Software Engineering
Correct Tests Are Not Enough: Measuring and Training Oracle Conversion in Specification-Based Test Generation
Yunhao Liang, Chengguang Gan, Ruixuan Ying et al. (+3 more)
Generating tests from a natural-
language specification requires both an input that exposes faulty behavior and a correct expected output. These requirements need not improve together: a model can increase test correctness by choosing easier inputs, or discover useful inputs whose expected outputs it cannot predict. We study this interaction through executable reward decomposition and suite-level oracle-conversion measurement. Our generator jointly emits five input–output tests in one response. During training, audited reference programs provide correctness feedback, while a fixed bank of faulty programs provides two utility signals: potential input kill and effective kill after checking the generated output. An additive GRPO objective preserves both signals without requiring execution at inference time. On an audited TC-Bench split with 506 training and 142 evaluation tasks, three independently trained Qwen3.5-9B runs at step 75 increase full-test correctness from 28.59% to 42.54%, input kill from 24.06% to 25.27%, and effective full kill from 12.23% to 14.15%. Matched 50-step ablations reveal a trade-off: removing kill rewards yields higher correctness and slightly higher full kill, but lowers input kill to 21.60%. A fixed-input source–oracle crossover on 64 training-pool tasks attributes the principal NoKill-to-FullKill difference to harder input selection rather than worse output prediction on identical inputs. These results identify oracle conversion as a measurable bottleneck and show the benefits and limits of preserving input-utility feedback in joint test generation.
Read the full abstract ▾
All authors Yunhao Liang, Chengguang Gan, Ruixuan Ying, Hanjun Wei, Zhe Cui, Shiwen Ni
Machine Learning
A budget-dependent crossover between coverage- and response-based training-set selection for machine-learned interatomic potentials
Jia Bi and Alin-Marin Elena
Selecting compact training sets for machine-learned interatomic potentials requires deciding whether to preserve structural diversity or target configurations on which models disagree. The better choice can depend on how much data is retained, making a comparison at one training-set size insufficient. Here we link selection criteria to prediction accuracy through a budget-resolved comparison of retrained MACE models on GAP-20 Carbon and pooled revised MD17. Structural coverage is compared with a response-guided selector that targets disagreement between a coverage-trained model and a full-data reference. This retrospective response witness tests the value of model disagreement for compressing an already labelled pool. At 5%, coverage gives smaller absolute deviations from the full-data error than random sampling across four force endpoints in both
datasets. The witness has larger deviations than coverage at 1% and 5%, but the ordering reverses at 20%. At 20%, witness-selected models also lower direct held-out force errors by 0.46–5.89% relative to coverage, with all eight paired training-seed intervals favouring the witness. Six errors fall below the full-data reference. Mean force-error reductions are 0.164–0.167~meV~$\text{\AA}^{-1}$, with larger gains for tail and masked endpoints. Complementary analyses show that learned similarity preserves the coverage ranking, while selecting by frozen-model error gives higher error than
embedding coverage. These findings establish retained-data budget as a deciding variable in atomistic training-set selection and provide a direct test of when response-guided compression improves on structural coverage.
Read the full abstract ▾
Computer Vision
FACT: A Forensic Agent with Compiled Tool-Use Trajectories for AI-Generated Image Detection
Jiaoyang Chen, Bin Hu, Jingyu Hu et al. (+3 more)
AI-generated image detection is increasingly open-world: new image generators produce highly realistic images that make visual artifacts harder to identify. Existing detectors usually rely on a fixed set of forensic cues, so a detector that works well for one generator family may fail on another. We introduce FACT (Forensic
Agent with Compiled Tool-use Trajectories), which learns an image-conditioned tool-use policy for forensic analysis. Instead of applying a fixed detector, FACT decides which forensic tools to call, interprets the returned evidence, and stops when sufficient evidence has been collected. FACT follows an Evolve–Distill–Refine pipeline: it evolves an execution-verified forensic skill, compiles the skill into action–observation tool-use trajectories, distills them into a compact
agent, and refines the policy with cost-aware GRPO. Across two internal and four public
benchmarks, FACT achieves the best performance among all compared methods, including on recent unseen generators, deepfakes, and manipulated images.
Read the full abstract ▾
All authors Jiaoyang Chen, Bin Hu, Jingyu Hu, Kun Zhou, Qin Zhang, Zhengzhe Liu
Computational Complexity
Ulam Rank Aggregation Is Hard to Approximate for Four Rankings
Sk Ruhul Azgor, Diptarka Chakraborty, Le Van Cuong et al. (+3 more)
We study the approximability of rank aggregation under the Ulam metric. In the Ulam median problem, the goal is to find a ranking (permutation) minimizing the sum of its Ulam distances to the input rankings, while in the Ulam center problem, the objective is to minimize the maximum such distance. We prove that, for every $0<\varepsilon< 1/34$, it is $\mathrm{NP}$-hard to approximate either Ulam median or Ulam center within a factor of $35/34-\varepsilon$, even when the input consists of only four rankings. We further show that unless P = NP, neither problem admits a
polynomial-time additive approximation scheme. Prior to our work, only the exact versions of both problems were known to be $\mathrm{NP}$-hard, and that too only when the number of input rankings is unbounded [Fischer et al., ESA'25 and Bachmaier et al., J. of Discrete
Algorithms'15]. Furthermore, our inapproximability results are optimal in terms of the number of input rankings since for three inputs it is already known to be
polynomial-time solvable [Chakraborty, Das, Krauthgamer, SODA'21]. En route, we introduce a new general framework for reducing Boolean constraint satisfaction problems (CSP) to the Ulam median with only four inputs. As a specific instantiation of the reduction framework, we obtain our hardness-of-approximation results. The corresponding hardness for the Ulam center follows from a reduction from the Ulam median.
