Peer Revenue Equity in science

Papers

Fresh from arXiv, typeset properly. Highlighted words trade on the market.

Computer Science

cs Watch Computer Science

Wednesday 30 September

Sound · Cross-listed from eess.AS

Pitch Smoothing Using Relative Interval Networks

Chin-Yun Yu, Chi-Jen Peng, Li Su et al. (+1 more)

Pitch tracking systems typically couple a per-frame fundamental frequency ($F_0$) estimator with a temporal smoothing stage to obtain continuous trajectories. Conventional Viterbi smoothers enforce first-order continuity but lack long-term temporal awareness and could lock into octave errors across corrupted frames. We propose Relative Interval Networks (RIN), a trajectory smoothing framework that reconciles per-frame pitch estimates with data-driven multi-hop pitch differences. We extract robust relative pitch intervals across arbitrary frame offsets using Variable-Q Transform cross-correlation. We formulate pitch smoothing as an $L_1$-norm optimization problem and prove its equivalence to a minimum cost circulation problem, solved efficiently via linear programming. Evaluations across speech, singing, and instrumental datasets show that RIN substantially improves weak estimators, matches or outperforms Viterbi decoding at a comparable computational cost, and provides superior robustness under certain acoustic degradation.
Read the full abstract

All authors Chin-Yun Yu, Chi-Jen Peng, Li Su, György Fazekas

Distributed Computing

EPR Count for Runtime Prediction in Distributed Quantum Computing

Fatih E. Bilgen and Ozgur B. Akan

EPR-pair consumption is commonly used as a communication-cost objective in distributed quantum computing, but minimizing EPR cost does not necessarily minimize distributed execution time. Despite its widespread use, the reliability of EPR count as a runtime surrogate has received limited direct characterization across different workloads and communication conditions. This work addresses this gap by systematically evaluating the relationship between EPR cost and runtime over all balanced two-QPU mappings of QFT, QAOA, and CDKM circuits. The results show that mappings with identical EPR cost can have substantially different execution times, lower-EPR mappings can be slower, and minimum-EPR mappings can be runtime-suboptimal. The reliability of EPR count is strongly workload dependent and also changes with the communication operating regime: increasing EPR-generation latency amplifies runtime differences hidden by equal EPR cost, while increased communication concurrency can reduce the ability of EPR count to preserve runtime ordering. These results show that EPR count and execution time are related but distinct optimization objectives, and identify regimes in which explicit communication-aware runtime modeling is necessary.
Read the full abstract

Systems and Control

Concept and Rationale for Stratospheric Balloon-Based Laser Debris Removal

F. Ruiz Vincueria, M. Divoky, E. Sanchez-Laulhe et al. (+9 more)

Laser-based deorbiting has long been proposed as a promising technique for mitigating the growing population of orbital debris, particularly decimetre-scale objects – large enough to cause catastrophic collisions yet too small to be economically captured by current active debris removal (ADR) missions. However, the practical implementation of laser-based approaches remains constrained by several factors. Ground-based systems suffer from atmospheric absorption, scattering, and turbulence, which degrade beam quality and limit effective energy delivery. Space-based platforms benefit from superior propagation efficiency, enabling more compact and lower-power systems, but require complex, high-cost orbital infrastructure with demanding deployment, maintenance, and scalability challenges. This paper proposes a stratospheric balloon-based concept for laser debris removal, exploring the intermediate regime between ground-based and space-based approaches. Operating above 99% of the atmospheric mass, such a platform may enable ultraviolet (3HG/4HG) laser operation – largely precluded from ground level by ozone absorption – with improved propagation and ablation coupling efficiency, while presenting a different set of engineering constraints than orbital systems. Rather than a fully developed system, this work focuses on the conceptual design and underlying rationale. We describe the main architectural elements, discuss the expected operational envelope, and position the concept against existing ground- and space-based solutions, identifying potential advantages as well as key challenges and limitations requiring further investigation. Developed within STRATOLASER, an EIC Pathfinder Challenge project targeting TRL 4 through stratospheric balloon experiments.
Read the full abstract

All authors F. Ruiz Vincueria, M. Divoky, E. Sanchez-Laulhe, H. Yang, J. Rodr{\i}guez, V. Montero, V. Canales, E. Alcalde, A. Zavoli, M. Rico, J. Klare, A. Ollero

Systems and Control

Flatness-based Neural Network Control of DC-DC Converters

Kamakshi Tatkare, Ufuk Topcu and Brian Johnson

In this work, we present a simple method to engineer a small neural network to control power converters. Learning-based methods have emerged as a promising approach for power converter control over wide operating point variations. However, many of these approaches do not exploit the analytical structure of converter dynamics and, therefore, search over unnecessarily broad policy classes. We address this gap with a systematic framework that combines model-based trajectory generation with policy learning. First, we use differential flatness of averaged converter models to generate feasible state and constrained duty cycle trajectories where duty acts as a control effort signal. Then for policy learning, we use these model-generated trajectories to construct labeled state-to-duty data and train a multilayer perceptron to approximate the control law with a static map. We apply this framework across buck, boost, and buck-boost converters through topology-dependent flat output selection and a common learning pipeline. Simulations show close agreement between learned and model-generated duty trajectories and stable regulation under large synchronous input and load disturbances.
Read the full abstract

Machine Learning

Predicting Multi-View Rashomon Representation: Can We Learn Where Models Disagree?

