Computation and Language
MemoryLACE: Memory Lifecycle-Aware Consolidation and Evidence Retrieval
Meriem Yacoubi, Pia Schmidt, Nenad Petrovic et al. (+3 more)
Long-term LLM
agents must preserve information across interactions while distinguishing repeated evidence, historical states, updates, and unresolved contradictions. Existing textual
memory systems retrieve semantically relevant
memories efficiently but often leave these relationships implicit, whereas richer structured approaches model them through global graphs, hierarchical abstractions, or reflection at greater complexity. We introduce MemoryLACE (MemLACE), a lightweight
memory framework that explicitly models the lifecycle of textual evidence through sparse merge, supersession, and contradiction relations while preserving atomic natural-
language memories and their provenance. Rather than retrieving
memories independently, MemLACE reconstructs relation-aware evidence units that expose current, historical, supporting, and conflicting evidence for downstream
reasoning. Across BEAM and StructMemEval, using open-weight and proprietary LLM backbones, MemLACE achieves the highest overall performance in same-backbone comparisons while reducing end-to-end runtime on BEAM by 66.6% relative to Hindsight, the strongest reported reflective-
memory baseline. Ablation studies identify lifecycle expansion and temporal awareness as the principal contributors to these gains. Together, the results demonstrate that explicitly modeling the local lifecycle of textual evidence is sufficient to substantially improve long-term
memory reasoning without requiring comprehensive knowledge graphs or global reflection.
Read the full abstract ▾
All authors Meriem Yacoubi, Pia Schmidt, Nenad Petrovic, Ahmed Frikha, Martin Kirchhoff, Alois Knoll
Computer Vision
RoboTok: A Scalable Data Engine for Internet Demonstration Video Retrieval and Dexterous Manipulation Learning
Howard Qian, Yiting Chen, Yunfei Xie et al. (+6 more)
Robot learning increasingly depends on broad and diverse demonstrations, yet collecting
robot data remains expensive and difficult to scale across the wide range of real-world tasks. To address this bottleneck, we introduce RoboTok, a scalable data engine that uses a query human manipulation
video to retrieve manipulation-relevant internet demonstrations for training dexterous
robot policies. Specifically, we learn a latent motion space from 3D hand trajectories expressed in estimated actor-centered reference frames. This representation enables manipulation behaviors to be compared across variations in camera viewpoint, scene appearance, and actor occlusions, while remaining compact enough for efficient indexing and
retrieval over internet
video collections. We evaluate RoboTok against existing
robot-data
retrieval approaches using
retrieval metrics and downstream
robot policy performance, showing that RoboTok retrieves more manipulation-relevant demonstrations and improves downstream task success. In real-world
robot experiments, RoboTok-guided policies achieve a mean success improvement of 42.2 percentage points over the strongest baseline for each task, establishing hand-pose trajectory-aware
retrieval as a scalable way to leverage continuously growing web
video for
robot learning.
Read the full abstract ▾
All authors Howard Qian, Yiting Chen, Yunfei Xie, Kejia Ren, Podshara Chanrungmaneekul, Gaotian Wang, Bowen Wen, Chen Wei, Kaiyu Hang
Machine Learning · Cross-listed from cond-mat.stat-mech
Generative Nested Sampling of Atomistic Thermodynamic Landscapes
Alessandro Coretti, Nico Unglert, Sebastian Falkner et al. (+2 more)
Nested sampling (NS) resolves the thermodynamics of an atomistic system from a single simulation, but its practical reach is limited by the Markov-chain updates needed to decorrelate walkers within each likelihood-constrained ensemble. Flow-based NS has reduced this bottleneck for gravitational-wave (GW) inference, yet its transfer to atomistic systems is not merely a change of application. Comparing a GW150914-like binary-black-hole likelihood with an eight-particle two-dimensional Lennard-Jones (LJ) system of comparable dimensionality, we show that the two landscapes differ fundamentally: atomistic multimodality is discrete and combinatorial, generated by particle permutations separated by hard collision walls, and its coordinate coupling is dense and collective, whereas the GW posterior exhibits smooth degeneracies and localized parameter coupling. Guided by this diagnosis, we introduce NS-Flows: a single conditional normalizing flow, conditioned on the NS energy bound and trained on a sliding window of recent live sets, that replaces MCMC by direct parallel draws corrected by importance-weighted rejection resampling. Live sets supply the data self-consistently, allowing flow training without structured priors or a pre-existing
dataset. For LJ disks under periodic boundary conditions, the
algorithm can reduce energy evaluations by two orders of magnitude and wall-clock time by about 30%, an advantage that becomes more favorable as the cost of the potential grows. The flow's generative efficiency further acts as a physical diagnostic: it varies non-monotonically along the annealing trajectory, is lowest in the dense disordered regime, and is quantitatively captured by the constrained ensemble's internal mode complexity together with target drift across the training window, identifying
liquid-like ensembles, not prior-target separation, as the hard case for current flow architectures.