Read the full abstract ▾
All authors Sk Ruhul Azgor, Diptarka Chakraborty, Le Van Cuong, Debarati Das, Mursalin Habib, Tien Long Nguyen
Sound
Where Does the Sound Go? Tracing Acoustic Information Loss in Audio-Conditioned LLMs
Song-ha Jo, Sehyun Lee, Soyoon Kim et al. (+2 more)
Audio-conditioned
language models often underuse acoustic cues such as prosody,
emotion, and non-
speech sounds, raising the question of whether ASR-supervised frontends discard this information before it reaches the LM. We test whether the frontend is responsible by comparing Whisper-Tiny and Whisper-Small with EnCodec, DAC-VAE, and WavTokenizer in a shared Qwen3.5-4B audio-LM pipeline on ASR,
emotion recognition, and sound captioning. Encoder replacement alone does not resolve this underuse: Whisper variants remain strongest overall, including on
emotion and environmental sound captioning. To localize the failure, we trace task-relevant information through the encoder, projector, LM layers, and LM head. Linear probes and
geometric analyses show that discriminative acoustic structure remains recoverable at the final LM layer, even when MCQA accuracy trails probe accuracy by up to 83 points. Because the answer format and decoding procedure are controlled, this task-dependent gap points to content-specific readout failure rather than generic format bias. LogitLens analyses and a targeted LM head intervention support the conclusion that acoustic underuse is not explained solely by encoder-side information loss and that readout
alignment can be a dominant bottleneck.
Read the full abstract ▾
All authors Song-ha Jo, Sehyun Lee, Soyoon Kim, Jaesik Choi, Sanghyuk Choi
Information Theory
Algorithmic List Decoding at Capacity and Optimal Proximity Gaps for Reed-Solomon Codes
Fernando Granha Jeronimo
We give a unified hidden-derivative framework for list decoding and mutual correlated agreement of ordinary Reed–Solomon codes over
prime fields, on arbitrary prescribed evaluation sets. For every fixed slack $\gamma>0$, every sufficiently large block length $n$, every
prime $q\ge n$, and every dimension $1\le k\le(1-\gamma)n$, a deterministic
algorithm finds all codewords within relative distance $1-k/n-\gamma$ in $q^{O_\gamma(1)}$ time. The final list has size $n^{O_\gamma(1)}$, independently of $q$. Both statements extend to bounded-input-list recovery, with constants depending additionally on the input-list bound. For every fixed curve degree $\ell$, at most $n^{O_{\gamma,\ell}(1)}$ parameters on a curve $f_0+zf_1+\cdots+z^\ell f_\ell$ admit a nearby codeword whose exact agreement support is not a maximal jointly explained support of the coefficient words. For lines this gives MCA error $n^{O_\gamma(1)}/q$, with no proximity loss. The interpolation stage reparameterizes and optimizes the hidden-derivative construction of Brakensiek, Chen, Putterman, Zhang, and Zheng; differential root enumeration uses Kopparty's
algorithm. We then prove that a specialization-safe differential equation has a cover by constant-dimensional varieties of
polynomial cumulative degree, outside polynomially many parameter values. Intersecting these varieties with equations from the full agreement support yields both the field-size-independent list bound and exact-support MCA.
Read the full abstract ▾
Human-Computer Interaction
Tactile Search: Enhancing Targeting in 3D Space
Amber Maimon, Iddo Yehoshua Wald, Jonas Keppel et al. (+5 more)
Visual search is crucial in daily life, from scanning for relevant information to spotting signs of danger. When sensory channels are overloaded or degraded, cognitive tasks can be supported by crossmodal information representations through vibrotactile cues. We introduce Tactile Search, an approach that uses modulation of frequency and amplitude of vibrations to the hands, for guiding
attention to the location of objects in 3D space. We evaluated this approach in a competitive VR game where participants searched for targets using both
vision and touch. Across two studies – an in-the-wild demonstration (n=55) and a controlled laboratory experiment (n=28) – we found that vibrotactile feedback significantly improved performance and increased user confidence. In the combined haptic condition, performance did not differ across target heights. We further analyzed participants' subjective experiences and search strategies highlighting the benefits of the tactile cues. Our findings suggest that Tactile Search can enhance interaction and provide design considerations for integrating haptic search into interactive systems.
