Papers

Functional Web generation is increasingly optimized with executable rewards, yet existing methods largely focus on the quality of the final page and leave the process of diagnosing and repairing imperfect implementations underexplored. We identify a central challenge in this setting: the Generator and Refiner produce executable artifacts with direct environment rewards, whereas the intermediate Critic influences downstream behavior without a directly executable outcome. We introduce WebLoop, an execution-grounded framework that jointly learns generation, critique, and refinement within a shared policy. WebLoop trains an execution-free Critic with complementary signals for requirement-level discriminability and downstream helpfulness, first establishing reliable diagnosis and then introducing consequence-aware credit, while all three roles are jointly optimized with group-relative policy learning. With Qwen3.5-9B, WebLoop reaches 41.5 Overall on WebRise and 38.9% accuracy on WebGen-Bench, improving the base model by 11.3 and 15.4 points, respectively. The gains transfer to first-pass generation, persist at 27B scale, and generalize from text-only training to multimodal inputs. Controlled analyses further show that the improvement cannot be explained by an additional refinement pass alone, highlighting the importance of learning the Critic and the loop itself.

Multi-agent LLM systems negotiating with a stateful counterpart waste model calls in three ways: polite loops that never meet the counterpart's hidden acceptance condition, malformed outputs that trigger retries, and compliance deadlocks in which the counterpart demands something the agent must refuse. We study a three-part control stack - a 5-Pillar runtime constitution, a 4-tier swarm (Director, three-agent majority vote, Monitor, schema hard gate) and Cognitive Annealing (deterministic deadlock detection, atomic purge of the agent-side context, a canonical recovery message) - against a released adversarial Gatekeeper whose acceptance rules are fixed regular expressions and whose LLM only renders reply text. The testbed has a known solution: it measures whether the stack executes a constitution-aligned strategy against swarm drift and recovers from deadlock, not whether it discovers anything. In five runs per configuration (30 runs; Gemini 2.5 Pro agents, Claude Haiku 4.5 Gatekeeper) we find: (i) the constitution and Director make an acceptable framing possible but not reliable - 0/5 baseline unlocks versus 1/5 and 2/5 with the constitution; when the swarm unlocks it does so in one turn with 7-8 calls and about 15k tokens (67-73% below baseline); when it does not, it costs 17-38% more; (ii) the Monitor and hard gate do not reduce unlocks and leave an audit trail; (iii) under a honeytrap-to-compliance deadlock, LLM-only steering escapes 0 of 5 times while atomic purge plus a canonical strike escapes 5 of 5 (Fisher $p = 0.008$) at the same call budget, with zero calls for the strike. LLM-written strikes failed the deterministic pre-flight 5 of 5 times although an LLM Monitor had approved four. Pre-registered hypotheses on average call and token reduction were not supported. Cost is bounded in every arm by deterministic stop rules; the stack adds recovery at no extra model cost.

As leading LLMs improve on capability evaluations, their limitations in producing creative outputs on design tasks remain insufficiently characterised. Our work introduces Design Creativity Bench, a benchmark that evaluates diversity and appropriateness in UI designs. It measures distinctiveness among models on the same prompt (originality), how much a model's designs change between two prompts for the same UI goal in different product domains (creative range), and the share of a brief's acceptance criteria each design meets (appropriateness). Originality is 0.592 for same-prompt design pairs from different models (95% CI [0.582, 0.602]), far below the 0.764 for same-prompt human-model pairs (95% CI [0.751, 0.778]). Creative range is 0.581 across models (95% CI [0.567, 0.597]), against 0.902 for human designs (95% CI [0.884, 0.919]). Appropriateness is above 90% for every model in our test set, and the best model reaches 99.2%, slightly above the 98.0% for human designs. Our work shows that the default output of LLMs, though generally appropriate, is substantially more repetitive than the human baseline. This calls for strong measures to address the issue. A live leaderboard with newer models is maintained at https://kombai.com/benchmark/.

