Papers

Segment routing (SR) involves routing requests through shortest paths or a limited number of segments, which are themselves the shortest paths between their endpoints. While this flexibility makes segment routing attractive for traffic engineering (TE) in IP/MPLS networks, it becomes challenging to update when some links become unavailable (e.g., maintenance) and routing decisions must adapt while limiting the number of configuration changes between two consecutive periods. We address the problem of dynamically selecting SR-TE configurations for multiple traffic demands with the objective of minimizing the maximum link utilization (MLU). In addition, a reconfiguration budget is imposed in order to limit the number of segments modified. Based on this representation, we develop a MLU column-generation optimization framework. It allows to jointly consider traffic distribution, SR configuration complexity, and temporal reconfiguration costs in time-varying IP/MPLS networks. The numerical results are obtained using realistic datasets provided by Orange, with topologies containing up to 1,263 nodes and 15,000 traffic demands. The optimality gap (i.e., accuracy) is below 5% in 80.45% of the 133 instance-time evaluations.

Machine-learning classification is now used across medicine, the social sciences, economics, education and engineering, very often by researchers who do not program. Two errors recur in such work: preprocessing learned on all rows before cross-validation (data leakage), and reporting the cross-validation score of the classifier that won a comparison (selection bias). Both make published scores optimistic. We present EasyClassifier, an open-source Python package that guides a user from a CSV or Excel file to a finished analysis through a sequence of plain-language, multiple-choice questions, with no code. It makes the correct procedure the default rather than an option: every preprocessing step is learned inside each training fold; the selected classifier receives a separate final score from nested cross-validation (up to 2,000 rows), or an untouched 20% test set; classifiers run with fixed default parameters and fixed random seeds; and every run writes a report with a ready-to-adapt Methods paragraph, the references to cite, and figures prepared for publication. On ten public datasets from seven fields and a random-label control, split in half so that one half served as an untouched external test, the common practice overstated balanced accuracy by 2.9 percentage points on average (up to 9.9 on small data), whereas the score EasyClassifier reports had a mean signed bias of +0.7 points and a mean absolute error of 2.3 points (one-sided Wilcoxon p = 3.2 x 10^{-4}, 30 runs). On random labels it reported 48.8% balanced accuracy against a chance level of 50%. EasyClassifier is available under the MIT license from PyPI (pip install easyclassifier), GitHub, and Zenodo.

In recent years, flow-matching models have produced significant improvements in zero-shot text-to-speech synthesis. Conditioned on an audio prompt and text, these models learn a velocity field and generate speech by iteratively solving an ODE. During inference, the solver evolves a single state spanning both the prompt and the region to be generated, although only the generated region is ultimately retained. The discarded prompt state, however, still matters; its intermediate values influence generation through the velocity field that couples the two regions. As sampling continues, this state can drift away from the prescribed conditional path, introducing a discrepancy into subsequent generation updates. Unlike the unknown generated trajectory, the prompt path is available in closed form from the reference audio and the initial noise. We exploit this observation with Prompt-Consistency Inference (PCI), a training-free rule that restores the prompt block to its analytic value before each velocity evaluation, while leaving the generated block unchanged. PCI improves speaker similarity and intelligibility across the evaluated flow-matching TTS backbones without additional network evaluations. Our ablation studies further show that PCI keeps post-step prompt discrepancies smaller and that corrections covering the later sampling stages recover much of the observed similarity gain.

Self-evolving large language model (LLM) systems repeatedly propose, evaluate, and incorporate updates to prompts, skills, or other persistent artifacts. Despite their growing effectiveness, these systems typically operate under a predetermined iteration or compute budget, without a principled criterion to determine when further evolution is no longer worthwhile. This can lead to two undesirable consequences: unnecessary computation after performance has saturated and the risk of returning late updates that overfit or exploit the evaluation signal. These issues motivate us to study two fundamental questions: when should a self-evolving system stop, and what should it output once it stops? We address the first by formulating an online sequential testing problem and constructing an anytime-valid restart detector using the per-item paired evaluation outcomes already produced by self-evolving LLM systems. We address the second by formulating a change-point estimation problem and using the estimated transition to select an earlier artifact for output. The resulting procedure is plug-and-play and requires no modification of the underlying self-evolving algorithms. Across two self-evolving frameworks, three LLM model families, and five benchmarks, our method substantially reduces computation costs while maintaining comparable unseen-test performance. For example, on SearchQA with SkillOpt and DeepSeek V4 Flash, our method stops at round 4 rather than the full budget of 40, reducing token usage by 91.6% while achieving 82.43% unseen-test accuracy versus 82.00% under the full-budget run.

