Artificial Intelligence
Positive Ratings, Hidden Concerns: Employee Voice Disclosure in AI-Mediated Organizational Listening
Thilo Tamme, Michael Saatkamp, Alma Bonte et al. (+3 more)
Organizations started listening to employees through conversational AI
agents alongside structured surveys. Little is known about what these channels change in what employees say when disclosure carries hierarchical risk. We report a field study inside a global management consulting firm whose process pairs a pre-survey with an adaptive AI voice interview on the same themes within one session. Across 44 first-session interviews (132 matched theme observations), 20-41% of sessions showed a favorable rating co-occurring with a substantive concern voiced later, depending on the favorability threshold. The Gioia analysis drew on 158 protective quotes from 65 eligible sessions. Disclosure rarely arrived unguarded: employees softened concerns, deflected accountability, and bounded how far they went, and this protective work tracked the perceived legitimacy of the listening structure. We develop a grounded model of bounded disclosure and derive four propositions for voice, channel and listening research. Silence, we argue, can persist inside expression.
Read the full abstract ▾
All authors Thilo Tamme, Michael Saatkamp, Alma Bonte, Daniel Weiss, Anton Hantel, Andrej Levin
Dynamical Systems
Rigidity from the strong unstable foliation for higher-dimensional Anosov and partially hyperbolic flows
Andrey Gogolev
Recently Burniol Clotet proved that two transitive Anosov flows on 3-
manifolds with equivalent strong unstable foliations must be conjugate up to time rescaling. We further explore the mechanism of Burniol Clotet's argument and treat the case of higher-dimensional Anosov flows and their isometric extensions by compact Lie groups.
Read the full abstract ▾
Machine Learning
Same Loss, Different Gradients
Ningkang Peng, Xiaoqian Peng, Yifan He et al. (+4 more)
Differentiable learning typically assumes that the scalar objective evaluated in the forward pass and the gradient supplied to the optimizer in the backward pass describe the same mathematical object. We show that this correspondence can fail when probabilistic objectives rely on finite special-function recurrences, custom backward rules, and numerical clipping. In high-dimensional von Mises-Fisher learning, real numerical implementations can produce identical forward scores and losses at the same learning state while supplying different gradients and following different optimization trajectories. We characterize the structure of this mismatch in finite-start Bessel recurrence and show that classwise radial mismatch can compose through probabilities into a locally nonconservative update field. Evaluating the accuracy of special-function values and derivatives separately is therefore insufficient to characterize the realized learning objective. Motivated by this observation, we introduce AR/FR, a fixed-depth analytic realization that constructs a potential and its derivative jointly, ensuring forward-backward coherence by construction. We establish a uniform cubic-order error bound relative to the exact Bessel ratio over the entire nonnegative concentration axis and propagate this guarantee to learning scores and objectives. As representation dimension increases, the original finite recurrence becomes sequentially deeper, whereas the worst-case AR/FR error guarantee tightens cubically, jointly providing coherence, certified fidelity, and fixed-depth computation. These results suggest that a differentiable numerical primitive is defined by both the values it realizes and the derivatives it actually supplies to the optimizer; together, they constitute the numerical realization of the learning
algorithm.
Read the full abstract ▾
All authors Ningkang Peng, Xiaoqian Peng, Yifan He, Anjie Hu, Chao Tan, Peirong Ma, Yanhui Gu
Machine Learning
How Accurate Is Accurate Enough?
Ningkang Peng, Qianfeng Yu, Jingyang Mao et al. (+2 more)
How accurate must a numerical approximation be within a learning system? Primitive error alone cannot answer this question: errors of the same magnitude can have very different consequences for losses, predictions, and gradients at different learning states. We study this question through the learning objective itself. The objective weights classwise numerical errors nonuniformly according to the current state, so the importance of an error depends not only on its magnitude but also on the class it affects and the weight that class receives. For softmax cross-
entropy, we characterize this coupling between class weights and errors and derive the exact extrema of the signed loss change over pairings of fixed non-target probability and score-error multisets, with the target probability and target score error held fixed. Building on this structure, we establish finite-error guarantees that propagate primitive error to losses, probabilities, predictions, and feature gradients, then invert these guarantees to obtain a certified primitive tolerance for the current state under prescribed learning-level error requirements. We give a complete instantiation of the framework in high-dimensional von Mises-Fisher learning. Controlled interventions and a large collection of saved learning states show that identical primitive error can produce substantially different learning consequences, while certified numerical tolerances vary by orders of magnitude across states under the same learning-level requirements. These results show that the adequacy of a numerical approximation must be assessed in relation to the current learning state and the quantity to be preserved; numerical accuracy should itself be treated as part of the learning objective.
Read the full abstract ▾
All authors Ningkang Peng, Qianfeng Yu, Jingyang Mao, Xiaoqian Peng, Yanhui Gu
Machine Learning
Does a Shared Temperature Imply a Shared Angular Scale in Probabilistic Contrastive Learning?