Deep randomized models fix hidden-layer parameters through random initialization and learn only closed-form readouts, typically adding depth by stacking random trans formations without target-aware control of hidden-state evolution. We propose LAIR Net, the Leaky Alignment-Impulse Residual Network, which mixes a shallow learned anchor into each hidden state through a leaky residual transition. We derive a depth uniform bound on input-perturbation sensitivity and use controlled simulations to attribute gains over a randomized baseline to the anchor rather than recursion or added capacity. Benefits emerge when a nonlinear target structure is learnable at the available noise level and diminish for nearly linear targets or dominant noise. Across 23 benchmark datasets, LAIR-Net achieves the best average rank among eight randomized networks and twelve conventional models, with relative performance associated with the same nonlinear-structure and noise quantities identified in simulation.

Flow models generate trajectories from an initial distribution to a target distribution by solving an ordinary differential equation defined by a velocity field. Flow matching learns this velocity field by modeling the transport dynamics between the two distributions. Wavefunction flow establishes a formal connection between flow models and quantum dynamics by introducing a continuity Hamiltonian, which drives the Schrödinger evolution of quantum states. In this paper, we investigate accurate and efficient quantum simulation of the wavefunction flow, thereby realizing the efficient implementation of flow models on quantum computers. We first leverage a quantum read-only memory (QROM)-based phase kickback framework for the wavefunction flow simulation, generating probability densities that closely match those produced by the corresponding conventional flow model. To address the high circuit-resource cost, we further incorporate a trained quantum neural network (QNN) into the phase kickback framework, replacing QROM for data encoding. Numerical experiments demonstrate that our proposed method implements flow models on quantum computers more efficiently, since it maintains the accuracy of wavefunction flow simulation compared with the QROM-based framework, and significantly reduces the circuit resources.

DNNs exhibit robustness and generalisation issues not seen in humans. They are also far less data-efficient learners, requiring considerably more training samples to accurately classify novel exemplars. Inductive bias could help with these issues by providing in-built mechanisms to improve generalisation, and hence, reduce reliance on learning from data. We incorporate a biologically-inspired inductive bias into a new activation function, HAND (Homeostasis, Accelerating Nonlinearity, and Divisive-nomalisation), and show its effectiveness with CNNs trained on image classification. Using HAND a ConvNeXt-tiny required 25 training epochs to reach the same accuracy on ImageNet1k as the unmodified model achieved after 200 epochs. Consistent with the effects of an inductive bias, the performance gap reduced with training time and increased data augmentation. When the volume of training data was reduced and unevenly distributed between classes (Long-tailed ImageNet) the improvements in accuracy were even larger and did not reduce with increased training time. Generalisation performance with the common-corruptions data, and the ability to reject samples from unknown classes, were unaffected or improved by HAND. Results generalised across CNN architectures and training data-sets. HAND can, therefore, reduce the required training time and/or the required volume and variety of training data, helping to improve sample efficiency.

This paper develops a unified framework for establishing original energy dissipation of equilibrium-preserving exponential time differencing Runge–Kutta (ETDRK) methods for $L^2$ gradient flows. The analysis is based on an auxiliary continuous evolution representation of a general ETDRK method, which allows the energy variation to be controlled without imposing positive-definiteness conditions on the Butcher coefficients or introducing stabilization terms. For a broad class of equilibrium-preserving ETDRK schemes, we derive an explicitly computable time-step restriction. Explicit restrictions are obtained for representative schemes of orders two through five, and the framework also applies to arbitrarily high-order interpolation-based ETDRK methods. For the latter class, the resulting admissible step sizes are substantially larger than those obtained from the previous interpolation-specific analysis. Numerical experiments for the Allen–Cahn equation confirm the theoretical dissipation restrictions.

