Quantitative Methods
Toward Controlling Biology with Language:Offline Learning of Prompt-Conditioned Interventions for Cells, Organoids, and Biobots
Nam H. Le, Douglas Blackiston, Michael Levin et al. (+1 more)
Artificial intelligence increasingly serves as a natural-
language interface to complex technical systems, letting people accomplish sophisticated tasks by describing what they want rather than specifying how to do it. Extending this interface to living systems is harder: unlike code or images, a biological intervention has no closed-form linguistic meaning, and the paired
language-intervention-outcome data needed to learn such a mapping is expensive to collect, since each example requires its own wet-lab experiment. One way around this is to treat an existing archive of interventions and their already-observed outcomes as a fixed, offline
dataset, and use a
vision-
language model to judge, without any new experiments, whether an archived outcome matches a natural-
language description. But whether that judgment is reliable enough to train a
language-to-intervention mapping on – without new experiments and without human validation – has remained unclear. Here we show that a natural-
language interface for a living organism – a xenobot, a synthetic multicellular construct with no nervous system – can be learned entirely offline this way, using a
vision-
language model's own judgment as the sole training reward: an instruction is mapped to the intervention already on record as producing the described behavior. This mapping generalizes to entirely new instructions, evaluated against archive data withheld from training (80.0% held-out accuracy vs a $66.7% chance baseline, matching a network trained directly on ground-truth labels).
Read the full abstract ▾
All authors Nam H. Le, Douglas Blackiston, Michael Levin, Josh Bongard
Data Analysis and Statistics
Developing Multi-Dimensional, Sequential sWeighting Routines for Reaction Selection with Emphasis on Kaon Identification at CLAS12
A. Acar, M. Bashkanov, D.P. Watts et al. (+1 more)
In
nuclear and particle physics experiments, event selection is often a multidimensional classification problem which involves several correlated observables. The widely used sWeight technique provides a statistically rigorous means of separating signal and background using discriminating variables, but its conventional formulation is generally applied to a single discriminating variable or a simultaneous multidimensional fit. In this paper we present a novel extension of the widely used sWeight technique, in which an arbitrary number of discriminating variables can be used to construct a multidimensional sequential framework. This method allows the extraction of signal distributions in controlled variables in a rigorous way. Particular
attention is given to the treatment of signal and background correlations, as well as the propagation of statistical uncertainties through successive weighting stages. The method yields high-purity signal samples even in the presence of large and correlated backgrounds. As a
benchmark application, the validity of this technique is demonstrated on kaon identification in a
dataset collected with the CLAS12 detector at the Thomas Jefferson National Laboratory for cascade
baryon searches. The final workflow is shown to be robust and consistent across different run periods and beam/detector conditions, demonstrating minimal systematical uncertainties associated with the method. Statistical uncertainties are proved to be evaluable using a bootstrap-based procedure. The method provides a general framework for multidimensional event weighting and is intended for application in complex analyses requiring robust signal selection in the presence of complex correlated backgrounds.
Read the full abstract ▾
All authors A. Acar, M. Bashkanov, D.P. Watts, M. Nicol
Instrumentation and Methods for Astrophysics
A Missing Latent, Not a Missing Simulator: Radius-Augmented Inference for Real JWST Retrieval
Angshuman Chakravertty and P V V Raj
Amortized simulation-based inference (SBI), which is trained on radiative-transfer simulators, recovers
exoplanet atmospheres accurately on synthetic James Webb Space
Telescope (JWST) spectra but collapses when it comes to real reduced spectra. The flow posterior collapsed on real WASP-39b (importance-sampling effective sample size (ESS) = 1, best-fit ḩi}2/N = 301), and such a failure is usually blamed on missing the forward-model physics, but ruling these levers out with nested sampling first (a temperature gradient, SO2 opacity, and a high-fidelity opacity set) leaves the fit unchanged, meaning the collapse is not from the simulator misspecification but instead from a missing latent, the
planet radius. To fix this, we build MIRAGE, a radius-augmented flow-matching posterior calibrated against an independent nested-sampling reference with importance sampling and an optimal-transport map, which yields a physical and literature-consistent
retrieval of real WASP-39b (with ḩi}2/N from 301 to 0.06). This same method transfers unchanged across two instruments and three real JWST targets, including one spectrum self-reduced end-to-end from raw Mikulski Archive for Space
Telescopes(MAST) data. The lesson is cross-domain, as a latent the simulator encodes but the inference omits can masquerade as misspecification.
Read the full abstract ▾
Number Theory
Multiplicative recurrence of Piatetski–Shapiro and nil-Bohr sets
Sun-Kai Leung
Adapting the circle method developed by Frantzikinakis–Klurman–Moreira, we establish a sufficient criterion for a set of natural numbers to be multiplicatively recurrent. We apply this criterion to Piatetski–Shapiro sets with small exponents and to regular irrational nil-Bohr sets. We also study the value distribution of unimodular completely multiplicative functions along these sets.
