Computation and Language
Rubric Rewards from Item Response Theory
Milad Yazdani, Yaser Souri, Xiren Zhou et al. (+4 more)
Many
language tasks have no single answer that can be checked automatically. Rubrics provide criteria for judging responses to these tasks. For reinforcement learning, the resulting verdicts must be combined into a scalar reward. A common approach sums the points assigned to satisfied criteria. Distinct verdict patterns can thus receive the same reward, and the fixed points encode how much each criterion should count, not how strongly its verdict distinguishes the current rollouts. Beyond this aggregation problem, judging the full rubric needs more judge requests as the criterion count grows. To address these limitations, Rubric Response Theory (RRT) measures quality and selects criteria when rubric criteria are monotone indicators of a shared target. Rather than adding assigned points, RRT uses a two parameter item response model that treats the verdict pattern as evidence about scalar quality specific to the rubric. Under this model, its likelihood score maximizes the local signal-to-noise ratio for quality. Its Response Parameter Network (RPN) reads the
prompt and criterion text to predict criterion difficulty and discrimination. As the policy distribution changes during training, RRT uses online expectation maximization to update the RPN from current rollout verdicts. With Qwen3.5-4B as the policy, RRT's macro criterion score across Medical, Science, Rubrics as Rewards Science, and RubricBench is 1.7 points above that of group relative policy optimization (GRPO). On hard and very hard criteria in Medical and Science, RRT gains 2.8 to 5.6 points over GRPO. At half the criterion budget, adaptive Fisher selection with a frozen RPN keeps the macro criterion score across four
datasets within 0.1 points of GRPO with full judging. These results show RRT can reduce judge requests while remaining competitive with GRPO.
Read the full abstract ▾
All authors Milad Yazdani, Yaser Souri, Xiren Zhou, Pranit Chawla, Dena Shahriari, Subhojit Som, Xia Song
Sound
CoSE-E: A Benchmark for Code-switched Speech Evaluation in Enterprise Settings
Shama Gupta, Hoang H Nguyen, Chelsea Huang et al. (+2 more)
Code-switching (CS), a seamless alternation between
languages within a single utterance, remains a critical challenge in automatic
speech recognition (ASR). While prior works focus on conversational CS-ASR, enterprise settings demand evaluation of operational impact beyond edit-distance errors: how code-switching transcription errors propagate to downstream voice
agent task failures. In this work, we propose (1) a CS-ASR synthetic
benchmark and multidimensional evaluation framework tailored to enterprise domains, (2) systematic evaluation of frontier ASR systems across 5
language pairs, (3) diagnostic analysis of the additional transcription errors that code-switching introduces across
language pairs and models. We release COSE-E to support enterprise-focused CSASR evaluation for
multilingual voice
agents in enterprise deployment.
Read the full abstract ▾
All authors Shama Gupta, Hoang H Nguyen, Chelsea Huang, Lindsay Devon Brin, Fanny Riols
Networking
Energy-Driven Evaluation of Network Digital Twinning Applied to mmWave Beam Management
João Borges, Bruno Castro, Kleber Cardoso et al. (+2 more)
Network Digital Twins (NDTs) are important enablers of 6G and future networks. However, there is a lack of studies regarding practical aspects, such as the impact of simultaneously changing twinning rate, fidelity, and other NDT operational parameters. For instance, works often consider the impact of operational parameters in isolation or with physical twin (PTwin) implementations relying on simulations. Therefore, the main contribution of this work is to provide an in-depth study on the performance impacts of both twinning rate and fidelity, using PTwins that rely on measurements obtained from hardware. We also investigate the optimization of these two operational parameters to minimize energy consumption, exploring a
Bayesian Optimization (BO) method and what-if analysis. In this work, the NDT models an indoor propagation environment to optimize beam management. The virtual twin (VTwin) was implemented with the Sionna ray tracing (RT) simulator and the PTwin with our in-house setup composed of customized Wi-Fi
radios with 32
antenna elements operating at 60 GHz. The study reveals important aspects of wireless channel NDTs, suggesting that fidelity levels vary throughout the experiment and with the what-if difficulty. Moreover, the twinning rate can be adjusted using sample-efficient methods, such as BO, even at lower fidelity levels.
Read the full abstract ▾
All authors João Borges, Bruno Castro, Kleber Cardoso, Andrey Silva, Aldebaro Klautau
Artificial Intelligence
Not All Thinking is Created Equal: Latent Reasoning Discovers a Recurrent Search Algorithm for Depth Generalization
Huzi Cheng and Zhewei Zhang
Large
Language Models can perform multi-step
reasoning and improve task performance through different forms of intermediate computation, from
token-based traces to computation carried out in latent space. However, a question remains open: do these different forms of thinking rely on the same underlying mechanism? To address this, we train and compare five variants of the same GPTNeoX backbone from scratch on an extended multi-hop
reasoning task (ProsQA-Ext): a vanilla model, a Chain-of-Thought (CoT) model, a Pause
Token model, and two latent-
reasoning models that are optimized end-to-end without intermediate
reasoning traces. We find that, strong in-distribution (ID) performance does not guarantee depth generalization. Vanilla, CoT, and Pause
Token models solve ID problems well, but rely largely on local graph features and generalize poorly to out-of-distribution (OOD) problems with longer hops. In contrast, latent variants generalize better and show internal dynamics consistent with forward reachability propagation on the graph. Causal interventions and
circuit analysis localize this computation to a sparse recurrent search
circuit in the bottleneck latent model: an
attention head retrieves graph relations, an MLP and the residual stream update the reachability state across recurrent steps, while multiple
attention heads together then do the candidate matching. Together, these results show that different thinking mechanisms can learn distinct computational solutions, even at similar ID performance. In this setting, latent recurrence supports a reusable forward-search
algorithm that generalizes beyond the training depth.