Read the full abstract ▾
All authors Amber Maimon, Iddo Yehoshua Wald, Jonas Keppel, Eunhee Chang, Yoshifumi Kitamura, Stefan Schneegass, Rainer Malaka, Donald Degraen
Computer Vision
Hierarchical Prompt Injector for Domain Generalization Segmentation
Xin Kun Lin, Ruoyu Guo, Jiaqi Guo et al. (+2 more)
Domain Generalized Semantic
Segmentation (DGSS) is a challenging task, as
vision models often rely on low-level appearance cues that change across domains. In contrast, structural attributes exhibit cross-domain stability, motivating the use of structural priors for DGSS. Existing methods use
prompt learning to transfer such priors into DGSS models, but typically encode each class as a single holistic
prompt. Moreover, these methods apply
prompts uniformly to all pixels, offering no mechanism to adapt when only a subset of object regions is visible due to viewpoint changes, occlusion, and environmental variation. We address this with Spatial Hierarchical
Prompts (SHP) that enrich each class with region-level
geometric anchors capturing structural appearance from distinct viewing angles, ensuring complementary coverage under arbitrary viewpoints. Additionally, we propose the Hierarchical
Prompt Injector (HPI), which enables spatially adaptive
prompt injection in foundation models. HPI spatially grounds
prompts by modeling their semantic relevance and spatial influence with visual features. Considering the difficulty of learning spatially and semantically aware
prompt injection, we further introduce auxiliary supervision to align hierarchical
prompts with their corresponding object regions. We achieve 70.62% and 72.74% mIoU on synthetic-to-real and real-to-real
benchmarks, respectively. Code and checkpoints are released at https://github.com/MosukFate/HPI
Read the full abstract ▾
All authors Xin Kun Lin, Ruoyu Guo, Jiaqi Guo, Maurice Pagnucco, Yang Song
Machine Learning
Budgeted Task-Aware Acquisition of Dynamic Networks
Zihe Zhou
Learning on dynamic graphs is difficult when changes in the underlying network are only partially observed. Acquiring current graph information incurs observation and computational costs, making complete updates impractical under limited resources. This paper focuses on budgeted task-aware acquisition on dynamic networks, where a model needs to decide which stale graph information to refresh for a downstream task. We propose Scout, a lightweight framework that learns the task value of querying each node from the maintained graph and observation history. Our evaluation covers one synthetic and 4 real-world dynamic networks, 2 downstream tasks, 9 acquisition baselines, and several query budgets. Scout achieves the highest mean downstream performance in 15 of the 21
benchmark-budget settings and ranks within the top two in 19 of them. Its average rank is 1.4, compared with 3.2 for the strongest baseline. On the same dynamic network, task-matched acquisition improves link-prediction AUC by 0.005-0.007 and node-classification accuracy by 0.007-0.017 over task-mismatched acquisition. These results show that useful graph observations depend on the downstream task and that limited observation budgets can be allocated effectively by learning directly from downstream utility.
Read the full abstract ▾
Networking
SLA-Safe Energy Control for AI-Native NG-RAN Using Stability-Aware Constrained PPO
Dharmendra Kumar
One important AI-for-RAN use case is energy saving, in which
radio resources and
cell energy modes must be dynamically controlled without violating user quality-of-service (QoS) or service-level agreement (SLA) requirements. However, aggressive sleep-state or deactivation decisions may reduce energy consumption at the cost of throughput degradation, delay increase, SLA violations, and unstable mode switching, especially under time-varying and bursty traffic conditions. This paper proposes a stability-aware constrained reinforcement learning framework for SLA-safe energy control in 5G NG-RAN. The problem is formulated as a constrained Markov decision process in which an AI-native controller selects closed-loop energy-saving actions based on
cell load, queue status, active-user information, current energy mode, and SLA-related indicators. The proposed framework uses constrained proximal policy optimization with adaptive Lagrangian penalties to account for throughput, delay, and SLA constraints. To improve operation under traffic distribution shift, the controller is trained using mixed nominal and stress traffic regimes, while a switching-stability penalty is introduced to reduce oscillatory transitions between active and low-power modes. Simulation results in a seven-
cell NG-RAN environment show that the proposed controller reduces energy consumption by approximately 41.4% under nominal traffic, 10.5% under stress traffic, and 22.9% under unseen-stress traffic relative to the Always-On baseline. Under stress and unseen-stress traffic, the controller preserves zero SLA violation and zero throughput loss, indicating service-preserving operation under challenging conditions. The proposed method also reduces switching activity compared with basic threshold-based energy saving.
Read the full abstract ▾
Machine Learning
Beyond Arbitrary Geometry: Topology Generalization In neural PDE Operators
Peiyao Chen, Zhouyuan Xu, Ran Ding et al. (+2 more)
Neural operators that accept arbitrary meshes are often treated as geometry-general, but unseen domain
topology changes both the invariant and decaying subspaces of a PDE operator. We use Hodge heat flow as a controlled lens on this distinction and introduce TopoBox-3D, where tunnels and cavities vary Betti support while the exact Hodge decomposition separates the harmonic kernel from the positive spectrum. Across six architectures, models that infer
topology implicitly suffer excess matched degradation in 37 of 45 model–task
topology-OOD
cells, yet cases that change harmonic dimension are not more strongly penalized on average. The dominant difficulty is instead spectral: the initial Rayleigh quotient is the most stable predictor of error, and spectral broadening adds information for edge and face cochains. Most strikingly, controlled probes show that explicit incidence and harmonic coordinates do not yield the best kernel-identity accuracy; nevertheless, TNO ranks first in mixed-input nonharmonic accuracy on all six tasks with nontrivial harmonic support. Together, these results establish
topology as a distinct generalization axis beyond arbitrary-geometry compatibility and show that its influence extends across the Hodge spectrum rather than remaining confined to the harmonic kernel. More broadly, they suggest that global, low-frequency structural priors may help organize predictions in the faster-decaying complementary component, offering a new perspective on how neural operators may generalize across
topology as well as geometry.