Legged robots offer superior mobility in unstructured environments, but reliable operation in such conditions requires robust state estimation. To address the vulnerability of proprioceptive estimators in rough terrain, recent methods have incorporated radar to provide velocity measurements. However, their limited yaw observability still leads to drift, and failure-aware fusion for adverse environments remains underexplored. In this letter, we present RAGNAROK, the first radar-visual-kinematic-inertial SLAM designed for robust operation in challenging environments. It integrates slip- and rolling-contact-aware leg velocity estimation, a kinematics-aware radar factor, and degradation-aware image enhancement. We further incorporate a B-spline-based radar-aided proprioceptive backbone, adaptive weighting, and online extrinsic calibration. Extensive experiments on public and self-collected datasets demonstrate that RAGNAROK achieves robust performance under challenging conditions and outperforms state-of-the-art baselines. The source code and dataset are available at https://github.com/hanjun815/RAGNAROK.

Self-distillation can improve reasoning without a separately trained, more capable teacher, but its effectiveness depends on how the self-teacher gains an advantage over the student. Conditioning the teacher on reference answers or solutions can provide such an advantage, but this information may be unavailable. Reflection offers a way to derive explicit error diagnoses and revision guidance from self-generated attempts, yet existing reflection-based methods often combine it with reference information, rich task feedback, or persistent memory. We introduce ReTeach, a Reflective self-distillation framework that constructs its self-Teacher through multi-round reflection and retry using only self-generated attempts and outcome-level verification. Starting from an unsuccessful student rollout, the teacher alternates explicit reflection with renewed attempts until success or the retry budget is exhausted, without reference answers or solutions, external diagnostic feedback, or cross-example memory. Each failed retry informs subsequent reflection, while successful correction provides outcome-level evidence for the potential utility of the resulting teacher context. An outcome-aware selection and weighting strategy distinguishes initially correct, reflection-corrected, and unresolved examples, assigning separate weights to their category-normalized distillation losses. Through on-policy distillation, the student matches the teacher's context-conditioned token-level predictive distributions at prefixes of its own rollouts, transferring the benefits of iterative correction while retaining single-pass inference. Across six benchmarks spanning mathematical reasoning, science question answering, and tool use, ReTeach improves average accuracy over GRPO by 1.39 percentage points.

Atomistic simulation is a fundamental tool for studying long-term materials evolution, from diffusion and defect dynamics to interfacial reactions and fracture. Yet conventional simulators typically advance at microscopic resolution, spending substantial computation on low-impact local updates before reaching structurally consequential states, an evolutionary-resolution bottleneck that limits long-horizon simulation. We propose AtomWorld-Mirror, a time-aware macro-step world model for the critical evolution backbone of atomic systems. For Step-Wise atomistic simulation, AtomWorld-Mirror distills short micro-event segments into physically reachable transitions between key states, jointly predicting sparse structural edits and accumulated physical time through latent macro-step dynamics. Local reachability, inventory conservation, and continuous-time consistency constrain each transition. By amortizing local atomic physics into a reusable latent macro model and replacing explicit micro-event replay with macro-step inference, this formulation provides a path toward substantially faster prediction of long-term materials evolution while preserving structural validity and time semantics. Across five atomic systems, spanning Cu-rich RPV steel irradiation aging, Cu-Zr metallic glass, and Li$_3$N-based anti-perovskite solid electrolyte, macro-step inference delivers a speed up of $10^3$ to $10^4$ times over event-by-event simulation.