Read the full abstract ▾
Applied Physics
Finite Element Simulation of Microwave Technologies for PowderBased Volumetric Additive Manufacturing Processes
Brendan Mackey, Levi Smith and Bosco Yu
Volumetric additive manufacturing (VAM) is a promising field of advanced manufacturing which reduces production times of conventional 3D printing while retaining application to complex geometries. Current VAM systems are limited to photo-polymerizing resins, which reduces the general applicability across industry. This work presents exploratory finite element simulations of microwave-based VAM systems compatible with ceramic, metal, or composite powder feedstock. The simulations explore three devices – a linear resonator, a cross interferometer, and a boundary heater – each with potential uses in manufacturing processes for one or multiple of the considered feedstock materials. We find that with precise control of input frequency, power, and waveguide transverse modes, thermal
lattices can be engineered to selectively heat regions of the feedstock, allowing direct application to VAM. Resultant
lattices have cellular dimensions on the order of 1 mm, and require input powers on the order of 100 kW. With advances in microwave technologies allowing for this precision, such devices could open pathways to VAM with materials currently outside the scope of optical methods.
Read the full abstract ▾
Quantum Physics
Neural Fourier Surrogates for Data Reuploading Quantum Neural Networks
Oliver Knitter, Jonathan Mei, Sang Hyub Kim et al. (+3 more)
For
quantum machine learning, the exact boundary between classical and
quantum advantage is still poorly understood. Direct comparison between
quantum neural networks (QNNs) and existing classical models, which encompass fundamentally different function classes, often fails to provide broader insight into the difference between the two. Inspired by the techniques of Neural
Quantum States and Random
Fourier Features, this work introduces Neural
Fourier Surrogates (NFS), a
stochastic classical neural network architecture for efficiently learning coefficients over the same finite
Fourier series support as
quantum neural networks. Testing on a selection of tabular
benchmark datasets, we find that NFS is an effective classifier architecture broadly competitive with established classical baselines, including a comparable Random
Fourier Features model, and possessing comparable performance to data-reuploading QNNs; combined with additional analysis comparing the learned
Fourier spectra of QNNs and NFS on synthetic data, these results establish NFS as a natural classical baseline for evaluating QNN performance.
Read the full abstract ▾
All authors Oliver Knitter, Jonathan Mei, Sang Hyub Kim, Chi Chen, Masako Yamada, Martin Roetteler
Computation and Language
Contextual trajectory and incremental contextual displacement: Towards using LLMs to understand dynamic, utterance-specific meaning construction
Grayson Wycliffe Storer and Julia Witte Zimmerman
Transformer-based large
language models (LLMs) such as RoBERTa represent text using contextual word
embeddings (CWEs), which alter the
embeddings associated with each
token based on surrounding context. We construct
token-wise incremental trajectories by repeatedly recomputing a
token's CWE as successive words are added to a sentence, yielding a representation of how contextualized
embeddings evolve as the utterance unfolds. We evaluate this approach using garden-path sentences as a test case with characteristic features.
Token-wise trajectories reproduce known features of garden-path processing, including disruption around the critical region, and reliably distinguish garden-path sentences from matched disambiguated controls. We introduce several metrics for quantifying representational displacement across contextual increments and show that trajectory information can be highly predictive of sentence type. We find that ambiguity-related information is recoverable not only from the sentence-level CLS representation but also from ordinary vocabulary
tokens, suggesting that utterance-level information is distributed across multiple representational scales. In exploratory analyses, we find qualitatively similar trajectory structures in other ambiguity- and misdirection-related linguistic phenomena. Together, these results establish
token-wise incremental trajectories as a promising framework for studying utterance-specific meaning construction using LLMs.
Read the full abstract ▾
Cryptography and Security
Do Defenses Against LLM Extraction Work Across Attacks? A Lifecycle Benchmark of Black-Box Model Extraction
Shuze Liu, Kaixiang Zhao, Runyang Xu et al. (+4 more)
Large
language models (LLMs) deployed through text-only APIs face model extraction risks, as adversaries can collect their responses to train surrogates that reproduce their capabilities. While prior work has developed diverse attacks and defenses, evaluations remain fragmented across access assumptions, model configurations, query budgets, and security objectives, limiting comparability across methods. To address this gap, we introduce a unified
benchmark covering six extraction attacks, ten defenses, and two adaptive attacks that paraphrase or back-translate protected responses before surrogate training. The
benchmark controls model configurations, query data, budgets, and held-out evaluation conditions within each comparison while preserving attack-specific querying and training procedures. We measure surrogate capability, fidelity to the victim, output quality using Rep-4, and query-budget sensitivity; defenses use their own security metrics paired with surrogate performance. For the adaptive attacks, we jointly measure provenance-detector scores and the capability and fidelity of surrogates trained on rewritten responses. The
benchmark thus provides a reproducible basis for comparing extraction methods and their interactions with defenses under text-only access. Code and artifacts are available at https://github.com/sliu11-byte/MEA-Bench.
