arXiv:2606.04671v1 Announce Type: new
Abstract: Exceptional points appear in non-Hermitian systems as degeneracies, where not only eigenvalues but also eigenvectors coalesce. They are of great theoretical and experimental interest due to their exotic topological properties and enhanced sensitivity to perturbations. Experimental realizations of higher-order exceptional points, where more than two eigenvectors coalesce, rely on highly fine-tuned setups. Recently, stimulated Brillouin scattering has been employed to generate second-order exceptional points in a fabrication-free setup by leveraging off-resonant scattering. In this work we generalize this approach, and we develop an off-resonant, multimode theory for stimulated Brillouin scattering as an avenue towards realizing symmetry-induced exceptional points of any order. We present the experimental implementation of our program in an accompanying paper. Our multimode theory could also be employed in applications in optoacoustic sensing, synthetic neuromorphic computing, microwave photonic filters, and optoacoustic quantum signal processing.
Science Journals
arXiv:2606.04272v1 Announce Type: new
Abstract: The standard LLM training pipeline applies reinforcement learning (RL) only after pre-training and supervised fine-tuning (SFT). We question this status quo by training a LLM from scratch and applying RL, SFT, and SFT followed by RL directly to intermediate pre-training checkpoints. We find that RL is effective very early, and often matches the full SFT$\to$RL pipeline early as well. Through experiments on harder problems, we find that targeted pre-training data composition is a strong lever for RL effectiveness, even more so than model scale. Beyond reasoning accuracy, applying RL directly to base checkpoints expands the model's distribution; the sharpening effect reported in recent work arises only when RL follows SFT. The general capabilities of the model remain essentially unchanged by RL, while they degrade following SFT. Finally, we merge RL and SFT objectives by parallel averaging, which outperforms across all other training methods discussed, across metrics, while preserving general capabilities. Together, these results suggest that LLM training might benefit from an expanded use of RL.
arXiv:2606.04285v1 Announce Type: new
Abstract: The $G3W2$ approximation -- the second-order self-energy beyond $GW$ -- is known to violate some fundamental analytic properties of the self-energy. In particular, its lack of positive semi-definiteness leads to unphysical features such as negative spectral functions. In this work, we reformulate the $G3W2$ approximation within the algebraic-diagrammatic construction (ADC) framework. The resulting ADC-$G3W2$ formalism enforces the same analytic form as the exact self-energy, namely a sum-over-state representation, and, consequently, guarantees positive semi-definiteness. Starting from the $GW$ self-energy, we construct a hierarchy of ADC-based approximations of increasing sophistication, including ADC-2SOSEX, ADC(3)-$G3W2$, and a full ADC-$G3W2$ scheme. These methods can be interpreted as nonperturbative resummations of vertex corrections to the self-energy, yielding Hermitian effective Hamiltonians whose diagonalization provides quasiparticle and satellite energies. This establishes a formal bridge between many-body perturbation theory formulated in terms of the screened interaction $W$ and conventional ADC schemes based on the bare Coulomb interaction. The performance of these ADC-based approximations is gauged for valence ionization potentials and benchmarked against their parent method.
arXiv:2606.04286v1 Announce Type: new
Abstract: Online reviews provide valuable insights into the perceived quality of facets of a product or service. While aspect-based sentiment analysis has focused on extracting these facets from reviews, there is less work understanding the impact of each aspect on overall perception. This is particularly challenging given correlations among aspects, making it difficult to isolate the effects of each. This paper introduces a methodology based on recent advances in text-based causal analysis, specifically CausalBERT, to disentangle the effect of each factor on overall review ratings. We enhance CausalBERT with three key improvements: temperature scaling for better calibrated treatment assignment estimates; hyperparameter optimization to reduce confound overadjustment; and interpretability methods to characterize discovered confounds. In this work, we treat the textual mentions in reviews as proxies for real-world attributes. We validate our approach on real and semi-synthetic data from over 600K reviews of U.S. K-12 schools. We find that the proposed enhancements result in more reliable estimates, and that perception of school administration and performance on benchmarks are significant drivers of overall school ratings.