Spiking neural networks (SNNs) offer an energy-efficient computing paradigm through sparse event-driven computation, showing great potential for efficient multimodal learning. However, applying SNNs to high-level multimodal tasks, such as image-text retrieval (ITR), remains challenging, since sparse spike representations make it difficult to capture semantic structures required for cross-modal alignment. Existing spiking ITR methods rely on local alignment and additional soft-label supervision during training, while lacking awareness of structural and multi-granularity relationships. To address these issues, we propose a Multi-head Spiking Graph Attention Network (MSGAT) for structural modeling and equip it with dynamic attention heads to capture complementary relational patterns and enable spike-driven graph reasoning and aggregation. However, within a two-branch multi-granularity fusion framework, the fine-grained spike representations generated by MSGAT are sparse and discrete, whereas the global representations are continuous, making conventional feature-level fusion susceptible to interference across heterogeneous representations. Therefore, we introduce Sim-Fuse, a similarity-space fusion alignment strategy integrating coarse- and fine-grained matching relations while avoiding direct fusion of heterogeneous representations. Experiments on Flickr30K and MSCOCO show our method outperforms ANN methods under matched settings and existing SNN retrieval baselines. Moreover, with only two time steps, our SNN achieves comparable or superior performance to its ANN counterpart while reducing theoretical module-level energy by 55%. The code is provided in the Supplementary Materials.

Database systems are predominantly developed in unsafe languages (e.g., C/C++) to meet performance requirements through low-level memory management, yet this reliance renders them prone to systemic memory-safety issues that compromise reliability, consistency, security, and durability. Through an extensive bug analysis of prominent database systems, we show that these memory-safety issues persist in production environments despite advancements in database testing tools. While emerging hardware-assisted extensions, such as Arm's Memory Tagging Extension (MTE) and CHERI, offer a promising mitigation path for memory safety, their practical applicability and performance overhead within the specialized constraints of database systems remain largely unexplored. In this paper, we evaluate hardware extensions through the lens of what we define as the database trilemma: the fundamental trade-off between safety, performance, and portability, to determine their viability for production-grade database systems. We provide the first side-by-side comparison of MTE and CHERI across a database workload suite. Our bottom-up study spans microarchitecture, the compiler/runtime/OS stack, core data structures (ART, B+Tree, hash table, skip list, linked list, and queue), and full database systems (Redis, LevelDB, SQLite, MySQL, DuckDB, and LadyBugDB) to characterize their performance, safety guarantees, and ease of adoption. We find that while hardware extensions can offer near-deterministic protection, they introduce non-uniform performance taxes: MTE provides high portability with modest overheads (~10%), while CHERI delivers superior safety guarantees with higher performance penalties (20-60%) and significant porting effort. Our study equips database systems architects with an actionable guide toward building the next generation of reliable, secure databases using hardware-assisted safety mechanisms.

Existing singing pitch correction approaches rely on target melodies or accompaniment tracks, which may be unavailable in practice. We formulate reference-free singing pitch correction as a music-constrained sequence editing task that determines whether and how each note should be corrected from the input performance alone. A pretrained symbolic music encoder with lightweight singing-domain adapters produces contextual representations of the singing MIDI. Based on these representations, two lightweight correction heads jointly model correction necessity and signed pitch modification through a factorized pitch-editing distribution. An input-dependent tonal prior derived from the estimated key distribution then reranks the candidate offsets, favoring tonally compatible corrections without target melodies or ground-truth key annotations. Experiments on real paired amateur and professional singing recordings show that the method improves note-level pitch accuracy from 73.25 to 83.43, outperforming a context-based baseline by 3.99 percentage points while balancing error correction and preservation of correctly performed notes.

Federated reinforcement learning (FRL) enables distributed agents to collaboratively train decision-making policies, but its decentralized training process also exposes global policy learning to Byzantine manipulation. Existing poisoning attacks primarily focus on how to construct malicious updates, while trajectory-level intervention timing remains largely implicit. In sequential decision making, however, where an intervention is applied can alter subsequent trajectories and learning signals. Through controlled experiments, we find that changing the selected trajectory states materially alters attack efficacy even when the malicious-update construction is fixed. We therefore identify when as a distinct attack dimension and introduce the Viability-constrained Behavioral Steering Attack (V-BSA), which uses local policy uncertainty to select sparse intervention states and applies envelope-constrained behavioral steering. Across discrete-action benchmarks, V-BSA achieves substantial degradation against robust aggregators and ensemble defenses with only a fraction of the interventions used by dense poisoning, while revealing task- and aggregation-dependent boundaries. Overall, our results highlight intervention timing as a distinct dimension of sequential robustness in FRL. The code is available at https://github.com/Yodeesy/V-BSA