Read the full abstract ▾
All authors Alessandro Coretti, Nico Unglert, Sebastian Falkner, Georg K. H. Madsen, Christoph Dellago
Multiagent Systems
Where Reliability Lives: Experimental Localisation of Behavioural Properties in an Agent System
Timothy Marsden, Matthew Collecutt and James Marsden
Reliability claims about agentic systems implicitly locate each property somewhere: in the model, or in the machinery around it. We built a system where that location is an experimental question. The subject is a persistent simulated settlement whose authoritative append-only ledger adjudicates every attempted act against world state; accepted history is the only reality. Mind, institution and world were separated before any experiment. Holding cognition fixed, we intervened on the institution's epistemic mechanisms (evidence provenance, belief availability, physical-evidence legibility); preregistered experiments refuted our central prediction twice, in opposite directions. A registered falsifier then supplied the input the geometry had denied the belief channel, a staged veridical first-hand witness, and its marginal value, non-positive throughout the witness-free phases, turned positive: 9 of 11 seeds, zero added false attribution. Holding institutional enforcement fixed, we intervened on cognition four ways: ablating the native minds' machinery, killing and resetting them mid-task, substituting a frozen frontier-LLM panel for the entire native cognition, and corrupting beliefs with trusted false testimony. Behaviour changed dramatically: one falsehood cost each trusting run about 900 futile actions and the distrusting arm none. Five pre-declared properties did not move in any tested trajectory: accepted reality stayed singular, invalid attempts were refused with typed reasons, duties outlived their processes, no work was accepted twice, and no false completion was ever accepted (2,581 substituted-panel claims, none false). Our claim is limited to this setting: measured behavioural properties were separable from substantial changes to cognition, established by intervention. One designed world, not a population of institutions; no test of an
agent optimising against the institution.
Read the full abstract ▾
Numerical Analysis
Coupled Tensor-Tensor Completion Method with Applications in Drug Repurposing
Maryam Bagherian, Albert Hung, Ivo Dinov et al. (+1 more)
Many biomedical challenges can be posed as tensor completion problems where the observed entries of a multidimensional array (a tensor) are used to impute the missing values. In such settings, incorporating side information about the modes of the tensor, such as gene-gene similarity, can significantly enhance the solutions of the completion problem. Most existing tensor completion methods can only incorporate side information in the form of
matrices. In this study, we introduce a novel framework to incorporate side information in the form of tensors. Our new approach, called Coupled Tensor-Tensor Completion (CTTC), leverages the hidden connections among
multimodal tensors to improve tensor completion performance. In addition to practical utility, CTTC has theoretical foundations in distance metric learning and group theory. We derive an alternating
algorithm to solve the CTTC optimization problem and establish its convergence to a stationary point. Finally, we show that CTTC outperforms state-of-the-art tensor completion methods at predicting drug effects. Results: Compared with other tensor completion methods, including HaLRTC, CTRC,
Cell, and NTDDR, CTTC demonstrates superior run-time and RSE tensor completion accuracy on two
benchmark datasets, DTD and LINCS.
Read the full abstract ▾
All authors Maryam Bagherian, Albert Hung, Ivo Dinov, Joshua Welch
Computers and Society
Reducing Catastrophic Risk from AI with Systematic Monitoring and Evaluation of Rogue AI Progression
T. Bauer, W.P. Kegelmeyer, E. Begoli et al. (+15 more)
This article presents a structured framework of behavioral indicators that may signal progression toward potentially catastrophic threats from artificial intelligence systems. We adopt a pragmatic approach, inspired by established methodologies in cybersecurity and national security. By establishing clear metrics, indicators, and thresholds across multiple dimensions of AI capability and behavior, this framework enables researchers and policymakers to implement evidence-based monitoring protocols.
Read the full abstract ▾
All authors T. Bauer, W.P. Kegelmeyer, E. Begoli, A. Sadovnik, T. Emerson, C. Corley, N. Generous, J. Moore, B. Bartoldson, R. Goldhahn, M. Goldman, M. Greaves, M.J.D. Vermeer, B. MacLennan, D. Schulker, N. VanHoudnos, J. Bansemer, Y. Bengio
Computer Vision · Cross-listed from physics.med-ph
Improving Clinical Target Volume Segmentation Accuracy using Anatomical Priors and Active Learning for the AGITG TOPGEAR Clinical Trial
Phillip Chlap, Mark Lee, Trevor Leong et al. (+11 more)
Training deep learning-based medical image
segmentation models is challenging with limited curated
datasets. For AGITG TOPGEAR, a gastric cancer trial, the Clinical Target Volume (CTV) is complex and defined by multiple anatomical landmarks, making upfront training data preparation difficult for an automated contour QA
segmentation model. We investigate anatomical priors, derived from surrounding organ
segmentations, to provide spatial context and improve TOPGEAR CTV
segmentation accuracy. We also evaluate active learning, iteratively expanding the training
dataset by selecting cases expected to improve performance. One hundred TOPGEAR CT scans were retrospectively analyzed. An initial set of 10 expert-contoured cases was used to train an nnU-Net model. TotalSegmentator generated a voxel-wise anatomical prior map from surrounding structures as an additional input channel. Active learning was simulated over four iterations, selecting cases by model uncertainty and
segmentation performance. All models used five-fold cross-validation for an ensemble uncertainty measure. Evaluation used a hold-out testing set of 50 cases. The anatomical prior improved CTV
segmentation accuracy, increasing mean Dice Similarity Coefficient (DSC) from 0.84 to 0.86. Active learning similarly improved performance to 0.86, with greatest benefit in the final round. Combining the anatomical prior with active learning achieved the highest accuracy, with a DSC of 0.87. Model uncertainty correlated with DSC, supporting its use in identifying suboptimal predictions and guiding active learning. Anatomical priors and active learning each improved CTV
segmentation accuracy and generalizability, with their combination achieving the best performance, supporting integration into
segmentation model development for automated contour QA in radiotherapy clinical trials.