Read the full abstract ▾
All authors Shuze Liu, Kaixiang Zhao, Runyang Xu, Jingzhi Chen, Nathan Wu, Yu Wang, Yushun Dong
Machine Learning
SHARPO: Segment-Level Credit Assignment for Agentic Reinforcement Learning
Xinchen Du, Zhengze Zhou, Wenhui Zhu et al. (+4 more)
Agentic reinforcement learning (RL) trains a large
language model (LLM) to act over long, multi-step interactions. However, a single localized error can cause task failure, while trajectory-level rewards provide limited guidance for assigning credit to individual decisions. To address this limitation, we introduce Segment-level Hindsight Advantage Reweighting for Policy Optimization (SHARPO), a credit-assignment mechanism that refines Group Relative Policy Optimization (GRPO) at the level of environment-facing segments. Inspired by the existing on-policy self-
distillation (OPSD) method, SHARPO computes teacher-student log-probability gaps within each segment and uses the resulting signal to compute a bounded multiplier on the GRPO advantage. This multiplier is shared by all
tokens within the segment, allowing credit to vary across different segments. With Qwen2.5-7B-Instruct, SHARPO outperforms existing baselines on the ALFWorld and WebShop
benchmarks, including GRPO, SDAR, RLSD, and StepOPSD.
Read the full abstract ▾
All authors Xinchen Du, Zhengze Zhou, Wenhui Zhu, Han Yu, Sen Na, Rohit Jain, Alborz Geramifard
Computational Complexity
A Degree–Size Relation for Resolution over Polynomials
Shuo Pang
For every constant-width CNF, we show that linear degree in
polynomial calculus (PC) implies exponential size in resolution over constant-degree
polynomials, over the same
prime field. Applications include exponential lower bounds for CNFs in $\operatorname{Res}(\operatorname{PC}_r/\mathbb{F}_p)$ and hence in $\operatorname{Res}(\oplus_p)$, separations between different
moduli, improved lower bounds for $\operatorname{Res}(k)$ up to $k=\varepsilon\log n$, proof-search consequences, and an implication of super-
polynomial $AC^0[p]$-Frege bounds from very strong PC degree lower bounds. The proof uses the common-multiplier idea isolated from Braun [arXiv:2609.23015] to construct a Razborov–Smolensky approximation that preserves inferences, without introducing extension variables. The approximation errors are measured by ranks of the multiplication maps induced by the error-witness
polynomials, modulo bounded-degree PC consequences.
Read the full abstract ▾
High Energy Astrophysical Phenomena
Optimal Pulsar Timing Array Strategies with the Deep Synoptic Array
Tyler Cohen, Paul B. Demorest, Jeremy G. Baier et al. (+50 more)
Pulsar timing array (PTA) experiments have seen evidence for a nanohertz-frequency gravitational wave background (GWB) through coordinated
radio timing observations of millisecond
pulsars. The significance of this evidence is expected to grow and the GWB's progenitor(s) uncovered with continued observations, especially as the next generation of
radio telescopes comes online. We present a framework for simulating timing observations of a set of
pulsars with a set of
telescopes to predict the per-
pulsar optimal instrument, observing frequency, and integration time. We apply this methodology to the NANOGrav source list of 85
pulsars observed with the Green Bank
Telescope, Very Large Array, Canadian Hydrogen Intensity Mapping Experiment (CHIME), and the Deep Synoptic Array (DSA), a
radio dish array under construction in Nevada. We discuss the dominant noise contributions to the timing precision of these
pulsars and show that most are not dominated by intrinsic pulse phase jitter noise. We also determine which
pulsars should be timed with DSA and CHIME or only DSA and show that an ab initio DSA PTA is more sensitive than one which uses current instruments. Finally, we solve for the optimal integration time per
pulsar, subject to a fixed time budget, that maximizes the GWB signal-to-noise. We show that there is little benefit to optimizing over integration time for these
pulsars observed with DSA. Through Monte Carlo realizations of the PTA, we show that the GWB sensitivity improves by $\sim 5\%$ for time-optimized observations compared to the equal-time-per-source scenario with an early-science, fixed source list.
Read the full abstract ▾
All authors Tyler Cohen, Paul B. Demorest, Jeremy G. Baier, Ryan S. Lynch, Scott M. Ransom, Michael T. Lam, Gabriella Agazie, Akash Anumarlapudi, Anne M. Archibald, Zaven Arzoumanian, Paul T. Baker, Ava L. Battaglia, Paul R. Brook, Anabelle L. Brueggemann, H. Thankful Cromartie, Kathryn Crowter, Megan E. DeCesar, Timothy Dolch, Elizabeth C. Ferrara, William Fiore, Emmanuel Fonseca, Gabriel E. Freedman, Nate Garver-Daniels, Peter A. Gentile, Joseph Glaser, Deborah C. Good, Jeffrey S. Hazboun, Ross J. Jennings, Megan L. Jones, David L. Kaplan, Matthew Kerr, Duncan R. Lorimer, Jing Luo, Hugo E. Marquis, Alexander McEwen, Maura A. McLaughlin, Natasha McMann, Bradley W. Meyers, Cherry Ng, David J. Nice, Timothy T. Pennucci, Benetge B. P. Perera, Nihan S. Pol, Henri A. Radovan, Paul S. Ray, Ann Schmiedekamp, Carl Schmiedekamp, Brent J. Shapiro-Albert, Ingrid H. Stairs, Kevin Stovall, Abhimanyu Susobhanan, Joseph K. Swiggum, Haley M. Wahl
Machine Learning
TrueMuse: A Benchmark for Data Attribution in Text-to-Music Models
Jiawei Yu and Jian Liu
Text-to-music generation models are trained on massive music collections, creating a growing need for data attribution methods that can quantify the contribution of individual training samples. However, existing attribution methods are difficult to rigorously evaluate due to the lack of reliable ground truth, making it challenging to reliably assess their actual effectiveness. To address this gap, we introduce TrueMuse, a controlled
dataset and
benchmark for text-to-music data attribution. TrueMuse is constructed by fine-tuning three
diffusion-based text-to-music models on carefully curated attribution samples, whose known inclusion in fine-tuning provides controlled attribution targets for evaluation. The
benchmark covers four attribution settings, spanning melodic structure, timbral characteristics, artist-level stylistic signatures, and genre-level shared patterns, and includes 133 attributes, 648 fine-tuned models, and 95,456 generated samples across two
prompt types. Using TrueMuse, we systematically evaluate existing black-box attribution methods along four dimensions: fine-tuning improvement,
prompt-type difficulty, multi-task training, and fine-tuning data size. Our results show that attribution remains challenging, with existing methods exhibiting substantial variation across evaluation settings, highlighting the need for more reliable and generalizable attribution methods for text-to-music generation.