Even a few action perturbations can substantially degrade the performance of a deployed decision policy. Certifying the resulting return loss is challenging in stochastic environments, where returns vary even without an attack. We introduce FoSeRL, a framework for certifying deployed RL policies against precommitted, temporally sparse action attacks. The deployed policy is unchanged, with no smoothing or retraining. Certification requires a resettable simulator supporting shared randomness and independent one-step successor queries, but no analytical dynamics model. FoSeRL certifies that an attacked episode loses no more than a prescribed amount of return relative to the same episode unattacked, with at least a target probability and at a user-specified confidence level. Both runs share the initial state and randomness, so the measured loss reflects the attack, not the episode; carrying the running return gap as a state coordinate makes it the terminal value, reducing trajectory-level certification to terminal safety. Time-dependent barrier conditions on the augmented state bound the terminal failure probability: satisfied exactly, they certify every admissible precommitted attack; learned from sampled trajectories and verified on held-out data, they certify the same guarantee under a specified attack-episode setting. Across six stochastic continuous-control environments and three RL policy families (TD3, SAC, and PPO), FoSeRL certifies non-trivial cardinality–magnitude robustness frontiers, achieves substantially larger certified budgets than policy smoothing, and reveals marked robustness differences among policies with comparable nominal performance.

Autoregressive generation in Large Language Models (LLMs) is constrained by the memory and computational demands of attention mechanisms. Sparse attention methods mitigate this cost by selecting only high-probability entries of the attention matrix. We observe that in many such methods, this renders the probability-value multiplication negligible, shifting the bottleneck to the query-key step. Key heads can therefore be reduced to accelerate inference, while retaining more value heads preserves capacity with limited additional decoding cost. We introduce Sparse Asymmetric Group-Query Attention (SAGA), which decouples key and value head counts to exploit this principle, and pair it with approximate top-N (Atop-N) attention, a simple sparse attention method designed to study the interaction between sparsity and head-count asymmetry. We formalize the benefits of this asymmetry theoretically and validate them empirically through latency measurements and quality evaluations on models up to 1.5B parameters. Together, SAGA and Atop-N achieve end-to-end decoding speedups exceeding $2\times$ over our full-attention GQA baseline at long contexts. Models trained from scratch with SAGA nearly match the quality of comparable GQA variants on the evaluated benchmarks. To facilitate adoption, we introduce an efficient fine-tuning method that converts pretrained models to the SAGA architecture, enabling practitioners to benefit from our approach without costly retraining.

We study model-free online reinforcement learning (RL) for episodic tabular Markov decision processes, focusing on both gap-dependent regret and policy switching cost. While fine-grained gap-dependent analysis has been established for model-free RL algorithms using Hoeffding-type exploration bonuses, such results for model-free algorithms with variance-based exploration bonuses remain unknown, despite their superior worst-case and coarse-grained gap-dependent guarantees. In this paper, we resolve this open problem by establishing the first fine-grained gap-dependent regret upper bound for UCB-Bernstein+, a refined UCB-Bernstein algorithm, in variance-aware model-free online RL. Moreover, by integrating a stage-wise policy update design into our fine-grained framework and using refined variance-based bonuses, we achieve the best-known gap-dependent local switching cost to date. In addition, our analysis yields improved worst-case guarantees for both regret and local switching cost over the original UCB-Bernstein algorithm. Numerical experiments further demonstrate that UCB-Bernstein+ achieves favorable empirical performance in both regret and local switching cost.