Ningkang Peng, Qianfeng Yu, Jingyang Mao et al. (+4 more)
In probabilistic contrastive learning, a shared temperature is commonly interpreted as a shared similarity scale, but this interpretation does not hold for high-dimensional distributional class representations. We study the exact von Mises-Fisher (vMF) probabilistic score used by ProCo when representation dimension and class concentration grow jointly. We prove that the score retains a class-dependent leading angular gain $g_c=A_c/\tau$, where $A_c$ is the mean resultant length. This gain enters Softmax competition, pairwise decision boundaries, and feature gradients. On real CIFAR-LT, ImageNet-LT, and iNaturalist representations, the theory accurately predicts boundary movements and local gradient changes under the full vMF score. Classwise temperature adjustment also changes the cosine-zero intercept and finite-dimensional response. We construct intercept-preserving and Pure Angular controls to separate the leading gain from these accompanying changes. Complete gain equalization yields a shared-scale cosine prototype rule at leading order; a finite-dimensional margin condition guarantees agreement of the two classifiers. Across 16 frozen representation settings, prediction agreement is 98.43-99.99%, with disagreements concentrated at small cosine margins. In controlled contrastive-only training with the training-frequency prior, Pure Angular editing improves both learned representations at all tested CIFAR-10/100 imbalance factors and retains positive changes on ImageNet-LT. Thus vMF concentration not only describes class distributions, but also forms a decision and learning scale in high-dimensional probabilistic contrastive learning.
Read the full abstract ▾
All authors Ningkang Peng, Qianfeng Yu, Jingyang Mao, Xiaoqian Peng, Tingyu Lu, Peirong Ma, Yanhui Gu
Statistical Methodology
Policy evaluation and learning with partially identified utility under truncation by death
Zi Huang and Zhichao Jiang
Policy evaluation and learning aim to assess and optimize treatment assignment rules based on individual characteristics. A fundamental challenge arises when outcomes are truncated by death, rendering conventional policy utilities undefined. To address this challenge, we study policy evaluation and learning under truncation by death within the principal stratification framework. We propose the survivor average utility and subgroup survival rates for evaluating treatment policies. Under treatment ignorability and monotonicity, the subgroup survival rates are point identified, whereas the survivor average utility is only partially identified through sharp bounds. We then develop semiparametrically efficient
estimators for the subgroup survival rates and hybrid
estimators for the survivor average utility bounds. Building on this evaluation framework, we formulate a constrained minimax optimization problem for policy learning that minimizes worst-case regret relative to a
benchmark policy while requiring the learned policy to achieve a survival rate no lower than that of the
benchmark. We show that the optimal policy admits a threshold representation based on two priority scores. We develop an estimation procedure for the optimal policy and establish regret and feasibility guarantees. Simulation studies and an application to the MIMIC-III clinical
dataset demonstrate the practical performance of the proposed methods.
Read the full abstract ▾
Artificial Intelligence
Persona and Persuasive Framing in AI Voice Agents: A $2\times2$ Field Experiment with Children
Thilo Tamme, David Steck and Anton Hantel
Conversational
agents increasingly interact with children, yet evidence on how their design shapes children's susceptibility to persuasion comes almost entirely from the lab. We report a $2\times2$ randomized field experiment embedded in a public German Santa Claus telephone hotline. Children's calls were randomly routed to one of four LLM voice
agents varying persona (Santa, high authority, vs. Helper, low authority) and framing (persuasive nudges toward prosocial wishes vs. neutral). Of 1,072 logged calls, 89 conversations (median age 6) met inclusion criteria. Persuasive framing raised the probability of a prosocial wish from 11.6% to 45.7%, robust to controls. Persona authority showed a near-zero effect: Santa did not outperform the Helper. Persona instead shaped engagement; children hung up on the Helper far more often within the first minute (65% vs. 39%). Where context already lends an
agent legitimacy, how it speaks shapes children's compliance more than who it claims to be.
Read the full abstract ▾
Machine Learning
ChartDensity-Bench: Benchmarking MLLMs for Numerical Data Reconstruction under Visual Density
Xinhe Wu and Yadong Jin
Multimodal large
language models (MLLMs) offer a promising approach for recovering numerical data from scientific charts, but their ability to reconstruct chart data from visually dense figures remains poorly understood. Existing chart understanding
benchmarks primarily evaluate question answering or chart-level
reasoning and provide limited support for evaluating structured numerical reconstruction from scientific figures. We introduce ChartDensity-Bench, a
benchmark for evaluating MLLMs on structured numerical data reconstruction from compound chart figures under controlled visual density. Built from charts paired with source-level ground-truth data, ChartDensity-Bench systematically varies the number of simultaneously presented charts ($k\in{1,3,6,9}$), enabling controlled evaluation of density-induced degradation. We further propose a multi-dimensional evaluation framework covering structural reliability, reconstruction completeness, parseability, and numerical fidelity. Experiments on five recent MLLMs show that numerical reconstruction generally degrades as visual density increases, while the magnitude of degradation varies substantially across models. Chart-level paired comparisons further show that the same source chart can incur higher reconstruction error when embedded in denser visual contexts. These findings highlight visual density as an important and previously underexplored factor in MLLM chart data reconstruction and provide a systematic
benchmark for evaluating model robustness in this setting.