Read the full abstract ▾
Number Theory
Heavenly Elliptic Curves over Cubic Number Fields
Suzanne O'Hara
The study of heavenly abelian varieties is motivated by a question of Ihara. When an
elliptic curve $E/K$ is heavenly at $\ell$, the extension $K(E[\ell^\infty])/K(\mu_\ell^\infty)$ is pro-$\ell$ and unramified away from $\ell$. These are the same arithmetic conditions as the fixed field of the kernel attached to pro-$\ell$ étale covers of the projective line over $K$ minus three points. In this paper we study heavenly
elliptic curves defined over cubic number fields. Following the work on McLeman and Rasmussen in the quadratic case, we find that there is a wider range of possible behaviors for the trace of a balanced
elliptic curve in the cubic case. In this setting, we introduce a further distinction between balanced and totally balanced curves. With this, we show that trace of the representation on the $\ell$-torsion for totally balanced
elliptic curves is non-surjective. We also compute a bound on
primes $\ell$ after which any heavenly
elliptic curve defined over a cubic number field must be balanced. Finally we compare the trace behavior of totally balanced
elliptic curves with CM
elliptic curves.
Read the full abstract ▾
Artificial Intelligence
Verifiable Visual Rewards Transfer from Synthetic Scenes to Natural Prompts
Shuyue Stella Li, Xiaochuang Han, Yulia Tsvetkov et al. (+1 more)
Precise instruction following in image generation, such as satisfying object counts and spatial relations, remains an open challenge at least in part because it is learned using unreliable reward models such as object detectors and
vision-
language models. We introduce Verifiable Visual Rewards (VVR), the first framework for programmatically verifiable image rewards, and show that training on it generalizes to natural
prompts. Each VVR task is a scene of
geometric objects and relations among them, from which we derive both the
prompt and a deterministic verifier, so tasks can be generated in any number and at any chosen complexity. We release VVRBench, with 10,000 tasks over 32 constraint types, and VVRBench-Challenge, with 720 more complex tasks; the strongest model we evaluate—GPT-Image-2.5—solves 21.4% of VVRBench-Challenge. Using VVR scores as rewards for reinforcement learning (RLVVR) raises the accuracy of Stable
Diffusion 3.5 Medium on VVRBench from 2.8% to 28.3% and demonstrates consistent easy-to-hard generalization. These gains extend to out-of-domain
benchmarks, and mixing VVR into existing objectives further improves overall performance and human preference, motivating the adoption of VVR into standard image generation post-training recipes.
Read the full abstract ▾
All authors Shuyue Stella Li, Xiaochuang Han, Yulia Tsvetkov, Luke Zettlemoyer
Differential Geometry
Deformations of generalized complex structures on $\mathcal{G}$-flat transitive Courant algebroids
Vicente Cortés and Paula Naomi Pilatus
We extend Gualtieri's deformation
theorem for generalized complex structures on exact Courant algebroids to the more general class of $\mathcal{G}$-flat transitive Courant algebroids. Our approach relies on the framework of tame Fréchet spaces and Hamilton's Nash-Moser (implicit function)
theorem. As an important step of the proof of our deformation
theorem, we endow the group of autoequivalences of a $\mathcal{G}$-flat transitive Courant algebroid with a tame Fréchet Lie group structure and show that there exists a tame subgroup of exact autoequivalences, whose infinitesimal action is by inner derivations.
Read the full abstract ▾
Distributed Computing
GPUPhysBench: Benchmarking Coding Agents for Correct and Efficient GPU Physics Simulation
Yuchen Sun, Jinjin He, Sinan Wang et al. (+1 more)
Writing fast GPU code for physical simulation is difficult: implementations must preserve numerical accuracy while handling irregular data access, synchronization, and iterative solvers. We introduce GPUPhysBench, a
benchmark of 50 tasks testing whether coding
agents can meet these demands. Tasks cover fluids, deformable solids, and granular materials, from individual simulation operators to complete simulators.
Agents write, compile, test, and optimize GPU code with access to a NVIDIA GPU under fixed time budgets. We report pass rates and runtime performance relative to expert-optimized reference implementations. In a single-attempt evaluation of six frontier model-harness pairs, the two strongest pass all 50 tasks, but even the fastest reaches at least 0.9 the reference speed on only 22% of them, and no submission is more than 5% faster than the reference. The largest gaps arise in collision detection, constraint solving, and iterative solvers. GPUPhysBench brings physical simulation workloads to coding-
agent evaluation, testing both the ability to implement numerical methods correctly and the ability to make them run efficiently.