AutoCFD5 Case 2 compares the participants' CFD results with wind tunnel measurements of surface pressure, velocity profiles and force coefficients. To place each run on a diagram of error against computational cost, these quantities must be combined into one number. This note defines that number, the joint error. Following the practice of interlaboratory proficiency testing (ISO 13528), every deviation from the experiment is expressed in units of the spread of the participants' results at that point, estimated robustly with the standard's Algorithm A. The point-wise spreads are pooled into one scale per component, so that every measurement point carries equal weight within its component. The joint error is a fixed weighted sum of five component errors. The scales are computed once from all AutoCFD5 submissions received by 5 October 2026 and then frozen, so that scores are reproducible and comparable across workshops.

End-to-end speech parsing, a task recently proposed, consists in predicting both the transcription and the syntactic tree for a spoken utterance. Existing architectures for speech parsing often utilise intermediate neural networks. In this work, we examine the effectiveness of intermediate neural networks (NN) for parsing, and, specifically, what role do they play. We introduce a simpler end-to-end architecture for speech parsing, where we remove these intermediate NN units, reducing the parameters by 12%, while achieving comparable or better performance than prior method on both automatic speech recognition (ASR) and parsing. We demonstrate that intermediate NN units help reduce the representational gap when the pre-trained encoder is frozen. We do a comprehensive evaluation of speech parsing on French, and medium-low resource languages Slovenian and Naija. We further investigate the impact of the training data size and intermediate layers of the pretrained speech encoder on speech parsing.

Reinforcement learning with verifiable rewards (RLVR) and on-policy distillation (OPD) have become two main paradigms for post-training reasoning models. RLVR gives each response a single outcome label, leaving the steps inside it without separate credit. OPD provides token-level guidance at student-visited prefixes, but its pointwise signal does not directly reflect the pattern of teacher–student disagreement across the vocabulary. Dense, unbounded log-ratio supervision can amplify the teacher's influence, yet a strong solver is not necessarily a suitable guide when the student's solution paths depart from the teacher's. We propose Residual Advantage (\RA{}), which treats the teacher–student probability residual as a bounded one-step reward, subtracts the corresponding state value under the student policy to form a standard advantage, and centers the result within each response before adding it to the verifier advantage. The guidance term has zero mean within each response, so the verifier advantage remains the response's mean label and the teacher only redistributes credit among the steps within it. \CoRA{} further updates a teacher LoRA with verifier advantages on the same scored student batch and uses the updated teacher in the next iteration's residual, adapting guidance to the student's attempts. With Qwen3-1.7B-Base and Qwen3-4B-Base students and a Qwen3-8B teacher, \RA{} combined with GRPO or REINFORCE++ improves the underlying sequence-advantage algorithm in all 24 comparisons on three mathematical benchmarks, raising macro Avg@8 by 1.7–3.6 points and Pass@8 by 3.9–6.3 points. Both combinations surpass teacher-only OPD, and \CoRA{} adds a further 1.0–1.5 Avg@8 points.

We study the matroid secretary problem with a nonnegative monotone submodular objective. Elements arrive in uniformly random order, and every acceptance decision is immediate and irrevocable. We give an $8.699$-competitive algorithm for arbitrary matroids. The algorithm first rejects a randomly sized initial segment of the arrival sequence as a learning set and runs submodular greedy on all elements in this set. Each later element receives a fixed weight equal to the marginal value it would have when inserted into the stored greedy sequence. On these fixed weights, the algorithm runs a linear-objective matroid secretary subroutine. The algorithm uses $O(nr)$ value queries and $O(n^2)$ independence queries, where $r$ is the matroid rank.