Mingyue Ma, Zongbo Han, Changqing Zhang et al. (+1 more)

Foundation models are increasingly adopted across a wide range of applications, often serving as core blocks within AI systems. Yet different foundation models may encode the same input from multiple different views, leading to substantial representation disagreement, which we term Rashomon Representation. Such disagreement often signals inputs that a given model encodes in a way inconsistent with other models, offering a valuable yet underexplored signal for input reliability estimation. While prior work has largely focused on measuring disagreement across multiple models with a representation set, we instead focus on predicting disagreement from a single representation. We hypothesize that this disagreement follows some consistent, input-dependent patterns rather than occurring at random. To test this, we quantify disagreement by comparing each sample's nearest neighbors across different models' representation spaces, then train a lightweight predictor that estimates disagreement from a single model's representation. At inference time, given a new input, the predictor uses that input's representation to tell whether it aligns with or diverges from those of other models. Extensive experiments across diverse foundation models and datasets show that representational disagreement is indeed input-dependent, predictable, and generalizable, enabling efficient reliability estimation of foundation models.
Read the full abstract

All authors Mingyue Ma, Zongbo Han, Changqing Zhang, Guangyu Wang

Sound

SEAR: Spoofing Evidence-Grounded Audio Reasoning Benchmark for Audio Language Models

Rong Wan, Suliu Qin, Jiaxi Li et al. (+5 more)

Audio language models (ALMs) are increasingly used for audio deepfake detection (ADD), yet existing benchmarks assess their verdicts or rationale plausibility without verifying the underlying acoustic evidence. To address this issue, we first introduce spoofing evidence-grounded audio reasoning (SEAR), a four-task AQA benchmark to evaluate ALM-based ADD through acoustic evidence identification and quantification, deepfake detection, and forensic rationale generation. We further propose a bona-fide-based acoustic evidence agent (BAEA), which equips a frozen ALM with controlled acoustic tools under fixed or adaptive evidence-acquisition policies. Experiments with six ALMs reveal a clear gap between plausible rationales and verifiable acoustic evidence reasoning, while BAEA-Fixed improves final verdicts and forensic rationales on both evaluation partitions. Controlled interventions further show that misleading evidence degrades both detection and grounding performance.
Read the full abstract

All authors Rong Wan, Suliu Qin, Jiaxi Li, Wei Xie, Wenwu Wang, Xiaolong Han, Lu Yin, Xilu Wang

Computation and Language

When a Kindergartener Solves Calculus: Measuring Capability Leakage in Role-Prompted Reasoning Models

Pakhapoom Sarapat, Saksorn Ruangtanusak, Kunat Pipatanakul et al. (+1 more)

We investigate the problem of role-capability leakage (RCL), in which a role-prompted reasoning model generates convincing in-role text while continuing to exhibit capabilities on benchmarks that exceed those implied by the assigned role. For example, when a model is prompted to assume the role of a kindergarten student, one might expect its performance on a mathematics benchmark to reflect kindergarten-level ability rather than expert-level proficiency in solving calculus problems. We introduce RoleCapBench, a curriculum-grounded benchmark for evaluating RCL across six educational roles and four assessment levels spanning elementary school through A-level, and use it to evaluate three open-weight reasoning models. We find that although the models can generate stylistically convincing in-role responses, they consistently fail to align their underlying capabilities with their assigned roles. Naive role prompting yields strong role-voice scores of 1.218–1.389 while retaining above-role accuracy of 0.811–0.898. RCL persists across a range of prompting conditions, including prompts that explicitly instruct the model to match the role's capability level. To mitigate this problem, we propose Injection, an inference-time intervention that combines explicit, role-specific capability guidelines with a guiding prefilled response prefix. Injection improves role-capability alignment across models, reducing above-role accuracy by up to 0.562 while preserving in-role accuracy with a marginal drop of less than 0.058 across most models. All artifacts, including scripts and evaluation data, will be released upon acceptance.
Read the full abstract

All authors Pakhapoom Sarapat, Saksorn Ruangtanusak, Kunat Pipatanakul, Pittawat Taveekitworachai

Cryptography and Security · Cross-listed from quant-ph

Natural Barriers to Quantum Extraction: On the Post-Quantum (In)security of (O)EKE and Masny-Rindal OT

James Bartusek and Jake Januzelli

Encrypted key exchange (EKE), introduced by Bellovin and Merritt (IEEE S&P 1992), and Masny-Rindal OT, introduced by Masny and Rindal (ACM CCS 2019), are highly-efficient methods for compiling essentially any KEM into advanced cryptographic protocols, namely password-authenticated key exchange (PAKE) and oblivious transfer (OT), by relying only on idealized symmetric-key primitives. They have become leading candidates for practically-implementable PAKE and OT due to (1) their simplicity, (2) their plug-and-play nature, allowing for flexibility in the choice of KEM, and (3) existing proofs of UC-security (in the classical adversarial model). Due to point (2) above, these compilers yield attractive candidates for efficient post-quantum PAKE and OT, especially given the recent post-quantum KEM standardization efforts. This motivates the question of whether the (UC-)security of these compilers translates to the quantum adversarial model. In this work, we show that it does not. In particular, we prove that a general family of (O)EKE protocols, as well as Masny-Rindal OT, are not UC-secure against quantum polynomial-time adversaries, even when instantiated with a post-quantum KEM. To establish UC-insecurity, we devise an adversarial strategy that provably thwarts any attempt by the simulator to extract its input (the password in the case of PAKE, and the receiver's choice bit in the case of OT). To complement these negative results, we establish that both compilers yield certain notions of game-based security. Along the way, we establish a novel “advantage-tight” one-way to hiding lemma that may be of independent interest.
Read the full abstract

