Artificial Intelligence
VIBE-Bench: Evaluating Personalized Large Language Models When Profiles Don't Mean Preferences
Yiwen Jiang, Yang Deng, Stephanie Fong et al. (+9 more)
Personalized Large
Language Models (PLLMs) aim to tailor responses to individual users, where a central challenge is preference
reasoning: inferring query-relevant preferences from user-related history. Existing
benchmarks, however, largely assume that such preference can be retrieved from semantically related history. We study an underexplored but practically important regime, profile-preference conceptual misalignment (PRCM), where observable profile cues and query-specific preferences lie in different concept spaces, making semantic
retrieval inconsistent for personalization. We introduce VIBE-Bench, a
benchmark with two psychology-grounded tasks, 3,504 personas and 12,239 dialogues, including a manually verified gold test set, and requires cross-concept preference
reasoning beyond surface semantic overlap. Experiments with several personalization methods show that current PLLMs largely rely on shallow semantic correlations and fail to acquire robust cross-concept mappings. These findings establish PRCM as a distinct failure regime in PLLMs and position VIBE-Bench as a focused testbed for advancing preference
reasoning beyond semantic matching.
Read the full abstract ▾
All authors Yiwen Jiang, Yang Deng, Stephanie Fong, Zimu Wang, Yaling Shen, Wei Feng, Hongxi Yang, Xiangyu Zhao, Zhongxing Xu, Deval Mehta, Xuelian Cheng, Zongyuan Ge
Robotics
Zhixin Wang, Chengzheyi Yao, Leyuan Liu et al. (+2 more)
Vision-and-
Language Navigation (VLN) requires an
agent to navigate through unseen 3D environments according to natural-
language instructions. Explicit
reasoning can improve instruction understanding and semantic grounding, but autoregressive generation at every step accumulates large decision-stage latency over multi-step
navigation. We propose VerNav, a verifier-first framework for low-latency LLM-based VLN. The verifier reduces decision-stage latency by replacing per-step autoregressive generation with batched action verification, while an
entropy-based adaptive generator is invoked only for uncertain decisions to produce compact state evidence. To further improve
navigation performance with the verifier, we introduce a two-stage
alignment scheme: (i) VPO improves local action-preference
alignment in static verifier training, and (ii) step-level reinforcement fine-tuning provides dense progress rewards over multi-step
navigation rollouts during dynamic task execution. Experiments on the Room-to-Room (R2R)
benchmark show that the verifier-only decision path of VerNav achieves competitive
navigation performance among representative LLM-based VLN
agents while reducing average decision-stage LLM latency per step by more than $10\times$ compared with autoregressive methods.
Read the full abstract ▾
All authors Zhixin Wang, Chengzheyi Yao, Leyuan Liu, Xiaosong Zhang, Yongzhao Zhang
Graphics
HyperSketch: Controllable Video Sketching in a Style Hyperspace
Xinding Zhu, Xinye Yang, Yingping Yang et al. (+3 more)
Vector sketch animation offers tremendous advantages for multimedia and creative design through concise line expressions and flexible editing. Learning-based generation methods of sketch animation have made significant progress in the last decade, but still suffer from limited style diversity and controllability. This paper presents a controllable
video sketching method that automatically converts
videos into multi-style vector sketch animations. A continuous style hyperspace is constructed by multi-dimensional sketch styles (fidelity, simplicity, text guidance strength) and the timeline. With this hyperspace, stroke control points are parameterized as 4-variable Bernstein
polynomials, ensuring smooth and differentiable style transitions. A multi-task, multi-stage optimization framework is designed to learn stroke hyperparameters accurately and efficiently. We further developed a web-based interactive interface that allows real-time style manipulation via editable curves. Experiments show the style controllability, high-quality, and user-friendliness of our method, which outperforms SOTA methods.
Read the full abstract ▾
All authors Xinding Zhu, Xinye Yang, Yingping Yang, Mengjian Li, Fei Gao, Jiazhou Chen
Artificial Intelligence
RPCBench: A Benchmark for Proactive Premise Critique in LLM-based Recommendation
Zhongru Chen, Yuan Wu and Yi Chang
Large
language models are increasingly used as interactive recommender assistants. Their evaluation should therefore go beyond plausible item recommendation and test whether they can recognize flawed recommendation requests. Existing recommender
benchmarks mainly assess ranking, generation, or preference satisfaction, while existing error-detection
benchmarks are usually not grounded in recommendation-specific user and candidate evidence. To address this gap, we introduce RPCBench, a
benchmark for evaluating Recommender-Premise Critique: the ability to detect, diagnose, and properly handle faulty premises in natural-
language recommendation requests. RPCBench contains evidence-grounded test instances from five recommendation domains and covers ten types of premise failures. Each instance provides a visible recommendation context and a corrupted user query. We further design a fine-grained evaluation framework that measures proactive detection, error localization, post-detection handling strategy, and evidence faithfulness. Through a systematic evaluation of 11 LLMs, we find that proactive detection is the main bottleneck in Recommender-Premise Critique, and models perform worst on underspecified-premise errors. We also observe that target-critical information density matters more than redundant evidence, and that longer
reasoning does not monotonically improve critique quality: performance peaks at intermediate
reasoning length, while overly long
reasoning is accompanied by an overthinking penalty. The code is available at https://github.com/ZhongruChen/RPCBench.