Read the full abstract ▾
Artificial Intelligence
Kepler: Auditable World Models for ARC-AGI-3
Wensen Wu
ARC-AGI-3 evaluates
agents in interactive environments whose rules and objectives must be inferred from observation. We present Kepler, an open-source harness that represents hypotheses as executable world models and validates them through retrospective transition checks and conditional prediction checks. Under one frozen Claude Opus 5 configuration, Kepler obtained a server-verified 100.00 RHAE on all 25 public games, with no per-game model selection or score-conditioned reruns. On 181 of 183 completed levels, the final Opus attempt used no more actions than the corresponding median-human baseline. The retained board runs used 8,256 environment actions, of which 7,292 occurred in scored levels. Retained local provider-session records yield 858.0 million
tokens, 97.37% cache reads, and a \$777.72 cost at September 1, 2026 API list-equivalent rates. We also report three evaluation failures: source-code leakage that produced an invalid perfect run,
agents reconstructing a removed harness in a control condition, and autonomous repair masking a broken planner. A single-game observation case study showed that animation frames contained task-relevant information absent from settled text grids. Across the final Claude Opus 5 and GPT-5.6 Sol boards, 48 of 50 game-model
cells reached 100. These results indicate that public-set score alone has limited discriminative value and motivate first-attempt, cost-conditioned, and verification-aware reporting.
Read the full abstract ▾
Computation and Language
VERITYGATE: A Four-Gate Schema-Level Faithfulness Framework and Paired Benchmark for Grounded LLM Narrations over Structured Evidence
Sachin Gupta
Fluent LLM explanations may not follow the evidence from a structured system. We present VERITYGATE, a four-gate checker for declared evidence IDs, entities, numbers, and claim types. It checks a fixed schema; it does not verify every fact in the prose. At r=0 and r=1, we test 900 instances per setting (450 grounded-ungrounded pairs) with GPT-4o-mini, Llama-3.3-70B, and Claude Sonnet 4.6. Under this schema-level contract and before repair, 80.3% of mini claims and 47.9% of Sonnet claims fail. These are verifier rejection rates, not prose-
hallucination rates. One repair pass raises claim survival from 19.7% to 28.0% for mini and from 52.1% to 54.3% for Sonnet. Verified claims per example change by +0.14 for mini, -0.71 for Llama, and -0.47 for Sonnet, so survival and output volume must be reported together. A second Sonnet pass gives no clear gain. At r=1, Gate 4 covers 97.0%, 98.7%, and 100% of failing claims for mini, Llama, and Sonnet. Small human studies support the rules but show gaps between schema checks and correct prose. A domain-specific GPT-4o judge test shows an order effect, so it is only a usefulness check. We release the code and data.
Read the full abstract ▾
Quantum Physics
Remote state preparation of a single-spin state via hybrid spin-photon entanglement
Yunge Jiang, Yehan Yu, Lili Song et al. (+3 more)
Remote state preparation is a fundamental
quantum-information protocol that exploits prior knowledge of a target state to reduce the communication resources required for
quantum-state distribution. Here, we propose a remote-state-preparation protocol for a stationary electron-
spin qubit using hybrid
entanglement between the electron
spin and a coherent-state light pulse, generated by
spin-dependent reflection from a
spin-cavity system. Unlike single-
photon encodings, the coherent-state encoding is naturally tolerant of
photon loss: a loss channel attenuates the coherent-state amplitude, leading to a gradual reduction in preparation fidelity rather than probabilistic protocol failure. The central difficulty—implementing the projection onto superpositions of nonorthogonal coherent states required in the conventional protocol—is circumvented by coupling the transmitted pulse to an auxiliary
spin, followed by homodyne detection and a
spin projective measurement. The protocol therefore avoids direct coherent-state-superposition measurements and
photon-number-resolving detection. In the ideal limit, deterministic preparation of a particular class of target states can be achieved using only one bit of classical communication. We analytically quantify the effects of coherent-state nonorthogonality, fiber loss, and
spin dephasing on the preparation fidelity. Increasing the coherent-state amplitude improves state distinguishability and hence the preparation fidelity, but also enhances which-branch information leakage caused by
photon loss, resulting in an optimal amplitude at each transmission distance. For experimentally relevant parameters, the optimized average fidelity remains well above the classical
benchmark over distances of tens of kilometers. We also discuss a possible implementation using diamond nitrogen-vacancy centers.