Read the full abstract ▾
All authors Yuchen Sun, Jinjin He, Sinan Wang, Bo Zhu
Combinatorics
Parking with Frustrated Drivers
Joshua Hallam, Jenson Molebash and Chris Porter
Imagine there are $n$ cars lined up along a one-way street containing $n$ spots. Each car contains a group of friends, including a reluctant driver. Each car has a preferred spot and cars enter one by one. The cars drive to their preferred spot and if it is empty park there. If it is not empty, a friend in the back yells out “Hey! You should have driven faster!". Frustrated by this, the driver continues down the road until they find the last unoccupied spot (if one exists) and parks there. We say a sequence $(a_1,a_2,\dots, a_n)$ of preferred spots is a frustrated parking function if all cars can park under this rule. In this paper, we study the enumerative properties of frustrated parking functions. In particular, we show that the number of frustrated parking functions of length $n$ is $(2n-1)!!$. This is done by associating frustrated parking functions with height labeled Dyck paths. Using this association, we are then able to better understand the sets of lucky cars and lucky spots for frustrated parking functions. We show that the frustrated parking functions of length $n$ where the first $k$ cars (or first $k$ spots) are lucky is given by $k!S(n,k)$ where $S(n,k)$ is the Stirling number of the second kind. This in turn implies that the number of frustrated parking functions where once a car (or spot) is unlucky, the remaining cars (or spots) are unlucky is counted by the $n^{th}$ Fubini number. We also show that the number of frustrated parking functions with $k$ lucky cars (or spots) is given by the second order Eulerian number.
Read the full abstract ▾
Computer Vision
RT-Super: Learning Tumor Segmentation from Longitudinal Images and Reports
Pedro R. A. S. Bassi, Wenxuan Li, Hanxue Gu et al. (+11 more)
Multi-tumor
segmentation is important for early cancer detection and allows radiologists to visualize, verify, and understand AI predictions. However, tumor
segmentation masks are expensive, time-consuming, and unavailable for many tumor types in public data. Instead, hospitals have vast, readily available data that can guide
segmentation: radiology reports, longitudinal images, and multi-phase images. We use this readily available data to substitute for tumor masks in training AI for tumor
segmentation. To this end, we propose a new architecture, RT-Super. It has a teacher network, which analyzes the patient's longitudinal images and reports to create high-quality tumor masks. These masks train a student network, which sees a single image and no report. At inference, when longitudinal images and reports are unavailable, we use the student. RT-Super uses a new CNN-
Transformer architecture and novel Consistency Losses that exploit tumor location consistency across longitudinal images. We train RT-Super to segment esophagus, uterus and spleen tumors, which have few or no public masks. Even without training masks, RT-Super can segment these tumors and surpass public AI models. Overall, we demonstrate that learning from longitudinal images, multi-phase images, and reports can overcome mask scarcity and advance multi-cancer detection and
segmentation. Code: https://github.com/MrGiovanni/RT-Super
Read the full abstract ▾
All authors Pedro R. A. S. Bassi, Wenxuan Li, Hanxue Gu, Jieneng Chen, Xinze Zhou, Zheren Zhu, Sezgin Er, Ibrahim E. Hamamci, Bjoern H. Menze, Gulhan E. Akan, Kang Wang, Yang Yang, Alan L. Yuille, Zongwei Zhou
Numerical Analysis
Isogeometric multi-patch preconditioner for time-dependent electromagnetic problems
Bernard Kapidani, Gabriele Loli, Giancarlo Sangalli et al. (+2 more)
We build upon recent developments in high-order spline-based
geometric methods for the three-dimensional initial boundary value problem of Maxwell's equations. Previous schemes were limited to single-patch geometries. To address the computational challenges associated with inverting mass
matrices in multi-patch settings, we extend existing techniques for 0-forms mass and stiffness
matrices to 1-forms mass
matrices. Our approach leverages the tensor-product structure of spline spaces to construct a spectrally equivalent preconditioner for the 1-forms mass
matrices. Specifically, we design an efficient preconditioner for single-patch domains that is robust with respect to both mesh size and spline degree. In the multi-patch case, this methodology is extended by integrating the single-patch preconditioner with an additive Schwarz approach, ensuring robustness and scalability with respect to the number of patches. We analyze the computational complexity of the proposed preconditioner and validate its robustness and efficiency through comprehensive numerical experiments.