Artificial Intelligence · Cross-listed from stat.ML

BayesNDE: Bayesian Generative Modeling for Neural Density Estimation

Chenglin Li and Qiao Liu

Density estimation is a fundamental problem in statistics and machine learning. In this work, we introduce BayesNDE, a neural density estimator based on Bayesian generative modeling. BayesNDE learns a Bayesian generative model and evaluates its density without requiring invertible networks or Jacobian-determinant computation. For each observation, it infers a sample-specific latent posterior to construct an adaptive proposal that focuses computation on regions contributing most to its density. Bridge sampling then combines samples from this proposal with separate posterior samples to estimate the density. Experiments on nonlinear and multimodal synthetic datasets show improved estimation of density values and better recovery of the density structure compared to the state-of-the-art neural density estimators. Applications to real-world datasets further demonstrate improved anomaly detection. Together, these results highlight BayesNDE as a flexible and effective neural density estimator, demonstrating how posterior inference can turn generative models into tools for density estimation. The code and tutorials are available at https://github.com/liuq-lab/BayesNDE.
Read the full abstract

Computer Vision

DyRAD: Radar Novel View Synthesis for Dynamic Driving Scenes

Merav Keidar, Tomer Borreda, Rajalakshmi Nandakumar et al. (+1 more)

Reconstructing dynamic driving scenes from recorded sensor data supports closed-loop evaluation of autonomous driving systems by synthesizing observations beyond the original trajectory. Unlike cameras and LiDAR, radar measures radial velocity directly through Doppler. Yet existing radar novel-view synthesis fails to exploit this capability: methods addressing dynamic scenes reconstruct only range-azimuth tensors, while methods that render Doppler assume static scenes. Moreover, because radar processing spreads each reflection across multiple bins, existing representations absorb this spread into scene geometry, causing it to render incorrectly when the viewpoint moves. We present DyRAD, which models dynamic driving scenes using static background reflectors and motion-tracked dynamic point reflectors to render complete range-azimuth-Doppler (RAD) tensors. Reflector velocities are derived from object tracks and projected onto the line of sight, making Doppler both a rendered output and supervision for those tracks. Crucially, we render reflectors through a fixed analytic point-spread function (PSF) derived from the radar's signal-processing chain, preventing sensor-induced spread from being baked into the scene representation. Beyond improving scene reconstruction, this separation also enables zero-shot sensor-configuration transfer, allowing the same reconstructed scene to be rendered under different radar specifications without refitting. We evaluate DyRAD on RADIal, Boreas, and a synthetic benchmark across both on-path poses and displaced viewpoints untested by prior work. On RADIal, DyRAD recovers radar detections in 90.7% of reference-detected objects, compared with 26.9% for the strongest baseline.
Read the full abstract

All authors Merav Keidar, Tomer Borreda, Rajalakshmi Nandakumar, Or Litany

Systems and Control

Parameter-Robust Sensorless Control of IPMSM Drives With Adaptive Flux Observer

Fobao Zhou, Zhenxiao Yin, Xueyan Wang et al. (+3 more)

To address parameter sensitivity commonly in interior permanent magnet synchronous motor (IPMSM) sensorless control, this paper proposes a parameter-robust control framework by extending the adaptive flux observer from surface-mounted PMSM to salient-pole machines. First, parameter mismatches are interpreted as an equivalent flux vector with $d$-axis and $q$-axis components. Permanent magnet flux, $d$-axis inductance, and resistance mismatches are mapped to the $d$-axis flux component, whereas $q$-axis inductance mismatch is mapped to the $q$-axis component. Then, a scalar adaptive flux update is designed to compensate the $d$-axis flux variation by tracking the equivalent flux magnitude. Since the $q$-axis component rotates the observed flux direction and cannot be eliminated by scalar adaptation, a high-frequency $q$-axis voltage injection method is introduced to identify the $q$-axis inductance. Meanwhile, stability analysis proves bounded equivalent flux magnitude tracking and identifies the residual tangential flux responsible for position error. Finally, experimental results show that the proposed method does not require precise resistance, $d$-axis inductance, or flux, while the $q$-axis inductance is provided by online identification.
Read the full abstract

All authors Fobao Zhou, Zhenxiao Yin, Xueyan Wang, Yang Shen, Yuanfeng Qu, Hang Zhao

Machine Learning

Dynamic LoRA-Experts and Prototype-Ensemble Matching for Class-Incremental Learning

Hongwei Zhao, Rui Liu and Yansong Liu

Class-Incremental Learning (CIL) aims to continuously learn new classes without forgetting previously acquired knowledge. Parameter-efficient fine-tuning with pre-trained models reduces parameter overhead but can suffer from cumulative interference and suboptimal alignment between inference samples and specialized modules. We propose Dynamic LoRA-Experts and Prototype-Ensemble Matching (DLEPEM), a two-stage rehearsal-free framework. DLEPEM allocates a task-specific LoRA-Expert for each incremental task to reduce cross-task interference, then combines frozen pre-trained-model prototypes with task-adaptive LoRA-Expert prototypes for reliable task-level discrimination. Experiments on standard CIL and Few-Shot CIL benchmarks demonstrate strong performance under the evaluated protocols.
Read the full abstract