arXiv:2606.04479v1 Announce Type: new
Abstract: Recent text-to-image (T2I) models can render highly legible and well-structured text within images, enabling applications including document generation and slide generation. However, it remains unclear whether such systems faithfully preserve reasoning ability when complex solutions must be expressed directly through rendered text, or whether they merely imitate surface-level patterns. We investigate this question by evaluating reasoning fidelity in visual text generation, where models must express complete reasoning processes as images. Our evaluation includes long text rendering, factual knowledge probing, context understanding, and multi-step reasoning. Across these settings, we find that current T2I models frequently produce semantic errors, logical inconsistencies, and incorrect intermediate steps, even when the rendered text appears visually clear. These failures contrast with the strong reasoning performance of text-only models on the same tasks. Our findings reveal a substantial gap between visual text generation and procedural reasoning, motivating more reliable visual text reasoning.
arXiv:2606.04684v1 Announce Type: new
Abstract: The real-time hardships of video processing seriously limit the usage of Automatic License Plate Recognition (ALPR) with application in dynamic traffic monitoring settings. High-fidelity recognition of unconstrained variables, e.g. drastic variations in illumination, acute camera scans, high vehicle speeds, and harsh physical concealment, is a problem that often leads to disjointed tracking paths and poor Optical Character Recognition (OCR) rates. In order to mitigate these weaknesses, the study proposes a 5 stage, end-to-end algorithmic pipeline, encompassing a smooth transition between deep learning based object detection, multi-object tracking which is kinematic in nature, and geometry temporal data interpolation. The suggested architecture takes advantage of a very powerful YOLOv8 nano model to localize the vehicle at the first stage and then Simple Online and Realtime Tracking (SORT) algorithm is used to build spatial-temporal links between frames. Another, more specific typology of YOLOv8 object detectors the license plate area, channeling the sliced array to an EasyOCR chain under the limitations of positional syntax verification. More importantly, an offline interpolation mechanism of temporal bounding box is initiated to recast fragmented paths.
arXiv:2606.04475v1 Announce Type: new
Abstract: A mobile phone vibrating on a hard surface often sounds qualitatively unlike ordinary audiovisual recordings when reproduced through laptop loudspeakers. We propose that part of this perceptual distinctiveness can be described as a nested periodicity: a first-order cepstral structure reflecting the vibration period and its multiples, and a second-order cepstral structure reflecting repeated spacing within the first-order cepstrum. Treating the perceptual effect as real and using a deliberately transparent synthetic signal chain, we model six stages: mechanical generation, surface and air propagation, microphone capture, encoding and decoding, laptop-speaker playback, and re-recording or post-processing. The synthetic analysis shows that the first-order cepstral periodicity is preserved across the chain, whereas a cleaner bimodal or quasi-bimodal second-order cepstral signature is most evident at the mechanical source and at laptop-speaker playback. The result supports, but does not prove, the hypothesis that laptop reproduction can re-emphasize a latent contact-vibration periodicity that is less cleanly expressed in intermediate recorded and encoded forms. We frame second-order cepstral bimodality as an exploratory descriptor of contact-vibration playback rather than as a completed perceptual metric. Required validation includes recordings of real devices, controlled playback transfer functions, perceptual judgments, and comparisons against ordinary speech, music, and environmental recordings.
arXiv:2606.04687v1 Announce Type: new
Abstract: Directed Acyclic Graph (DAG) based BFT protocols have demonstrated the capability to achieve significantly high throughput in practice. Recent advancements focused on minimizing the good-case latency of these protocols, approaching the theoretical lower bound. However, the high communication complexity inherent in existing DAG-based protocols limits their scalability. This primarily arises because each vertex in the DAG must include a linear number of edges (references) to vertices from previous rounds.