Read the full abstract ▾
All authors Bernard Kapidani, Gabriele Loli, Giancarlo Sangalli, Mattia Tani, Rafael Vázquez
Machine Learning
Bounding Retraining Equivalence and the Deletion Floor in Materials Machine Unlearning
Can Polat, Mustafa Kurban, Erchin Serpedin et al. (+1 more)
In materials machine learning, closely related retained structures can sustain accurate property predictions even after removing a specific record, rendering post-deletion prediction error an ambiguous metric for machine unlearning. To resolve this ambiguity, we define the deletion floor as the expected target loss under a specified retraining procedure at the deleted request. Standard indistinguishability constraints yield a sharp interval bounding an update's target loss around this baseline reference. Theoretically, a conditional neighbor bound links a low deletion floor directly to retained fit, prediction regularity, and local label agreement, while an exact ridge identity isolates residual fit from the prediction change induced by record deletion. Empirically, controlled redundancy sweeps show an $\approx 8\times$ drop in median normalized retraining loss when one retained relative remains after deletion. Across two distinct fitting regimes in a paired Materials Project study, the lower-floor regime also exhibits a larger prediction change on more than 50% of the shared requests. Systematic comparisons against approximate updates and the original model decouple deliberate target suppression from preserved overall model utility. Consequently, request-level unlearning evaluations should report reference loss, prediction change, and retained utility together, interpreting post-deletion accuracy against what retraining itself leaves behind.
Read the full abstract ▾
All authors Can Polat, Mustafa Kurban, Erchin Serpedin, Hasan Kurban
Machine Learning
DR-net-Mamba: Selective State-Space Modeling for Long-Range ECG Time-Series Denoising
Basile Morel, Samuel Ruiperez-Campillo, Andreas P. Streich et al. (+2 more)
Electrocardiogram (ECG) recordings are corrupted by non-stationary noise sources that degrade diagnostic reliability, particularly in ambulatory and long-duration recordings. Deep learning denoisers exist, but convolutional architectures are limited by their receptive field,
transformer-based models scale quadratically with sequence length, and
diffusion-based approaches incur prohibitive inference cost. We propose a Mamba-augmented model that inserts selective state-space blocks at the convolutional bottleneck, combining local feature extraction with long-range temporal modeling at linear complexity. We comprehensively evaluate the proposed model with respect to reconstruction fidelity, noise robustness, recording-length scaling, and downstream diagnostic classification across over 40 pathology classes. On synthetic and real
datasets, our model achieves the highest SNR and lowest RMSE, with the Mamba advantage increasing with sequence length and in low-SNR regimes. On classification with two independent classifiers, the proposed Mamba-based models achieve the best macro AUROC among all denoisers and improve over their convolutional base models. Calibration is more nuanced and classifier-dependent: denoising improves Binary Cross-
Entropy and Brier score on Inception1D but often fails to beat the noisy input on ResNet1D-Wang, and the lead-specific Mamba variant is the only denoiser to improve both calibration metrics over the noisy baseline on both classifiers. Per-class analysis reveals a morphology-dependent benefit: Mamba substantially improves ST/T-change diagnoses, which depend on broad, context-sensitive waveforms.
Read the full abstract ▾
All authors Basile Morel, Samuel Ruiperez-Campillo, Andreas P. Streich, Julia E. Vogt, Thomas Hofmann
Quantum Physics
Succinct Arguments for QMA from Collapsing Hash Functions
James Bartusek and Giulio Malavolta
We prove the existence of succinct arguments for QMA, assuming only the existence of collapsing hash functions. This is the first scheme that relies only on unstructured “Minicrypt” assumptions, which are not known to imply public-key encryption. Our main technical contribution is a
quantum-succinct claw-state generation protocol that allows us to bootstrap a small number of
quantum correlations into an arbitrarily large number of claw-state correlations, using classical communication only. This improves upon the work of [Zhang, STOC 2021], having better round complexity, a proof in the standard model, and being overall much simpler. This yields a
quantum-succinct blind delegation of
quantum computation protocol from one-way functions, which we plug into the communication-compression compiler of [Bartusek, Liu, and Malavolta, EUROCRYPT 2026] to obtain succinct arguments for QMA.
Read the full abstract ▾
Dynamical Systems
Uniqueness of four-body convex central configurations for all masses
Tejasvi Singh Tomar
We prove that for every choice of four positive masses and every cyclic ordering of the bodies there is exactly one strictly convex planar central configuration with that ordering, up to similarity. This proves a
conjecture of Albouy, Fu and Sun, also called the Simó-Yoccoz
conjecture, and answers Problem 10 in the list of Albouy, Cabral and Santos. Uniqueness follows by a covering argument from the nondegeneracy of these configurations. We prove nondegeneracy by hand: Dziobek's relations reduce it to an
eigenvalue bound for a $2\times2$
matrix that depends only on the shape of the configuration. As consequences, every degenerate planar four-body central configuration is concave, and the known
symmetry theorems for kites, isosceles trapezoids and rhombi follow in a few lines.
Read the full abstract ▾
Optimization and Control
Accelerated Algorithms for Stochastic Monotone Inclusions with Fixed Queries
Sucheol Lee and Donghwan Kim
We study acceleration of
stochastic first-order methods for monotone Lipschitz inclusions with a fixed number of oracle queries per iteration, measured by the expected squared residual. Existing methods either have suboptimal oracle complexity, require an increasing number of oracle queries per iteration, or both. Under same-sample oracle access and mean-square Lipschitz
stochastic noise, we develop variance-reduced anchored forward-backward (VRAF), an anytime composite method that makes two oracle queries per iteration. VRAF attains the first log-free $O(1/\epsilon)$ complexity under these oracle assumptions and hence the optimal $O(\sigma^2/\epsilon)$ variance dependence. This, however, leaves open whether the optimal $O(1/\sqrt{\epsilon})$ deterministic term can also be attained. We establish an impossibility result by extending the lower bound of Foster et al. (2019) to
stochastic oracles permitting repeated queries to sampled
stochastic operators, showing that $O(LD/\sqrt{\epsilon}+\sigma^2/\epsilon)$ oracle complexity is unattainable in general. We therefore develop recentered regularized
stochastic extragradient (RRSEG), which attains the near-optimal oracle complexity previously achievable only with an increasing number of queries per iteration, while using a fixed number of queries.