Artificial Intelligence

Learning Steganography Is Easy, Learning Steganographic Reasoning Is Hard

Julian Schulz, Lukas Fülle and Rieke Fruengel

Chain-of-thought monitoring as an approach for AI oversight and control is threatened by the possibility of steganographic reasoning, where LLMs conceal their reasoning inside innocuous-looking text. Two neighbouring capabilities, steganographic messaging (passing a concealed message) and encoded reasoning (reasoning in an illegible but unconcealed format), have already been shown to emerge under training pressures that occur in real pipelines, such as reinforcement learning against monitors. This suggests that steganographic reasoning too might arise as an unintended side effect of training. Here, we compare how easily models learn steganographic reasoning and these two neighbouring capabilities across three elicitation methods: reinforcement learning, in-context learning, and supervised fine-tuning (SFT). For most tasks, models learn steganographic reasoning only under SFT, while they learn steganographic messaging and encoded reasoning under all three elicitation methods. Even under SFT, steganographic reasoning requires at least twice as much training as messaging, and for several model-task combinations it is not learned at all. However, on a cover task that makes hiding information especially convenient, steganographic reasoning can be successfully learned under all three elicitation methods. Steganographic reasoning is thus much harder than steganographic messaging and encoded reasoning, and learning the latter two does not imply learning the former. Yet it lies within reach: an easy version is learned under every elicitation method, when the cover task is convenient for hiding information.
Read the full abstract

Machine Learning

Fast Regularized Policy Mirror Descent with One-Step TD Updates

Qipei Chen, Wenye Li, Yule Sun et al. (+1 more)

Policy mirror descent (PMD) enjoys fast convergence in regularized Markov decision processes (MDPs), but existing guarantees often rely on exact or increasingly accurate policy evaluation. We analyze PMD coupled with a persistent critic advanced by one temporal-difference (TD) update. For finite discounted MDPs, we establish global linear convergence in value for exact coordinate-wise Bellman updates, with any positive constant actor stepsize and arbitrary finite critic initialization. The proof combines a resolvent-based auxiliary distribution with a decaying Bellman-violation correction and a potential weighted by inverse coordinate weights. We then study stochastic TD-PMD with general strongly convex mirror maps under a single off-policy Markov trajectory. With suitably chosen constant stepsizes and a finite-batch TD update, the method achieves an expected value gap of $\epsilon$ after $\widetilde{O}(1/((1-\gamma)^5 \widetilde{\sigma}_b \epsilon))$ transitions. The stochastic analysis relies on the trajectory-wise Lipschitz continuity of the regularizer, derived from uniform bounds on vertex Bregman divergences, together with a visitation-weighted resolvent estimate for signed critic-error propagation that yields an inverse-linear dependence on behavior coverage $\widetilde{\sigma}_b$. In contrast to many prior guarantees for regularized policy optimization, our sample-complexity guarantee holds without trajectory resets, generative-model access, or nested policy-evaluation loops. Numerical results are consistent with the theoretical convergence analysis.
Read the full abstract

All authors Qipei Chen, Wenye Li, Yule Sun, Ke Wei

Computer Vision

Spherical Interpolation for Backward-Compatible Multimodal Representations

Simone Ricci, Niccolò Biondi and Federico Pernici

Contrastive vision-language models map visual and textual representations into a shared normalized embedding space, making cosine similarity the natural metric for cross-modal retrieval. A practical challenge arises during model upgrades: independently trained models generally produce incompatible representation spaces, so replacing a deployed model typically requires recomputing embeddings for the entire gallery, which is prohibitively expensive at scale. Orthogonal post-hoc alignment can partially mitigate this problem by mapping new-model queries into the old-model gallery space. However, because independently trained models can differ in fine-grained representation structure, the orthogonal alignment remains approximate, leaving a residual angular discrepancy between the old-model query and the aligned new-model query. We study whether interpolation along the spherical geodesic between these two normalized query representations can improve retrieval without re-indexing the gallery. We characterize when this path contains an interior query direction closer to an idealized retrieval-optimal direction than either endpoint, and connect this characterization to Recall@$K$ through a local margin-based certification result. Experiments across multiple benchmarks and model families show that post-alignment spherical interpolation improves over orthogonal alignment alone, recovering backward-compatibility in most evaluated settings. Consistent with our geometric characterization, per-query oracle analysis shows that retrieval-favorable interior points occur frequently in practice. Code is available at https://github.com/miccunifi/SLERP_backward_compatibility .
Read the full abstract

Machine Learning

RainAtlas: A Multi-Continental Dataset for Precipitation Downscaling

Pierre-Louis Lemaire, Luca Schmidt, Wietze Suijker et al. (+2 more)

Extreme rainfall events are increasing in intensity and frequency as climate change accelerates. While kilometer-scale precipitation forecasts are critical for supporting local decision-making, the limited availability of high-resolution precipitation observations hinders their accuracy, especially in under-resourced regions. Machine learning models are widely used to downscale precipitation data to km-scale, but their application to unseen geographies presents challenges. First, processing raw high-resolution precipitation datasets across regions requires significant engineering and domain expertise. Second, generalization across regions remains difficult. To help overcome these barriers, we release RainAtlas, a large-scale, ML-ready and multi-continental dataset for precipitation downscaling. Covering three continents, RainAtlas harmonizes heterogeneous hourly km-scale observations to a common 2-km grid. Each regional partition contains around 210,000 aligned low- and high-resolution precipitation pairs, respectively from ERA5 reanalysis and direct observations. We benchmark state-of-the-art ML-based downscaling models across RainAtlas using a wide range of metrics. Our evaluation reveals substantial variance in out-of-domain generalization depending on the training regions. This underscores the need for cross-regional, multi-source km-scale evaluation, establishing RainAtlas as a well-positioned benchmark for precipitation downscaling research.
Read the full abstract