Read the full abstract ▾
All authors Yunge Jiang, Yehan Yu, Lili Song, Yan Mu, Chaoyun Peng, Mingfeng Wang
Machine Learning
AnyJev Technical Report
Jiamu Zhang, Tianze Yang, Yucheng Shi et al. (+6 more)
A typed decision is a choice among a fixed set of options, returned as a probability rather than as text. Systems that need typed decisions today use models trained for that purpose. This report describes AnyJev, which reads a typed decision from one prefill of a pretrained instruction-tuned
language model. The readout restricts the next-
token distribution at the answer position to the option
tokens. It has two defects: the model assigns higher probability to some labels whatever the input, and to some positions in the option list. AnyJev corrects both with no gradient steps and no parameter changes: it divides out a label prior estimated from unlabelled inputs, and it averages log-probabilities over the K cyclic rotations of the option list. On two 20-option tasks the rotations lower the order-flip rate from 0.33 to 0.14 and from 0.33 to 0.18, and raise accuracy on 11 of 11 models on both. Reading every rotation requires K prefills. A stopping rule selected against the full-rotation decision on unlabelled states cuts that. Selecting the threshold on one unlabelled split and bounding its disagreement on a second, it reads 10.6 rotations of 18 at a verified 0.008 bound on two of four
cells; selected and bounded on one split, as our serving run did, it reads 7.3 and serves 2.2 times as many decisions per second on vLLM. The code is open source.
Read the full abstract ▾
All authors Jiamu Zhang, Tianze Yang, Yucheng Shi, Evan Chen, Zixiang Nie, Kelly Wan, Liangjie Hong, Ninghao Liu, Liang Wu
Numerical Analysis
Matrix-Free Moment-Matching Method for Reduced-Order Modeling of Quadratic-Bilinear Descriptor Systems with Multiple Inputs
Zeyuan Gao, Cheol W. Lee and Oleg Zikanov
This paper develops a new moment-matching order reduction method for multi-input quadratic-bilinear descriptor systems. The proposed approach accommodates multiple inputs acting on both the differential and algebraic equations and constructs separate projection spaces associated with different input channels and input combinations. The method uses a
matrix-free
algorithm, allowing the reduced-order
matrices and tensors to be computed without explicitly assembling or storing the full-order system
matrices and tensors. This feature makes the proposed method particularly suitable for large-scale computational fluid dynamics problems. Numerical experiments carried out for two-dimensional flow problems demonstrate that the resulting reduced-order models accurately reproduce the transient responses of the full-order models under multiple time-varying inputs.
Read the full abstract ▾
Combinatorics
Reversing the Mostar line-graph inequality with long pendant paths
Kerem Kat
Let $K_t$ be obtained by attaching a pendant path of length $t$ to a fixed rooted graph $K$. We prove that $\mathrm{Mo}(L(K_t))-\mathrm{Mo}(K_t)$, the difference between its line-graph and original Mostar indices, is exactly affine on each parity class beyond a sharp uniform cutoff. The slope depends on distance-level neighbor counts and is positive for every non-bipartite core. Triangle chains give infinitely many graphs with $\mathrm{Mo}(L(G))>\mathrm{Mo}(G)$ at every positive cyclomatic number $c$, with maximum degree three, resolving Alex-Indulal Problem 3.3. At zero slope, a cactus branch-mass formula decides equality; identical vertex profiles need not give identical intercepts. Fixed cores giving equality for all sufficiently long attachments exist exactly when $c$ is odd. Maximum degree three suffices, and at every even $c$ a binary-tree construction gives equality on one eventual parity class.
Read the full abstract ▾
Computational Complexity
Entrywise Logarithmic Matrix Algebra and Dichotomy of Planar Graph Homomorphisms
Jin-Yi Cai and Zhuxiao Tang
We prove a complexity classification of counting planar graph homomorphisms with non-negative weights. For a real symmetric
matrix $M$ with nonnegative entries, the problem $\GH(M)$ is either (1) P-time computable over all graphs, or (2) \#P-hard in general but P-time computable over planar graphs, or (3) \#P-hard over planar graphs. Furthermore, $\GH(M)$ in (2) consists of precisely those that involve the P-time FKT
algorithm to count planar perfect matchings with a
holographic transformation. The dichotomy for entrywise positive and full rank
matrices lies at the core of the proof. In this setting, the main tool is a
matrix algebra ${A}$ (a vector space with bilinear multiplication) formed by taking the real span of (centered) entrywise logarithms of all realizable
matrices from $M$ using planar edge gadgets and
polynomial interpolation. We then define a multiplicative homomorphism $F(\cdot)$ on the
algebra ${A}$, which turns out to be a
polynomial map. The exponents of
prime factorization of component
polynomials of $F$ induce graph distances for distance-regular graphs. These graphs are either the Hamming graphs or Doob graphs using a classification result for distance-regular graphs (if $\PlGH(M)$ is not \#P-hard), after obtaining their intersection numbers. The dichotomy is proved in either case. We then extend the complexity classification to all nonnegative
matrices.