Read the full abstract ▾
All authors Phillip Chlap, Mark Lee, Trevor Leong, Matthew Field, Jason Dowling, Hang Min, Julie Chu, Jennifer Tan, Phillip K. Tran, Tomas Kron, Annette Haworth, Martin A. Ebert, Shalini K. Vinod, Lois Holloway
Computation and Language
Jina-OCR-v1: Efficient Document Parsing with Speculative Decoding and Dense Verifiable Rewards
Alejandro Barón García, Feng Wang, Emilia Garcia Casademont et al. (+1 more)
We present Jina-OCR-v1, an end-to-end document parsing model built to serve on low-budget GPUs. It combines the compressed-
vision encoder and the 3B mixture-of-experts decoder of DeepSeek-OCR, which activates about 570M parameters per
token, with a FastMTP speculative decoding head that shares a single draft block recursively across K=3 prediction steps. Greedy verification makes decoding lossless. Post-training combines instruction
alignment, robustness fine-tuning on difficult documents, and GRPO under dense verifiable rewards: deterministic formula, table, and structural checks that award partial credit. The training data mixes cleaned public corpora with targeted synthetic pages. At the default dynamic-resolution setting, Jina-OCR-v1 scores 91.14 on OmniDocBench v1.6 and 83.4 on olmOCR-Bench, and reaches the highest page throughput in our comparison at 2.57 pages per second. On a low-budget GPU such as the NVIDIA L4, FastMTP doubles decoding speed over greedy autoregressive decoding. The model is publicly available at https://huggingface.co/jinaai/jina-ocr-v1.
Read the full abstract ▾
All authors Alejandro Barón García, Feng Wang, Emilia Garcia Casademont, Han Xiao
Machine Learning
Portable Causal Fairness Across Synthetic Data Generator Families
Steven Golob, Sikha Pentyala and Martine De Cock
When a statistical agency or regulator releases synthetic data in place of sensitive records, it chooses the generator that produces the table, and can shape that generator so unfair pathways are absent. DECAF made this concrete on one non-private GAN: three
fairness definitions become three sets of edge cuts on the generator's causal graph. Whether the mechanism belongs to DECAF, or to causal factorisation itself, was untested. We port all three definitions to nine generators from three unrelated families (marginals-based, GAN, and
diffusion, each with differentially private variants), across three levels of formal
privacy guarantee, over 2,520 matched-pair runs on Adult and COMPAS
datasets. The mechanism transfers everywhere, and our new causal
diffusion backbone yields the fairest release of any family we tested, at fidelity close to the marginals tier. Applying the cut barely moves fidelity, only costs a downstream classifier about $0.07$ to $0.15$ AUC on average, and adding
privacy guarantees don't make the data less fair.
Read the full abstract ▾
Computers and Society
Open Problems in AI Risk Modeling: Insights from a Workshop on the Technical Foundations of AI Risk Modeling
Krystal Jackson, Deepika Raman, Jakub Kryś et al. (+8 more)
We investigate the design of robust risk models to assess societal risks posed by advanced AI systems, an emerging area in AI governance. Many regulatory proposals increasingly require systemic risk assessment, but in the absence of rigorous quantitative methods, the question remains what state of the art risk modeling should look like in practice. We identify the key methodological and institutional challenges that currently limit the adoption of risk modeling. We review five research traditions that inform this problem: probabilistic risk assessment, catastrophic AI risk analysis, cybersecurity risk quantification,
Bayesian causal inference, and threshold-based governance. We compare two leading proposals, scenario-based risk estimation and
Bayesian network-based threshold setting. Drawing on a workshop with 22 experts and subsequent analysis, we identify a structured agenda of open questions concerning model structure, scope, evidence integration, validation, and governance. We close by outlining priorities for progress, arguing that it will depend on integrating quantitative modeling with independent evaluation, transparent and tiered disclosure, and institutions capable of maintaining and updating risk models over time.
Read the full abstract ▾
All authors Krystal Jackson, Deepika Raman, Jakub Kryś, Sean P. Fillingham, Jack Kengott, Andrew J. Lohn, Nada Madkour, Henry Papadatos, James Sykes, Anna Katariina Wisakanto, Malcolm Murray
Machine Learning
Frontier LLMs are effective batch optimizers: Assessing reasoning models in continuous and discrete settings
Frank Hu, Shriram Chennakesavalu and David Graff
Frontier large
language models (LLMs) have become attractive priors for optimization due to their large-scale pretraining that enables them to navigate a variety of optimization settings. However, the effectiveness of modern
reasoning LLMs in batch optimization settings remains underexplored. Here we investigate the performance of the current generation of frontier LLMs as batch optimizers in both continuous and discrete settings. We find that while LLMs are competitive zero-shot batch optimizers for numerical test functions, their performance is brittle compared to classical non-LLM optimization approaches. However, LLM priors are significantly better in semantically rich settings, indicating that their batch optimization behavior is highly effective when navigating and
reasoning over the discrete spaces most similar in structure to their pretraining data.