All authors Pierre-Louis Lemaire, Luca Schmidt, Wietze Suijker, Alex Hernandez-Garcia, David Rolnick

Computer Vision

P-SRM: Selective Recovery of Rejected Predictions in Visual Tracking

Youbin He and Siwei Wang

Many visual tracking methods use rejection mechanisms to suppress unreliable predictions. However, these mechanisms can also reject correctly localized candidates, leaving useful information unused. We investigate how to identify and recover these candidates while preserving native accepted outputs and candidate coordinates. To this end, we propose P-SRM (Post-rejection Selective Recovery Method), which combines spatial responses, past accepted states, and native decision margins to reassess candidates and selectively restore reliable predictions. We evaluate P-SRM on six trackers and four datasets spanning category-specific, point, and generic object tracking. Across all nine configurations, P-SRM improves rejected-candidate ranking and overall tracking performance. These results show that post-rejection verification can identify and recover useful predictions discarded by native rejection, demonstrating the value of reusing rejected information. Project repository: https://github.com/PalestyHR/P-SRM.
Read the full abstract

Machine Learning · Cross-listed from stat.ML

Estimation of the Label-Noise Transition Matrix with Performance Guarantees via Selective Classification

Xabier de Juan, Santiago Mazuelas, Yilun Zhu et al. (+1 more)

Modern machine learning depends heavily on massive datasets, but obtaining high-quality annotations at scale is often expensive. As a result, learning from noisily-labeled data has become common, making accurate estimation of the label-noise transition matrix crucial. However, existing transition matrix estimators rely on the fragile estimation of class-posteriors and do not provide finite-sample performance guarantees. In this work, we propose a novel methodology to estimate the transition matrix based on one-sided selective classification. This approach bypasses class-posterior estimation, provides finite-sample performance guarantees, and leverages flexible learning methods for binary classification. Moreover, we introduce effective algorithms to implement the proposed methodology and provide their refined finite-sample performance bounds.
Read the full abstract

All authors Xabier de Juan, Santiago Mazuelas, Yilun Zhu, Clayton Scott

Information Retrieval

KUAISHOU Explorer LLM-Rec Challenge 2026: Reasoning Generative Recommendation

Jiangxia Cao, Hao Peng, Wenlong Xu et al. (+97 more)

Generative recommendation, has been attracted a surge of attentions in industrial and academic research community, towards to build more smart system to build next-generation recommender. Under the significant developing wave of large language model, our team have been developed Semantic ID based OneRec/OneRec-V2. These models have been widely deployed in production and demonstrate the scaling potential of the autoregressive next-item prediction paradigm for industrial recommender systems. Building on the success of OneRec, we further explored a series of models, including OneRec-Think, OpenOneRec, and OneReason, that connect item Semantic IDs with natural language in a unified representation space and seek to unlock the potential of natural-language chain-of-thought (CoT) reasoning for recommendation. However, our preliminary works found that introducing reasoning CoT does not always improve the recommendation performance. To address this issue, OneReason strengthens the semantic alignment between items and language, introduces structured template-based supervision for interest reasoning, and applies advanced reinforcement learning techniques to make reasoning more beneficial to recommendation. As a frontier topic to building recommendation foundation models, we believe this topic has significant research value and hope to encourage more researchers to explore it together. To this end, together with the SIGIR 2026 community, we organized the KUAISHOU Explorer LLM-Rec Challenge 2026: Reasoning Generative Recommendation.
Read the full abstract

All authors Jiangxia Cao, Hao Peng, Wenlong Xu, Jiaxin Deng, Zhixin Ling, Xingmei Wang, Kun Shang, Can Tang, Zhihuai Cai, Jun Du, Fang Su, Xiaojuan Liu, Yiling Li, Chenglong Yu, Chongling Rao, Haixuan Gao, Haitao Xu, Jian Liang, Ruiming Tang, Chenglong Chu, Guohong Mu, Honghui Bao, Hui Wang, Jialong Chen, Jiao Ou, Muhao Wei, Peng Zhang, Renpu Liu, Ruochen Yang, Shugui Liu, Xinqi Jin, Yan Sun, Yifan Wang, Yingzhi He, Yufei Ye, Yusen Huo, Tingkuo Wang, Jihong Zhang, Lanxi Zhu, Pengyuan Liu, Zhipeng Yi, Luankang Zhang, Hang Lv, Xuyang Zhi, Tianyu Li, Bintao Wu, Chuang Ou, Siyue Su, Ziyuan Wang, Yuliang Sun, Baiyan Che, Feiyang Xu, Shiwen Zhang, Shiteng Cao, Chongcong Jiang, Yuan Fang, Xiangwu Yang, Hao Deng, Zijian Du, Pengxun Wang, Xiaoming Wang, Shun Qin, Yingqi Song, Tianyi Li, Naixiao Peng, Chenyu Zhou, Qiliang Jiang, Quan Zheng, Cheng Jin, Siying Zeng, Hongjia Xu, Junwu Hu, Teng Fu, Zhengkang Mei, Haijun Yu, Kai Li, Shengyang Zhou, Zhijia Wei, Siyi Xiong, Bo Liu, Zichun Guo, Zhubin Han, Jinpeng Fu, Bingqian Liu, Yuyi Wang, Yu Liu, Qinghai Tan, Ruijie Zhou, Zhuohang Li, Zhijia Zhong, Xiangnan He, Jirong Wen, Min Zhang, Wenwu Ou, Peng Jiang, Han Li, Kaiqiao Zhan, Yanan Niu, Lantao Hu, Kun Gai