Read the full abstract ▾
Computation and Language
Verbalized and Internal Probabilities Are Coupled in Large Language Models
Sinead Williamson, Jiaxuan Li, Nick Foti et al. (+2 more)
Large
language models carry an internal notion of uncertainty in their sampling distribution, i.e., the probabilities they place on generating one answer rather than another. They can also be asked to state a confidence, in words or as a number: a verbalized uncertainty. Prior work suggests that internal probabilities track relative frequencies in the training data, and that verbalized probabilities track explicit probabilistic assertions in the training data. However, we do not know whether these two readouts are aligned, except when frequencies and probabilistic assertions in the training data happen to align. This limits our understanding of when we can use verbalized uncertainties as a proxy for either training data frequencies, or a model's internal distribution. We resolve this gap by systematically exploring how LLMs probability readouts are impacted by training and in-context data, via intervening on the underlying uncertainty sources in the data. We find that both internal and verbalized probability readouts are impacted by both distributional and asserted uncertainty in the training data. Further, we find that verbalized and internal probabilities are aligned beyond what would be expected by independently tracking the same uncertainty sources, suggesting that verbalized probabilities can be used to probe a model's internal distribution.
Read the full abstract ▾
All authors Sinead Williamson, Jiaxuan Li, Nick Foti, Russ Webb, Masha Fedzechkina
Statistical Mechanics
Thermal entanglement transitions from strong $\textrm{SU}(2)$ symmetry
Siqi Mo, Ehud Altman and Samuel J. Garratt
We study mixed-state
entanglement in Gibbs states with strong $\mathrm{SU}(2)$
symmetry, focusing on locally interacting
spin systems with
ferromagnetic interactions. Our main result is to show that finite-temperature ordering transitions are associated with
entanglement transitions, and therefore the steady state of strongly symmetric Lindbladians can exhibit
entanglement transitions. While in the paramagnetic phase it is known that the distillable
entanglement and logarithmic
entanglement negativity between two halves of a large system grow logarithmically with the square root of the number of
spins, we show that in
ferromagnetic phases these quantities grow parametrically faster with system size. To arrive at this result we first establish relations between these mixed-state
entanglement measures and
spin correlations in states that are singlets under global $\rm{SU}(2)$
symmetry transformations. We then introduce a semiclassical theory for $\rm{SU}(2)$ singlet thermal states. While the global singlet constraint generally enters this theory as a complicated function of the full semiclassical
spin configuration, we show that in large $S$ limit it simplifies drastically to a Gaussian suppression of total magnetization in disordered phases as well as in ordered phases in the vicinity of continuous thermal phase transitions. This leads us to a field theory describing
spin correlations in the singlet sector. Using this we determine the behavior of various probes of mixed-state
entanglement at low and at high temperatures, supporting our analytical results using numerical Monte Carlo simulations of a three-dimensional
lattice realization of our semiclassical theory. We also use exact numerics in one-dimensional
spin-$1/2$ systems to confirm our predictions for the scaling of mixed-state
entanglement with correlation length and system size in the paramagnetic phase.
Read the full abstract ▾
Computer Vision
Align Then Reason: A Multimodal Lip-Sync Judge for Dubbing
Rui Liu, Bhavin Jawade, Haoqi Li et al. (+4 more)
Dubbing quality control requires a reference-free judge that can determine whether a candidate text line matches a speaker's visible articulation in both content and timing, using only silent
video and text because dubbed audio may not yet exist. Existing visual
speech recognizers and
video-
language models are poorly suited to this setting: even when fine-tuned to recover spoken content from lip motion, they remain largely insensitive to temporal errors. We introduce $\textit{Align Then Reason}$ (ATR), a
multilingual lip-sync judge that first establishes a monotonic
alignment between frame-level lip representations and the phonetic units of the candidate line, then reasons over this
alignment to make the final judgment. An
alignment scorer provides the LLM with both local evidence for each phonetic unit and a calibrated global
alignment score, enabling it to reason jointly about content and timing. On a seven-
language benchmark, our method improves mean AUC over the corresponding Qwen3.5 SFT baselines by 59.4%, 50.2%, and 50.8% with 2B, 4B, and 9B reasoners, respectively. The gains generalize across LLM families, reaching mean AUC improvements of 45.9% and 46.6% over the best baseline for LLaMA-3.1-8B and Mistral-7B, respectively. They also transfer across
datasets to three unseen MuAViC
languages. Furthermore, we evaluate on two downstream tasks built from real dubbing lines. On dub-line reranking, ATR-9B outperforms the best lip-reading baseline by 52.0%, while on script-to-clip assignment, ATR-9B improves over the best lip-reading baseline by 17.7%.