Read the full abstract ▾
Sound
RAST: Resolution-Aware Privileged Structure Transfer for Low-Resolution Audio Activity Recognition
Ji Hwan Park, Gautham Krishna Gudur, Yufei Shen et al. (+2 more)
Audio is increasingly used for human activity recognition (HAR) because it captures object interactions, environmental events, and contextual cues in everyday environments. High-resolution (HR) audio provides rich acoustic information for model development but incurs substantial energy and storage costs and may expose sensitive
speech content. Low-resolution (LR) audio offers a more
privacy-preserving and resource-efficient alternative for deployment, but reduced sampling rates can remove acoustic cues essential for activity recognition, leading to significant performance degradation. We formulate this training-deployment mismatch as sensor-resolution privileged learning, in which HR audio is available during training, while inference relies exclusively on LR audio. We propose RAST, a resolution-aware transfer framework that compresses HR teacher representations by preserving
token-level information and neighborhood structure before performing localized HR-LR
alignment. Experiments on the SAMoSA and AudioIMU
datasets show that RAST consistently outperforms LR-only training and direct teacher-transfer baselines, improving LR-only recognition by up to approximately 7.8% while requiring only LR audio at inference.
Read the full abstract ▾
All authors Ji Hwan Park, Gautham Krishna Gudur, Yufei Shen, Dawei Liang, Edison Thomaz
Quantum Physics
Motzkin-Straus Optimization on an Entropy-Computing Platform
PoJen Wang, Sutapa Samanta, Yuntai Song et al. (+1 more)
We introduce a framework for combinatorial optimization using sum-constrained continuous quadratic programs solvable by QCi's Dirac-3S photonic
entropy computer. This is enabled by the Motzkin-Straus
theorem which provides a powerful bridge between discrete clique problems and optimization over the probability simplex. We demonstrate this framework's versatility by solving constraint satisfaction problems, providing extensive
benchmarks on the DIMACS suite. The Dirac-3S platform matches or outright leads two independently implemented classical baselines on more than four-fifths of the
benchmark instances, reaching the best known solution on nearly all structured graph families, even outperforming both classical solvers on several of the largest instances tested. On the other hand, well-tuned classical continuous optimizers retain an edge only on the hardest planted-clique instances. This work establishes a viable pathway for solving combinatorial optimization problems using natively analog unconventional computing platforms, while positioning
entropy computing as a competitive approach for navigating non-convex landscapes and providing rigorous baselines for an emerging computational paradigm.
Read the full abstract ▾
All authors PoJen Wang, Sutapa Samanta, Yuntai Song, Mohammad-Ali Miri
Artificial Intelligence
Action Conditioned Bisimulation For GUI Agent Memory
Hongbo Zhang, Liuyang Song, Quanquan Li et al. (+3 more)
An
agent that remembers what it did on a web page must decide when two pages count as the same.
Memories built on observation similarity merge pages that look alike but behave differently, and GUIs are full of such pages: two tabs of one widget or two rows of one menu answer the same click differently. We define the merge rule as an action-conditioned bisimulation over the empirical predictive state graph a frozen
agent fills as it acts. Two states merge only when their shared actions lead to agreeing outcomes and successor blocks under an affordance label. Observation similarity never enters the rule, and nothing is trained. It replaces the merge rule of an existing outcome-value
memory, so a closed-loop comparison isolates it. On MiniWoB++ it raises success rate over a memoryless
agent, while a control taking identical exploratory detours, the prior successor-representation merge, and the same criterion without action conditioning change nothing.
Read the full abstract ▾
All authors Hongbo Zhang, Liuyang Song, Quanquan Li, Daqian Yang, Yan Wen, Zhengtao Yao
Computer Vision
Distill the Visual Evidence, Not Just the Answer: Cross-World On-Policy Distillation for Vision-Language Models
Yuanhao Sun, Huawei Ji, Jiaxin Ding et al. (+2 more)
A central goal of
vision-
language model (VLM)
distillation is to transfer both the teacher's
language capabilities and its visual understanding. However, existing methods primarily supervise the student's output, leaving visual understanding implicit. Our analysis reveals that a student can match the teacher's answer without relying on the same visual evidence, raising the question: how can we ensure the student responds to the visual information that actually determines the answer? To this end, we propose Cross-World On-Policy
Distillation (CW-OPD), which explicitly supervises the student's response to changes in visual evidence. For each example, CW-OPD constructs two visual worlds that share the question and scene context but differ in answer-critical evidence, yielding different answers. We perform on-policy
distillation in both worlds and distill the teacher's cross-world belief transition, encouraging the student to match not only what the teacher predicts but also why its prediction changes with the evidence. A gradient analysis shows that this term is invariant to errors shared by both worlds and supplies a corrective signal invisible to endpoint matching alone. In this way, CW-OPD makes reliance on the relevant visual evidence an explicit
distillation target rather than an implicit consequence of output matching. To diagnose whether a model truly grounds its answers in visual evidence, we introduce CWBench, which measures cross-world consistency via Cross-World Pair Accuracy (CWPA). Experiments on Qwen3.5-4B show that CW-OPD outperforms the strongest baseline by 1.2 points on average, and the 4B student exceeds DeepSeek-V4.1 (552B) by 22.4 CWPA points on CWBench. Code is released in https://github.com/baokou-fw2/CWAD.