Beyond scaling their parameters and data, large language models currently gain versatility on new problems along a single axis: the tokens they spend on chain-of-thought (CoT). We investigate whether dynamic routing programs, which execute a subset of the model's layers or iterate some of them, can open a second axis of test-time adaptation, complementary to CoT and free of any gradient update. Prior work showed that such programs exist and bring accuracy and efficiency gains on problems similar to those they were trained on; we ask whether they can also be identified rapidly, from a handful of demonstrations (3 or 10), by a strategy that transfers across models and tasks without training. First, we find that strategies that select programs by the probability they assign to the demonstrations' labels, arbitrated by the model's own confidence, bring consistent gains: on average over the 49 tasks of MMLU and substantially on four of seven models, and most of all on far out-of-distribution tasks such as ARC-AGI, where programs double the accuracy of a 7B model whose CoT fails. Second, on MMLU across the seven post-trained models, routing complements CoT in practice: the two succeed on different queries, and their composition exceeds CoT alone. Despite these gains, our analyses show that confidence-based selection leaves much of the potential untapped, in two places in particular: (1) in surfacing the routing potential that is already present early in pre-training but becomes harder to select after post-training, and (2) in making models robust to the refinements routing introduces, since unsuccessful routes tend to drive the residual stream out of the distribution that the following layers expect. Together, our results point to dynamic routing as a paradigm for extending the plasticity of existing and future LLMs in rapid test-time adaptation.

We prove a complexity dichotomy theorem for $\mathrm{Holant}^*$ problems over complex-valued symmetric constraint functions $\mathcal{F}$ on domain size $3$. We give a decidable tractability criterion and prove that if $\mathcal{F}$ satisfies the criterion, then $\mathrm{Holant}^*(\mathcal{F})$ is solvable in polynomial time, and otherwise it is #P-hard. This is the first Holant dichotomy for a set of complex-valued constraint functions on higher domains. We show that complex-valued constraint functions have a rich structure not observed in real-valued constraint functions. This structure is only revealed when we analyze them in a bipartite Holant setting with a non-standard bilinear form and provides the backbone to the proof of the dichotomy. We use group actions and the spin representation of $\mathrm{SL}(2, \mathbb{C})$ in $\mathrm{SO}(3, \mathbb{C})$ and the generalized orthogonal group to facilitate this analysis. We also introduce $\textit{frames}$ and $\textit{shells}$. Frames linearize the group action and provide a unified framework for proving #P-hardness when used in conjunction with shells. Furthermore, we characterize the lower dimensional constraint functions by $\textit{annihilators}$, which makes it possible to analyze $\textit{essentially Boolean domain}$ functions in domain size $3$.

Routine complete blood counts (CBCs) could yield new biomarkers, but the private records needed to evaluate candidates cannot be shared with frontier language model agents that excel at discovery. We distilled the evidence held in the Clalit Health Services panel of over 5.4 million patients into a released scoring tool: for each of 13 immune-mediated diseases, a graph attention network was trained inside the data boundary to predict the case-control AUC of candidate CBC expressions, and only the trained weights were released. The tool grounds an agent's propose-score-refine loop in real-world data without exposing any patient data. In external validation, agent-discovered expressions improved on their literature-seeded starting points by a median of 4.18 AUC percentage points, and across three independent cohorts, reranking the candidates of three frontier research tools improved on their first choices in most comparisons, with gains that varied by cohort. The released scorer supports privacy-preserving biomarker hypothesis generation.

In life-or-death situations, a benevolent lie may appear more moral than telling the truth. Yet such lies can backfire, producing unintended and sometimes fatal consequences. This tension, famously disputed by Kant and Constant in 1797, applies not only to lying but to assertive speech acts in general, raising the question of which utterance should be chosen when moral stakes are high. We present a logical framework for the ethical evaluation of speech-act utterances based on agents' beliefs. Implemented in Answer Set Programming (ASP), the framework assesses utterances under deontologism, consequentialism, and principialism, and is illustrated on Sartre's The Wall (1939), a reworking of that controversy in which lying leads alternately to rescue and to death. Our setting is general by design: as two variants show, accommodating a new moral situation amounts to adjusting parameters, not rules.