Read the full abstract ▾
All authors Rui Liu, Bhavin Jawade, Haoqi Li, Shivam Mehta, Karan Saxena, Yinghong Lan, Cameron R. Wolfe
Instrumentation and Detectors
Impact of Gain Layer Doping Concentration on the Performance of 4H-SiC LGADs
Tao Yang, Ben J. Sekely, Yashas Satapathy et al. (+12 more)
Gain layer doping concentration determines the electric field profile in a low gain avalanche detector (LGAD) and is therefore the primary design parameter. This work reports its influence on the charge collection and timing behavior of 4H-SiC LGADs. Two mesa-isolated devices with cold He ion implanted termination differ only in gain layer doping, $4\times10^{17}$~$cm^{-3}$ and $5\times10^{17}$~$cm^{-3}$, and are characterized together with a PIN diode from the same wafer using the ultraviolet transient current technique (UV-TCT). The 25% doping difference produces a gain difference of about two orders of magnitude at the same applied bias, representing a gain range of 3.7 to 216. Under UV-TCT excitation, the higher-doped device exhibits time-resolution minima within a gain range of about 20–80 across the investigated
laser intensities, consistent with the transition from electronic-noise-dominated jitter to multiplication shot noise. The non-monotonic timing response is also observed under a $^{90}Sr$ beta source with a Si LGAD reference: the higher-doped device reaches 51.7 ps at 280 V and then degrades, while the lower-doped device requires more than 450 V. This measured gain window provides an experimental reference for further optimization of the gain-layer design.
Read the full abstract ▾
All authors Tao Yang, Ben J. Sekely, Yashas Satapathy, Abraham Tishelman-Charny, Greg Allion, Gil Atar, Philip Barletta, Gabriele Giacomini, Carl Haber, Steve Holland, John F. Muth, Spyridon Pavlidis, Chengxi Yang, Haichen Wang, Stefania Stucci
Robotics
Reactive Humanoid Multi-Contact Using Learned Stability Models
Stephen McCrory, Beomyeong Park, Nicholas Kitchel et al. (+2 more)
We present a planning and control approach to reactively use hand contacts to stabilize a
humanoid in low stability scenarios, where only using feet contacts may result in a fall. Candidate contacts are sampled within the
robot's reachable workspace, and a preview is computed by rolling out the centroidal dynamics through pre-impact, impact and post-impact phases. Sampled points are scored based on the Center of Pressure (CoP) control authority at the post-impact phase. Central to our approach is a learned model of the
robot's CoP region during post-impact, which enables rapid evaluation of candidate contact points compared to traditional optimization-based methods. The presented planner has two stages: the first selects an optimal bracing region and the second computes an optimal bracing point within the region. Our simulation results demonstrate an average increase in impulse resilience of 89% over recovery without hand contacts and 17% over a naive planning strategy (closest reachable region). We validate our framework on hardware, performing push tests while standing and walking. The standing trials show an average 43% reduction in stabilization time compared to naive hand placement and the walking trials demonstrate a 18% reduction compared to baseline recovery (without hand contacts).
Read the full abstract ▾
All authors Stephen McCrory, Beomyeong Park, Nicholas Kitchel, Nehar Poddar, Robert Griffin
Robotics
TRACE: Privacy-Preserving Next-Best-View Selection over Distributed 3D Gaussian-Splat Maps
Amirhossein Mollaei Khass, Athanasios Cosse, Qiyu Sun et al. (+1 more)
Share the light, not the map. We study next-best-view selection for a team of
robots, each of which builds its own 3D Gaussian Splatting map and keeps it private. A
robot picks the view with the largest expected information gain (EIG) about the splats along its own path. This gain depends on the other maps. Their splats occlude its own and shine behind them, so the gain has to be evaluated against the pooled map. No
robot has this map. We show that the coupling passes through only two ray quantities, the transmittance in front of a splat and the radiance behind it, and that both are sums over the hits of the ray. Hence, they decompose across the
robots, and each
robot sums them over depth bins in its own map, along the rays of a candidate view, and sends the sums with their pose derivatives. The
robot planning the view turns them into its EIG and gradient on SO(3). Transmittance and Radiance Aggregates, communicated for the EIG, give the protocol its name: TRACE. No
robot shares its splats, and the message size does not grow with a map. We prove that the reconstruction is exact unless a depth bin behind a splat mixes hits of two
robots, and we bound the error otherwise. Over 100 next-best-view decisions in Habitat-Sim, TRACE picks a heading within 15 degrees of the centralized one in 83.3% of the cases, and its views reach 97.9% of the centralized EIG.
Read the full abstract ▾
All authors Amirhossein Mollaei Khass, Athanasios Cosse, Qiyu Sun, Nader Motee
Robotics
Getting Out and Getting Back: World and Behavior Grounding in Real2Sim2Real Co-Training
Samuel Liu, Youngsun Kim, Martin Matak et al. (+1 more)
Simulation can expand scarce real demonstrations for co-training, yet how world fidelity and similarity to human behavior affect policy performance remains unclear. We distinguish world grounding, which aligns simulation with the real system, and behavior grounding, which aligns simulated trajectories with human motion. We build a real2sim2real pipeline that varies these axes independently to generate data for co-training. On a dynamic dexterous pick-and-sort task, fully grounded co-training raises success from 52% to 86%; averaged across configurations, world grounding improves success by 18 percentage points and behavior grounding by 10. Deployed policies behave like a mixture of real-derived and simulation-derived policies, imitating real demonstrations in covered states and relying on simulated behavior elsewhere, which we examine through latent-space analysis. Together, these results suggest complementary roles: world grounding lets policies use simulated experience beyond real-data coverage, while behavior grounding matters mainly when world grounding is imperfect. Grounded simulation remains beneficial when co-training foundation models.