Computation and Language

Synthetic Pre-pretraining Survives Scale, but Not as a Grammatical Prior

Atsuki Yamaguchi, Tatsuro Inaba, Joel Niklaus et al. (+3 more)

Pre-pretraining (PPT) on synthetic non-natural language data improves token efficiency during language model pre-training (PT). Prior work attributes this gain to a grammatical prior, i.e., a structural inductive bias learned during PPT that transfers to natural language grammar. However, PPT has only been tested on models of at most 1B parameters and PT budgets below 2B tokens on predominantly web text. It is unknown whether PPT is effective at larger scales and under more realistic PT data mixtures that combine diverse sources (e.g., code and math). We therefore present a comprehensive study on PPT spanning five PPT tasks, four PT data mixtures, four parameter scales (500M to 7B), and PT budgets of up to 100B tokens. Our results demonstrate that the downstream performance and token efficiency gains of PPT persist at scale, e.g., saving at least 21B PT tokens at the 3B scale. However, in contrast to prior work, we find no consistent evidence that these gains stem from a grammatical prior. Downstream performance does not consistently align with grammatical acceptability across model sizes. Instead, we find that downstream gains arise from PPT tasks that improve long-range retrieval. Finally, PPT performance gains are robust to how PT data mixtures are composed and diminish only when web text is absent. Overall, PPT is a low-cost addition to PT, and future PPT task design should target long-range retrieval rather than natural language grammar.
Read the full abstract

All authors Atsuki Yamaguchi, Tatsuro Inaba, Joel Niklaus, Michal Štefánik, Aline Villavicencio, Nikolaos Aletras

Computational Complexity · Cross-listed from quant-ph

Improved Quantum Random Self-Reduction for Linear Problems

Vahid R. Asadi, Shuichi Hirahara and Nobutaka Shimizu

We study quantum random self-reductions for linear problems over finite fields. Let $M\in\mathbb{F}^{n\times n}$ be an arbitrary matrix, and let $\mathcal{O}$ be an oracle that agrees with the linear map $x\mapsto Mx$ on an $\varepsilon$-fraction of inputs $x\sim\mathbb{F}^n$. Given coherent access to $\mathcal{O}$ and coherent entry access to $M$, we give a uniform quantum reduction that computes $Mx$ on any prescribed input $x$ with probability at least $2/3$ in time $\widetilde{O}(nT^{1/3})$, for $n\le T\le n^{3/2}$ and constant field size and $\varepsilon$, where $T$ is the cost of one coherent query to $\mathcal{O}$. In particular, when $T=\widetilde{O}(n)$, the reduction runs in time $\widetilde{O}(n^{4/3})$, improving the $\widetilde{O}(n^{3/2}+T)$ reduction of Asadi, Golovnev, Gur, Shinkar, and Subramanian (SODA 2024). Our reduction uses the Bogolyubov–Ruzsa subspace guaranteed by additive combinatorics, but it avoids learning this subspace explicitly, which was computationally expensive for the previous reduction; in particular, it does not recover a basis for its orthogonal complement. The main technical step is to decompose the inputs into sparse pieces and find a vector that lies outside the Bogolyubov–Ruzsa subspace via a quantum search based on amplitude amplification. This yields a tunable tradeoff between the cost of querying the average-case oracle and the cost of verifying matrix-vector products.
Read the full abstract

Robotics

Toward Real-Time VLAs: Stage-Aware Two-Step Flow Denoising and System-Level Evaluation

Di Wu, Rongtian Shen, Ping Liu et al. (+8 more)

Vision-language-action (VLA) models face a timing gap between low-rate inference and high-rate robot execution. We characterize this gap through end-to-end latency measurements of model inference and the robot execution chain. Repeated Flow Matching denoising contributes substantially to inference cost, while robot-side delays mainly arise from perception acquisition, communication scheduling, and physical response. Analysis of the velocity field shows relatively stable magnitude and direction in early integration, followed by stronger directional correction near the terminal steps. Based on this stage heterogeneity, we propose two-stage non-uniform denoising, reducing the number of steps from 10 to 2 and model-inference time from 61.557 ms to 21.956 ms. We also develop a distributed real-time VLA framework with independent inference, action-publication, and robot-control rates, modular observation acquisition, and action-provenance logging. Using {\pi}0.5 as the baseline, we evaluate six real-time execution methods on a long-horizon physical garment-folding task. Legato performs best overall among training-based methods, while Temporal Smoothing leads among training-free methods; both perform strongly in task success, completion time, action continuity, and acceleration smoothness. Combining two-step denoising with representative execution methods substantially reduces inference cost with a small reduction in task performance. These results motivate joint optimization of model-inference efficiency and robot-system timing.
Read the full abstract

All authors Di Wu, Rongtian Shen, Ping Liu, Yan Shen, Zhenhan Yin, Shun Zuo, Xuhua Chen, He Zheng, Lingfeng Zhang, Jianglin Zhang, Tao Zhang