Simulation is a fundamental computational primitive in reinforcement learning (RL), yet conventional simulation explicitly generates individual trajectories even when downstream procedures use only aggregate statistics. To address this, we develop an exact fast-simulation framework for finite-horizon tabular Markov decision processes. Our framework has two complementary modes. In direct batch simulation, a batch is represented by its aggregate Markov flow. With sufficient parallel simulation resources, this flow can be obtained by trajectory aggregation; when such simulation is unavailable or costly but the initial state and transition distributions are directly accessible, we instead generate an identically distributed flow through forward Markov-flow sampling without materializing individual trajectories. The latter reduces the simulator-side computational dependence on batch size $m$ from $O(m)$ to $O(1)$. In adaptive batch simulation, when batch length is determined by a data-dependent condition, exact multivariate-hypergeometric splitting recursively refines a candidate Markov flow while preserving the conditional law, reducing the cost dependence on $m$ from $O(m)$ to $O(\log m)$. Together, these modes accelerate simulation by keeping trajectories aggregated whenever possible and refining flows only when required to locate data-dependent boundaries. The framework applies broadly across simulator-based, offline, and online batch or stage-based RL, as illustrated with representative algorithms from each setting.

When a team of robots delegates planning and mutual trust to LLM agents, a single compromised robot can corrupt the shared outcome. We study this threat in a grounded task: a multi-robot survey in which measurements can be verified against the physical world, but every verification costs budget that would otherwise advance the mission. We treat the compromised robot as a stealthy adversary in the system-theoretic sense: it is limited not by an energy bound but by the team's own detectors. We then derive two bounds. First, the probability that the adversary's reports are verified is bounded below in terms of the degrees in the communication graph and the verification budget. Second, the map error caused by any stealthy adversary is bounded above by the value of a linear program over the adversary's bias distributions; its solution is an exchange rate between stealth budget and damage: below a critical verification level the worst stealthy attack tells rare, full-magnitude lies on the records least likely to be verified, and above it the better purchase is small biases hidden in the noise. In experiments where the honest robots are LLM agents, both bounds hold at the budget the attack actually spent. The experiments also show that which records an LLM robot re-checks is unbiased, but how much it re-checks is unpredictable.

Accurate thyroid ultrasound segmentation is often challenged by low contrast, speckle noise, and unclear boundaries. Although recent methods have improved segmentation accuracy, many rely on resource-intensive architectures or lack explicit integration of multiscale features with global biomedical visual guidance. In this paper, we introduce ThyCLIPNet, a lightweight semantic-guided hybrid encoder-decoder framework that integrates BiomedCLIP-derived biomedical semantic guidance into a lightweight multi-scale CNN segmentation pipeline. The encoder integrates MobileNetV2 with efficient channel attention, while atrous spatial pyramid pooling and a custom convolutional block attention module enrich bottleneck features. The decoder combines hierarchical skip connections and lightweight attention refinement with a BiomedCLIP-guided gated fusion pathway that projects vision-only global biomedical embeddings into decoder feature space and selectively integrates them through semantic-local fusion and spatial gating. To the best of our knowledge, ThyCLIPNet is among the first lightweight thyroid ultrasound segmentation frameworks to use BiomedCLIP's vision encoder alone for image-only global semantic guidance without text prompting. Experiments on TG3K, TN3K, DDTI, and PKTN achieve dice similarity coefficients of 96.22%, 87.58%, 84.73%, and 80.70%; intersection over union scores of 92.72%, 77.91%, 73.51%, and 67.64%; and 95th-percentile hausdorff distances of 3.75, 16.38, 18.23, and 10.86, respectively. ThyCLIPNet uses 8.55M parameters and 22.99G FLOPs. Overall, the results support integrating global biomedical semantic guidance with lightweight multi-scale CNN representations for robust and computationally efficient thyroid ultrasound segmentation. Source code: https://github.com/Tasnim-Jahan/ThyCLIPNet. [Abstract shortened for arXiv. See PDF for full abstract.]

This paper illustrates the properties of a new CBF-governor that enables simultaneous input and output constraint satisfaction for a class of multi-input linear time-invariant systems with state feedback and integral action. This governor is designed so as to modify the reference command while preserving the nominal feedback controller. Necessary and sufficient conditions are derived in a companion paper under which the governor is shown to be feasible at every state of a prescribed operating set and ensures simultaneous input and output constraint satisfaction, forward invariance, and bounded closed-loop solutions. These theoretical results are illustrated in this paper through several numerical examples. In each of these examples, we show how the goals of the CBF-governor are met and corroborate the corresponding necessary and sufficient condition. The computational burden associated with the proposed CBF-governor is also articulated, with its online component comparable to either that of a QP solver or determined using an exact closed-form solution. The offline component is related to the checking of the SIOCF condition which is of $O(N)$.