Read the full abstract ▾
Robotics
Real-Time Shape Control of Multi-Segment Soft Robotic Arms Using Koopman Operators with Global and Local Observables
Jiahe Wang, Eron Ristich, Sultan Haidar Ali et al. (+5 more)
Multi-segment soft robotic arms can continuously reconfigure their body shapes for safe interaction, but tip control alone is insufficient for constrained-space tasks. Therefore, shape control is a more important task for multi-segment soft arms than tip control, but remains challenging due to the high dimensionality and nonlinear dynamics of continuum deformation. In existing work, shape control accuracy is defined by the error in the global frame (global shape error). For multi-segment soft arms, using only global shape error as the control objective is insufficient, as segment coupling,
gravity-induced loading, and inertial effects become more significant. This difficulty increases with the number of segments. In this paper, we present a Koopman-based model predictive control framework that combines global and local observables, enabling real-time shape control on multi-segment soft robotic arms. The framework is evaluated through numerical and physical experiments. Numerical experiments demonstrate the scalability of the proposed controller by achieving shape control on
robots with up to 10 independently actuated segments. The physical experiments demonstrate that the controller is capable of (1) real-time shape control of 3- and 5-segment robotic arms with tip speeds up to 0.6 m/s, (2) robust tracking without retraining, including distal payloads up to 400~g and recovery from a 7~N lateral disturbance, and (3) the potential for future inspection applications through a confined-space demonstration. These results demonstrate that the proposed framework enables dynamic, scalable, and accurate real-time shape control on multi-segment soft robotic arms.
Read the full abstract ▾
All authors Jiahe Wang, Eron Ristich, Sultan Haidar Ali, Eric weissman, Lei Zhang, Wanxin Jin, Yi Ren, Jiefeng Sun
Systems and Control
A Reconfigurable Hybrid Convolutional-Fully Connected Neuromorphic Core for Biomedical Edge Inference
Sarah Johari, Suman Kumar, Abhishek Mishra et al. (+2 more)
This work presents a programmable FPGA-based architecture for spiking convolutional neural network (SCNN) inference, with real-time hypoxia classification serving as a biomedical edge application. The architecture implements a hybrid spiking convolutional-fully connected (CNN-FC)
topology on a programmable, quantized, layer-based neuromorphic hardware core. Early layers perform spiking convolution using receptive-field connectivity with support for multi-channel kernels and stride, while deeper layers use fully connected spiking layers for classification. A PyTorch-based hardware-software co-design flow enables deployment of trained parameters with quantization and configurability support. The design is first validated on MNIST and Fashion-MNIST, achieving hardware accuracies of up to 98% and 86%, respectively, at 16-bit precision. It is then applied to hypoxia classification using red and infrared photoplethysmography (PPG) signals acquired from a shoulder-mounted sensor, with skin tone included as an additional input channel. The resulting classifier achieves an average hardware accuracy of 88.26% across five folds at 16-bit precision while consuming 1.455 W of dynamic power, demonstrating the feasibility of low-power neuromorphic biomedical classification at the edge.
Read the full abstract ▾
All authors Sarah Johari, Suman Kumar, Abhishek Mishra, Anush Lingamoorthy, Nagarajan Kandasamy
Human-Computer Interaction
Signal-Driven Pervasive Game Design: The LifeSync-Games Framework as a Player Experience Integration Layer
J. Macías-Cáceres, F. Gutiérrez-Vela, P. Paderewski-Rodriguez et al. (+1 more)
Pervasive games extend the magic circle across spatial, temporal, and social dimensions, yet treat the player's physiological and cognitive state as a passive receptor rather than an active signal. This paper presents LifeSync-Games (LSG), a framework that (unlike proposals treating player signals as an additional dimension), operationalizes them as a Player Experience Integration Layer (PEIL) acting transversally across the three existing pervasive dimensions through verified real-world signals: physical activity, sleep quality,
memory, and decision speed. The framework introduces a gamified integration artifact (the LSG portal) that mediates the real <-> virtual exchange through redeemable points, real-world missions, and structural gamification. Five HCI design principles grounded in Self-Determination Theory and Flow Theory are proposed, instantiated across six commercial
video games, together with a study protocol (n = 70-80 participants, quasi-experimental design). This paper reports the design stage of LSG: rule thresholds and portal parameters are design decisions pending empirical calibration; no data collection has yet been conducted. The main contribution is a theoretically grounded framework and validation protocol positioning player-sensitive integration as the mechanism enabling pervasive games to respond to the player's actual biological and cognitive state.
Read the full abstract ▾
All authors J. Macías-Cáceres, F. Gutiérrez-Vela, P. Paderewski-Rodriguez, R. González-Ibáñez
Emerging Technologies
QSVT-Based Three-Phase Unbalanced Power Flow
Kamini Shahare and Peng Zhang
This letter introduces QSVT-3PF, a
quantum singular value transformation (QSVT) based solver for three-phase unbalanced power flow with embedded single-phase grid-forming inverter (GFM) operation. The contributions are threefold: 1) reformulating the Newton correction as a QSVT-compatible inverse problem using a normalized block-encoded phase-domain Jacobian; 2) introducing a regularized singular-value filter to improve robustness under ill-conditioned and stressed operating conditions; and 3) validating the proposed solver on an IEEE 5-bus system and the IEEE 123-node test feeder with single-phase GFM integration. Test results show that QSVT-3PF matches the classical Newton
benchmark in residual convergence, voltage profile, and final operating point, demonstrating the feasibility of QSVT-3 for large, unbalanced distribution systems.