Data Structures and Algorithms · Cross-listed from quant-ph

Learning Random Quantum Circuits and the Emergence of Pseudorandomness

Srinivasan Arunachalam, Qizhao Huang and Makrand Sinha

We give an efficient algorithm for learning $k$-dimensional brickwork random quantum circuits using only copies of the output state obtained by applying $U$ to the all-zero input. For a depth-$d$ circuit on $n$ sites with random $2\ell$-qubit gates, the algorithm learns the original circuit $U$ with high probability in $\text{poly}(n,2^{\ell d})$ time for every constant dimensional lattice. In particular, the algorithm is polynomial time as long as $\ell d = O(\log n)$. In one dimension, this reaches the natural boundary suggested by pseudorandomness: pseudorandom states require $\ell d=\omega(\log n)$, and structured cryptographic constructions suggest that this scale may be achievable from above. In higher dimensions, it remains plausible that the same $\ell d=\omega(\log n)$ scale remains the threshold for ancilla-free pseudorandomness, and our results help clarify the conditions under which pseudorandomness can arise in this setting. The main idea behind our algorithm is a local correlation criterion that identifies gates in the final layer without learning their entire backward light cones, avoiding a bottleneck in the previous approaches. A key technical ingredient is a Carbery-Wright type anticoncentration inequality for low-degree polynomials of Haar random unitaries whose small-ball exponent is independent of the matrix dimension. The dimension-independent exponent is crucial for handling gates of growing locality. This anticoncentration result may also be of independent interest.
Read the full abstract

Robotics

Learning from Runtime Feedback through Failure-Bank Self-Evolution for Vision-Language-Action Models

Mingyue Cui, Zheyuan Liu, Yihan Zhu et al. (+2 more)

Vision-language-action (VLA) models generalize broadly across robotic manipulation tasks, but complex environments require balancing task success with unintended contact. Runtime shields can correct individual actions, but they leave the underlying policy unchanged, so repeated disagreements may create a persistent policy-shield mismatch that blocks task progress. To address this challenge, we introduce FailBank, a four-stage self-evolving framework that converts runtime feedback into persistent policy improvement. During collection, a fixed CBF-based safety module serves as an observe-only teacher, producing counterfactual corrections while the policy remains in control. Outcome-aware admission then converts useful proposals into corrective targets and retains successful uncorrected actions as quiet anchors for guarded LoRA updates. We evaluate FailBank on the VLA-Arena benchmark across two difficulty levels and two VLA backbones. Compared with the base policies, FailBank improves the joint success-cost operating point. Across the two backbones, FailBank improves task success rate by 8.5 and 6.9 percentage points, while reducing policy-induced cumulative cost by 35.6% and 23.8%, respectively. Compared with runtime shielding, FailBank raises task success rate by 25.4 and 9.5 percentage points, while maintaining comparable policy-induced cumulative cost. These results show that runtime feedback can serve as persistent policy supervision rather than only as a temporary action constraint.
Read the full abstract

All authors Mingyue Cui, Zheyuan Liu, Yihan Zhu, Zheyuan Zhang, Meng Jiang

Operating Systems

Capture the lifecycle: KV Cache management in ReAct Agents with KVTether

Kaihua Fu, Yukun Zhou, Chaokun Chang et al. (+6 more)

Efficient serving of long-context reasoning-and-acting (ReAct) agents relies on KV cache reuse to reduce large language model (LLM) prefill latency and monetary cost. However, a semantic gap exists between agent harnesses and the underlying serving stack. Through context mutation, tool execution, and subagent coordination, context messages may become actively engaged, permanently discarded, and temporarily unused, while the serving stack only observes accesses to the corresponding KV cache. This lifecycle blindness prevents recency-only policies such as LRU from reclaiming dead KV promptly and from preserving older KV that will be reused sooner than newer entries. We present KVTether, a lifecycle-aware KV cache management framework for ReAct agents. By tracing semantic primitives embedded in agent harnesses, KVTether captures runtime lifecycle semantics during highly dynamic execution. KVTether then translates message-level semantics into KV-level lifecycle states and uses these states to drive state-prioritized cache management without exposing physical complexities to agent harnesses. After reclaiming dead KV, KVTether preferentially preserves live-but-idle KV that is waiting for reuse, reducing premature eviction before reuse. Across agent benchmarks and production workloads, KVTether reduces end-to-end request latency by up to 26.3% and 17.4% relative to LMCache and MORI, respectively, and lowers estimated task cost by 40.0% and 33.2% on average.
Read the full abstract

All authors Kaihua Fu, Yukun Zhou, Chaokun Chang, Yinghao Yu, Luping Wang, Guodong Yang, Jiuchen Shi, Quan Chen, Wei Wang

Machine Learning

Should I stay or should I show? Learning to selectively disclose information

Carlotta Giacchetta, Alessando Bogani, Cesare Barbera et al. (+4 more)

In many high-stakes settings, human decision-makers can acquire support information before making a decision. However, acquiring information is costly, and disclosure may fail to improve human decisions or may even impair them. We tackle this problem by studying selective disclosure, i.e., the problem of learning when to reveal support information to a human decision-maker under a budget constraint. We first show that the optimal policy is a threshold rule on the Value of Information (VoI), i.e., the expected reduction in human decision risk induced by disclosure. Since VoI is unknown in practice, we estimate the regime-specific human risks and bound the possible degradation of the resulting plug-in policy relative to lack of disclosure, as well as its regret relative to the optimal policy. Experiments on benchmark datasets show that selective disclosure outperforms both no disclosure and full disclosure, regardless of whether the support information is beneficial or harmful. Two user studies show that human-AI team performance can improve when disclosure is led by our learned policy and not human-selected, although this advantage varies across tasks. A counterfactual benchmark, which replaces participants' predictions with a machine-learning prediction when disclosure occurs, suggests that these differences might depend on lower adherence to advice when the information is automatically provided rather than self-requested.
Read the full abstract

All authors Carlotta Giacchetta, Alessando Bogani, Cesare Barbera, Giovanni De Toni, Michele Caprio, Andrea Pugnana, Andrea Passerini