Read the full abstract ▾
All authors Yuanhao Sun, Huawei Ji, Jiaxin Ding, Luoyi Fu, Xinbing Wang
Machine Learning
Distilling Diffusion Score Discrepancy for Efficient Training Data Attribution
Shixuan Liu, Joan Serrà, Kin Wai Cheuk et al. (+6 more)
Training data attribution for
diffusion models aims to identify the training samples that influence a generated instance, but existing methods either require costly per-sample gradient computation or query-specific model optimization. Moreover, most methods attribute changes in a proxy loss rather than changes in the actual model's generative behavior. We address these limitations by formulating attribution directly with a local score discrepancy measure, which applies to any
diffusion variant (including DDPM, EDM, and flow matching), and by showing that such measure can be estimated without retraining, as a preconditioned gradient similarity. We instantiate this
estimator as Training-data Influence via score Discrepancy (TID), which uses Kronecker-factored
curvature to avoid random projections and per-sample gradient storage. We then distill TID into TIDE, a forward-only student trained online to reproduce the teacher's rankings from the
diffusion model's internal activations. Under counterfactual evaluation on CIFAR-10, ArtBench-10, and MS-COCO, TID matches or outperforms state-of-the-art approaches, while TIDE retains most of TID's accuracy at four to five orders of magnitude lower per-query cost, attributing generated samples in milliseconds and faster than the generation itself.
Read the full abstract ▾
All authors Shixuan Liu, Joan Serrà, Kin Wai Cheuk, Jinju Kim, Woosung Choi, Yukara Ikemiya, Wei-Hsiang Liao, Jiaqi W. Ma, Yuki Mitsufuji
Quantum Physics
Experimentally Testable Quantum Advantage in Shallow Circuits
Kishor Bharti and Adán Cabello
Experimental tests of shallow-
circuit quantum advantage require explicit classical bounds at finite
circuit sizes. We refine the finite-size classical soundness bound of Aasnaess's graph-distributed construction to depend linearly on the number of players. Combined with standard disjoint-player repetition, this gives a two-round test on a single processor for any finite nonlocal game with a finite-dimensional perfect
quantum strategy and classical winning probability $\gamma<1$, with arbitrarily small classical success. The method is based on playing $m$ copies with disjoint players and teleporting each player's register to a uniformly random one of $N$ sites before the questions are revealed. Each answer bit of a depth-$D$, fan-in-$K$ classical response with fixed wiring depends on at most $K^D$ question wires, making cross-player dependencies unlikely when $N$ is large. The
quantum implementation has constant depth per round and wins with certainty. A classical device wins with probability at most $\gamma^m+O(mK^D/N)$, which vanishes as $O(\log N/N)$ for $m=\lceil\log_{1/\gamma}N\rceil$ at fixed $D$ and $K$. We present an explicit proposal for an experimentally testable
quantum advantage with 99
qubits.
Read the full abstract ▾
Solar and Stellar Astrophysics
Magnetic field diagnostics of a solar active region filament
Daiki Yamasaki, Yu Wei Huang, Yuki Hashimoto et al. (+2 more)
We performed spectropolarimetric observations of an active region filament in He I 10830 angstrom and Si I 10827 angstrom lines to investigate its
magnetic field structure. We carried out full-Stokes inversions with the HAZEL code, which takes into account the Zeeman and Hanle effects. As a result, we yielded a mean field strength of 101 $\pm$ 33 G and a horizontal field nearly parallel to the filament axis, such that the distinction between the two classical normal- and reverse-polarity models becomes physically insignificant. In addition, we found Zeeman-like signatures in the linear polarization, characterized by double-peaked symmetric profiles, in some pixels of our observations. Since these profiles could not be well reproduced by modeling that included both the Zeeman and Hanle effects, we performed inversions assuming only the Zeeman effect. The inversion yielded a strong
magnetic field of approximately 500 G. However, simultaneous observations of Si I 10827 angstrom indicate a photospheric
magnetic field weaker than 100 G. Therefore, the scenario proposed by Diaz Baso et al. (2016), in which the strong field inferred from He I 10830 angstrom originates from contamination by the underlying photosphere, does not apply to our filament. The Zeeman-like profiles are preferentially found in optically thick regions ($\tau$ ~ 1.4-2.5), where the simplifying assumptions adopted in HAZEL are expected to become less reliable. Our results suggest that these profiles reveal limitations of the current inversion framework in optically thick regions and motivate future radiative-transfer modeling incorporating self-consistent radiation fields, differential illumination of the multiplet components, and possibly partial frequency redistribution.
Read the full abstract ▾
All authors Daiki Yamasaki, Yu Wei Huang, Yuki Hashimoto, Satoru UeNo, Kiyoshi Ichimoto
Differential Geometry
Steady gradient Ricci Yang-Mills solitons on 2-orbifolds
Peng Lu and Jiuru Zhou
Following the recent work of M. Womack \cite{Wo26} we consider the steady
Ricci Yang-Mills
solitons on 2d surfaces containing orbifold point $\mathbb{R}^2/\mathbb{Z}_p$. We show that there are a family of
solitons depending parameter $\lambda$ which approaches to $\mathbb{Z}_p$-quotient of Hamilton cigar
soliton as $\lambda \to (-2/p)^+$ and approaches to (after rescaling) $\mathbb{Z}_p$-quotient of round sphere as $\lambda \to \infty$. For any integer $q >p$ there is a $\lambda$ whose corresponding
soliton metric is defined on football orbifold $S_{p,q}^2$.