Read the full abstract ▾
Computational Geometry · Cross-listed from math.DG
The Discrete Harmonic Center of a Quadrilateral
Marc Alexa
Triangulate a simple quadrilateral by connecting all vertices to an additional point. If the vertices carry values, the piecewise linear function can be assigned a Dirichlet energy. We show that the minimal Dirichlet energy as a function of the location of the inserted point is convex, and the location of the minimum is independent of the values at the corners - a quadrilateral has a discrete harmonic center, characterized by an
equilibrium of currents across the inserted edges. It turns out that the fixed points of the Möbius involution swapping opposite corners of the quadrilateral are critical points of this energy, so the discrete harmonic center is Möbius-covariant. For tangential and cyclic quadrilaterals the center admits simple closed forms related to the circle centers. The center and its data-independence generalize to polytopes with d + 2 vertices in dimension d, but the conformal characterizations are special to four points in the plane.
Read the full abstract ▾
Information Retrieval
SwapRec: Warming Up Cold Items Through Training-Time Swaps
Marta Moscati, Jan Malte Lichtenberg, Davide Abbattista et al. (+3 more)
Interactions with cold items negatively impact real-time personalization of ID-based recommender systems. This is because the use of such interactions degrades user preference estimates, whereas excluding cold items from the user profile prevents real-time recommendation updates. In industrial scenarios, one heuristic often applied to address this shortcoming at inference time is to replace, i.e., "swap", cold-start items by their most similar "warm" neighbor, where similarity is inferred from the items' side information. In this paper, we demonstrate that sequential models, most often used for real-time personalization, are not robust to such swaps, and propose SwapRec, an approach to address this issue. SwapRec relies on using the same swap heuristics already at training time. We apply SwapRec to state-of-the-art models for sequential recommendation and analyze its impact by means of quantitative experiments in three recommendation domains (online shopping, movie, music). The experimental results show that, irrespective of the underlying sequential architecture, our easy-to-implement SwapRec approach allows for substantially more accurate recommendations when in presence of interactions with cold items, simultaneously leading to a larger percentage of cold items in the recommendation lists.
Read the full abstract ▾
All authors Marta Moscati, Jan Malte Lichtenberg, Davide Abbattista, Antonio De Candia, Laura Boggia, Matteo Ruffini
Cryptography and Security · Cross-listed from q-fin.RM
Pricing the DeFi Tail: Do Protocols or Depositors Price Operational Risk?
Nils Bundi
Similar to banks, DeFi protocols expose depositors to operational risk (USD 9.45 billion across 1,075 events since 2020). Unlike banks, they are not required to hold capital against it. A protocol may maintain a buffer voluntarily. Absent one, the risk falls on the depositor, who should then demand a risk premium in the supply yield. I quantify the underlying tail on one
benchmark, a per-sector Basel loss-distribution approach fitted to a new operational risk event
dataset, and test both margins against it. Tails in the four core sectors are no heavier than the Moscadelli banking band $[0.85, 1.39]$. Bridge, Derivatives, and the residual Other sector exhibit cyber-loss-level tails ($\hat\xi \approx 1.6$), with point estimates past the infinite-mean boundary. The Lending tail implies a $\mathrm{VaR}_{99.9}$ capital buffer of 18% of TVL and of the ten largest Lending venues, the four holding a buffer cover on average 5% of it. Under
market discipline, depositors should demand a higher yield in compensation where a venue does not maintain a buffer. I find that venues without a buffer pay a higher premium than those with (a 125-bps gap in medians): evidence the
market discriminates in the right direction. However, the premium falls far short of an adequately priced tail. This unpriced tail falls disproportionately on the retail depositor, who sees only the posted rate but lacks the information and skills to price it. Because these products are not bank-regulated, I recommend disclosure over capital mandates: protocols, and any service providers that front access to it, should publish standardized losses, existing capital buffers and tail coverage.
Read the full abstract ▾
Computer Vision
A multicenter benchmark and clinically structured metric for coronary CTA report generation
Zhiyu Ye, Yue Sun, Limiao Zou et al. (+7 more)
Reliable evaluation of automated coronary computed
tomography angiography (CCTA) report generation requires standardized multicentre
benchmarks and clinically structured metrics. We established a four-centre
benchmark comprising 3,021 CCTA series from 818 patient-report pairs to evaluate seven open-source three-dimensional
vision-
language models. We developed CSM$_{\text{CCTA}}$, a clinically structured metric for CCTA report evaluation, with patient-, vessel-, and segment-level variables defined according to clinical guidelines. Report pairs are compared at the finest shared anatomical level, and the contributions of different clinical components are weighted based on expert assessments. We estimated these weights using 70 expert-scored cases and evaluated clinical
alignment in a non-overlapping set of 30 cases. CSM$_{\text{CCTA}}$ showed a strong correlation with radiologist scores (Pearson's $r=0.97$, $p<0.001$), exceeding the next-best metric, FORTE ($r=0.70$), by 0.27, and agreed with expert preferences in 115 of 160 pairwise comparisons (71.9%). Under controlled perturbations, CSM$_{\text{CCTA}}$ remained stable to clinically equivalent wording and decreased monotonically with progressive information omission. In the multicenter
benchmark, the CCTA-trained C2RG model achieved the highest CSM$_{\text{CCTA}}$ scores across all four hospitals, although its performance remained far from optimal. In contrast, CCTA-irrelevant reports accounted for up to 98.7% of the outputs from generalist models. Together, the
benchmark provides a standardized setting for model comparison, while CSM$_{\text{CCTA}}$ enables clinically structured evaluation of finding agreement and anatomical specificity. These results support a more clinically aligned and anatomically resolved approach to evaluating CCTA report generation. Code is available at https://openi.pcl.ac.cn/OpenMedIA/CSM_CCTA.