We present Clownfish, a partially synchronous DAG-based BFT protocol designed to address the scalability bottleneck. Clownfish achieves lower communication complexity by selectively reducing the number of edges in DAG vertices. When using a communication-optimal consistent broadcast, Clownfish attains quadratic total communication complexity per round, outperforming prior DAG-based protocols. Clownfish also reduces the additional latency in failure cases by optimizing the round advancement rule. Additionally, Clownfish supports multiple leaders per round to reduce average latency while maintaining its lower communication complexity. Our experimental evaluation demonstrates that Clownfish provides significantly better scalability than existing DAG-based protocols.
arXiv:2606.04688v1 Announce Type: new
Abstract: Autoregressive mesh generation has gained attention by tokenizing meshes into sequences and training models in a language-modeling fashion. However, existing approaches suffer from two fundamental limitations: (i) low tokenization efficiency, which yields long token sequences and prevents scaling to high-poly meshes, and (ii) absence of geometry-aware guidance, as generation is conditioned only on global shape embeddings rather than local surface cues. We introduce MeshWeaver, an autoregressive framework that treats mesh generation as a surface weaving process by directly predicting the next vertex instead of independent coordinates. At its core is a multi-level sparse-voxel encoder that injects geometric context into the generative process in three complementary ways: providing voxel features as vertex representations, guiding token prediction via cross-attention to voxel features, and serving as a structural scaffold that constrains generation around the input surface. Our hierarchical design enables coarse-to-fine vertex prediction in a single decoding step, while tightly coupling the generative model with 3D geometry. Extensive experiments demonstrate that MeshWeaver achieves a state-of-the-art compression ratio of 18%, can generate meshes with up to 16K faces, and significantly improves geometric fidelity over prior approaches.
arXiv:2606.04699v1 Announce Type: new
Abstract: Early and accurate detection of Alzheimer's disease (AD) is important for timely intervention and disease management. Generalized Eigenvalue Proximal Support Vector Machine (GEPSVM) and its Universum-based variants have shown promising results for AD classification. However, existing methods treat Universum samples as independent points and do not consider the geometric relationships among them. This paper proposes two graph-guided Universum learning models, namely UG-GEPSVM and IUG-GEPSVM, for AD versus cognitively normal (CN) classification using structural MRI data. In the proposed framework, mild cognitive impairment (MCI) subjects are used as Universum data to provide intermediate information between AD and CN classes. A graph is constructed over the Universum samples using Gaussian similarity, Minimum Spanning Tree connectivity, and multi-hop propagation. From this graph, a Laplacian matrix is derived that captures the geometric structure of the MCI samples. This Laplacian-based regularization is incorporated into the learning process in place of the conventional independent Universum penalty term. UG-GEPSVM integrates this regularization into the generalized eigenvalue formulation, while IUG-GEPSVM extends the numerically stable improved GEPSVM framework using a standard eigenvalue formulation. Experiments on ADNI MRI dataset variants using ICA- and PCA-based features at five different noise levels show that both proposed models consistently outperform existing GEPSVM and Universum-based methods. UG-GEPSVM achieves the highest average AUC of 88.07% and maintains stable performance under increasing noise levels. Statistical tests further confirm the significance of the observed improvements.
arXiv:2606.04701v1 Announce Type: new
Abstract: GUI agents today assume a static screen, where the world is frozen between two actions. However, real interfaces such as short-video applications violate this assumption, as their content keeps playing, and a competent user must decide what to watch and for how long. We formalize this task as Living-Screen-Native GUI agents and introduce LivingScreen, the first benchmark instantiating it on short-video platforms, with a faithful browser-based environment, a three-tier task suite, and metrics that jointly score accuracy and information efficiency. Evaluating extensive frontier models, we find that none reaches the human cost-accuracy performance, and that their dominant failure mode is over- and under-observation, pointing to observation control as a missing capability axis for future GUI agents. All data and code will be available at https://github.com/BITHLP/LivingScreen.