Fault injection is increasingly used to evaluate the reliability of tool-using LLM agents. However, there has been limited study of how fault-adoption results should be interpreted when the agent itself determines which authoritative observations become visible during execution. In this paper, we present a systematic study of this evidential validity problem in agent fault-injection evaluation. We develop a measurement protocol that specifies what observations can refute an injected assertion, determines whether they can become visible before the affected fact is first used, and records whether the evaluated execution actually realizes this condition. We construct SSCBench as an instantiation of the protocol and evaluate four fault operators and five agent configurations over 1,191 faulted executions in two $\tau$-bench environments. Our experiments show that an admitted fault case and agent configuration can realize substantially different evidential conditions across executions, and that aggregate adoption can remain well defined even when the population supporting a timely-counterevidence claim is sparse or absent. For example, among 44 adopted runs in which counterevidence eventually became visible, only 17 received it before first use, while 27 received it afterward. We also find that first-error timing and later stance revision need not coincide, and that automated trajectory analysis can recover adoption without reliably recovering the first faulty-reliance event needed for temporal diagnosis. We argue that the evidential condition realized by an execution and the population supporting a claim-specific interpretation are part of fault-injection evaluation itself and should be reported before adoption is interpreted as failure under pre-use counterevidence.

Large-scale Quantum Key Distribution (QKD) networks rely on trusted repeaters, making the security properties of the selected communication path an essential part of end-to-end assurance. At the same time, network operators may be unwilling to disclose their internal topology. We present a topology-certification mechanism that lets a provider prove in zero knowledge that policy-compliant routes between two endpoints exist, without disclosing the routes in an individual presentation. Our main idea is to certify nodes and edges independently using multi-message signatures, rather than signing the complete graph as one object. For a route chosen in advance, proof size and cryptographic proving and verification work depend on its positions and attributes, independently of the overall network size, whereas route discovery and certification have separate graph-dependent costs. The construction hides the actual route length up to a public bound and sketches extensions to node-disjoint routes and monotonic additions within a graph epoch. We give a formal security model and conditional proofs of unforgeability and graph hiding for a generic construction. For a BBS-based construction, presentation-size estimates remain below $300$\,KiB at $\ell=m=n=50$, and measured generation and verification times remain below $400$\,ms at $\ell=64$ and $m=n=8$.

Federated learning (FL) typically adopts a server-client architecture, where the server aggregates clients' local models (i.e., the input) and returns the aggregated global model (i.e., the output) to clients. An untrusted server may return a tampered global model to compromise the output integrity. Some existing schemes focus on verifying the output integrity. However, these schemes mostly rely on clients pre-negotiating a set of identical auxiliary information among themselves and keeping it confidential from the server. This dependency creates a verification vulnerability: if the auxiliary information is leaked, the verification method may be circumvented. Additionally, in some privacy-sensitive scenarios (e.g., commercial federated learning), the global model may need to be kept confidential from an untrusted server. However, only a few works achieve the output privacy while addressing verification vulnerability, at the cost of prohibitive computation and communication overhead. To address these challenges simultaneously, we propose a novel provably privacy-preserving FL method called EIFL. Specifically, we adopt a two-stage aggregation and combine it with symmetric encryption to protect the output privacy. To address the verification vulnerability, we propose an efficient verification method based on vector inner product for output integrity, and a random vector generation method for clients to agree on auxiliary information. EIFL innovatively binds the auxiliary information to the output integrity, and eliminates the need to keep the auxiliary information confidential from the server. Moreover, EIFL is robust against client dropout during the verification phase through a simple resending operation. Evaluation results validate the advantages of EIFL over state-of-the-art schemes in terms of computation and communication overhead.