Read the full abstract ▾
Networking
Bringing dApps to OCUDU: An E3 Controller for Real-Time Open RAN Intelligence
Angelo Feraudo, Ravis Shirkhani, Stefano Maxenti et al. (+3 more)
Real-time control loops in Open RAN are increasingly colocated with the gNB, where dApps, i.e., programmable applications with sub-millisecond access to PHY- and MAC-layer signals, enable latency-critical use cases such as spectrum sharing, channel-aware scheduling, and integrated sensing. What turns dApps from a per-vendor mechanism into a portable component of the emerging AI-RAN ecosystem is the E3 interface: this interface defines how a dApp subscribes to RAN telemetry, receives indications, and issues control actions back to the RAN. To date, E3 has been implemented on OpenAirInterface and on NVIDIA Aerial; OCUDU, i.e., the Linux Foundation's open-source CU/DU project, has lacked a comparable E3 path. We close that gap with an open-source E3Controller for OCUDU, designed as a sidecar daemon so the E3 protocol stack, service-model logic, and dApp interactions live entirely outside RAN core processing, which is touched only through well-defined instrumentation hooks. We validate the controller with two reference service models, Spectrum and L1, that deliver O-RAN fronthaul I/Q samples to a spectrum-sensing dApp over a Foxconn RPQN 4800
radio unit. Our evaluation shows that the E3 path imposes no measurable throughput penalty on the OCUDU RAN, demonstrating that a real-time dApp/RAN interface is feasible on production-grade open RAN software.
Read the full abstract ▾
All authors Angelo Feraudo, Ravis Shirkhani, Stefano Maxenti, Andrea Lacava, Michele Polese, Tommaso Melodia
Formal Languages and Automata
NeuroSTAR: Automata-guided Neuro-symbolic Specification Formalization
Joy Saha, Trey Woodlief, Sebastian Elbaum et al. (+1 more)
Automated translation of natural
language (NL) descriptions into Linear Temporal Logic over finite traces (LTLf) is a prerequisite for automated formal verification of a system's dynamic behavior. Several LLM-based methods have recently shown potential for this task. However, they struggle with the nuance of natural
language descriptions, which can lead LLMs to only partially capture the intended meaning. To address this limitation, we propose NeuroSTAR (Automata-guided Neuro-symbolic Specification Formalization), an NL-to-LTLf framework that builds on two insights. First, it leverages multiple generators to obtain diverse LTLf candidates. Second, it uses an automata-theoretic semantic comparison based on DFA traces to identify behavioral disagreements that guide formula refinement. We evaluate NeuroSTAR and show that it improves NL-to-LTLf translation performance by 8-18 percentage points relative to the prior state-of-the-art (SoTA) on unambiguous
benchmarks. We further study its applicability to a body of driving law text, a complex, realistic, and reference-free domain critical for autonomous-vehicle specification. This study shows that NeuroSTAR can capture the necessary temporal semantics in 83.9% of the driving law sections, which demonstrates the effectiveness of automata-guided reference-free refinement in formalization.
Read the full abstract ▾
All authors Joy Saha, Trey Woodlief, Sebastian Elbaum, Matthew B. Dwyer
Computation and Language
No country for old linguists: LLM-brain alignment underdetermines neural computation
Elliot Murphy
Nastase et al. (2026) argue that large
language models (LLMs) may illuminate
language processing because both rely on distributed, context-sensitive representations shaped by statistical learning. Their rejection of simple cortical "boxology" is persuasive, and they articulate a strong case for the value of LLM-
brain alignment research. The key question is what kind of inference LLM-
brain alignment licenses. My claim here will be narrow: representational
alignment can in principle constrain mechanistic hypotheses, but it does not by itself identify a mechanism. Nastase et al. acknowledge that an encoding model can capture features represented in neural activity without establishing a shared architecture or
algorithm. Yet the authors sometime move from
alignment to "shared computational principles" and ultimately to LLMs as mechanistic models of natural
language. Indeed, their methodological caveat that
alignment does not establish a shared architecture or
algorithm sits uneasily with their conclusion that LLMs might instantiate the same computational principles as biological
brains and provide a "fully mechanistic model" of
language. I discuss what I consider to be problems of logical, causal, and computational underdetermination in Nastase et al.'s (2026) proposal.
Read the full abstract ▾
Computer Vision
Who Speaks for the Pruned? Visual Token Pruning as Coverage Optimization
Qingchan Zhu, Weihang You, Hanqi Jiang et al. (+3 more)
Visual
token pruning reduces the inference cost of
vision-
language models (VLMs), but most methods only ask which
tokens to keep. This retained-
token view can keep redundant high-scoring
tokens while leaving discarded evidence without a close representative. We propose CoverPruner, a training-free pruner that asks the complementary demand-side question: after a
token is removed, which surviving original
token represents it for the target VLM? CoverPruner formulates
pruning as Representational Coverage Maximization (RCM), covering the full projected visual-
token set with query-weighted demand. It instantiates RCM with projector-space coverage and a lightweight first-layer
attention probe. Across multiple VLM architectures and compression rates, CoverPruner achieves the best average accuracy among all compared methods, with the largest gains usually appearing under aggressive compression.