arXiv:2606.04703v1 Announce Type: new
Abstract: Experience internalization converts contextual experience from past interactions into reusable parametric capability, offering a promising path toward continual learning in large language models (LLMs). While prior work has predominantly focused on single-iteration transfer, we discover that under multi-iteration experience learning, existing methods suffer from a progressive capability collapse rather than compounding improvement. We systematically examine this failure through three vital dimensions of experience internalization: (1) Experience Granularity: We find that principle-level experience is more durable than instance-level experience, as it effectively abstracts transferable strategies away from trajectory-specific details. (2) Experience Injection Pattern: Our analysis reveals that step-wise injection significantly outperforms global injection by aligning experience with intermediate decision states, a property that is critical for long-horizon tool use. (3) Internalization Regime: We demonstrate that off-policy context-distillation on high-quality teacher trajectories provides a substantially more stable training signal than on-policy context-distillation, which is inherently limited by local corrections on student-induced flawed states. Together, these insights yield a simple yet robust recipe for stable and sustainable experience internalization, providing concrete guidance for engineering self-evolving and continually learning LLMs.
arXiv:2606.04704v1 Announce Type: new
Abstract: Rocq (Coq) are now widely used in various fields, including software verification and mathematical proofs. When proving a new theorem, users often need to search and apply proven theorems to assist the current proof process. However, the current search command is limited to the environment of imported modules and cannot search for theorems outside of this scope. Furthermore, tool developers and researchers may want to obtain detailed information about theorems, such as theorem's names, statements, and dependencies. But there are currently no user-friendly and efficient tools available for extracting comprehensive information from Rocq projects. We introduce a Rocq theorem extraction and analysis tool, TheoremExtr, which is capable of analyzing theorem composition and extracting theorems, dependencies, and definitions from both parsing phase and runtime. We extracted 71,795 theorems and their dependencies from 32 open-source projects from the Rocq community. In addition, we extracted 27,481 definitions and their types among these projects. We also developed a website that supports cross-project similarity search for theorems and definitions. The tool is available at https://github.com/Rw1nd/TheoremExtr, and the search website is available at https://lemmasearch.com.
arXiv:2606.04394v1 Announce Type: new
Abstract: Large language model chatbots are increasingly deployed in organizational settings such as healthcare, finance, and public services. Evaluating policy alignment is therefore critical to reliable chatbot deployment. By analyzing real-world user queries, we identify composed-policy violation is prevalent in various chatbots but overlooked by existing benchmarks. This paper present COPAL, an automated tool for evaluating composed-policy alignment in chatbots. COPAL efficiently generates queries that trigger composed-policy failures in chatbots via empirically derived interaction patterns and explicit handling contracts. Queries generated by COPAL expose substantial query handling failures: across 9 served models, composed-policy queries yield a 33.1% error rate on average, indicating that composed-policy alignment warrants further investigation.