Read the full abstract ▾
All authors Zhiyu Ye, Yue Sun, Limiao Zou, Cheng Xu, Keting Xu, Tong Hu, Yue Yu, Hairong Zheng, Yining Wang, Tong Zhang
Robotics
Knowing When to Stop: Adaptive Action Chunking via Internal Cross-Attention Dynamics in VLAs
Runze Xu, Xiaolong Shan, Shuang Dai et al. (+2 more)
Action chunking is a standard execution strategy in modern
Vision-
Language-Action (VLA) frameworks, but fixed execution horizons impose a trade-off between efficiency and accuracy. Short chunks require frequent inference and may cause oscillatory behavior, whereas long chunks can become misaligned with newly observed states. We address this limitation with an adaptive action chunking approach based on internal cross-
attention dynamics in the action expert. We observe that, as the prediction horizon extends, action-to-observation cross-
attention becomes increasingly dispersed and its
entropy rises toward a plateau. This pattern is associated with higher action prediction error and provides an online signal that the current observation offers limited grounding for further open-loop execution. Based on this observation, we introduce a training-free truncation mechanism that detects sustained high-
entropy plateaus and dynamically selects the execution horizon during inference. The method uses
attention weights already computed by the policy and introduces negligible additional overhead. Evaluations on $\pi_{0.5}$ and X-VLA across RoboTwin 2.0, LIBERO, and three real-world manipulation tasks show improved average task success over fixed-horizon and adaptive chunking baselines, while preserving efficient closed-loop control. These results show that cross-
attention dynamics can provide a practical internal signal for adaptive action execution in VLAs.
Read the full abstract ▾
All authors Runze Xu, Xiaolong Shan, Shuang Dai, Yu Wang, Jincheng Yu
Robotics
Peg-in-Bench: A Modular Benchmark for High-Precision Robotic Insertion
Yosel Delgado, José G. Buenaventura-Carreón, Floris Erich et al. (+4 more)
High-precision insertion remains a fundamental challenge in robotic manipulation due to the strict
alignment requirements and contact-rich interactions involved. Although peg-in-hole tasks are widely used for evaluation, existing bench- marks often rely on fixed task configurations, limiting their ability to assess robustness and generalization across different insertion scenarios. This paper introduces a reconfigurable peg-in-hole
benchmark designed to evaluate task generalization in high-precision insertion. The
benchmark consists of a set of fully 3D-printable modular components, including multiple peg geometries, tolerance levels, and configurable base structures that can be combined to generate a large variety of insertion and assembly tasks. By varying object layouts, orientations, and task structures while maintaining controlled physical conditions, the
benchmark enables systematic evaluation of adaptation to unseen scenarios. To support reproducibility, we additionally provide a scenario generation tool capable of producing standardized task configurations and machine-readable task descriptions. The scenario generation tool and the STL files of the
benchmark pieces are available through the project repository: https://github.com/aistairc/peg-in-bench.
Read the full abstract ▾
All authors Yosel Delgado, José G. Buenaventura-Carreón, Floris Erich, Roman Mykhailyshyn, Tomohiro Motoda, Koshi Makihara, Yukiyasu Domae
Machine Learning
When Metropolis and Hastings Meet Bradley and Terry: Exact MCMC From Preference Voting
Ariel Smogorghevski, Nir Rosenfeld and Yaniv Romano
Sampling from distributions conditioned on desired semantic properties is an emerging challenge in modern generative modeling. Metropolis-Hastings (MH) provides a principled route to conditional sampling, but requires access to exact pointwise target-density evaluations, which are not available in generative settings. Meanwhile, pairwise comparisons by humans or model "judge" are highly accessible and have proved valuable across diverse applications. We introduce Pref-MH, a general exact MH sampler for judge-induced conditional distributions using only
stochastic binary pairwise comparisons. Our key observation is that the MH unnormalized density ratio matches the preference odds of the Bradley-Terry (BT) choice model. The central challenge is that while MH requires precise ratio computation, BT judges provide only sampled binary feedback. To this end, we develop a valid accept/reject rule whose resulting Markov chain provably converges to the target distribution. We further show that, for a fixed proposal kernel and budget, Pref-MH is optimal in the Peskun-Tierney sense among this class of exact reversible acceptance rules. Experiments on text generation and molecular design with LLM judges, as well as image generation with VLM judges, demonstrate that Pref-MH provides a practical and flexible approach to conditional sampling when comparative feedback is relatively easy to obtain.
Read the full abstract ▾
Artificial Intelligence
In-Context Neurofeedback: Can LLMs Control Their Internal Representations through Privileged Access?