Read the full abstract ▾
All authors Qingchan Zhu, Weihang You, Hanqi Jiang, Changdi Yang, Tianming Liu, Geng Yuan
Software Engineering
Compound Prompt Constraints in LLM Code Generation: A Factorial Study of Format, Persona, and Urgency
Shrenik Jadhav, Nickalsa LaPlaca, Caleb Stone et al. (+3 more)
Large
language models (LLMs) are increasingly used in software engineering pipelines for code generation, where production
prompts often combine multiple constraints. This paper presents a full-factorial empirical study of how output formatting, persona assignment, and urgency framing jointly affect LLM code-generation reliability. We evaluate all 27 combinations in a controlled 3x3x3 design and decompose each compound condition into an additive prediction and a residual interaction term that captures super-additive degradation. The study uses all 164 HumanEval+ problems across five OpenAI models from the GPT-4o family, GPT-4.1 family, and o3-mini, yielding 22,140 greedy-decoding evaluations. A format-aware extraction pipeline separates formatting failures from
reasoning failures, and significance is assessed with McNemar's test, odds ratios, and 95% confidence intervals. Results show that compound constraints can produce architecture-dependent degradation not predictable from single-factor experiments. The GPT-4o family exhibits consistent super-additive effects, with pass@1 reductions 3-12 percentage points beyond additive predictions; the largest interaction is -12.2 pp on GPT-4o-mini for JSON + expert persona + moderate urgency. JSON combinations generally produce larger interactions than XML. In contrast, the GPT-4.1 family is largely resistant, while o3-mini shows a qualitatively different pattern in which structured output constraints can improve performance. These findings show that vulnerability is architecture-dependent rather than size-dependent, that individually neutral or beneficial constraints can combine to cause substantial degradation, and that compound-
prompt testing should be standard in reliability assessment for LLM-assisted engineering pipelines.
Read the full abstract ▾
All authors Shrenik Jadhav, Nickalsa LaPlaca, Caleb Stone, Ashok Raja, Omar Ochoa, Vidhyashree Nagaraju
Numerical Analysis
A discontinuous Petrov-Galerkin finite-element framework for the simulation of microwave-heated flows
Oreste Marquis, Matthias Maier and Bruno Blais
We present a high-order multiphysics solver for the simulation of microwave-heated flows. The solver couples a discontinuous Petrov-Galerkin (DPG) finite element method for the time-harmonic Maxwell equations with continuous Galerkin finite element methods for the heat equation and the incompressible Navier-Stokes equations. We validate the electromagnetic solver against multiple
benchmark problems: wave propagation in a rectangular waveguide, a cavity problem with a singular solution, and a microwave-heated obstacle problem, comparing our results against numerical and experimental data from the literature. The results confirm the validity of the implementation and demonstrate its ability to perform adaptive mesh refinement using the DPG method's built-in error
estimator. The final part of the study showcases the capabilities of the multiphysics framework through simulations of microwave-heated flow around obstacles with singular
geometric features. These results highlight the potential of the proposed framework for the simulation and optimization of microwave-assisted chemical processes. Finally, the developed high-order multiphysics solver has a low
memory footprint, since the electromagnetic solver relies on a Conjugate Gradient (CG) iterative solver and the fluid solver is implemented in a
matrix-free fashion, making the overall approach scalable and well-suited for large-scale parallel simulations.
Read the full abstract ▾
Emerging Technologies
Embedding Single-Phase Grid-Forming Inverters in Three-Phase Unbalanced Power Flow
Kamini Shahare, Fei Feng and Peng Zhang
This letter introduces Grid-Forming Three-Phase Power Flow (GFM-3PF), an augmented Newton power flow framework for modeling single-phase grid-forming inverters (GFMs) in three-phase unbalanced networks. The contributions of this work are threefold: 1) developing a two-stage solution strategy based on positive-sequence initialization followed by three-phase phase-domain lifting for robust large-scale computation; 2) incorporating phase-domain unbalanced network modeling with controlled mutual coupling together with phase-dependent, frequency-sensitive load representation; and 3)
embedding a single-phase GFM bus model directly into the three-phase unbalanced Newton power flow equations through a common droop-governed frequency state. GFM-3PF is validated on 3-bus, 33-bus, and 118-bus test systems, demonstrating its accuracy, robustness, and scalability.
Read the full abstract ▾
Computer Vision
VeriPhy: Agentic Physical Reasoning for World Model Evaluation and Refinement
Wenzhuo Xu, Yuchen Zhu, Chongjian Ge et al. (+8 more)
Visual fluency in generated
video does not imply physical reliability, and a scalar quality score alone is incapable of indicating the obligation a clip violates or the moment it fails. We present VeriPhy, an auditable physical-verification system in which a text-only planner compiles the
prompt into typed physical obligations and a statically validated execution plan before any frame is observed. During execution, observations gate and scope only declared calls to frozen low-level experts (e.g.,
segmentation and tracking, counting, eleven typed physical measurements over the resulting tracks, depth, OCR, and audio-event detection). Each action returns a provenance-carrying evidence record whose payload, when usable, is either a typed measurement or an explicitly tagged learned state. Typed resolvers and fixed composition map usable records to a three-valued state (supported, contradicted, or unknown, surfaced as plausible, implausible, or abstain) with full provenance, so that every verdict is traceable to the evidence that produced it. We anchor evaluation in a 1,500-clip corpus of human-annotated flaw records that localize real generation failures in
prompt reference, space, and time. On a 149-clip core carrying 304 such records, VeriPhy accounts for 228, against 164 for a published question-decomposition evaluator given the same clips and the same claims. Recall alone does not separate it from prompting the same backbone monolithically, which reaches 222; what separates them is that each decision retains its evidence record and provenance, making the traces auditable one verdict at a time and usable as the interface through which a critic verdict could be written back into generation.