Force-aware manipulation typically relies on specialized force or tactile sensors. We show that force-aware manipulation can instead be achieved through visual force prediction from the deformation of a compliant Fin Ray gripper. Our approach trains two models. First, we train a visual force estimator on calibration data and use it to annotate task demonstrations with force estimates. Second, we train an action–force proposal policy on these force-augmented demonstrations to jointly generate candidate robot actions and their associated forces. At test time, we sample candidate actions and the forces they are expected to produce, then execute the action whose predicted force is closest to a target from the demonstrations. We evaluate our approach on berry picking, empty-can grasping, in-hand reorientation, and plug insertion. Our results show that visual force prediction can guide inference-time action selection for contact-rich manipulation without requiring force or tactile sensors at deployment.

Early-stage computational drug discovery requires coordinating heterogeneous scientific tools across multi-step workflows. We present a lightweight, tool-augmented framework in which a locally deployable compact language model plans calls to 18 modular tools. A Unified Molecular Schema maintains shared molecular records, while a plug-in interface supports tool replacement and extension. We construct 1,263 manually refined query-plan pairs through workflow-graph path coverage and apply LoRA fine-tuning to three compact model families. Under the query-level split, all fine-tuned models generate fully parseable and schema-compliant plans on 47 held-out cross-group queries. Llama 3.2-3B achieves a tool-selection F1 of 0.998, sequence exact match of 0.979, and argument F1 of 0.960. Under the stricter workflow-grouped split, which excludes identical ordered tool sequences across partitions, sequence exact match reaches 0.452 to 0.548, highlighting the remaining difficulty for compact models in generating complete workflow paths unseen during training. These results demonstrate the feasibility of compact, locally deployable planning while identifying compositional generalization as an important direction for further improvement.

Reasoning models can remain capable of solving a task while still defaulting to cheaper but misleading shortcuts. This creates a central oversight problem: when a model gives an answer with plausible but incomplete reasoning, can an overseer determine whether that output should be trusted? To study this problem, we introduce PyINE, a framework for scalable elicitation and oversight using instrumented Python programs as a verifiable execution substrate. In PyINE, programs define task environments, execution traces provide authoritative labels for outcomes and intermediate facts, and task variants can be generated mechanically rather than through static human annotation. We instantiate the framework in PyINE-v1, a first release built from nearly one million deterministic execution traces and over 500,000 matched LLM-generated code variants used for counterfactual evaluation. Using standard RL with verifiable rewards on cue-varied tasks, we train a shortcut-following model that improves substantially at predicting execution outcomes while still making systematic errors when misleading human-facing cues conflict with the program's realized behavior. We then evaluate activation probes, trained text classifiers, prompted judges, and a lightweight debate protocol as overseers of this model. We find that performance pooled at the dataset level can hide weak coverage of the failures that matter most: cheap learned overseers often miss rare shortcut-driven errors, while stronger model-based checks are more balanced but substantially costlier and harder to turn into reliable thresholded decisions. PyINE-v1 turns this failure-mode coverage problem into a reusable experimental setting for developing oversight methods that are verifiable, failure-mode-aware, and cost-sensitive.

A pedestrian with a phone could be shown where nearby people will be in the next few seconds, if the forecast stays on the ground while the phone moves, is calibrated, and needs only a monocular camera and visual-inertial odometry (VIO). We build and evaluate such a system. People are detected, lifted onto the floor by ray-plane intersection, tracked in a gravity-aligned metric frame, and forecast as per-step probability maps by a small U-Net trained with a negative log-likelihood (NLL) loss on bird's-eye (SDD) and first-person (EgoTraj-Bench) trajectories. On handheld ADVIO recordings, vertical VIO drift and the user's own changes of level silently rescale monocular ground positions (by 87% within 90 s on one clip; on another, all tracks are lost for the last 31% of the clip); keeping the camera's height above the floor constant under a low-pass-filtered altitude avoids this, though it lags on escalators. On the SDD and EgoTraj-Bench test splits, the final forecaster lowers the NLL at 4.8 s by 1.51 and 1.37 nats relative to a fitted constant-velocity Gaussian. Lacking ground truth for people in handheld video, we score forecasts against the tracker's own later raw measurements. In an internally pre-registered evaluation on seven held-out clips, the final forecaster's NLL is lower than the benchmark-fitted baseline's at 1.2, 2.4 and 4.8 s (by 0.17, 0.23 and 0.47 nats; 95% intervals over people exclude zero), but by less than half as much as on the development clips. Exploratory analyses cut both ways: resampling clips instead of people widens the intervals to include zero at 1.2 and 2.4 s, and once both forecasters are recalibrated on the development clips the network is significantly better only at 1.2 s; but two clips run with ADVIO's reference poses favour the network much more when re-run with the phone's own poses. We discuss what such self-consistency scores can and cannot show.