Koshiro Aoki, Ryota Takatsuki, Gouki Minegishi et al. (+2 more)
Whether large
language models (LLMs) can control their own internal representations matters for both machine metacognition and AI safety. A recent study applied neurofeedback to LLMs and claimed that they can control their internal representations. However, the reported control may rely on superficial mechanisms rather than genuine internal access because the control targets in that study are not privileged, meaning that a third party can infer them from the
prompt. We redesign the neurofeedback paradigm for LLMs so that the control target satisfies the privileged access requirement, which is closer to neurofeedback experiments in human cognitive neuroscience. Under this stricter setting, the models do not demonstrate reliable control over privileged internal representations, suggesting that previously reported control cannot exclude the possibility that it relies on superficial mechanisms. Our results indicate that rigorous assessments of metacognition in LLMs require evaluation methods that demand privileged access.
Read the full abstract ▾
All authors Koshiro Aoki, Ryota Takatsuki, Gouki Minegishi, Yusuke Haruki, Daisuke Kawahara
Systems and Control
Lightweight CFR-Based Modulation Adaptation in a Real-Time MIMO-OFDM SDR Testbed
Luca Borst, Maryam Ansarifard, Ankith Vinayachandran et al. (+2 more)
Conventional link adaptation typically relies on scalar link-quality indicators such as signal-to-noise ratio (SNR), while richer channel state information (CSI) can improve adaptation at the cost of higher processing complexity. This paper investigates a compact alternative for modulation selection in a real-time multiple-input multiple-output orthogonal frequency-division multiplexing (
MIMO-OFDM) system using channel frequency response (CFR) magnitude descriptors. A
dataset of 87,817 over-the-air (OTA) samples is collected using a USRP-based testbed, with CFR measurements extracted at the base station (BS) from received uplink pilots. Decision tree (DT), random forest (RF), and k-nearest neighbours (KNN) classifiers are evaluated using BS-side SNR, CFR features, and their combination. SNR-only classifiers achieve 35%-42% test accuracy, whereas CFR-only features achieve 73.6%, 81.4%, and 80.0% for DT, RF, and KNN, respectively. CFR-based performance is maintained near the 10% BLER reliability thresholds, with RF reaching 82.8%. A depth-7 DT with 123 leaves is further integrated into the LabVIEW C Node for real-time inference. The results show that compact BS-side CFR descriptors provide more discriminative information than the available scalar BS-side SNR while remaining suitable for lightweight SDR implementation.
Read the full abstract ▾
All authors Luca Borst, Maryam Ansarifard, Ankith Vinayachandran, Kishor C. Joshi, George Exarchakos
Computer Vision
HELIOS: From midnight to noon, continuous outdoor urban scene relighting
Hala Djeghim, Nathan Piasco, Luis Roldão et al. (+4 more)
Modifying the illumination of driving images is a fundamental challenge, as most
datasets are captured at specific times of day. Existing methods rely on synthetic data or paired multi-illumination supervision, which limits their generalization to the diverse and challenging conditions of real-world scenarios. To address this, we propose HELIOS, a novel image relighting approach that relies on unlabeled real-world
datasets without requiring any paired images for training. Our approach integrates albedo-based conditioning into a cycle-consistent
diffusion pipeline to prevent identity collapse and ensure accurate domain translation. To handle low-visibility nighttime conditions, we introduce a robust albedo
distillation strategy that transfers structural stability from the daytime domain. Additionally, we replace traditional text
prompts with a fine-grained control mechanism based on GPS-derived
solar angles, enabling smooth and continuous lighting manipulation across the day-night cycle. Through extensive evaluation and a user study, we demonstrate that HELIOS produces structurally consistent and realistic results in both night-to-day and day-to-night tasks, outperforming state-of-the-art methods.
Read the full abstract ▾
All authors Hala Djeghim, Nathan Piasco, Luis Roldão, Moussab Bennehar, Dzmitry Tsishkou, Céline Loscos, Désiré Sidibé
Machine Learning · Cross-listed from physics.optics
Direct Optimization of a 3D Finite-Source Reflector via Neural-Network Parameterization
Roel Hacking, Lisa Kusch, Martijn Anthonissen et al. (+1 more)
We present a direct optimization method for three-dimensional freeform reflectors that transform the light of a finite-étendue source into a prescribed far-field angular intensity distribution. The reflector profile is represented by a small neural network (a multilayer perceptron), which is trained end-to-end through a differentiable ray-tracing objective. We furthermore parameterize the emission directions in gnomonic coordinates, and show how we use this to ensure that every emitted ray intersects the reflector. At each iteration, the network is converted to a bicubic spline representation for ray-tracing efficiency, and intersections with this smooth surface are solved by a damped Newton solve, with gradients computed via the implicit function
theorem. The traced output distribution is compared with the desired target on a 'soft' histogram, under an $H^{-1}$-type spectral weighting that emphasizes long-range transport of flux to improve convergence. Optimization is performed using a BFGS method with self-scaled Broyden updates and a plateau-perturbation rule to prevent stalling. The method converges reliably within seconds on a single GPU for all examples tested.