Read the full abstract ▾
All authors Wenzhuo Xu, Yuchen Zhu, Chongjian Ge, Xuan Shen, Jing Shi, Jason Kuen, Yongxin Chen, Molei Tao, Christopher McComb, Noelia Grande Gutiérrez, Jiuxiang Gu
Distributed Computing
BASP: Communication-Efficient Batch-Aware Sequence Parallelism for LLM Training
Bigyan Ghimire and Jon C. Calhoun
Long-context
reasoning for large
language models (LLMs) is becoming increasingly important, but training over long sequences remains challenging due to massive
memory and communication requirements. Sequence parallelism has emerged as an essential technique for addressing bottlenecks in long sequence LLM training. However, we observe that existing sequence parallelism methods are batch-agnostic and apply uniform sequence partitioning across all batch sizes, resulting in inefficient communication. In this paper, we introduce Batch- Aware Sequence Parallelism (BASP), a sequence parallelism approach that leverages batch structure to reduce communication overhead. BASP exploits batch structure by partitioning GPUs into disjoint sequence-parallel groups according to the micro- batch size. This design reduces the all-to-all communication group size, thereby localizing communication and improving training efficiency. Experimental results on an NVIDIA A100 cluster show that BASP improves end-to-end training time by up to 1.17 - 1.31x in Llama and Qwen models compared to standard sequence parallel baselines, while preserving identical model accuracy and
memory usage.
Read the full abstract ▾
Machine Learning
Routing Is Not Enough: Diagnosing Intra-Adapter Subspace Contention in MoE+LoRA Fine-Tuning
Mehreen Hossain Chowdhury, Nowshin Mahjabin, Ahmed Shafin Ruhan et al. (+3 more)
Multi-domain fine-tuning often combines MoE routing with LoRA, assuming that
token-level routing separates domain-specific updates. We test this assumption in MoE+LoRA using Python code paired with biomedical text and mathematical
reasoning. Although these domains show near-disjoint expert routing, adding biomedical data substantially increases code perplexity, indicating that routing separation alone may not prevent negative transfer. To localize the failure, we introduce Jaccard routing overlap and adapter-gradient cosine similarity, which measure expert sharing and update compatibility, respectively. These diagnostics indicate that interference arises mostly from nearly orthogonal domain gradients competing within the same low-rank adapter subspace. We address this issue with SpawnLoRA, which dynamically adds gated sub-adapters inside MoE experts when adapter-level contention is detected, while keeping the router fixed. We evaluate SpawnLoRA on Phi-tiny-MoE-instruct and OLMoE-1B-7B across multiple mixture settings and find that it effectively reduces negative transfer compared with standard and rank-adaptive LoRA. These results demonstrate that structural separation inside experts provides benefits beyond routing or rank expansion alone.
Read the full abstract ▾
All authors Mehreen Hossain Chowdhury, Nowshin Mahjabin, Ahmed Shafin Ruhan, Md Azam Hossain, Abu Raihan Mostofa Kamal, Md Tahmid Rahman Laskar
Emerging Technologies
RACE-AIMC: Selective Inference for Heterogeneous Analog In-Memory Accelerators at the Edge
Osama Yousuf and Martin Lueker-Boden
Analog in-
memory computing (AIMC) speeds up neural-network inference by doing the arithmetic directly inside a
memory array, instead of shuttling weights back and forth between
memory and a processor. This saves energy, but the physical devices that store the weights are imperfect: programming errors, electrical noise, limited-resolution converters, and outright broken
cells all distort the computation, and every physical chip is distorted in its own way. A designer with several such chips available faces an uncomfortable choice: run all of them and combine the answers (safe, but wasteful of energy), or trust a single chip blindly (cheap, but with no guarantee on how often it is wrong). This paper introduces RACE-AIMC (Risk-Aware Certified Ensemble for AIMC), a framework that resolves this choice with statistics rather than guesswork. Offline, RACE-AIMC studies a pool of physical accelerators, picks the single best one for a given energy budget, and computes a mathematically exact upper bound on how often that accelerator will be wrong when it chooses to answer. Online, only that one accelerator is switched on; a lightweight check decides whether to accept its answer or defer to a fallback. In our simulations using a noisy weight mapping and multiple independent test runs, every certified bound stayed under a 10% error target (mean bound 7.83% +- 0.89%, with 70.88% +- 0.98% of inputs answered directly). The resulting system matches the accuracy of a clean digital baseline while cutting modeled energy use by 69.02% relative to always running every accelerator in the pool.