Concerns are growing about the water use of data centers. To quantify the scales, drivers, and possible trajectories of this water use, analytical estimation is required. However, current estimation methods and results vary widely due, in part, to a lack of established best practices. This variance is illustrated with a U.S. case study demonstrating a roughly 24-fold difference in possible outcomes based on observed U.S. study design variations. To address these challenges, this review synthesizes 14 best practices for advancing the science and utility of data center water use estimation. These proposed best practices can also be used as critical questions for stakeholders when evaluating the quality of any study. Using a structured rubric, it further assesses the literature to date (31 studies) against these best practices. The assessment reveals important opportunities for improvement and standardization in future study scopes, clarity, methods, and scientific knowledge building utility. It further finds that, while many studies provide reasonable transparency, scope definitions, replicability, and uncertainty treatment, many fall short in scientific justification, data representativeness, transparency on hydropower assumptions, terminological consistency, and meaningful discussions of limitations and future work. Recommendations are proposed for analysts, stakeholders, and policymakers to leverage these findings to help build a vibrant data center water use research community. Finally, this review can also be used as a basic primer for understanding data center cooling configurations and water use among analysts, policymakers, the media, and the public.

Safety filters for large-scale multi-agent systems can be formulated at the population level through mean-field control barrier functions (MF-CBFs). Most existing formulations are first-order, requiring the control input to appear explicitly in the first time derivative of the safety functional; higher-order extensions have only very recently been proposed for a single, directly actuated population. We consider instead the case in which the population to be kept safe is not directly actuated, but is transported by a field generated by a second, controlled population. For this setting, we develop a higher-order MF-CBF framework for density dynamics with safety constraints of relative degree larger than one. Our main result is that indirect actuation of this kind does not destroy the control-affine structure of the safety condition: the control-dependent part of the second time derivative of the safety functional is obtained in closed form and is linear in the control velocity field, with a kernel given by the actuated density times the spatial gradient of a first variation. Relative-degree-two safety constraints can therefore still be enforced through a convex quadratic program. We then specialize the framework to a leader-follower density control problem in the presence of dangerous regions, where safety is ensured by keeping the overlap of each population with those regions below a prescribed threshold. We show that the follower safety functional has relative degree two with respect to the leader control field, while the leader safety functional has relative degree one. We further characterize exactly when the two simultaneous constraints are infeasible, introduce an adaptive-gain mechanism to mitigate potential conflicts, and bound the tracking errors induced by the safety correction. Numerical simulations in cluttered environments illustrate the resulting safety-performance trade-off.

In worker-equipment proximity monitoring, video transformers are widely used for collision anticipation and have demonstrated strong performance. However, their accuracy comes with substantial computational demands, creating a tension with the need for low-latency inference on mobile robots and the pursuit of lower-carbon computation in construction. To address this, this study investigates where computation within an established video transformer is redundant and whether that redundancy can be removed without materially degrading predictive performance. Using VideoMAEv2-Base on the Nexar Collision Prediction dataset, we first examine how collision-relevant information evolves across network depth and then investigate two complementary forms of redundancy: structured capacity redundancy in multilayer perceptrons (MLPs) and spatiotemporal redundancy in the token stream. Linear probes show that interpretable motion cues, including flow magnitude, looming, and approach versus retreat, are most accessible at intermediate layers, whereas collision-label discrimination strengthens toward the final layer. Token redundancy is axis-specific: adjacent temporal-token similarity reaches 0.970 in later layers, while spatial similarity falls to 0.297, indicating substantially greater redundancy across time than across space. Exploiting this asymmetry, temporal token merging reduces backbone computation from 356.99 to 178.50 GFLOPs and latency from 12.69 to 7.21 ms per clip, a 1.76x speedup, while mean average precision changes only from 0.7478 to 0.7443. Importance-guided retention of 50% of MLP units preserves an AUC of 0.753, compared with 0.529 under matched random retention, and reveals that pruning alters score calibration before discriminative ranking collapses. These findings establish a new pathway for pursuing faster algorithms through targeted temporal token compression and neuron pruning.