Read the full abstract ▾
Combinatorics
Counting hypergraphs without linear cycles of fixed length
József Balogh, Ramon I. Garcia and Abhishek Methuku
Let $C_{k}^{(r)}$ be the $r$-uniform linear cycle on $k$ hyperedges. An $r$-graph is $C_{k}^{(r)}$-free if it contains no copy of $C_{k}^{(r)}$. Let $\operatorname{ex}_{r}(n,C_{k}^{(r)})$ denote the maximum number of hyperedges in an $n$-vertex $C_{k}^{(r)}$-free $r$-graph. Balogh, Narayanan and Skokan asked whether, for every pair of integers $r,k\ge 3$, the number of $C_k^{(r)}$-free $r$-graphs on $n$ labelled vertices is \[ 2^{(1+o(1))\operatorname{ex}_{r}(n,C_{k}^{(r)})}. \] While the analogous statement is known to fail for graphs ($r=2$), by a construction of Morris and Saxton, the general question remained open for hypergraphs. Very recently, Jiang and Longbrake answered the question affirmatively when $r\geq5$. In this paper, we completely resolve the problem by proving that the answer is affirmative for every pair of integers $r,k\geq3$. When $(r,k)\neq(3,3)$, we establish balanced supersaturation results for linear cycles and combine them with the hypergraph container method. For $(r,k)=(3,3)$, we instead decompose $3$-graphs according to their pair-codegrees, encode the subhypergraph formed by the hyperedges that contain a large codegree pair as a directed graph, and apply a multicolour
entropy theorem.
Read the full abstract ▾
Signal Processing
Memory in Behavioral Models as Motion on a Slow Invariant Manifold
Nicholas B. Tufillaro
A single-tone large-signal operating point of a nonlinear two-port is a periodic orbit of a periodically forced
circuit. When the device has
memory, the Floquet exponents of that orbit separate into fast (electrical) and slow (thermal and trapping) modes, and long-term
memory is motion on the invariant
manifold attached to the slow modes. Existence and uniqueness of that
manifold follow from the parameterization method of Cabré, Fontich and de la Llave, applied to the stroboscopic map at the orbit; the
manifold is the spectral submanifold of Haller and Ponsioen, without a small-forcing parameter. The
manifold is a bundle over the circle of drive phase, its fiber dimension the number of slow exponents, and the dynamic X-parameter kernel of Verspecht et al. identifies its reduced dynamics from the transient after a step in drive amplitude. An exact reduced model has one
memory state per slow exponent, the memoryless X-parameter surface is the fixed-point family of the reduced dynamics, and the envelope-domain model is the reduced dynamics driven by the envelope. The hypotheses are verified and the
manifold constructed for a GaN HEMT compact model with thermal and trap
memory: the trap's time constant at the operating point, $14 \mu$s, is set by the linearization, not by its $6$ ms emission time, and the trap's nonlinearity confines a
polynomial representation of the
manifold to a few millivolts of drive, so trap
memory must be represented over the operating range, by tables or a fitted network, not by an expansion about the operating point.
Read the full abstract ▾
Quantum Physics
Rayleigh-Wood Threshold Controls Nonlinear Photon Correlations in Atomic Arrays
Zai-Cheng Xiao and Li-Ping Yang
Two-dimensional atomic arrays provide versatile free-space
quantum optical interfaces by coupling
photons to collective
lattice modes. Here we show that the Rayleigh–Wood anomaly provides a sharp control mechanism for both linear and nonlinear
photon scattering from a single atomic layer. As the
lattice spacing crosses the diffraction threshold, newly opened radiative channels rapidly broaden the dominant collective modes, converting specular reflection into diffuse off-axis scattering. In transmission, momentum-space Fano interference between the incident and collectively scattered fields produces single-
photon intensity zeros that strongly reshape the momentum-space correlation pattern. An integrated nonlinear contrast reveals a pronounced channel asymmetry: upon opening the Rayleigh channels, the connected two-
photon contribution drops by several orders of magnitude relative to the factorized background in transmission, while remaining close to unity in reflection. These results establish radiative diffraction thresholds as a control principle for
quantum nonlinear optics in single-layer atomic arrays.
Read the full abstract ▾
Instrumentation and Detectors
Kilopixel Performance of the Kinetic Inductance Detectors for the Terahertz Intensity Mapper
Justin S. Bracks, Reinier M. J. Janssen, Steve Hailey-Dunsheath et al. (+32 more)
We characterize a flight-grade array of lumped-element kinetic inductance detectors (LEKIDs) developed for the long-wavelength module of the Terahertz Intensity Mapper (TIM). From an 864-pixel science array, we select 490 well-isolated resonators spanning the focal plane and readout band and measure their thermal response, optical responsivity, and noise using a ZCU111-based multitone readout system intended for flight. The thermal and optical response of the detector population is consistent with previous single-pixel measurements and can be described by Mattis-Bardeen theory under the influence of a change in Cooper pair or quasi-particle density. Under increasing blackbody loading, the measured noise transitions from a thermal generation-recombination-dominated floor to
photon-noise-limited scaling, with the majority of detectors achieving
photon-noise-limited operation by approximately 400 fW incident power, with a median detector noise-limited NEP of $NEP_{det} = 1.1\times10^{-17} W \sqrt{Hz}$. The measured
photon-noise scaling implies a median optical efficiency of approximately 0.67, indicating additional unknown loss sources between the detectors and blackbody radiator, tentatively attributed to losses in the waveguide. These results demonstrate that the TIM LEKID architecture retains the required sensitivity when scaled to kilopixel-class arrays; the principal remaining challenges are array-level tone optimization, resonator tracking, and identification of frequency-domain collisions.
