting LLMs intellectual property (IP), ensuring client privacy, and mitigating performance loss on heterogeneous data. Existing methods like Offsite-Tuning (OT) secure the LLMs IP by having clients train only lightweight adapters, yet our analysis reveals they suffer from a fundamental performance bottleneck, leaving a significant gap compared to centralized training. To bridge this gap, we introduce FedProxy, a new federated adaptation framework. FedProxy replaces weak adapters with a unified, powerful Proxy Small Language Model (SLM), compressed from the proprietary LLM, to serve as a high-fidelity surrogate for collaborative fine-tuning. Our framework systematically resolves the trilemma through a three-stage architecture: (i) Efficient Representation via server-guided compression to create a resource-friendly proxy; (ii) Robust Optimization through an interference-mitigating aggregation strategy to handle data heterogeneity; and (iii) Effortless Fusion via a training-free "plug-in" mechanism to integrate learned knowledge back into the LLM. Experiments show FedProxy significantly outperforms OT methods and approaches centralized performance, establishing a new benchmark for secure and high-performance federated LLM adaptation.

Although deep research agents (DRAs) have emerged as a promising paradigm for complex information synthesis, their evaluation remains constrained by ad hoc empirical benchmarks. These heuristic approaches do not rigorously model agent behavior or adequately stress-test long-horizon synthesis and ambiguity resolution. To bridge this gap, we formalize DRA behavior through the lens of category theory, modeling deep research workflow as a composition of structure-preserving maps (functors). Grounded in this theoretical framework, we introduce a novel mechanism-aware benchmark with 296 questions designed to stress-test agents along four interpretable axes: traversing sequential connectivity chains, verifying intersections within V-structure pullbacks, imposing topological ordering on retrieved substructures, and performing ontological falsification via the Yoneda Probe. Our rigorous evaluation of 11 leading models establishes a persistently low baseline, with the state-of-the-art achieving only a 19.9\% average accuracy, exposing the difficulty of formal structural stress-testing. Furthermore, our findings reveal a stark dichotomy in the current AI capabilities. While advanced deep research pipelines successfully redefine dynamic topological re-ordering and exhibit robust ontological verification -- matching pure reasoning models in falsifying hallucinated premises -- they almost universally collapse on multi-hop structural synthesis. Crucially, massive performance variance across tasks exposes a lingering reliance on brittle heuristics rather than a systemic understanding. Ultimately, this work demonstrates that while top-tier autonomous agents can now organically unify search and reasoning, achieving a generalized mastery over complex structural information remains a formidable open challenge.\footnote{Our implementation will be available at https://github.com/tzq1999/CDR.

Federated graph learning (FGL) has attracted significant attention for enabling privacy-preserving collaborative model training based on multiple participants’ local graphs. However, a node’s neighbors in one participant’s local graph may be distributed across the local graphs of other participants, resulting in topological nonindependence and creating a specific non-independently and identically distributed (non-IID) problem. Existing unsupervised federated learning approaches under topological nonindependence encounter two problems: 1) the lack of encoding the distances between nodes across different participants’ local graphs reduces the models’ accuracy; and 2) the volume of data transmitted escalates with the increase in inter-edges or nodes, leading to great communication overhead. To tackle these problems, we propose a federated graph learning model based on contrastive encoding and cluster centroid sampling (FCECS), a novel unsupervised FGL model. First, we design a local contrastive strategy based on alignment masking and a global contrastive strategy based on cluster centroid sampling and integrate them to encode node distances both within and across participants’ local graphs, thereby improving the model’s accuracy. Second, the global contrastive strategy based on cluster centroid sampling guarantees a consistent volume of data transmission, thereby reducing communication overhead caused by an increasing number of inter-edges or nodes. Experimental results on real-world datasets demonstrate that FCECS achieves an average accuracy improvement of 57% compared to existing FGL models.

Token filtering has been proposed to enhance the utility of large language models (LLMs) by eliminating inconsequential tokens during training. While using fewer tokens is expected to reduce computational workloads, existing methods have not yet achieved a real-world efficiency boost. This is primarily due to two factors: (1) existing work has inadequate sparsity for speedup, and (2) token filtering operates within a sparsity range that is non-standard in existing machine learning (ML) libraries and thus cannot be efficiently supported. This paper presents Centrifuge, a system that leverages algorithm and system co-design to unleash the full efficiency of token filtering in LLM training. At the algorithm level, Centrifuge filters activations of inconsequential tokens in the attention backward kernel to amplify the sparsity in backward computation. At the system level, Centrifuge proposes an automatic workflow that transforms sparse GEMM into dimension-reduced dense GEMM for optimized efficiency using standard ML libraries. Evaluations on models with various scales—from 1.1B to 40B—demonstrate that Centrifuge reduces backpropagation time by up to 49.9\% and end-to-end training time by up to 34.7\% when filtering 50\% of tokens. Utility assessments indicate that Centrifuge preserves the utility benefits of token filtering and significantly enhances model performance by up to 26.6\% compared to standard training. Centrifuge is designed for seamless integration into existing LLM training frameworks, enabling systems already utilizing token filtering to accelerate training with just one line of code.

ByteScheduler partitions and rearranges tensor transmissions to improve the communication efficiency of distributed Deep Neural Network (DNN) training. The configuration of hyper-parameters (i.e., the partition size and the credit size) is critical to the effectiveness of partitioning and rearrangement. Currently ByteScheduler adopts Bayesian Optimization (BO) to find the optimal configuration for the hyper-parameters beforehand. In practice, however, various runtime factors (such as worker node status and network conditions) change over time, making the statically-determined one-shot configuration result suboptimal for real-world DNN training. To address this problem, in this paper we present a realtime configuration method (called AutoByte) that automatically and timely searches the optimal hyper-parameters as the training systems dynamically change. AutoByte extends the ByteScheduler framework with a meta network, which takes the systems’ runtime statistics as its input, dynamically adjusts the triggering threshold based on system environment characteristics, and outputs predictions for speedups under specific configurations. Evaluation results on various DNN models show that AutoByte can dynamically tune the hyper-parameters with low resource usage, and deliver up to 33.2% higher performance than the best static configuration method on the ByteScheduler framework.

We present Innovator-VL, a scientific multimodal large language model designed to advance understanding and reasoning across diverse scientific domains while maintaining excellent performance on general vision tasks. Contrary to the trend of relying on massive domain-specific pretraining and opaque pipelines, our work demonstrates that principled training design and transparent methodology can yield strong scientific intelligence with substantially reduced data requirements. (i) First, we provide a fully transparent, end-to-end reproducible training pipeline, covering data collection, cleaning, preprocessing, supervised fine-tuning, reinforcement learning, and evaluation, along with detailed optimization recipes. This facilitates systematic extension by the community. (ii) Second, Innovator-VL exhibits remarkable data efficiency, achieving competitive performance on various scientific tasks using fewer than five million curated samples without large-scale pretraining. These results highlight that effective reasoning can be achieved through principled data selection rather than indiscriminate scaling. (iii) Third, Innovator-VL demonstrates strong generalization, achieving competitive performance on general vision, multimodal reasoning, and scientific benchmarks. This indicates that scientific alignment can be integrated into a unified model without compromising general-purpose capabilities. Our practices suggest that efficient, reproducible, and high-performing scientific multimodal models can be built even without large-scale data, providing a practical foundation for future research.