Read the full abstract ▾
All authors Roel Hacking, Lisa Kusch, Martijn Anthonissen, Wilbert IJzerman
Computer Vision
Vision-Language-Guided Pseudo-Labels for Unsupervised Domain Adaptation in Semantic Segmentation for Waste Sorting
Udo Schlegel, Shubhangi, Gabriel Dax et al. (+3 more)
Obtaining labeled data for semantic
segmentation in applied settings (e.g., autonomous driving, industrial waste sorting) is expensive and often infeasible at scale. We present a cross-modal pseudo-labeling pipeline that enables unsupervised domain adaptation without any target-domain annotations. The pipeline is built on two core foundation models: SAM generates class-agnostic region proposals, and EVA-CLIP assigns semantic labels based on region-text similarity, with confidence filtering ensuring that only reliable pseudo-labels are used for self-training a
segmentation model. As an optional extension, BLIP provides
language-grounded verification for ambiguous regions, thereby improving pseudo-label quality without altering the overall pipeline. Evaluated on two domain shifts, synthetic-to-real autonomous driving and, with a primary focus, lab-to-factory industrial waste sorting, the pipeline consistently improves over source-only baselines. Our results demonstrate that pseudo-label quality, not quantity, is a decisive factor in self-training under domain shift, and that cross-modal
language grounding offers a practical path to reliable automatic annotation in deployment-critical applications.
Read the full abstract ▾
All authors Udo Schlegel, Shubhangi, Gabriel Dax, Sai Rahul Kaminwar, Florian Karl, Thomas Seidl
Machine Learning
Poisson-Gamma Dynamical Systems with Time-varying Transition Dynamics
Jiahao Wang, Yijun Wang, Nan Fang et al. (+1 more)
Bayesian methodologies for handling count-valued time series have gained prominence due to their ability to infer interpretable latent structures and to estimate uncertainties. Among these
Bayesian models, Poisson-Gamma Dynamical Systems (PGDSs) are proven to be effective in capturing the evolving dynamics underlying observed count sequences. However, the state-of-the-art PGDS still falls short in capturing the transition dynamics that are commonly observed in real-world count time series. To mitigate this limitation, a PGDS with time-varying transition kernel (TV-PGDS), is proposed to allow the underlying transition
matrices to evolve over time. Three specifically-designed Dirichlet Markov chains (Dir-Dir, Dir-Gam-Dir, PR-Gam-Dir) are constructed to accommodate heterogeneous structural mutations within these dependencies. Leveraging Dirichlet-Multinomial-Beta data augmentation techniques, a fully-conjugate and efficient Gibbs sampler is developed to perform posterior simulation. Experiments show that, in comparison with related models, the proposed PGDS achieves improved predictive performance due to its capacity to learn time-varying dependency structure captured by the time-evolving transition
matrices.
Read the full abstract ▾
All authors Jiahao Wang, Yijun Wang, Nan Fang, Sikun Yang
Numerical Analysis
Asymptotic behavior of Eckhoff's method for convergence acceleration of Dirac eigenfunction expansions
R.H. Barkhudaryan, G.G. Gevorkyan and L.D. Poghosyan
The current paper considers the problem of recovering a vector-function on $[-1,1]$ from a limited number of coefficients of its expansion into a series of eigenfunctions of a one-dimensional Dirac system. The Krylov–Lanczos–Eckhoff–Gottlieb acceleration method is examined in the situation when the boundary values it requires have to be computed from the generalized
Fourier coefficients themselves. This leads to a $2q\times 2q$ linear system whose
matrix is a block Vandermonde
matrix; its determinant and inverse are computed explicitly, and the asymptotic $L_2$-error constant of the method is found, paralleling the classical trigonometric case.
Read the full abstract ▾
Data Structures and Algorithms
Random-Priority Frontier Routing: Tight $\Theta(n^c)$ Bounds Against $c$-Node Cartels
Krišjānis Petručena
We study path diversification in trusted-node networks, where sensitive material is relayed through intermediate nodes, some of which may be compromised. Our randomized routing rule assigns each vertex an independent random priority and repeatedly expands the highest-priority vertex on the global frontier of the explored region. Let $G$ have $n$ vertices, let $s,t$ be honest endpoints, and let $C$ be a set of $c$ compromised intermediate vertices, called a cartel, whose deletion leaves $s$ and $t$ connected. For every fixed $c$ and every fixed target probability $q\in(0,1)$, we prove that $\Theta(n^c)$ independent executions are sufficient in the worst case for some route to avoid $C$ with probability at least $q$.