Read the full abstract ▾
All authors Justin S. Bracks, Reinier M. J. Janssen, Steve Hailey-Dunsheath, Talia Saeid, Bruce Bumble, Logan Foote, Elijah Kane, Lun-Jun Liu, Charles M. Bradford, Shubh Agrawal, James E. Aguirre, Hrushi Athreya, Brockton S. Brendal, Karia Dibert, Peter Dow, Jeffrey P. Filippini, Jianyang Fu, Karolina G. Garcia, Christopher E. Groppi, Bradley Johnson, Dylan Joralmon, Wooseok Kang, Garrett K. Keating, Ryan P. Keenan, Ian N. Lowe, Alex Manduca, Aashrita Mangu, Daniel P. Marrone, Philip D. Mauskopf, Evan C. Mayer, Sydnee ODonnell, Simon Tartakovsky, Mathilde Van Cuyck, Joaquin D. Vieira, Jessica A. Zebrowski
Machine Learning
Learning Under Forgetting: Statistical Support-Selective Retention in Stochastic Training Dynamics
Fujie Gao, Zuyue Zhang and Gang Sun
Prior work has shown that neural networks exhibit implicit biases toward low-complexity structure (e.g., spectral bias), memorization dynamics, and compression-like effects during training, but a unified dynamical account of selective retention remains incomplete. We propose Repeated Reinforcement with Persistent Forgetting (RPF) dynamics, a minimal framework in which repeated exposure reinforces patterns and structures that recur in the data, while persistent forgetting attenuates learned information. This view treats forgetting not merely as a failure mode, but as a selection mechanism. We build the theory in three successive layers. First, in an independent-feature model, we derive an exposure-selective survival law and a support-dependent retention boundary characterizing which patterns persist under forgetting. Second, in a shared-parameter model, we show that forgetting induces spectral filtering over covariance modes, preserving strongly supported shared components while suppressing weak ones. Third, under small-step and norm/coding approximations, we show how RPF dynamics induce an implicit trade-off between data fitting and the cost of stored information, yielding Minimum Description Length (MDL)-like compression. Controlled experiments provide evidence for this reinforcement–forgetting selection mechanism in scalar
memories and a nonlinear shared network. Joint reinforcement and attenuation interventions shift conditional retention, while matched exposure counts reveal forgetting-dependent effects of reinforcement timing and changes in the composition of the retained set. Together, these results show that repeated reinforcement and persistent forgetting jointly provide a controllable source of inductive bias beyond neural architecture and scale.
Read the full abstract ▾
Machine Learning
dattri-LLM: A Unified and Efficient Library for Training Data Attribution at LLM Scale
Shixuan Liu, Tongli Zhou, Junwei Deng et al. (+2 more)
Training data attribution (TDA) estimates the contribution of individual training examples to model outputs. Most scalable TDA methods rely on per-example gradients, whose computation and use at LLM scale pose challenges in efficiency, compatibility, and extensibility. We introduce dattri-LLM, a TDA library that makes gradient-based attribution more practical at scale. For efficiency, dattri-LLM uses compact gradient representations and dynamically routes gradient operations based on a cost model. For compatibility, its capture mechanism collects per-example gradients from existing training loops that call backward(), without requiring changes to the loop or its configuration. This includes distributed training with DDP and FSDP and pipelines built with HuggingFace
Transformers, TRL, and OLMo. For extensibility, dattri-LLM exposes reusable gradient operations and training-time callbacks for implementing attribution methods and applications. These interfaces support a variety of attribution methods, including gradient similarity,
curvature-based influence, and trajectory-based methods, as well as applications that act on gradients during training, such as online data selection. On the same hardware and workload, dattri-LLM achieves 3.2x the throughput of the fastest competing library on average, scales multiple attribution methods to 110B-parameter models across four H200 GPUs, and offers superior attribution fidelity-cost trade-offs across a range of models with different model families and scales. The source code of dattri-LLM is available at https://github.com/TRAIS-Lab/dattri-llm.
Read the full abstract ▾
All authors Shixuan Liu, Tongli Zhou, Junwei Deng, Pingbang Hu, Jiaqi W. Ma
Artificial Intelligence
PathAnchor: Path-Structured Evidence for Scientific Agents
Qiuhui Chen, Jiafan Lu, Shuaimin Tang et al. (+5 more)
Scientific
agents can retrieve relevant passages yet still lose functional order, mix evidence across sources, or state conclusions that exceed the retrieved record. We introduce PathAnchor, a bounded scientific
reasoning system built on path-structured evidence workspaces. Instead of treating passages or extracted concepts as independent units, the system retrieves source-linked Material-Sensor-Signal-System trajectories that preserve role, direction, and the evidence supporting each transition. A controller uses three read-only tools to search paper-specific trajectories, trace paths across candidate sources, and open exact evidence before producing a claim-cited answer and an explicit evidence boundary. On 120 single- and cross-paper flexible-sensor questions, PathAnchor scores 82.6% and leads six evaluated systems. Under a matched controller, corpus, and six-call budget, replacing unordered concept graphs with path-structured records raises source recall from 61.3% to 82.9%, increases answers whose claims all cite opened evidence from 69.2% to 90.0%, and reduces tool calls. These results show that evidence organization affects
retrieval and citation completeness under fixed
agent resources.