Read the full abstract ▾
All authors Peiyao Chen, Zhouyuan Xu, Ran Ding, Jiansheng Fan, Chen Wang
Machine Learning
Selective Posterior Margin Regularization for Forward-Corrected Classification
Zexing Zhang, Jichao Li, Tianyang Lei et al. (+2 more)
Learning with class-conditional label noise often relies on a transition model from latent clean classes to observed annotations. Forward correction embeds this transition in the likelihood, yet finite-sample networks may still memorize corrupted labels. The corrected likelihood also induces a reverse posterior over the clean classes that could explain each annotation. When its leading class differs from the annotation, the model and transition
matrix provide evidence against that annotation, but the leading alternatives can remain nearly tied. We introduce Selective Posterior Margin Regularization (SPMR), which preserves the Forward objective and converts this disagreement into a graded update on the clean classifier. SPMR selects the leading reverse-posterior class, scales a detached pairwise margin by the separation between the two leading posterior classes, and assigns correspondingly little influence to diffuse conflicts. The gap factorizes into transition- adjusted pairwise separation and the posterior mass carried by the leading pair. The active margin follows the locally minimum-norm logit direction that enlarges the selected pairwise margin. Across five known-transition
benchmarks, SPMR improves full-length Forward by 2.5-7.0 percentage points and remains 0.7-2.5 percentage points above Forward with Mixup and early stopping. Matched interventions support distinct gains from the posterior-space coefficient, transition-adjusted target, and pairwise action. The same design transfers to estimated transitions, human annotations, architectural changes, and stronger Forward recipes. The formulation uses latent-class evidence already available inside Forward correction without promoting every posterior conflict to a corrected label.
Read the full abstract ▾
All authors Zexing Zhang, Jichao Li, Tianyang Lei, XiongYi Lu, Yang Kewei
Information Theory · Cross-listed from quant-ph
Quantum Message Passing Convergence and Vanishing Block-Error Probability for Random LDPC Codes
Avijit Mandal, Christophe Piveteau, Joseph M. Renes et al. (+1 more)
Belief propagation with
quantum messages (BPQM) is a
quantum algorithm that decodes classical codes transmitted over classical–
quantum channels. It realizes optimal decoding on tree factor graphs over pure-state classical-
quantum channels. However, this tree-based analysis does not ensure vanishing block-error probability for LDPC Tanner graphs with cycles. In this work, we construct a two-stage BPQM decoder for random $q$-ary LDPC codes over symmetric $q$-ary pure-state channels, where $q$ is
prime, and prove that its ensemble-average block-error probability vanishes as the blocklength $N$ tends to infinity. For regular ensembles with $d_v\geq3$, fidelity bounds yield double-exponential decay of the average symbol-error probability throughout the BPQM success region. We apply depth-$\ell$ BPQM to coordinates with tree neighbourhoods and treat the remaining coordinates as erasures. With a suitable $\ell=\Theta(\log\log N)$, a noncommutative union bound controls the BPQM decoding errors, while the minimum-distance property guarantees erasure recovery. We also extend the analysis to finite-support irregular ensembles. These results are relevant to
quantum algorithms based on Regev's reduction, where coherent decoding uncomputes a codeword register. Decoded
quantum interferometry (DQI) uses a closely related
Fourier-based framework that reduces sparse max-LINSAT optimization problems to LDPC decoding problems on pure-state channels. Our results justify the use of BPQM in the decoding step of DQI and of coding-theoretic
algorithms based on Regev's reduction whenever the code is drawn from one of the random LDPC ensembles analyzed here and the induced memoryless symmetric pure-state channel lies in the BPQM success region. Curiously, our numerical results indicate that DQI+BPQM achieves a satisfaction ratio that closely matches that of simulated annealing.
Read the full abstract ▾
All authors Avijit Mandal, Christophe Piveteau, Joseph M. Renes, Henry D. Pfister
Numerical Analysis
High-order smoothness and high-order collocation approximation for Volterra integral equation with multiscale and nonlinear exponent kernel
Wenlin Qiu, Chuanwei Su and Xiangcheng Zheng
We consider a Volterra integral equation with multiscale and nonlinear exponent kernel. We propose the high-order smoothing conditions, under which the initial singularity of the solutions can be eliminated up to any prescribed order. This indicates the application of the multiscale nature of the kernel on local modification of the solutions. Then a discontinuous high-order collocation method with arbitrary
polynomial degree is developed and analyzed on uniform or graded meshes based on the solution regularity. This work serves as a comprehensive extension and complement to [Zheng, Qiu and Stynes, SIAM J. Numer. Anal., to appear] in both mathematical and numerical aspects.
Read the full abstract ▾
Machine Learning
SAFEGuard: Detect Optimization-Based Jailbreak Attacks Through Harmful Semantic Analysis and Fluency Measurement
Quoc Viet Vo, Trung Le, Damith C. Ranasinghe et al. (+1 more)
Despite the significant efforts devoted to aligning large
language models (LLMs) with human values and ensuring safe deployment, recent work has revealed that LLMs remain vulnerable to
adversarial jailbreak attacks that can bypass safety guardrails and elicit harmful responses. Many defense methods are proposed to detect jailbreaks but they are limited in their effectiveness to counter wide-range optimization-based jailbreak mechanisms that can yield highly fluency-optimized or harmful semantic obfuscated
prompts. To tackle this challenge, we propose a unified detection framework SAFEGuard which incorporates a hybrid fluency measurement based on cross-layer distribution distance and perplexity, and the analysis of harmful semantics through gradient matching. Our method is grounded in a paramount observation: high fluency
prompts maintain their malicious intention close to harmful
prompts while harmful semantic obfuscated
prompts often inject gibberish
token sequences. Our evaluation demonstrates that SAFEGuard consistently outperforms state-of-the-art baselines and achieves significant improvement in accuracy across different optimization-based jailbreaks. This underscores the effectiveness of SAFEGuard against evolving jailbreak attacks.