arXiv:2606.04705v1 Announce Type: new
Abstract: Semantic segmentation in medical imaging is a critical yet challenging task due to data scarcity and high variability across modalities. While foundation models like the Segment Anything Model (SAM) show promise, they often struggle with medical images without specific adaptation. Moreover, point prompts, despite being the most natural form of user interaction, provide insufficient spatial context for reliable segmentation, particularly when target structures are irregular or poorly contrasted. In this paper, we propose an enhanced segmentation framework that integrates a lightweight Box Predictor module into the MedSAM architecture. The Box Predictor estimates an approximate bounding box from a single user click using localized image embedding features, providing spatial guidance that reduces the ambiguity of point prompts, while introducing only 1.6M additional parameters and negligible inference overhead. We introduce a two-stage training pipeline where the Box Predictor is trained independently before being integrated into MedSAM. To validate the generalization capability of our method, we conduct extensive evaluations on four diverse datasets (FLARE22, BRISC, BUSI, LungSegDB) spanning distinct imaging modalities, including CT, MRI, and Ultrasound. Our method improves segmentation accuracy and robustness across varied anatomical structures and imaging domains, achieving Dice scores of 0.89 (BUSI), 0.93 (FLARE22), 0.88 (BRISC), and 0.98 (LungSegDB). Code is available at https://github.com/Amirhosseinmovahedi/MedSAM-BoxPredictor
arXiv:2606.04588v1 Announce Type: new
Abstract: Multimodal large language models have made rapid progress in video understanding, yet existing benchmarks largely rely on simple prompts and provide limited evidence about whether models can satisfy explicit output constraints. We introduce VCIFBench, a benchmark for evaluating complex instruction following in video understanding. VCIFBench constructs constraint-rich instructions from both benchmark-adapted and directly video-grounded prompts, covering content, format, style, and structure requirements, and evaluates model outputs with a hybrid verification pipeline. The benchmark contains 306 satisfiable test instructions, a 540-pair DPO preference dataset, and a 30-item conflict diagnostic subset. Experiments on 10 MLLMs show that joint constraint satisfaction remains challenging. We further show that DPO training on VCIFBench data can improve instruction-following performance.
arXiv:2606.04735v1 Announce Type: new
Abstract: Temporal credit assignment is central to both biological and artificial intelligence, yet its interaction with non-linear function approximation is poorly understood. We identify a systematic failure mode in deep reinforcement learning (RL) termed Trace-Mediated Peak Bias (TMPB). At intermediate eligibility trace depths, agents irrationally prefer trajectories with high-magnitude reward ``peaks'' over alternatives with higher cumulative returns. This provides a mechanistic account of the Peak-End Rule: a human memory bias where experiences are judged by their most intense moments rather than integrated utility. We show that TMPB emerges because traces amplify distal Temporal Difference errors into ``gradient shocks'' that fixed-step-size Stochastic Gradient Descent cannot normalize, leading to global overestimation. Conversely, adaptive optimizers mitigate this pathology via second-moment normalization. Our results suggest that human-like saliency distortions may emerge naturally from the mathematical constraints of credit assignment in distributed systems, and that adaptive optimization is a theoretical necessity for rational value estimation.
arXiv:2606.04395v1 Announce Type: new
Abstract: We propose ISS-BKNO, a unified framework that integrates Koopman operator identification, Neural ordinary differential equations (ODEs), fiber bundle geometry, and input-to-state stability (ISS) certification. Unlike prior approaches that address stability, extrinsic inputs, or environmental constraints in isolation, the proposed framework simultaneously learns controlled nonlinear dynamics while guaranteeing global convergence and a computable ISS gain. The architecture introduces a three-stage lifting pipeline: a bundle-aware encoder that separates environment-specific fibers, an environment-conditioned Koopman backbone whose matrix spectrum is constrained to lie in the left half-plane, and a residual neural ODE correction whose Jacobian satisfies a quadratic sector bound. Lyapunov-based ISS regularization turns the stability requirement into a differentiable penalty that is jointly optimized with the prediction objective. Theoretical results establish fiber invariance, ISS with an explicit gain formula, and an approximation error bound that scales with the EDMD residual. Experiments on a pendulum, cart-pole, a unicycle-based navigation task, and a Franka Emika manipulator demonstrate substantially improved prediction accuracy and robustness under matched disturbances compared with existing Neural ODE and Koopman baselines.