Read the full abstract ▾
All authors Qiuhui Chen, Jiafan Lu, Shuaimin Tang, Tao Dai, Suyuan Wang, Chenrui Ji, Zhenglei Zhou, Weimin Zhong
Computational Finance
Multiperiod bond portfolio optimization with transaction costs using a Markov Decision process
Balaji Ramachandran, Srikanth Iyer and Shashi Jain
Bank treasury portfolios must balance yield, liquidity, and interest-rate risk across bonds of different maturities. Static allocation rules are ill-suited to this task: portfolios concentrated in long-duration securities with no dynamic adjust- ment mechanism can accumulate large mark-to-
market losses and liquidity stress under rising interest rates, as illustrated by the failure of Silicon Valley Bank in 2023. We develop a tractable simulation-based framework for multi-period bond port- folio optimization under interest-rate risk and proportional transaction costs. Yield-curve dynamics are modeled using the Dynamic Nelson-Siegel parameter- ization with Vector Autoregressive factor dynamics, from which we construct a time-inhomogeneous discrete-state Markov chain approximating the joint yield process across bond maturities. This chain forms the state space of a finite- horizon Markov Decision Process in which the investor maximizes expected terminal wealth subject to proportional rebalancing costs. The optimal portfolio policy is obtained by backward induction. We also quantify the approximation error introduced by truncating the transition kernel, and show that it leaves mean terminal wealth almost unchanged while substantially distorting drawdown and tail statistics.
Read the full abstract ▾
Machine Learning
Lasting Effects of Abstract Pretraining Beyond Perplexity
Zachary Shinnick, Hemanth Saratchandran, Damien Teney et al. (+1 more)
Language models are typically pretrained from random initialization. Recent work challenges this convention, showing that a brief warm-up on abstract, algorithmically generated data can provide a better starting point for subsequent learning of natural
language. In this paper, we show that in small
language models, such a warm-up improves specific capabilities that are not reflected in
language-modeling perplexity. Our warm-up uses an abstract stack-manipulation task that requires compositional and state-tracking capabilities. Allocating as little as 1% of pretraining
tokens to this data improves multi-hop question answering by up to 3.9 F1 points on MUSIQUE, with additional gains on HOTPOTQA and 2WIKIMULTIHOPQA despite comparable
language-modeling perplexity. Controlled experiments show that the warm-up substantially accelerates the acquisition of deeper
reasoning chains. We also explore what drives this transfer. First, the structure of the data matters: replacing the stack task with a queue fails to produce the same gains. Second, the gains are specific: performance improves on sequential
reasoning chains, with no consistent benefit on tasks that combine or compare independent facts. Third, timing matters: mixing abstract data with natural
language is far less effective than an initial dedicated phase, and exposure after pretraining completely removes the benefits. The early advantage persists through billions of subsequent
language tokens. These results show that early abstract training can reliably shape the capabilities
language models later acquire.
Read the full abstract ▾
All authors Zachary Shinnick, Hemanth Saratchandran, Damien Teney, Anton van den Hengel
Atmospheric and Oceanic Physics
Long-term Variations of Cyclone-Influenced Exchange Flows in the Coastal Arctic in 2000-2025
Chunyan Li
Summertime synoptic weather regulates Arctic land-ocean exchange through meteorological tides during the ice-free period, when the predominantly easterly regime of the Beaufort High is punctuated by episodic Arctic cyclones, changing coastal winds to westerlies. We analyze wind-driven exchange flows in a northern Alaska lagoon using a 26-year reconstruction (2000-2025) validated by new in situ observations and FVCOM simulations. Phase-space EOF analysis of ERA5 data links deep cyclonic anomalies to 89% of Extreme Eastward Throughflow (EET) events (24 of 27). Coastal easterly occurrence frequencies show a consistent decrease (up to 38.3% at > 10 m/s) while low-to-moderate westerly occurrence frequencies increase by up to 30.9% but decrease by 40.6% at extreme westerly wind speeds (>10 m/s). Offshore ERA5 wind occurrence frequencies increased, with extreme offshore events (> 10 m/s) increasing by 82.4%. Local ASOS observations reveal a near-monotonic ~88% reduction in atmospheric forcing variability (East-West Wind Excursion variability drops from 2.98X10^4 to 0.35X10^4 km). The In-Out Flow Excursion variability declines steadily by 70% from 1.40X10^3 to <0.40X10^3 km, while the positive (outward) flow ratio variability contracts steadily from 7.4% to below 2.2%. Extreme Westward Throughflow (EWT) events were infrequent, but were concentrated in the most recent period, with five of six events occurring in 2022-2024 and one in 2013. These results indicate that the summertime coastal regime remains predominantly easterly, but exhibits progressively lower variability, while the wind distribution shifts toward more frequent low-to-moderate westerlies. This
evolution of weather-forcing state may impact flushing and biogeochemical connectivity of Arctic lagoon ecosystems.
Read the full abstract ▾
Software Engineering
Adaptive-GEPA: Make Your Harness Fit Heterogeneous Requests
Tianyu Chen, Yasi Zhang, Ruiyi Wang et al. (+3 more)
Reflective optimizers such as GEPA improve
language model
prompts from execution traces and evaluator feedback; full-program extensions can also rewrite tools and control flow. In practice, a user hands the same endpoint heterogeneous requests whose effective solutions require different tools,
reasoning modes, and control flow. Optimizing one shared program leaves this division of work implicit in source-code search, while optimizing a separate program per request family fixes it beforehand. We introduce Adaptive-GEPA, which learns both how to divide requests and how to solve them. It evolves a router and a library of specialist programs under one search budget. The router's instructions, each specialist's description, and its program code are plain, human-readable text, edited from feedback. To combine branches, it aligns specialists by the requests they handle and inherits descriptions together with programs. On a fixed mixture of four task families, the reported Qwen3-8B run evolves four experts without supplying family labels to the router or reflection model; its routing matches the task partition on all 651 test requests. Its family-mean test score (x100) rises from 52.6 to 70.6, compared with 62.5 for GEPA's full-program adapter and 54.0 for GRPO at a nominal budget of 18,000 scored calls. These counts do not equate total compute. Figure 1 summarizes the learning curves, final test scores, and routing agreement.