Read the full abstract ▾
All authors Quoc Viet Vo, Trung Le, Damith C. Ranasinghe, Ehsan Abbasnejad
Numerical Analysis
Efficient $C^1$ Bernstein Quasi-Trefftz Discretization for Heterogeneous High-Frequency Helmholtz Problems
Shelvean Kapita
We develop an efficient $C^1$ Bernstein quasi-Trefftz discretization for heterogeneous high-frequency Helmholtz problems on unstructured triangular meshes. The method first imposes strong $C^1$ continuity through Bernstein–Bézier smoothness relations on local macro-patches and then reduces the resulting conforming
polynomial space by enforcing projected Helmholtz residual constraints. On two-triangle patches, this compresses the local dimension from quadratic growth in the
polynomial degree to a trace-sized space of dimension $2p+1$ under the natural rank condition. Variable
matrix-valued coefficients are handled through local
polynomial projection, allowing coefficient-approximation effects to be separated from discretization error. A graph-residual formulation couples the reduced patch spaces and produces a sparse global system in compressed coordinates. The implementation combines explicit $C^1$ continuation, batched coefficient projection, batched residual construction, and stable local kernel extraction by QR factorization. Numerical experiments on heterogeneous unstructured meshes demonstrate high-order accuracy, roundoff-level $C^1$ conformity, robust high-frequency resolution, and substantial reductions in local and global computational cost. Turning-point and penetrable-scattering examples further illustrate the flexibility of the approach.
Read the full abstract ▾
Machine Learning · Cross-listed from stat.ML
Functional Attentive Interpretable Regression
Haixu Wang, Tianyu Guan and Jiguo Cao
In function-on-function regression, the coefficient surface $\beta(s,t)$ may exhibit complex support structure—from localized patches to global patterns such as disconnected regions, bands, or rings—where effect similarity does not align with Euclidean proximity. Projection-based methods that rely on fixed basis expansions can obscure such structure, while direct smoothing approaches risk oversmoothing the surface and its boundaries. We propose Functional Attentive Interpretable Regression (FAIR), which represents $\beta(s,t)$ directly through coordinate features and uses self-
attention to learn effect-adaptive neighborhoods, enabling information sharing at both local and global scales. A scalar compression network maps these learned representations to the coefficient surface. Sparsity and smoothness penalties applied over these neighborhoods promote localized support with coherent boundaries. We establish a sieve equivalence to tensor-product spline spaces and derive convergence rates. Simulations and applications to oceanographic and hydrological data demonstrate that FAIR recovers support geometry more accurately than existing methods while achieving superior prediction, particularly under sparse sampling.
Read the full abstract ▾
Networking
Biased Backpressure Routing for Multihop Wireless Networks with Heterogeneous Interfaces
Yujun Ming, Zhongyuan Zhao, Fikadu Dagefu et al. (+5 more)
Heterogeneous-interface multihop wireless networks (Het-MuNets) are emerging as a promising paradigm for tactical networks and for infrastructure-light applications such as vehicular communications, wireless backhaul, and non-terrestrial connectivity. To exploit the diverse profiles of heterogeneous communication technologies in penetration, interference, and bandwidth, packet-to-interface assignment must be determined on a per-hop basis, making routing and scheduling highly complex. In this work, we develop a unified framework for joint packet routing, link scheduling, and interface assignment in Het-MuNets with multiple concurrent flows. By modeling packet-to-interface assignment as transmission between virtual subnodes, we transform interface assignment into intra-device virtual routing, which is solved jointly with physical routing and scheduling under a unified multi-layer shortest path-biased Backpressure (SP-BP) scheme. Numerical results demonstrate that the proposed framework outperforms SP-BP operating on other baseline graph models and non-backpressure routing schemes in goodput, latency, and packet delivery rate.
Read the full abstract ▾
All authors Yujun Ming, Zhongyuan Zhao, Fikadu Dagefu, Justin Kong, Terrence Moore, Kevin Chan, Ananthram Swami, Santiago Segarra
Computation and Language
SinoGlyphBench: A Diagnostic Benchmark for Chinese Glyph-Level Obfuscation in Language-Model Moderation
Yifan Wang, Zimu Wang, Suliu Qin et al. (+10 more)
Glyph-level obfuscation can leave harmful Chinese content readable to humans while degrading automated moderation. We introduce SinoGlyphBench, a diagnostic
benchmark that identifies label-critical semantic anchors and creates matched original and glyph-obfuscated inputs in text and image modalities. By perturbing anchors, background context, or both, this design distinguishes corruption of moderation-relevant evidence from general surface variation. Across 176,916 paired evaluations of 12 LLMs and MLLMs, obfuscation increases harmful false-negative and false-positive rates by 6.1 and 4.7 percentage points, respectively, and reduces four-way accuracy by 5.0 points. Models retain 75.7% of the decisions that were correct on the matched original inputs. Full-scope perturbations cause the largest degradation, anchor-only perturbations are more damaging than background-only perturbations, and cross-script substitution is particularly difficult in the text modality. Analysis of structured outputs identifies observable mismatches in visible-form reading, intended-message recovery, and final safety judgment. The evaluated models, therefore, remain brittle to Chinese content written with non-canonical glyphs. Resources are available at https://github.com/fengshun124/SinoGlyphBench.