arXiv:2606.04093v1 Announce Type: cross
Abstract: In numerical experiments simulating Turbulent Radiative Mixing Layers (TRMLs) it is observed that as the cooling time in the mixed gas, $t_{\rm cool}$, becomes very short compared to the dynamical time of the turbulence, $t_{\rm eddy}/t_{\rm cool} \gg 1$, there is a change in the scaling behavior of the total energy radiated in the TRML as a function of this ratio, also known as the Damk\"{o}hler number, ${\rm Da} \equiv t_{\rm eddy}/t_{\rm cool}$, from $\dot{E}_{\rm cool} \propto {\rm Da}^{1/2}$ to $\dot{E}_{\rm cool} \propto {\rm Da}^{1/4}$. The latter, so-called "fast-cooling," regime is of particular interest as many astrophysical mixing layers lie in this regime. We demonstrate that the origin of this change is the suppression of turbulent folding of the surface by the ram-pressure of the inflowing gas, which becomes much greater than the turbulent pressure in this regime. We present an argument that reproduces the $\dot{E}_{\rm cool} \propto {\rm Da}^{1/4}$ behavior by appealing to the suppression of the fractal structure of the interface by the ram-pressure of the inflowing gas.
arXiv:2606.05094v1 Announce Type: new
Abstract: In this paper, we present Remote Access Memory Channels (RAMC), an explicit one-sided communication library designed to leverage the capabilities of HPE Cray Slingshot network hardware. Existing one-sided communication frameworks, such as MPI RMA and OpenSHMEM, rely on monolithic shared memory models that introduce scalability and usability challenges. These frameworks often assume symmetric memory regions or require blocking collective operations for window creation, which can mismatch user communication needs and hinder performance. Implicit models, such as PGAS and UPC, aim to simplify programming by treating local and remote memory as a unified region but ultimately rely on explicit mechanisms to implement data movement. MPI's recently-introduced partitioned communication API offers a persistent point-to-point interface but sacrifices the dynamic flexibility of RDMA. RAMC is designed to address these limitations. Based on the core concept of a persistent uni-directional communication channel, RAMC leverages Slingshot's unique memory region counters to enable efficient completion notification. Experiments with a RAMC-based heat diffusion code demonstrate RAMC has no difficulty scaling to 19.6 thousand processes across 250 nodes, and microbenchmark studies across multiple libfabric versions show RAMC can outperform Cray's proprietary MPI implementation (e.g., increases in bandwidth ranging from approx. 100%-130% for 1B-4KiB messages under libfabric 1.15.2, and from approx. 30%-45% under libfabric 2.3.1) while identifying additional areas for improvement, such as small message latencies.
DPDL: Towards Differential Privacy Preservation in Decentralized Stochastic Learning on Non-IID Data
arXiv:2606.04399v1 Announce Type: new
Abstract: In the paradigm of decentralized learning, a group of agents collaborate to train a global model using distributed datasets without a central server. Although the power of collaboration has been verified by many state-of-the-art studies, it entails extensive gradient information exchanging among the agents and thus induces high risk of privacy leakage for the individual agents. Moreover, in real-world applications, the training data are usually non-identically and independently distributed across the agents, inducing more challenges to enable privacy-preserved decentralized learning. To address these issues, we propose a privacy-preserved decentralized learning algorithm with non-IID data, DPDL, which leverages the notion of Differential Privacy (DP) in cross-gradient aggregation through a similarity-based calibration technique. Specifically, in each round, each agent perturbs the cross-gradients (i.e., the derivatives of its neighbors' local model in its private local data) by Gaussian noise mechanism before sharing them with its neighbors; it then adopt cosine similarity to calibrate the received perturbed cross-gradients such that the aggregation of the calibrated cross-gradients can be utilized to effectively update local model in a momentum-like manner. Our rigorous theoretical analysis not only reveals the minimum noise level required to achieve a specific level of privacy preservation, but also illustrates that our algorithm still achieves a linear speedup in training with non-IID data. We finally conduct extensive experiments on real-world dataset to validate the effectiveness of our algorithm in defending privacy attacks and in training accurate models.