Read the full abstract ▾
Computation and Language
Which the Eye Fears: Writing with Read-Blindness Explains Massive Activations in Transformers
Swagatam Mukhopadhyay, Vishal Vivek Saley, Vraj Parikh et al. (+1 more)
Massive activation features (MAs) in
Transformers are extreme-value residual-stream features that persist across layers despite the model's ability to suppress them. Why do they survive? Our investigation using an operator-level mechanistic analysis of
attention and feed-forward (FFN) blocks reveals that these blocks systematically ignore MA coordinates while reading, but not while writing; creating a read-write asymmetry that blocks corrective feedback while allowing continued accumulation. We find that both
attention and feed-forward layers have this read-blindness, and contribute to the emergence and persistence of MAs. To validate prior work that hypothesized that FFN's amplification abilities is the primary reason for MAs (Sun et al., 2026), we analyze the model checkpoints during learning. Contrary to our expectation, read-blindness emerges before FFN amplification, suggesting that it acts upstream in the MA mechanism. We further contribute gradient analysis to link this behavior to surprising asymmetries in the loss landscape, concluding that the model actively maintains this read-blindness. Finally, we find that removing read-blocking at different locations induces compensatory shifts elsewhere, but MAs still persist.
Read the full abstract ▾
All authors Swagatam Mukhopadhyay, Vishal Vivek Saley, Vraj Parikh, Mausam
Machine Learning
SANTA++: Sampling Attention through Representative Keys
Kyle Lee, Christian Z. Pratt, Ruoyu Fang et al. (+4 more)
Attention often concentrates on a small subset of
tokens in the context, but which subset matters changes from one query to the next. To exploit this changing structure, we introduce SANTA++, a training-free
stochastic attention method that uses representative keys for
memory-efficient selection without scanning the entire key-value (KV) cache. Cached keys are organized into teams, and the query scores one representative from each team to decide which teams to sample. We compute exact
attention scores within the sampled teams and reweight each team's contribution by the inverse of its inclusion probability. This importance sampling correction estimates
attention over the full cache, with a sampling budget that lets us trade
memory reads for accuracy. Remarkably, with 32 or 64 sampled teams, SANTA++ uses 16% to 22% of dense
attention's KV reads and retains 94% to 99% of the dense-
attention baseline's scores on LongBench v2 and HELMET's
retrieval-augmented generation subset, and 85% to 91% on RULER, with Qwen2.5-7B-Instruct at 32K context. With 31 sampled teams, our GPU implementation delivers a $1.69\times$
attention speedup over the dense FlashAttention baseline at 32K context. By reducing the number of cache entries read, SANTA++ in principle complements architectures with compressed KV representations, such as multi-head latent
attention. Our kernels are available at: https://github.com/OPUSLab/santapp-kernel-demo.git.
Read the full abstract ▾
All authors Kyle Lee, Christian Z. Pratt, Ruoyu Fang, Heekyung Lee, Avinash Lohitsa, Ryan Modafe, Kerem Y. Camsari
Machine Learning
Arbitrary-Accuracy Neural Approximation with Optimal Neuron Count and Near-Optimal Bit Complexity
Zilan Cheng, Li-Lian Wang and Zhongjian Wang
We study the minimum number of hidden
neurons required for arbitrary-accuracy approximation of multivariate Hölder-continuous functions on $[0,1]^d$ and the associated encoding complexity. For $d\geq 2$, we construct a fixed, explicitly defined activation function for which a closed-form network with two hidden layers of widths $d$ and $1$ achieves arbitrary accuracy in the uniform norm. We prove that $d+1$ is the exact minimum total number of hidden
neurons among standard feedforward networks with locally integrable activations and affine outputs. We further give a simpler construction using a single elementary activation that combines the floor and exponential functions. This construction requires three hidden layers of widths $d$, $1$, and $2$, only two
neurons above the minimum. If a skip connection is allowed, widths $d$, $1$, and $1$ suffice. These constructions use explicit grid addressing and integer encoding of quantized function values. For a bounded $\alpha$-Hölder class, they require $O(\varepsilon^{-d/\alpha}\log(1/\varepsilon))$ bits, matching the metric-
entropy lower bound up to a logarithmic factor.