Full actuation is commonly valued for independent position and attitude tracking. This work establishes a complementary principle: when only a position trajectory is prescribed, minimizing a positively homogeneous allocation cost selects underactuated-like behavior on a fully actuated multirotor. This cost class includes norms and power laws, including standard minimum-norm pseudoinverse allocation. We use fiberwise infimization to transport the actuator-output cost to wrench space, yielding a homogeneous infimal allocation cost. Restriction to unit zero-moment forces defines a directional landscape whose minimizing directions are independent of force magnitude. We show that aligning any fixed minimizing direction with the required world-frame force minimizes instantaneous and integrated zero-moment allocation costs. The resulting body force remains in a fixed one-dimensional subspace without transverse components. We derive an exact optimality-gap decomposition quantifying the additional cost of every other attitude trajectory and characterizing the minimum-cost class. When a disturbance or prescribed-motion change modifies the required-force direction, a fully actuated platform can immediately generate the modified force using transverse authority and reorient toward the new minimum-cost family. A structurally underactuated platform must reorient before generating that force. Thus, homogeneous allocation-cost minimization explains why optimal nominal motion can be underactuated-like, while full actuation remains valuable for disturbance rejection and transient force generation.

Structured prediction tasks pose unique challenges for in-context learning (ICL): their compositional outputs require modeling fine-grained, token-level patterns that sentence-level approaches fail to capture, and their task-specific annotation conventions are human-defined artifacts that cannot be acquired through pretraining alone. We propose Verb-ICL, a selective annotation framework for ICL-based structured prediction that addresses both challenges. Verb-ICL first selects representative examples using a token-level coverage strategy that captures local semantic patterns critical for structured prediction, then generates actionable error feedback that codifies task-specific annotation guidelines and incorporates this feedback into ICL demonstrations. We evaluate Verb-ICL on six structured prediction datasets spanning information extraction and semantic parsing. Experiments with recent LLMs show that Verb-ICL consistently outperforms strong selective annotation baselines under low-resource settings and continues to provide gains as the annotation budget increases. Extended analyses demonstrate that the generated feedback is predominantly useful across a four-category quality taxonomy, generalizes as task-level guidance beyond instance-specific corrections, and improves performance regardless of the underlying selection strategy.

This work proposes a latent Lagrangian-based framework for reduced-order modelling of forced nonlinear dynamical systems. In contrast with conventional Lagrangian or Hamiltonian neural networks, our approach learns a set of latent coordinates sufficient to capture the dynamics conjointly with two neural networks for the latent kinetic and latent potential energies, and leverages force supervision to eliminate the need for an ODE solver during training. Consistency of physical laws in the latent space is ensured through the principle of virtual work. Results show that the model effectively learns the subtle dynamics induced by the system's nonlinearity and non-convex potential energy, and generalizes well to unseen forces and initial conditions. These observations confirm the physical relevance of the proposed approach, and its interest for model reduction.

Sparse-view novel view synthesis is a central problem in 3D content creation, but diffusion-based approaches remain limited by iterative denoising, making multi-view generation expensive at inference time. We introduce NAMVIS, a diffusion-free framework that reformulates multi-view image synthesis as geometry-conditioned next-scale autoregression. Instead of generating target views through repeated denoising, NAMVIS predicts discrete visual tokens through a small number of coarse-to-fine scale steps, while sampling all tokens within each scale and across target views in parallel. To anchor this generation process to explicit camera geometry, we propose Multi-scale Projective Pose Encoding, which injects source and target camera transformations into both target-view self-attention and source-to-target cross-attention at every resolution. NAMVIS further combines global conditioning with dense geometry-aware cross-attention, enabling the model to preserve source-view appearance while maintaining target-view consistency. Across Objaverse, GSO, and OmniObject3D, NAMVIS outperforms diffusion-based baselines in PSNR, SSIM, and LPIPS, while running over 3 times faster than the evaluated diffusion baselines under the same evaluation setting. These results suggest that geometry-conditioned next-scale autoregression is a promising and efficient alternative to diffusion for sparse-view multi-view synthesis. Additional qualitative results, videos, and resources are available at https://corl-team.github.io/namvis/