arXiv:2606.04401v1 Announce Type: new
Abstract: The capabilities of large language models (LLMs) significantly depend on training data drawn from various domains. Optimizing domain-specific mixture ratios can be modeled as a bi-level optimization problem, which we simplify into a single-level penalized form and solve with twin networks: a proxy model trained on primary data and a dynamically updated reference model trained with additional data. Our proposed method, Twin Networks for bi-level DatA mixturE optiMization (TANDEM), measures the data efficacy through the difference between the twin models and up-weights domains that benefit more from the additional data. TANDEM provides theoretical guarantees and wider applicability, compared to prior approaches. Furthermore, our bi-level perspective suggests new settings to study domain reweighting such as data-restricted scenarios and supervised fine-tuning, where optimized mixture ratios significantly improve the performance. Extensive experiments validate TANDEM's effectiveness in all scenarios.
arXiv:2606.04405v1 Announce Type: new
Abstract: Modern Transformer architectures frequently employ normalization mechanisms such as RMSNorm and Query-Key Normalization, making parts of the model approximately scale-invariant with respect to weight magnitudes. In this regime, standard Frobenius-norm weight decay acts purely along the radial direction of the weight space and cannot directly simplify the function represented by the normalized layer. We study grokking in small algorithmic tasks through this lens and propose \emph{Low-Rank Decay} (LRD), a nuclear-norm-like spectral regularizer whose subgradient -- the polar factor $UV^\top$ -- retains a tangential component even in the scale-invariant setting. This distinction has a concrete dynamical consequence: after the model memorizes the training set and task gradients vanish, L2 decay can no longer reshape the weight spectrum, whereas LRD continues to compress singular values in an $\ell_1$-like fashion. On modular arithmetic tasks, we find that LRD induces rapid effective-rank collapse in Query/Key matrices and expands the data-fraction boundary at which delayed generalization (grokking) occurs. We further provide a spectral-geometric interpretation through the ``needle-to-fan'' expansion of the nuclear-norm subdifferential near low-rank strata.
arXiv:2606.04408v1 Announce Type: new
Abstract: High-dimensional and incomplete (HDI) data are prevalent in many real-world big data scenarios. Latent factor models serve as a common representation learning approach, capable of uncovering informative latent factors from such data. Nevertheless, most existing latent factor models rely solely on gradient descent for optimization, which may lead to insufficient and biased representations, particularly when dealing with heterogeneous HDI data. Thus, this study proposes an Ensembled Latent Factor Model via Differential Evolution and Gradient Descent Optimization (ELFM-DEGDO) with two-fold designed: 1) two diverse latent factor models are independently modeled via differential evolution and gradient descent optimization, respectively, and 2) the two diverse latent factor models are combined via a customized self-adaptive weighting mechanism to effectively fuse their strengths. By leveraging the complementary advantages of both optimization paradigms, ELFM-DEGDO is able to produce more comprehensive and less biased representations for HDI data. Three HDI datasets are tested to show that ELFM-DEGDO consistently performs better than related several latent factor models.
arXiv:2606.04425v1 Announce Type: new
Abstract: Modern agentic systems transform LLMs from session-bounded assistants into stateful systems that persist and evolve shared world state across sessions through memories, filesystems, tools, and other long-lived contextual artifacts. This shift fundamentally expands the attack surface of prompt injection. However, prior works on prompt injection have largely focused on model-level threats within a single session, overlooking how cross-session persistent system state fundamentally changes the system-level risk of agentic systems. Inspired by stored cross-site scripting in web systems, we introduce cross-session stored prompt injection, where a successful injection can persist within agentic system state and silently influence future executions long after the original attacker interaction has ended. To systematically study this threat, we formalize stored prompt injection and develop a taxonomy of how adversarial content persists and affects agentic systems across sessions. We further develop a benchmark and sandbox toolkit to evaluate the risks of stored prompt injection, enabling quantitative analysis of attack success across different models, attack goals, and persistence channels. Our findings highlight that persistence transforms prompt injection from an ephemeral model-level threat into a long-lived system-level vulnerability embedded within agent execution state. We hope this work draws broader attention to this emerging threat and motivates the community to systematically study and mitigate system risks arising from persistence in agentic systems.