Machine Learning

Beyond Accuracy: Prefix-Invariant Realizations of Low-Precision Fast Matrix Multiplication

Shuxiao Xie, Shuyang Xie, Yuan Cao et al. (+3 more)

Fast matrix multiplication saves multiplications through exact cancellation, but rounding sums that mix token rows can leave contributions from later tokens in earlier language model outputs. This threatens prefix invariance, which multiple-choice likelihood scoring relies on: a scored likelihood must depend only on its allowed prefix. On Qwen2.5-14B-Instruct, two fast FP8 realizations repaired to ordinary-looking accuracy still change the answers chosen by likelihood on 5.83% and 10.00% of 240 OpenBookQA items when only the text after the allowed prefix is replaced with the bf16 model's own greedy continuation. Both row-local controls, the bf16 model and a deployed FP8 matrix multiplication kernel, change none. Accuracy thus does not certify prefix invariance, and the stability criteria we analyze cannot tell realizations apart: across all 512 sign variants of two-level Strassen they stay constant while teacher-forced perplexities span a 772.4$\times$ range on the same model. We therefore construct certified realizations of two-level Strassen on bounded integer codes that quantize token rows independently, then mix and cancel exactly before rescaling, using 49 block multiplications instead of 64. Our certificate guarantees bitwise equality to a prescribed row-local classical int8 operator at the same quantization specification, so every certified realization inherits its prefix invariance. Certification thus turns realization choice into a pure cost decision: which certified realization runs can no longer change a single scored likelihood.
Read the full abstract

All authors Shuxiao Xie, Shuyang Xie, Yuan Cao, Dezhi Ran, Wei Yang, Tao Xie

Machine Learning

Backward-State Policy Is Part of the Learning Algorithm

Shuxiao Xie, Shuyang Xie, Dezhi Ran et al. (+2 more)

Low-precision training rounds tensors that the backward pass reads again, often for several gradients; each use can read the forward's rounded value, the original, or a new random rounding. This backward-state policy looks like a memory and precision detail, settled by copy accuracy and final loss. We argue that it is part of the learning algorithm, and that neither check shows whether it is right. Copy accuracy does not decide the outcome: in three pairs of 390M runs with an emulated FP8 backward, training fails when attention's backward reuses the forward's rounded output and succeeds with a new rounding from the same distribution. Even the most accurate copy, the original itself, can be wrong by our reference: the gradient of the forward pass as it actually ran, with gradients passed through rounding unchanged. For example, a normalization output stored in low precision feeds two gradients: the gain's gradient needs the original, but the next layer's weight gradient needs the rounded value that layer multiplied. Final loss, the other check, does not rule out the error of reading the original for both: it persists in models trained with such a store, while planned loss comparisons stay within a margin fixed in advance. We therefore derive from this reference which value each use must read, or which substitute gives the same gradient on average with the forward held fixed, and check these per-use requirements on single operators, without training. In three tests using PyTorch and Transformer Engine, the requirements predicted beforehand whether reuse changes what the backward computes on average relative to an independent copy, and every prediction held. Backward-state policy is thus part of the learning algorithm: it should be specified and checked use by use, not settled by copy accuracy and final loss.
Read the full abstract

All authors Shuxiao Xie, Shuyang Xie, Dezhi Ran, Wei Yang, Tao Xie

Machine Learning

A Comprehensive Benchmark of Source-Free Universal Domain Adaptation on Time Series Representations

Romain Mussard, Fannia Pacheco, Maxime Berar et al. (+2 more)

Source-Free Universal Domain Adaptation (SF-UniDA) extends Universal Domain Adaptation by removing access to source data at adaptation time while still handling label-set mismatches between domains. Despite growing interest in this setting for image data, no benchmark exists for time series, which are more challenging. We present the first SF-UniDA benchmark on time series. In addition, we provide the first study of pretrained foundation models as feature extractors for time series domain adaptation. In this context, we identify a critical and previously underexplored limitation of all existing SF-UniDA methods: the inference threshold for unknown-sample rejection is highly sensitive. We address this by proposing a plug-in auto-thresholding module that can be integrated into any SF-UniDA method. Experiments on three well-known time series datasets confirm the suitability of this module. They also highlight that foundation models do not systematically outperform classical backbones and that SF-UniDA tailored for time series is yet to be developed.
Read the full abstract

All authors Romain Mussard, Fannia Pacheco, Maxime Berar, Paul Honeine, Gilles Gasso

Human-Computer Interaction

Making Sense of Animal-to-Human Drug Development Evidence: Stakeholder Practices, Challenges, and Requirements for AI Tools

Rosni Vasu, Simona E. Doneva and Benjamin V. Ineichen

Animal models are widely used to study human biology and health interventions, yet translating findings from animal studies to humans remains challenging. Evidence across preclinical and clinical research informs experimental and translational decisions. Artificial Intelligence (AI) tools are increasingly reshaping how this evidence is searched, synthesized, and used, but it remains unclear how they should support the diverse stakeholders involved in assessing animal-to-human evidence. We conducted semi-structured interviews with 13 stakeholders to examine their evidence practices, challenges, and expectations for AI support. We found that stakeholders approach the same incomplete evidence base with different goals, expertise, and heuristics. Participants valued AI particularly for locating, screening, and extracting evidence, but were more cautious about automated interpretation and quality judgments. They emphasized transparency, source traceability, uncertainty communication, and human oversight. Based on these findings, we derive design implications for role-sensitive AI tools that support more systematic and transparent reasoning about animal-to-human translation.
Read the full abstract