Structural diagrams are widely used to represent complex systems and relational information across scientific, engineering, procedural, and spatial domains. Recent vision-language models (VLMs) have become increasingly capable of recognizing diagram elements and reasoning about their content, while complete diagram topology extraction remains comparatively underexplored. In this paper, we study diagram-to-graph topology extraction: extracting all diagram entities and the complete relations among them. To enable large-scale supervised training and systematic evaluation of this task, we introduce Knossos, a benchmark of 19,200 diagrams across six diverse domains, with 245,179 nodes and 439,740 edges. Its symbolic generation process provides exact alignment between rendered diagrams and annotations of complete topology, relation types, and connector geometry. To address the modeling challenge of complete topology extraction, we also present Ariadne, a structured framework that decomposes the task into node inventory extraction and source-conditioned edge prediction. Extensive experiments show that training on Knossos substantially improves complete topology extraction in smaller open-source VLMs. Ariadne further improves over one-step extraction under matched supervision, demonstrating the additional benefit of structured decomposition. It achieves the highest average Edge F1 among the evaluated methods on Knossos, while both backbone variants also improve over their unadapted counterparts on the real-world external benchmark. Code and benchmark are available at https://github.com/bangwayne/knossos_Ariadne_Public.

In a Transformer, token mixing is the step that lets each token draw information from other tokens, and it dominates the cost of encoding long sequences. Self-attention does this mixing very well: every token weighs every other token by content, which gives strong contextual modeling. That all-pairs comparison is also why its cost grows quadratically with sequence length. We introduce WNet, a Transformer encoder that replaces self-attention with token mixing based on the discrete wavelet transform (DWT). Three attention-free mixers recombine the scales: by linear fusion, by learned gating, or by letting each token choose its scales. A hybrid adds self-attention in the last layer only. A receptive-field analysis shows that wavelet mixers built from two-tap filters, such as Haar, never relate tokens outside fixed blocks, however deep the network, even when the filters are learned. Longer filters reach the whole sequence within two layers. We pre-train every model with masked language modeling on a fixed-token subset of C4 and fine-tune on GLUE, using one controlled setup with size-matched BERT and FNet baselines and a control that cannot mix tokens. The token-gated mixer trains as fast as attention at 256 tokens and 2.7 times faster at 4,096.

AI assistants receive requests that leave out information needed for a good outcome, for example about users' preferences or goals. They must then either speculate or ask for more information before proceeding. We reconceptualize this as a value-of-information problem: the assistant should acquire information whose absence causes the greatest avoidable loss in user utility. This is rarely known ex ante; rather, assistants must predict it in order to optimally allocate limited user interactions. We instantiate this problem in image generation and derive a reinforcement learning framework using multi-turn simulated users to maximize utility recovery under uncertainty. In a preregistered study with 456 interactive sessions across 76 human participants, this helped users significantly better match reference images with significantly fewer questions, less total interaction time, and lower cost. This points toward a simple and scalable framework for training language model assistants to better disambiguate user intent by asking more informative questions.

Most federated learning (FL) studies implicitly assume that training data is readily available at participating edge devices. In practice, this data must first be acquired through heterogeneous sensors with potentially different measurement modalities, and the quality and cost of this acquisition process can directly affect the learning performance. This dependence creates a coupled relationship among sensing, computation, and communication, necessitating a system model that captures their joint effect on resource efficiency and learning accuracy. This paper develops such a unified treatment for an over-the-air (OTA) FL system in which each device configures its sensing modality, power, resolution, and sample size, with aggregations conducted OTA at the server. We derive models for sensing noise, energy consumption, latency, and OTA aggregation distortion, and establish a non-convex convergence bound quantifying how these factors influence the global model trajectory. Guided by this analysis, we formulate a per-round optimization that balances energy, latency, and learning quality under physical device constraints. The resulting mixed optimization is addressed through an alternating procedure with closed-form and one-dimensional updates for continuous variables and a lightweight discrete search. Numerical experiments corroborate the advantage of jointly optimizing sensing, computation, and communication decisions.