Read the full abstract ▾
Computation and Language
Can LLMs Value the Right Evidence? Evidence-Value Misalignment in Dynamic Medical Diagnosis
Kehua Feng, Yunsheng Lu, Yitong Qiao et al. (+5 more)
A correct diagnosis reached from insufficient or misleading evidence can pose a clinical hazard, yet outcome-based accuracy may reward such lucky guesses. We call this mismatch between diagnostic decisions and the value of available evidence Evidence-Value Misalignment (EVM). To disentangle evidential grounding independently from diagnostic accuracy, we introduce MedEVM, a dynamic benchmarking environment comprising 1,050 cases across 24 disease systems. Observations arrive turn by turn, requiring models to continuously calibrate its decision by deciding whether to wait for more evidence or submit a diagnosis. Across 9 LLMs, four interesting patterns are observed. (1) Miscalibrated evidence tracking. Making a diagnosis often fails to calibrate evidence sufficiency, even in more capable models, and even worsens in
reasoning mode. (2) Misaligned diagnosis submission. Confidence in the correct diagnosis often fails to ensure timely submission despite sufficient evidence. (3) Evidence order matters. Reordering the same evidence changes diagnoses even when model confidence remains similar. (4) Misleading evidence remains influential. Added misleading evidence redirects diagnoses even after prior evidence becomes sufficient. We further verify that EVM predicts errors and that preventing premature submission improves accuracy. These findings motivate Evidence-Verified Diagnosis Harness (EVD-Harness). It decouples diagnosis generation from submission through an offline Contrastive Diagnostic Wiki and three online control stages, namely observation management, proposal and witness verification, and diagnosis submission control. Across five LLMs, EVD-Harness improves accuracy by 12.0–51.1 percentage points while mitigating EVM-related failures. Our results demonstrate that verifying evidential support before submission can make diagnostic decisions more reliable.
Read the full abstract ▾
All authors Kehua Feng, Yunsheng Lu, Yitong Qiao, Tiantian He, Lei Liu, Yue Shen, Jian Wang, Jinjie Gu
Number Theory
A difference formula of $p$-adic height pairings via the Bloch-Kato logarithm map
Taiga Adachi, Yu Katagiri and Ryota Shii
The construction of a $p$-adic height pairing for a
geometric $p$-adic representation of the absolute Galois group of a number field depends on a global $p$-adic logarithm and on local splittings of the Hodge filtrations at the
primes above $p$. We study the dependence on these splittings for suitable two-dimensional
symplectic self-dual representations, including self-dual twists of representations attached to even-weight newforms at non-ordinary
primes not dividing the level. We express the difference between the height pairings associated with the two splittings determined by Frobenius explicitly in terms of local Bloch–Kato logarithms. As an application over $\mathbb{Q}$, we prove that at least one of the two cyclotomic $p$-adic height pairings is non-trivial under the additional assumptions that the Frobenius
eigenvalues at $p$ are distinct and the localization map at $p$ from the Bloch–Kato Selmer group is non-zero.
Read the full abstract ▾
Algebraic Geometry
Integral Bott–Chern cohomology of Generalized Calabi–Eckmann threefolds
Federico Thiella
We compute the integral Bott–Chern
cohomology of a class of generalized Calabi–Eckmann
manifolds of complex dimension \(3\). Although these
manifolds are mutually non-biholomorphic, we show that their integral Bott–Chern
cohomology rings are nevertheless isomorphic. A cohomological distinction is then obtained via a characteristic map in
cohomology, connecting singular, Dolbeault, and integral Bott–Chern
cohomologies.
Read the full abstract ▾
Instrumentation and Detectors
Electrothermal behavior of superconducting nanowires buried in a commercial CMOS process
Mohamed Gharib, Leonid Popryho, Salma Abdelzaher et al. (+14 more)
Superconducting films native to a commercial CMOS process enable co-integration of superconducting devices (high-kinetic-inductance elements and nanowire detectors) with cryogenic CMOS on one monolithic die. Such films are buried in the front-end-of-line, beneath the back-end-of-line metal stack, an environment that might quench a self-heated hot spot. Current-voltage characteristics are reported for titanium nitride (TiN) nanowires in the polycrystalline-silicon resistor layer of a 22 nm FD-SOI process, and measured retrapping currents are compared with thermal models for longitudinal conduction to the contacts and interfacial
phonon cooling into the stack. The wires are strongly hysteretic, with switching-to-retrapping current ratios I_sw/I_r of 7.4-9.8, sustaining a normal hot spot. For two wires of length ell = 50 micrometers and widths W = 0.50 and 1.00 micrometers, over bath temperatures T_b = 0.15-1.19 K, I_r follows the interfacial form (T_hs^n - T_b^n)^(1/2) with n = 4 (T_hs, the effective hot-spot temperature), and is strongly inconsistent with the (T_hs - T_b)^(1/2) constant-kappa longitudinal-conduction limit; free-exponent fits give n = approximately 4.8-5.0. An effective cooling coefficient Sigma_K = 7.9(3) W/(m^2 K^4) describes the data: 31(21) times below the acoustic- (diffuse-) mismatch value for a single clean interface, and approximately 12 times below the lowest previously reported comparable nanowire coefficient. The model-based film thermal conductivity is kappa = approximately 20 mW/(m K). Across lengths ell = 1-50 micrometers, I_r matches neither ell^(-1) of pure longitudinal conduction nor ell^0 of pure interfacial cooling; a two-channel model reproduces I_r(ell) with crossover length ell* = approximately 3.5 micrometers predicted, to an O(1) prefactor, from Sigma_K and kappa.