Read the full abstract ▾
Computation and Language
Large Language Models in Resolving Contextual Knowledge Conflicts
Xinye Yang, Zhenyang Liu, Ruisi Li et al. (+1 more)
Most prior works focused on conflicts between an LLM's internal parametric knowledge and externally provided context. In contrast, we investigate how LLMs handle conflicts that arise within contextual knowledge itself. We introduce a taxonomy of six types of contextual conflicts (factual, inferential, temporal, granularity, perspective, and ambiguity) and contribute a comprehensive
dataset ContextConflict for this setting. The
dataset contains 5,781 samples, covers both
reasoning and summarization tasks, and includes both explicit contradictions and implicit conflicts that require multi-step
reasoning. Experiments on nine LLMs show that current models still fall short in resolving contextual knowledge conflicts. We further provide mechanistic interpretability insights into how LLMs process such conflicts, revealing their latent awareness of conflicts and the representational geometry underlying conflict processing. In addition, our analysis uncovers a consistent model bias towards earlier evidence, and this positional preference serves as a key obstacle to effective conflict resolution. Motivated by these findings, we further propose a simple training-free, label-free steering method that steers activations to encourage a more comprehensive incorporation of evidences for better conflict resolution. On our
dataset, the method consistently improves accuracy on
reasoning tasks and generates higher-quality, more balanced summaries for summarization tasks.
Read the full abstract ▾
All authors Xinye Yang, Zhenyang Liu, Ruisi Li, Yuanyuan Lei
Networking
Feasible but Not Safe: Constraint Violations and Report-Channel Attacks in Learned Cell-Free ISAC Association
Mehdi Zafari, Iman Mohammadi and A. Lee Swindlehurst
Learning-based schedulers have been proposed to provide real-time user, target, and access point (AP) association in distributed
cell-free integrated sensing and communication systems. In a typical approach, a graph neural network (GNN), trained on labels from a mixed-integer linear program, maps lightweight per-AP statistics to decisions on AP clustering, user and target scheduling, and mode selection in one forward pass. Such solutions assume that hard constraints, enforced only as soft training penalties, hold at inference, and that the self-reported statistics are truthful. Using our ASSENT
algorithm as an example, we find that despite high $F_1$ scores, many solutions violate at least one hard constraint, demonstrating that high prediction accuracy does not ensure joint feasibility. Projecting the GNN output onto a feasible solution restores constraint satisfaction with low utility loss, even with a simple greedy repair procedure. We further show that feasibility alone does not guarantee robustness to false data injection attacks. A single malicious AP that reports false information cannot substantially increase its user associations, but can greatly increase the rate of infeasible solutions. The effect of such attacks depends on the type of information being falsified. Misreporting information that affects the objective can largely be mitigated through feasibility projection, whereas falsifying information that affects the constraints cannot. The latter can, however, be detected using a low-complexity cross-AP consistency check. These results show that learned
ISAC schedulers should be evaluated using constraint-aware feasibility metrics in addition to conventional accuracy measures.
Read the full abstract ▾
Distributed Computing
Skywing: A Platform for Decentralized Mathematical Computing in Unreliable Environments
Alyson Fox, Colin Ponce, Annika Mauro et al. (+4 more)
Emerging edge, autonomous, and cyber-physical systems increasingly require mathematical computation across heterogeneous devices connected by unreliable communication networks. Traditional high-performance computing and distributed data-processing frameworks provide powerful abstractions for managed environments but are less suited to decentralized settings where centralized coordination, reliable communication, and global synchronization cannot be assumed. This paper presents Skywing, an open-source platform for decentralized mathematical computing in unreliable environments. Its programming model consists of three abstractions:
agents represent participants in a decentralized computation, processors encapsulate
algorithm-specific update rules, and iterations manage distributed execution. Skywing supports asynchronous operation, publish-subscribe communication, managed message handling, and the composition of independent
algorithms into complex decentralized workflows. We demonstrate Skywing using representative
algorithms from consensus, optimization, and numerical linear
algebra. Experiments on the native Skywing runtime include Push Sum and Max Consensus, a composed monitoring and control workflow, resilient Push Sum under delayed communication, and resilient asynchronous Jacobi under malevolent data corruption. These demonstrations show that Skywing supports diverse decentralized
algorithms while separating mathematical logic from communication and execution infrastructure. Skywing serves as both a deployment framework for decentralized applications and a research platform for developing resilient mathematical
algorithms.
Read the full abstract ▾
All authors Alyson Fox, Colin Ponce, Annika Mauro, Wayne Mitchell, Sarah Osborn, Tom Benson, Shayna Kapadia
Robotics · Cross-listed from math.AG
Equilibria for Networks of Linear Translational Springs
Luke Oeding, Ethan Clayton, Jackson Elsea et al. (+2 more)
We use tools from nonlinear
algebra to study the
equilibria of small linear translational spring networks. Specifically we use the techniques of homotopy continuation, monodromy, and parameter homotopy (a.k.a. cheater homotopy) to solve all rigid linear translational spring networks up to $5$ nodes in both $2$ and $3$ dimensions. We describe a method of implementing parameter homotopy that arises naturally from the physical structure of the system. We give precise total degree bounds on the maximum number of solutions for general planar spring networks. We discuss further efficiency gains obtained from polyhedral homotopy methods. We compare the computation efficiency of these techniques against a baseline of Newton's method.
Read the full abstract ▾
All authors Luke Oeding, Ethan Clayton, Jackson Elsea, Nicholas Wang, Adam Rutkowski