Read the full abstract ▾
All authors Tianyu Chen, Yasi Zhang, Ruiyi Wang, Xinran Zhao, Taoran Li, Mingyuan Zhou
Multiagent Systems
Where Do Multi-Agent Systems Fail? Evidence-Grounded Diagnosis of Collective Mechanisms
Zhengye Han
When a multi-
agent system answers correctly, it is tempting to conclude that its
agents shared, checked, and used information as intended. Yet a system can break one of its collective mechanisms, the rules that govern how
agents route, admit, store, and act on shared information, and still return the right answer, while a wrong answer rarely reveals which mechanism failed. We ask what evidence from an execution is sufficient to conclude that a particular mechanism was violated. Our answer is a diagnostic contract, which separates what counts as a violation from which execution records can establish one, and concludes that a violation is supported, ruled out, or unknown; removing records can make this conclusion unknown but never reverse it. We test contracts for four mechanisms by replaying executions from the step where a mechanism acts, once unchanged, once with the mechanism broken, and once with it restored. Broken mechanisms often left the answer correct. An LLM diagnoser detected many more violations from internal records than from public outputs, yet with identical records a generic
prompt often claimed certainty the records did not support, which
prompts stating the contracts largely avoided. The contracts also applied, in narrow form, to mechanisms in independently developed systems, but a diagnostic behavior that was nearly perfect on our
benchmark degraded on an independently developed workflow. A correct outcome is therefore no substitute for records of how collective mechanisms operated, and agreement on one
benchmark does not show that a diagnoser transfers to another system.
Read the full abstract ▾
Robotics
HALO: Heterogeneous Allocation Via Localized Observations for the Vehicle Routing Problem
Andrew Meighan, Hyungsub Kim and Or Dantsker
Scalable robotic fleets have become increasingly popular for various applications such as package delivery, warehouse management, and military operations. Prior fleet control
algorithms solve centralized routing problems with up to $1{,}000$ tasks in controlled environments, yet they fail to consider realistic constraints such as limited observation and communication ranges typical of decentralized fleets. Thus, deploying existing fleet control
algorithms into real-world settings is currently infeasible. To tackle this, we propose Heterogeneous Allocation via Localized Observations (
HALO) to solve the Vehicle Routing Problem (VRP).
HALO is a hybrid method that splits the VRP into allocation and routing portions to provide onboard, real-time solutions to
robots in dynamic environments. During the allocation phase,
HALO utilizes a heterogeneous graph neural network framework with unique message passing layers to explicitly separate the learning of spatial distributions and task-to-
robot compatibility. Evaluation results on a partially observable, online variant of the VRP show
HALO significantly outperforms the heuristic baseline while maintaining similar solution quality to an all-knowing offline variant of
HALO. While
HALO is explicitly designed for partially observable environments, it imposes no strict upper bound on the observation space allowing us to test
HALO on the traditional static, single-depot VRP. Here,
HALO outperforms state-of-the-art architectures strictly optimized for the static variant of the VRP by up to $14.06\%$. Throughout all testing, this framework maintains the quickest execution times which emphasizes its potential for large-scale, real-time deployment.
Read the full abstract ▾
Statistical Methodology
BOP2-ENR: a Bayesian optimal phase II design for adaptive enrichment with family-wise error rate control and a bound on joint-claim power
Kentaro Takeda, Belay B. Yimer and Masahiro Kojima
The
Bayesian optimal phase II (BOP2) design accommodates simple and complex endpoints but does not address biomarker-guided enrichment or multiplicity control across pooled and subgroup claims. We propose BOP2-ENR, which monitors biomarker-negative (NG) and biomarker-positive (PG) cohorts separately, permits interim enrichment to PG, and calibrates pooled and PG-restricted claims. Strong family-wise error rate (FWER) control is defined over a prespecified monotone-activity parameter space. Under the mixed configuration in which NG is inactive and PG is active, the probability of a false joint claim is bounded by the probability that NG survives futility monitoring. This exact single-cohort bound does not involve the final efficacy cutoff and also yields a ceiling on joint-claim power. For efficacy-toxicity monitoring, subgroup inactivity is defined by a union null and safety is established separately within each subgroup. Across the scenarios considered and binary, co-primary and efficacy-toxicity endpoints, all selected designs met the prespecified exact-bound and Monte Carlo verification criteria; in sensitivity analyses over odds ratios between component outcomes assumed independent in calibration, point estimates of the error-control quantities remained below the nominal level, although power was association-sensitive. The bound was nearly attained as PG efficacy approached the boundary of the nuisance space, whereas disabling enrichment produced mixed-configuration error rates near one. BOP2-ENR therefore extends the BOP2 framework to adaptive enrichment while clarifying that the NG futility gate governs the mixed-configuration component of strong error control and the attainable power of a joint claim.
Read the full abstract ▾