Read the full abstract ▾
Artificial Intelligence
CARE: Contrastive Anchor-based Rubric Evolution for Large Language Model Post-Training
Siyuan Li, Xinxin Song, Chen Ruinian et al. (+5 more)
Rubric-based reinforcement learning decomposes open-ended instructions into
prompt-specific, flexible rubrics, making it better suited than reinforcement learning with verifiable rewards for post-training LLMs on open-ended tasks. However, static rubrics are inevitably hacked as the policy evolves, and existing dynamic approaches introduce new problems: undirected rubric extraction, unreliable hack detection, and unbounded rubric proliferation. We propose $\textbf{CARE}$ ($\textbf{C}$ontrastive $\textbf{A}$nchor-based $\textbf{R}$ubric $\textbf{E}$volution), which grounds every rubric
evolution step in a high-quality anchor response generated by a frontier model conditioned on the
prompt and its rubrics. At each training step, CARE contrasts the highest-scoring rollout against the anchor, enabling two complementary mechanisms: an Adaptive branch that reactively repairs reward misspecification; and a Chase branch that proactively converts frontier-level quality gaps into sharper rubrics. Together, the two branches $\textbf{maintain discriminative accuracy in the high-reward region}$—the precise region where reward over-optimization mostly originates. Experiments on WildChecklist-9K with Qwen2.5-7B-Base and Qwen2.5-7B-Instruct show that CARE achieves state-of-the-art performance on Arena-Hard-2.0, InfoBench, and FollowBench, and is the $\textbf{only}$ method whose win rate against GPT-4.1 anchor responses shows sustained improvement throughout 300 training steps; additional results on Llama-3.1-8B-Instruct and Qwen3-8B further indicate that CARE generalizes across model families.
Read the full abstract ▾
All authors Siyuan Li, Xinxin Song, Chen Ruinian, Jingjing Fan, Tingxiong Xiao, Yangen Hu, Ke Zeng, Jinli Suo
Artificial Intelligence
CacheBridge: Efficient Cross-Model KV Transfer for Multi-Model LLM Serving
Xingyu Qu, Siyuan Lu, Zhiyu Chen et al. (+2 more)
Multi-model LLM serving uses routers, cascades and
agent workflows to move requests across models with different costs and capabilities. Existing serving systems reuse KV caches only within one model, so every switch forces the target to re-prefill history the source already processed, adding time to first
token (TTFT) that grows with history length and target size. Cross-model KV transfer can replace this prefill, yet prior methods are slower than re-prefill end to end. We present CacheBridge, the first serving system to accelerate heterogeneous multi-model workloads through cross-model KV transfer. It co-designs a compact,
attention-calibrated KV mapper with systematic optimization for online multi-model serving, which makes model switch efficient on latency and
memory cost. In Qwen3 14B$\rightarrow$32B
agent serving on $\tau$-bench, it reduces mean TTFT by $30$ to $98\%$ versus re-prefill at $1$ to $24$ QPS with comparable task success. Across five model pairs from three families, it uses $8$ to $12\times$ less mapper
memory than Closed-Form Mapper and retains on average $97.94\%$ and $102.94\%$ of re-prefill task quality on chat/
reasoning and RAG workloads, respectively. The promising results have opened new opportunities for efficient multi-model LLM serving through cross-model KV reuse.
Read the full abstract ▾
All authors Xingyu Qu, Siyuan Lu, Zhiyu Chen, Sheng Wang, Tao Lin
Information Theory
Algorithm-Hardware Co-Design of a Lightweight PCG Equalizer with a Fixed Step Size for Massive MIMO
Junshuo Wang, Shuting Liu, Shihan Wang et al. (+2 more)
Coarse quantization in massive multiple-input multiple-output (
MIMO) systems reduces power but causes clipping distortions. The
Bayesian Expectation-Maximization (BEM)
algorithm can recover clipped signals, but its
matrix inversion and dynamic step-size evaluation are hardware bottlenecks. We propose a hardware-friendly one-step correction that uses the initial Jacobi-preconditioned Conjugate Gradient (PCG) direction with a fixed relaxation parameter. The resulting symbol-level update has an ultra-lightweight $\mathcal{O}(U)$ feed-forward datapath and approaches high-resolution reference detectors in the evaluated massive-
MIMO setting. Our finite-dimensional analysis establishes the exact one-step descent law, proves that Jacobi normalization cancels the raw multiplicative near-far scaling while confining the loaded-system dependence to bounded attenuation factors, and gives verifiable sufficient conditions for fixed-step descent in terms of normalized channel coherence. System-level results indicate projected power savings for energy-efficient massive
MIMO uplinks.
Read the full abstract ▾
All authors Junshuo Wang, Shuting Liu, Shihan Wang, Yan Tian, Jienan Chen
Cryptography and Security
Using LLMs to Elicit Security Requirements for Service-Oriented Cyber Ranges
Michail Takaronis, Athanasia Kollarou, Georgios Kavallieratos et al. (+2 more)
Cyber ranges are complex environments comprising many interacting components and stakeholders with different security concerns. The Service-Oriented Cyber Range (SOR) is no exception, particularly when it comes to training scenarios targeting critical infrastructure. Security concerns are translated into security requirements, the elicitation of which is usually difficult and time-consuming. This work examines how large
language models can assist in eliciting security requirements for a service-oriented range and help produce a useful baseline for designers and developers. The approach follows a SEBoK-guided process in which security mission objectives and stakeholder needs were first identified and then provided as a
prompt context along with architectural guidelines to five LLMs: GPT-5.2, Gemini 3.1 Pro, Grok 4.1, Sonar, and Kimi K2.5. The models generated 84 security requirements in total, which were consolidated into a comprehensive set of 27 requirements and then mapped to the architectural layers of the service-oriented range. The final set was evaluated by five cybersecurity experts against the criteria of necessity, clarity, completeness, feasibility, and testability, with an additional rejection option. The results showed a high acceptance rate, specifically for necessity with 98.5%, clarity with 87.4%, completeness with 85.2%, feasibility with 78.5%, and rejection with 0.7%. Testability was lower at 44.4%, indicating a slight lack of information on how these requirements could be tested. These findings show that LLMs can support early stages of the elicitation of security requirements, although human review is still needed, especially to improve or adjust certain aspects of the requirements.