Read the full abstract ▾
All authors Samuel Liu, Youngsun Kim, Martin Matak, Gilwoo Lee
Machine Learning
On-the-fly Weight Generation: A Hypernetwork Proof of Concept on ARC-1D
Fabio J. Fehr and Philip Torr
General-purpose models can adapt to many tasks from context, while specialised models can execute individual functions with less capacity. Yet obtaining such specialists requires task-specific training or adaptation. We ask whether they can instead be generated directly from a few demonstrations. Using ARC-1D as a controlled testbed, we show that individual transformations can be represented by tiny specialist models, and that a hypernetwork can generate their parameters from context. The generated parameters form a structured weight space, while the resulting specialists show partial compositional generalisation and generalisation to transformations not seen during training. In both settings, removing explicit task identifiers improves generalisation beyond the training transformations. Together, these results provide a proof of concept that few-shot task context can be compiled on-the-fly into compact executable model parameters, and that the resulting weight space can support reuse and generalisation beyond known functions.
Read the full abstract ▾
Signal Processing
PI-AMFM: Permutation-Invariant Learning for Variable-Cardinality AM-FM Mode Decomposition in Biomedical Signal Analysis
Youngsun Kong and Ki H. Chon
Physiological recordings often contain nonstationary oscillatory components whose number and dynamics vary across signals. Amplitude- and frequency-modulated (AM-FM) representations are well suited to characterizing such dynamics and have shown broad utility in biomedical signal analysis. Recent approaches have incorporated neural networks to learn mode decomposition patterns from data, but component cardinality is often predefined or determined through separate stopping or selection mechanisms. We propose a permutation-invariant neural framework for variable-cardinality AM-FM mode decomposition (PI-AMFM). PI-AMFM combines a multiscale temporal encoder, Mamba backbone, and component-presence estimation, with permutation-invariant Hungarian matching during training. On synthetic AM-FM signals, PI-AMFM achieved lower decomposition, instantaneous-frequency, reconstruction, and mode-count errors than the compared methods while preserving the overall trajectory pattern in a crossing-chirp example. On photoplethysmographic recordings, recovered modes captured cardiac and respiratory dynamics despite training only on synthetic signals. These results support the feasibility of PI-AMFM for variable-cardinality decomposition of nonstationary biomedical signals.
Read the full abstract ▾
Machine Learning
Quantifying the Impact of Ambulance Ramping: A Multi-Year Analysis of Victorian Emergency Medical Services Cases
Ayesha Tanveer, Khandakar Ahmed, Assefa Teshome et al. (+2 more)
Ambulance ramping, the delay between hospital arrival and patient handover, is a critical operational bottleneck in Emergency Medical Services (EMS), yet its systemic magnitude and dynamics remain inadequately characterised at scale. This paper quantifies the scale, trajectory, and operational correlates of ramping across an entire statewide EMS system, analysing 2,850,575 ambulance attendances in Victoria, Australia from January 2020 to March 2024 using an Exploratory Data Analysis (EDA). After systematic preprocessing, an analytical cohort of 2,026,569 Emergency Department (ED) transports across 59 hospitals with ED and 79 Local Government Areas (LGA) was examined through interval decomposition, Pareto concentration, hourly cross-correlation, hospital arrival concurrency and priority-stratified operational comparisons. Cumulative Ambulance Hours Lost (AHL) totalled 1,491,127 hours, equivalent to approximately 96 ten-hour ambulance shift lost every day of the study window. Ten of 59 hospitals account for 57.8% of lost hours from 50.9% of cases. Annual losses rose 57% to a 2022 peak while transported demand fell 3.7%, indicating deterioration in per-case handover rather than growth in demand. Handover duration varies little with patient acuity, but rises monotonically with the number of ambulances arriving at the same hospital in the preceding hour, an effect persisting within every hour of the day. Hourly demand is moderately associated with ramping two to four hours later (r = 0.365). These findings establish the empirical preconditions for hospital-state aware ambulance routing.
Read the full abstract ▾
All authors Ayesha Tanveer, Khandakar Ahmed, Assefa Teshome, Oyetunde Gbadeyan, Ziad Nehme
Databases
TabJoinBench: A Benchmark for Joinable Table Discovery
Sandipan De, Jin Wang and Vivek Gupta
Join discovery aims to identify tables from large data repositories that can augment a query table with complementary information, enabling downstream tasks such as data exploration, feature engineering, and business intelligence. Although numerous join discovery methods have been proposed, existing studies rely on method-specific
benchmark construction, making reproducible and fair comparison difficult. We present TabJoinBench, a
benchmark for evaluating join discovery methods across semantic, relational, and hybrid data lake scenarios. TabJoinBench constructs query-candidate pairs using source-specific validation strategies, systematically introduces structural, representation, and semantic changes through composable perturbations while preserving reliable ground truth. We evaluate representative join discovery methods spanning set-based, feature-based, and learned approaches, together with general-purpose
language-model
embedding baselines, and publicly release the processed
datasets, ground-truth annotations, and generation pipeline to facilitate reproducible evaluation and future research.
Read the full abstract ▾