Read the full abstract ▾
All authors Yifan Wang, Zimu Wang, Suliu Qin, Changyu Zeng, Tong Chen, Siqi Chen, Yijie Lin, Lingyu Jiang, Jionglong Su, Yushan Pan, Haiyang Zhang, Wei Wang, Qiaoyu Tan
Artificial Intelligence · Cross-listed from quant-ph
Do Quantum AIs Dream in Paths? Path-Integral Slow Thinking through Grover Interference
Xiansheng Cai, Xiu-Hao Deng and Kun Chen
Reinforcement learning with verifiable rewards enables large
language models to think slowly, but the same training can induce policy collapse: probability concentrates onto a few successful trajectories and exploratory diversity erodes. We ask whether
quantum AI can realize slow thinking differently. We formulate slow thinking as coherent dynamics over
reasoning trajectories, a discrete path integral in which action sequences coexist in superposition and recombine before measurement. In our trainable realization, an exact verifier partitions the ensemble into collective accepted and rejected components that interfere under Grover amplitude amplification. A finite Grover
evolution is maximized when the pre-amplification success probability lies at an analytically determined value below one, so inference itself defines an interior training target and removes the monotonic pressure toward unit success. In exact statevector simulations of a 2x3 sliding puzzle, Grover training reaches accuracy 0.95 on a 32-question training set at one round, against 0.73 for the strongest classical control. On held-out questions specialization has a cost: an untrained uniform policy read out through the same amplification remains the strongest reference on this solution-dense
benchmark, and
quantum training preserves far more held-out accuracy than classical training - at four rounds with matched
circuit applications the two
quantum models reach 3.2 and 3.9 times the strongest classical controls. The number of training questions supported by fixed-size policies trained at each amplification budget also grows faster with the budget than with matched classical repetition. These results establish a Grover-based realization of path-integral slow thinking: the interior target preserves exploratory path diversity, and ensemble-level interference converts it into verified performance.
Read the full abstract ▾
Computer Vision
Map the Possibilities: Spatial Belief Fields for Language-Goal Aerial Navigation
Haotian Xu, Yue Hu, Zhengqiu Zhu et al. (+3 more)
Language-goal aerial
navigation requires an
agent to local- ize a potentially unobserved target from relational instruc- tions and partial observations, and translate this inference into metric actions in large-scale continuous environments. Existing methods often reduce
language grounding to one single waypoint or action, prematurely collapsing the spatial uncertainty inherent in incomplete evidence and ambiguous relations. To address this limitation, we introduce SBFNav, a closed-loop
navigation framework centered on a
language- conditioned Spatial Belief Field (SBF). Unlike ego-centric maps that primarily record what has been observed, SBF rep- resents a task-conditioned distribution over plausible target locations, preserving multiple spatial hypotheses under par- tial evidence. At each step, this distribution is updated from accumulated observations as new evidence becomes avail- able. Built on this representation, SBFNav selects the goal that best aligns with the instruction and observations as a met- ric waypoint for control. Experiments on both the original and revised CityNav
benchmarks achieve the best reported overall performance. On the Test Unseen split, our method improves SR from 25.91% to 32.29% and SPL from 19.63% to 30.43%. Ablation studies further confirm the advantages of spatial-belief modeling over single-point prediction.