Read the full abstract ▾
All authors Michail Takaronis, Athanasia Kollarou, Georgios Kavallieratos, Vasileios Gkioulos, Sokratis Katsikas
Artificial Intelligence
Denoising Diffusion Generative Models Secretly Calculate Attentions
Farzan Haddadi, Leila Monfared, Ebrahim Rezaii et al. (+3 more)
Denoising
diffusion models are the dominant architecture for image generation, whereas most natural
language generation and modeling are primarily handled by well-known
transformer architectures employing
attention mechanism. Here, we show that
diffusion models also inherently use an
attention mechanism very similar to that of
transformers. Therefore,
attention emerges as a universal machine learning principle, based on a general training objective. We also show similarities in basic functional principle of auto-encoders and
attention-based models. These equivalences allows us to interchange these designs based on practical requirements. As an example, we can reformulate the
diffusion framework to reduce the lengthy training process and computation-intensive image generation. Using this approach, a simplified
algorithm is proposed for image generation which is based on
attention mechanism. Results show that the
attention-based implementation achieves comparable performance with significantly less effort and computational resources.
Read the full abstract ▾
All authors Farzan Haddadi, Leila Monfared, Ebrahim Rezaii, Mohammadreza Malek-Mohammadi, Pejman Zakalvand, Narges Mokhtari
Machine Learning
iPINN for Broadband CARS Phase Retrieval: A Framework for Function Approximation and Inverse Modeling Problems in Nonlinear Spectroscopy
Ravi Teja Vulchi, Carl Messerschmidt, Mohammadsadegh Vafaeinezhad et al. (+4 more)
Phase
retrieval in broadband coherent anti-Stokes Raman spectroscopy (BCARS) is an ill-posed inverse problem. The Raman-like signal is encoded in the imaginary part of the resonant susceptibility, which mixes coherently with a non-resonant background (NRB) that varies across acquisitions. We introduce an inverse physics-informed neural network (iPINN) that predicts Lorentzian peak parameters from raw BCARS spectra and reconstructs the resonant susceptibility through a differentiable analytical forward model. A
transformer encoder assigns spectral features to 24 learnable peak slots, and a multi-view consistency loss enforces invariance across NRB pattern, NRB strength, and noise. Unlike direct spectral regression approaches, the method retains accuracy under varying acquisition conditions. On a public
benchmark, iPINN achieves the lowest error among the tested baselines (MAE 0.016 vs. next-best 0.046). On 28 zero-shot test spectra acquired across seven solvents and four focal positions, accuracy is depth-invariant in five of seven solvents. These results show that inverse parametric prediction with a differentiable physical decoder supports robust phase
retrieval across measurement conditions.
Read the full abstract ▾
All authors Ravi Teja Vulchi, Carl Messerschmidt, Mohammadsadegh Vafaeinezhad, Rajendhar Junjuri, Tobias Meyer-Zedler, Juergen Popp, Thomas Bocklitz
Artificial Intelligence
Towards reliable multimodal disaster severity assessment through preference optimization and explainable vision-language reasoning
Yuanjun Zhang, Fuzel Ahamed Shaik, Suvojit Acharjee et al. (+2 more)
Reliable disaster damage assessment requires models that provide both accurate predictions and transparent explanations. However, existing
multimodal approaches are limited by scarce annotated data and insufficient evaluation of
reasoning quality. This study proposes a two-stage training framework that integrates Supervised Fine-Tuning (SFT) and Direct Preference Optimization (DPO) within a unified data construction pipeline. From a single Human-in-the-Loop (HITL) annotation workflow, two complementary
datasets are derived, namely ReasoningSet, which contains validated rationales for SFT, and PreferenceSet, which comprises paired rationales for DPO-based
alignment. The framework evaluates both classification performance and explanation quality using automatic metrics, model-based scoring, and human ranking. Experimental results show that SFT improves accuracy from 73.64% to 78.29% and increases Macro-F1 by 29% compared to the baseline, while explanation quality improves by approximately 25%. Subsequent DPO
alignment further enhances interpretability on the PreferenceSet. Cross-model validation on InternVL-3-8B and LLaVA-1.5-7B demonstrates the robustness and generalizability of the approach. The proposed framework improves detection of underrepresented mild damage cases, reduces high-risk misclassifications, and strengthens
alignment between model
reasoning and human judgment. Overall, it provides a reproducible pathway to develop reliable
multimodal systems that deliver auditable, actionable disaster insights for emergency management.
Read the full abstract ▾
All authors Yuanjun Zhang, Fuzel Ahamed Shaik, Suvojit Acharjee, Fahad Khalid, Mourad Oussalah
Numerical Analysis · Cross-listed from math.OC
DOFFO_TR: a Decentralized Objective Function-Free Optimization method with Trust-Region
Stefania Bellavia, Greta Malaspina and Benedetta Morini
In this paper, we propose a novel objective function-free trust-region method designed to solve optimization problems over decentralized networks. Unlike traditional approaches that often rely on stepsize tuning, our framework employs a function-free trust-region procedure that enables adaptive selection of the step length. Our approach accommodates first- and second-order models and eliminates the need to share local function values and gradients among
agents, thereby enhancing
privacy and computational efficiency. On the theoretical side, we establish provable iteration complexity guarantees that, for some variants, match those established for classical centralized trust-region methods. Numerical evaluations demonstrate that our approach achieves a favorable trade-off between performance and efficiency, requiring only moderate communication overhead compared to state-of-the-art methods in the literature.