Read the full abstract ▾
All authors Mohamed Gharib, Leonid Popryho, Salma Abdelzaher, Tejas Guruswamy, Tomas Polakovic, Umeshkumar Patel, Orlando Quaranta, Thomas Cecil, Clarence Chang, Yu-Sheng Chen, Thomas H. Swift, Grayson M. Noah, Alberto Gomez-Saiz, John J. L. Morton, M. Fernando Gonzalez-Zalba, Antonino Miceli, Inna Partin-Vaisband
Artificial Intelligence
RIDE: Reference-Anchored Inference-Time Diffusion Editing for Scaffold Hopping
Ruoxi Gao, Frazier N. Baker, Trieu Nguyen et al. (+1 more)
Scaffold hopping is a critical task in drug discovery, which seeks to discover new, structurally distinct molecules that share key functional groups and similar 3D shape with a reference binding ligand. Existing
diffusion-based scaffold hopping methods formulate the problem as conditional generation of scaffolds given the functional groups. However, they lack a principled mechanism to jointly enforce 2D structural novelty and preserve the 3D shape of the reference ligand. Here, we introduce RIDE, a Reference-anchored Inference-time
Diffusion Editing framework for scaffold hopping. RIDE recovers the reference
diffusion noise trajectory conditioned on the binding pocket and functional groups, selects an optimal trajectory segment for editing via noise perturbation, and conducts a value-guided scaffold sampling to generate new scaffolds. Extensive experimental results demonstrate that, compared to baselines, RIDE consistently generates scaffolds with lower 2D similarity and higher 3D similarity to the reference, with an average improvements of 11.7% and 7.3%, respectively. Further analysis reveals that RIDE can accommodate various reward functions, and can preserve 3D similarity even when this is not explicitly included in the reward. Two case studies illustrate RIDE's ability to generate distinct scaffolds with different structures and properties, and its ability to introduce substantial 2D variation while maintaining very high 3D similarity. RIDE is publicly available at https://anonymous.4open.science/r/RIDE-C8A0.
Read the full abstract ▾
All authors Ruoxi Gao, Frazier N. Baker, Trieu Nguyen, Xia Ning
Machine Learning (Statistics)
Elicitation and Decision Geometry in Single-Index Bandits
Sakshi Arya and Cheng Soon Ong
We study two-arm contextual bandits with arm-specific single indices and a shared unknown monotone link. Monotonicity makes the optimal action depend only on the contrast between the index directions, hence arm-specific reward functions need not be estimated. We introduce Natural Boundary Learning (NBL), a greedy procedure that uses a sequential Stein contrast to learn the optimal boundary directly, without estimating the reward functions or the common link. We characterize the local Riemannian dynamics of NBL through a decision stability coefficient balancing arm separation, link geometry, and the context distribution. We show that this stability is connected to the elicitation geometry of the underlying convex potential. Under local decision stability, NBL contracts toward the optimal boundary and achieves $O(\log n)$ expected regret. Numerical experiments illustrate the predicted stability regimes and compare NBL with a parametric greedy
benchmark under link misspecification.
Read the full abstract ▾
Machine Learning
Cartridges++: KV Cache Compression without Off-Context Derailment
Sonia Laguna, Joao Monteiro, Marco Cuturi et al. (+2 more)
Serving long documents to a Large
Language Model (LLM) repeatedly is expensive: computations grow with context length, and the
memory footprint of the key-value (KV) cache balloons. Compressed KV (CKV) representations aim to mimic the cache of a document and are typically computed once and for all, ahead of inference time. Methods to obtain CKVs range from drop mechanisms that reduce their number of columns, to learned approaches. Among the latter, Cartridges have emerged as a leading compression method, learning compact KV representations through
distillation on relevant Q/A pairs. While existing evaluations focus primarily on whether Cartridges and other CKVs yield approximately similar responses to document-related, on-context queries, we investigate the crucial deployment question of whether they can handle off-context queries, something the native KV representation is particularly good at, thanks to the mechanics of
attention. We observe a fundamental trade-off: while Cartridges perform better for on-context queries, heuristic-variants preserve better the original LLM's ability to operate off-context. We measure this through their capability to avoid context contamination in their response, retain general knowledge, and follow instructions. We propose Cartridges++, simple modifications to cartridges that retain off-context abilities at small or negligible cost. The router variant decides at inference time whether the query should use the learned long-context
memory, while the data-mixing variant allocates a small fraction of training Q/As to queries outside the reference long document. Our study shows that assessing CKVs on document utility alone can mask substantial degradation in broader model capabilities, yet those issues can be fixed with benign changes to CKV inference or training.