Read the full abstract ▾
All authors Haotian Xu, Yue Hu, Zhengqiu Zhu, Yong Zhao, Wenguan Wang, Quanjun Yin
Artificial Intelligence
AgentBrew: Offline Tool-Use Agent Learning from Raw Real-World Trajectories
Zhiyi Lyu, Yewen Li, Longtao Zheng et al. (+7 more)
LLM-based
agents are increasingly deployed in real-world applications through tool-use APIs, yet training them for specific environments remains fundamentally difficult: real-world applications provide no pre-defined tasks or verifiers, no faithful simulators, and limited budget for large-scale environment interaction. In this paper, we propose AgentBrew, an offline training framework that learns effective tool-use policies from a single batch of raw interaction trajectories, without task verifiers or iterative on-policy rollouts. The
agent first explores the target environment to collect a raw trajectory corpus without quality filtering. To extract training signal from this noisy corpus, retrospective task inference reconstructs an aligned instruction for each trajectory based on its actual outcome, and PMI-Based credit assignment decomposes the trajectory's total information about the inferred instruction into additive per-action credits via pointwise mutual information (PMI). These credits weight the policy training objective, amplifying informative actions while suppressing ineffective ones. On three real-world MCP applications (GitHub, Notion, PostgreSQL), AgentBrew improves Qwen3-32B by +8.7 Acc / +9.7 Score on average, surpassing Qwen3-235B (+2.3 / +4.4) and outperforming rejection sampling (+5.9 / +10.3). These results demonstrate that fine-grained offline learning can recover useful supervision from raw trajectories that filtering-based approaches would discard. The code is available at https://github.com/alphatogo/AgentBrew
Read the full abstract ▾
All authors Zhiyi Lyu, Yewen Li, Longtao Zheng, Shengtian Yang, Lang Feng, Lei Feng, Peng Jiang, Kun Gai, Qingpeng Cai, Bo An
Artificial Intelligence
Learning Counterfactual World Models for Embodied Reasoning under Partial Observability
Todd Y. Zhou and Daniel Zhang
World models promise a general route to
embodied intelligence: learn predictive dynamics once, then reason, plan, and act with them. Increasingly, the representations beneath such models are pretrained on large-scale
video, interaction, and
multimodal corpora, which raises a question prediction quality alone cannot answer: when is a learned representation actually actionable? We identify a failure mode we call counterfactual collapse: a model predicts visually plausible futures while failing to distinguish interventions with different behavioral consequences. This arises whenever a representation is optimized for perceptual similarity rather than intervention structure, which is precisely the objective under which most large-scale pretrained encoders are learned. We introduce Counterfactual Latent World Models (CLWM), which combine a recurrent belief-state encoder, action-conditioned latent dynamics, and a contrastive counterfactual objective that separates futures induced by distinct interventions even when their observations look alike. Across occluded manipulation, aliased
navigation, and long-horizon manipulation, CLWM improves planning success over the strongest baseline (65.1% $\to$ 74.6% on Occluded Push and 67.3% $\to$ 78.9% on Aliased Maze) and reduces exploitative planning failures (18.4% $\to$ 9.7% on Deferred Kitchen), with ablations attributing the gains to hard counterfactual negatives, especially perceptual-alias negatives. Finally, our counterfactual separability metric, which tracks planning success across the five baseline model classes ($r \ge 0.94$), is representation-agnostic: given intervention-outcome labels, it can audit any encoder, pretrained or trained from scratch, before a planner trusts it. We do not yet measure it on large-scale pretrained encoders. Here we establish the metric and its relationship to planning success for world models trained from scratch.
Read the full abstract ▾
Robotics
CR-VLA-Force: Learning Control-aware Compliance VLA Model for Robust Contact-rich Robotic Manipulation
Zhaohong Mai, Chao Wang, Chao Zeng et al. (+4 more)
Integrating visuomotor policies or
Vision-
Language-Action (VLA) models with force/torque (F/T) perception has demonstrated significant progress in imitation learning for robotic manipulation. However, existing force-aware VLA models frequently exhibit limited capability in precise force tracking and rapid successive adjustments. This deficiency stems from the limitations of action-chunk execution strategies and the substantial latency between perception and real-time control. Such limitations can lead to task failures and safety risks, particularly when the execution of an action chunk exerts excessive interaction forces without timely adjustment. To overcome this challenge, we propose the Control-aware Compliance VLA (CC-VLA) framework for reactive control. The CC-VLA model employs a
multimodal mixture-of-experts (MoE) to encode force signal sequences and
vision-
language fused feature. Furthermore, it utilizes a multi-stage training strategy to ensure robust perception within the visual-semantic space and effective force perception under sparse sampling conditions. Additionally, a VLA-guided adaptive compliance controller is designed to facilitate precise position tracking during contact-free motion and optimal force-position tracking for contact-rich tasks. To facilitate high-precision F/T data acquisition, we also implement an adversaria shared teleoperation strategy for contact-rich demonstrations that bolsters system safety and interactivity. Extensive real-world experiments demonstrate that CC-VLA significantly improves success rates in challenging force-perception tasks and enhances force-control precision, while providing multi-level safety and robustness under the tested partial-OOD pose-shift settings.
Read the full abstract ▾
All authors Zhaohong Mai, Chao Wang, Chao Zeng, Sitong Mao, Heng Zhang, Shunbo Zhou, Chenguang Yang
Computational Complexity · Cross-listed from quant-ph
On the sample complexity of the generalized hidden shift problem over arbitrary finite groups
Juntai Zhou
We formulate the generalized hidden shift problem over arbitrary finite groups and obtain upper and lower bounds for the sample complexity in terms of group theoretic data.
Machine Learning
Scaling Optimal Classification Trees via Adaptive Feature and Sample Reduction
Jiancheng Tu and Wenqi Fan
Dynamic programming for optimal classification trees becomes computationally expensive as the numbers of features and training samples increase. We develop a joint feature- and sample-space reduction framework based on STreeD. Weighted STreeD merges duplicate records created after projection onto a fixed candidate set into weighted representatives. This reduces sample-dependent computation without changing the fixed-candidate optimization problem. Adaptive STreeD repeatedly refines a bounded candidate set, retains features used by the incumbent tree, rebuilds the weighted representation, and solves the resulting reduced problems. Each certified Weighted STreeD solution is optimal for its current candidate set, while the outer feature search remains heuristic over the full feature space. Experiments on five data sets show that Weighted STreeD achieves speedups of up to 121.41 times over standard STreeD. Adaptive STreeD reduces runtime in matched comparisons at depths 2 to 4 and continues to return feasible trees at greater depths where full-feature methods are limited by time or
memory. Under the same computational budget, its predictive performance remains comparable to the evaluated optimal classification tree baselines and is higher in some comparisons. These results show how joint feature- and sample-space reduction can scale dynamic-programming-based optimal-tree learning to more demanding instances.