Read the full abstract ▾
Distributed Computing
Prediction-Robust Service Deployment with Capacity-Aware Edge Admission
Hailiang Zhao, Ziqi Wang, Yifei Zhang et al. (+4 more)
Edge platforms instantiate executable services close to users to reduce request-serving cost, but each instance incurs a one-time deployment cost and remains useful only for a finite time-to-live (TTL). The resulting online decision is both prediction-sensitive and capacity-coupled: an optimistic forecast can waste deployment cost, whereas a delayed decision misses the burst it is intended to serve. We study this problem under a common TTL cost model and propose CAPSUM, a capacity-aware admission policy with an elastic specialization, CAPSUM-E. In the local elastic setting, every node-service trace is exactly a variable-price Bahncard instance. This reduction lets CAPSUM-E inherit PFSUM's tight prediction-error-dependent ratio, including $2/(1+\beta)$ consistency and $1/\beta$ robustness for $\beta>0$. A redirect-aware variant preserves the same local deployment schedule. For finite-capacity nodes, CAPSUM combines size-scaled break-even tests, a utilization-dependent shadow price, and evidence-density eviction; we prove capacity feasibility, scale invariance, and exact agreement with CAPSUM-E under an elastic configuration. We implement an exact local offline dynamic program and compare against direct common-model baselines and documented source-derived adapters for EDP-A, OREO, and uEDC-L. Experiments cover controlled prediction error, three synthetic demand regimes, a causal predictor on a public Globus Compute trace, and joint scaling to 1,024 nodes and 10,000 services. Under the common model, CAPSUM reduces normalized cost by 33.7-42.9% relative to the best source-derived adapter across the synthetic regimes and by 45.5% on the sampled trace.
Read the full abstract ▾
All authors Hailiang Zhao, Ziqi Wang, Yifei Zhang, Mingyi Liu, Xinkui Zhao, Kingsum Chow, Shuiguang Deng
Artificial Intelligence
FractalNet-Based Heterogeneous Federated Learning for Orbital Edge Intelligence in Satellite Mega-Constellations: A Wildfire Case Study
Sai Puppala, Bhabani P. Sinha and Koushik Sinha
Satellite mega-constellations are emerging as large-scale sensing, communication, and computation fabrics, yet their learning architectures remain largely inherited from terrestrial
federated learning and ground-centric mission operations— ill-suited to satellites that differ by orders of magnitude in Size, Weight, Power, and Cost (SWAP-C), radiation tolerance, link availability, and propagation delay. We propose a heterogeneous
federated learning method based on the FractalNet architecture for orbital edge intelligence. We formalize contact-window-constrained, depth-heterogeneous
federated optimization and introduce a distributed path scheduler that assigns model depth as a function of SWAP-C constraints, predicted inter-satellite contacts, and training statistics. To reduce message overhead and energy consumption, each tier pools updates periodically rather than at every contact opportunity, and a three-tier agentic control plane governs in-space scheduling, anomaly escalation, and policy-governed autonomy. As a case study, we apply the framework to wildfire detection, where each orbital shell naturally learns a different semantic level of situational awareness: pixel-scale thermal anomalies at low Earth orbit (LEO), regional fire-front dynamics at medium Earth orbit (MEO), and larger-scale risk propagation at geostationary or high Earth orbit (GEO/HEO). Experiments on simulated mega-constellations validate the approach across convergence, communication efficiency, energy adaptation, scheduled-pooling savings, robustness, and latency.
Read the full abstract ▾
Artificial Intelligence
A Score Is Not a Policy: Measuring the Value of Adaptive Revision
Ayushi Chadha
As agentic systems become compound systems, increasingly important decisions move above task execution itself: when should a higher-level controller preserve the strategy guiding another process, and when should it revise it? We study this meta-level control problem in a hierarchical latent reasoner whose manager can retain or replace a commitment governing lower-level computation. Across three precommitted training seeds, learned revision timing produces qualitatively different policies, ranging from an almost deterministic early clock to substantially more state conditioned schedule distributions, yet none outperforms the best forced timing policy evaluated on the same frozen checkpoint. This separates state dependence from decision value: a controller can vary its actions with internal state without turning that variation into a reproducible task-performance benefit. A deeper intervention study on the original checkpoint shows that timing itself is consequential and order-sensitive, while exhaustive enumeration reveals that a strong fixed schedule captures most of the measurable value available from timing at this decision budget. Counterfactual PERSIST/REPLAN diagnostics further show why score-level evidence can be misleading when predictability is dominated by decision position rather than within-position discrimination. Together, these results argue that learned meta-level control should be evaluated along three separate axes: whether its score depends on state, whether that dependence changes realized behavior, and whether those changes capture outcome value beyond a strong non-adaptive policy.
Read the full abstract ▾