Read the full abstract ▾
All authors Sonia Laguna, Joao Monteiro, Marco Cuturi, Pierre Ablin, Eleonora Gualdoni
Machine Learning
Attention Graphons: A Graph Limit Perspective on Graph Transformers
Caio F. Deberaldini Netto, Moshe Eliasof and Luana Ruiz
Graph
Transformers produce, for each
attention head, a dense $n\times n$
matrix of learned pairwise interactions. We ask a fundamental question: do these
attention-induced graphs converge to a stable limit object as $n$ grows, or does the learned interaction pattern remain unstructured and size-dependent? We answer this using dense graph limit theory, treating each
attention matrix as a finite sample from an underlying kernel—an
attention graphon—and studying concentration around this limit under the cut-distance. We derive a worst-case variance bound requiring no assumptions on the graphon, and a sharper regularity-aware bound based on nonparametric estimation theory. To operationalize the theory, we propose a canonicalize-then-block-average pipeline for estimating
dataset-level
attention graphons, and a variance-based diagnostic for testing whether
attention admits a stable continuum description. Experiments across multiple graph
benchmarks show that learned
attention stabilizes to
dataset-specific graphon structure on several
datasets; that empirical cut-distance and cut-norm variance decreases with $n$ consistent with our bounds; and that
attention graphons transfer to larger graph sizes with error decreasing in $n$.
Read the full abstract ▾
Robotics
CollisionSplatting: Collision-Aware Motion Planning in 3DGS Scenes with Image-Conditioned Objectives and Adjustable Conservatism
R. Khorrambakht, Joaquim Ortiz-Haro, Stephan Weiss et al. (+1 more)
Incorporating dense visual information into motion planning remains challenging, as
geometric planners rely on abstracted scene representations that discard visual richness, while learned visual models often lack
geometric interpretability and computational efficiency. This paper introduces CollisionSplatting, a simple, modular, GPU-accelerated, probability-inspired distance metric with tunable conservatism that operates directly on standard 3D Gaussian Splatting (3DGS) scenes. When combined with learned image-conditioned reward functions, this metric enables joint
geometric and visual planning by unifying collision-aware costs with image-space objectives. We integrate the metric into GPU-accelerated Model Predictive Path Integral (MPPI) and Rapidly-Exploring Random Tree (RRT) planners, and show on-par or better collision-classification performance compared to representative baselines while achieving substantially higher collision-checking throughput and significantly lower VRAM usage. Finally, we demonstrate the effectiveness of our metric in real-world
vision-guided
navigation and manipulation tasks, highlighting 3DGS as a practical bridge between rich perception and real-time motion planning.
Read the full abstract ▾
All authors R. Khorrambakht, Joaquim Ortiz-Haro, Stephan Weiss, Ludovic Righetti
Artificial Intelligence
From cacophony to hierarchy: a principled framework for assessing AI consciousness
Shamil Chandaria, Arvo Muñoz Morán, Fernando Rosas et al. (+11 more)
The question of AI consciousness is one of the most urgent pre-emptive problems in philosophy and computer science, yet progress is hampered by a cacophony of competing theories that often talk past each other. Separating the hard problem from the mapping problem allows the deepest metaphysical disagreements to be set aside: granting that experience supervenes on a system's organisation, the tractable question becomes at which grain of description that supervenience base sits. We extend Marr's three levels of analysis into a five-level hierarchy of functional descriptions (behavioural, computational, intrinsic causal-structural, organismic, and organism-environment) grounded in supervenience, coarse-graining, and multiple realisability. The major theories of consciousness are positioned within this hierarchy according to which level they take to be critical, and for each level we develop operationalisable indicators and assess current AI systems against them. A
Bayesian model then combines theoretical credences with indicator evidence into an overall credence in a system's capacity for consciousness. In illustrative assessments, the verdict for current LLMs is driven as much by where theoretical credence is placed as by how the evidence is read: under different stipulated readings and credence distributions, assessments range from below 0.01 to roughly 0.8, showing sensitivity to assumptions. Finally, the consciousness indicators at each level closely overlap with the architectural features needed for general intelligence, suggesting that increasingly capable AI may become a stronger candidate for consciousness. The framework supports a structured agnosticism, in which theoretical commitments are made explicit, credences are updated as evidence accumulates, and assessments take the form of aggregated probabilities rather than verdicts.
Read the full abstract ▾
All authors Shamil Chandaria, Arvo Muñoz Morán, Fernando Rosas, Anil Seth, Henry Shevlin, Marcus Hutter, Thore Graepel, Adam Bales, Iulia Comsa, Murray Shanahan, Ruben Laukkonen, Morten Kringelbach, Chris Frith, Shane Legg
Combinatorics
The classification of maximum scattered linear sets of $\mathrm{PG}(1,q^5)$
Giovanni Longobardi and Valentina Pepe
We classify the maximum scattered $\mathbb{F}_q$-linear sets of $\mathrm{PG}(1,q^5)$, proving that every such set is of pseudoregulus type or of Lunardon-Polverino type. Building on the reduction obtained by Lia, Longobardi and Zanella in [S. Lia, G. Longobardi and C. Zanella, Towards the classification of maximum scattered linear sets of $\mathrm{PG}(1,q^5)$, Algebraic Combinatorics 9 (2026), 327-355], we show that the two remaining candidate families contain no scattered linear sets. Our approach reduces these candidates to two normal forms and establishes their non-scatteredness through the existence of rational points on associated algebraic varieties.
Read the full abstract ▾