Read the full abstract ▾
Artificial Intelligence
Beyond Top-$k$ Skill Retrieval: Diversity-Aware Skill Routing for LLM Agents
Wang Wei, Tiankai Yang, Samyadeep Basu et al. (+7 more)
Large
language model (LLM)
agents increasingly rely on external skills, but routing user requests over large skill registries is difficult because many skills are functionally redundant while complex tasks often require complementary skill sets. Existing skill routers typically rank candidates independently by query relevance, which can waste context budget on redundant skills. We propose Diverse Skill Routing (DSR), a diversity-aware reranking framework that uses a Determinantal Point Process to balance relevance and non-redundancy. DSR introduces a query-residual diversity kernel that penalizes redundant skill overlap while reducing penalties caused only by shared query relevance. On the SkillRouter
benchmark, DSR improves recall and full coverage over a strong pointwise reranking baseline, with larger gains on multi-skill queries. These results suggest that skill routing should be treated not only as relevance ranking, but also as complementary set selection.
Read the full abstract ▾
All authors Wang Wei, Tiankai Yang, Samyadeep Basu, Hongjie Chen, Yue Zhao, Zhengzhong Tu, Xiyang Hu, Franck Dernoncourt, Ryan A. Rossi, Hoda Eldardiry
Machine Learning
Generalizing HVAC Control With Domain Randomized Reinforcement Learning
Pablo Boitel and Kun Zhang
Deploying advanced HVAC (Heating, Ventilation and Air Conditioning) controllers at scale remains difficult because performance often depends on accurate building models or per-site retuning. We propose NOMAD-RL (Neural Online Meta-Adaptation for Dynamics), a general-purpose Reinforcement Learning (RL) controller designed to transfer across heterogeneous thermal zones through a universal, non-invasive thermostat interface. The controller acts on temperature setpoints from zone measurements and forecasts, while a recurrent policy supports online adaptation under partial observability. Our main contribution is an adaptive domain randomization scheme based on physics-informed normalizing flows, which models correlated and
multimodal distributions of thermal-zone parameters while maintaining physical plausibility and controllability. This produces a realistic and progressively adaptive training curriculum that improves transfer across buildings. We evaluate NOMAD-RL against a constant-setpoint PID controller, RL without domain randomization, and MPC in single- and multi-zone settings. NOMAD-RL consistently outperforms the PID and non-randomized RL baselines, and approaches the performance of a well-tuned MPC, especially in the more challenging multi-zone case. These results highlight the potential of adaptive, physics-informed domain randomization for robust and transferable HVAC control.
Read the full abstract ▾
Computation and Language
CONDUIT: A Unified Residual-Stream Restoration Framework for KV Cache Reuse in Vision-Language Models
Pengan Chen, Kaisheng Zheng, Liang Hong et al. (+10 more)
Vision-
language models (VLMs) often answer new questions about recurring visual content, where reusing the key-value (KV) cache can avoid re-encoding expensive visual prefixes. Exact-prefix reuse, however, fails when the same visual content appears under a changed prefix. Selective recomputation can recover quality under a small visual-
token budget, but only when the right stale
tokens are refreshed. Raw-
attention selection can waste budget on high-
attention tokens with small value-norm proxy scores and on query-irrelevant images. To address these failure modes, we propose CONDUIT, a training-free refresh policy that unifies single- and multi-image reuse as residual-stream restoration. Building on norm-weighted
attention, CONDUIT ranks cached visual
tokens using cached-key query
attention and an accessible pre-output cached-value-norm proxy, then applies empirical image-level relevance amplification before one global selection. With one image, the coefficient is one and the rule reduces to intra-image
token selection. The method preserves model architecture and weights, adding only a single query-conditioned scoring pass at inference. At a 10% refresh budget, CONDUIT achieves 97.0-99.5% of the corresponding full-prefill five-
dataset average across three VLM backbones and leads budgeted methods on average; on the MMLongBench-Doc latency subset, it uses 13.5% of full-prefill FLOPs and achieves a 2.99x time-to-first-
token speedup.
Read the full abstract ▾
All authors Pengan Chen, Kaisheng Zheng, Liang Hong, Lixia Yi, Jiyue Jiang, Jiayang Chen, Yixuan Wang, Yimin Fan, Xinyuan Liu, Jiayi Li, Zhanqiu Zhang, Yiwen Guo, Yu Li
Machine Learning
Online Learning with LLM Experts from Limited Feedback
Wang Wei, Soumyabrata Pal, Koyel Mukherjee et al. (+4 more)
We study adaptive routing of
prompts to large
language model (LLM) experts to maximize response quality in an online setting with limited feedback. We formulate it as a bandit problem with $K$ actions that represent experts and $d$ features that encode
prompts, over a horizon of $T$ rounds. We propose
algorithms that strategically select and observe rewards to minimize regret. In the full-information setting, we achieve a regret of $\tilde{O}(d T / \sqrt{m})$, while in the bandit setting we achieve $\tilde{O}(d T \sqrt{K / m})$, where $m \ll T$ is a budget on feedback. Our experiments show that we efficiently learn high-quality routing strategies across diverse LLMs from limited feedback.
Read the full abstract ▾
All authors Wang Wei, Soumyabrata Pal, Koyel Mukherjee, Franck Dernoncourt, Ryan A. Rossi, Branislav Kveton, Hoda Eldardiry