Large language models (LLMs) often default to Modern Standard Arabic (MSA) when generating Arabic, even when prompted with dialectal Arabic. A natural explanation is that their internal representations are dominated by MSA. We test this hypothesis by adapting the language-dominance framework of Shani and Basirat (2025) (https://doi.org/10.18653/v1/2025.blackboxnlp-1.7) to 26 Arabic varieties. Across layers and model families, we find no evidence that MSA acts as a dominant internal representation for Arabic dialects. Instead, dialect representations form a dense and highly overlapping space: normalized mutual information drops sharply compared to patterns reported for more distinct languages. Moreover, the strongest separability effects are not confined to intermediate layers, but can shift toward later layers depending on the architecture. These findings challenge a common interpretation of MSA-biased generation: output preference does not necessarily reveal internal representational dominance. Analyses of multilingual and dialectal LLMs should therefore distinguish generation bias from the geometry of internal representations.

Despite recent advances in Vision-Language-Action models (VLAs) for robotic manipulation, their performance remains sensitive to changes in camera configuration. The problem becomes more evident in cross-setup deployment, as reproducing the exact camera pose used for training is nearly impossible. Unlike fixed external views, wrist views are more challenging because the camera moves with the robot, causing even small mounting variations to alter fine-grained geometric cues. To address this, we propose WARP-VLA, a camera-view robust VLA for diverse wrist camera configurations. WARP-VLA adopts a Mixture-of-Experts (MoE) architecture where individual experts learn view-specific feature transformations, and a router combines them based on implicit view information. This allows the policy to be deployed without requiring camera extrinsic parameters as additional input. Through experiments on the LIBERO benchmark, WARP-VLA improves the average success rate of pi-0.5 from 39.2% to 78.3% under wrist-view perturbations. The real-robot experiments further show that the feature-level adaptation learned in simulation successfully transfers to diverse deployment settings. To facilitate reproducibility and future research, we release our wrist viewpoint robustness benchmark and a plug-and-play implementation.

Graph clustering is a fundamental task in data analysis, aiming at grouping nodes with similar characteristics in the graph into clusters. This problem has been widely explored using graph neural networks (GNNs) due to their ability to leverage node attributes and graph topology for effective cluster assignments. However, representations learned through GNNs typically struggle to capture global relationships between nodes via local message-passing mechanisms. Moreover, the redundancy and noise inherently present in graph data may easily result in node representations lacking compactness and robustness. To address these issues, we propose a conjoint framework CoCo, which captures compactness and consistency in the learned node representations for deep graph clustering. Technically, our CoCo leverages graph convolutional filters to learn robust node representations from both local and global views, and then encodes them into low-rank compact embeddings, thus effectively removing the redundancy and noise as well as uncovering the intrinsic underlying structure. To further enrich the node semantics, we develop a consistency learning strategy based on compact embeddings to facilitate knowledge transfer from the two perspectives. Our experimental results indicate that our CoCo outperforms state-of-the-art counterparts on various datasets.

Humanoid robots possess the structural capability to traverse complex terrains. However, achieving stable t raversal without relying on perceived information remains challenging, particularly in complex environments. This paper introduces DAMP, a reinforcement learning framework aimed at achieving robust and naturalistic humanoid locomotion over challenging terrains, with the assumption that no perceived information is available. The framework leverages recurrent neural networks to capture temporal dependencies and implicitly infer privileged and other task-relevant latent information. By aligning the learned representations with the task objective, the method enables robust and goal-consistent policy learning. This end-to-end framework achieves transfer learning from simulation to real-world environments, demonstrating the proposed method's robustness and generalization capabilities. The video of the real-world demonstration can be found at the following link: https://youtu.be/AkI7TZB2DDM.

Modern autonomous driving systems rely on bird's-eye-view (BEV) perception models that fuse camera and LiDAR inputs to detect objects in 3D space. These models are accurate, but they cannot be deployed through standard inference runtimes. The reason is an operator mismatch between dense convolutions (which runtimes handle well), sparse 3D convolutions (which runtimes cannot represent), and geometric scatter operations (which runtimes have no vocabulary for). Today, every sparse convolution library is CUDA-only and PyTorch-coupled, locking BEV deployment to a single vendor's hardware and a single execution framework. We present BEVPIPE, a framework for deploying multimodal BEV perception pipelines using portable GPU compute APIs and integrating them with production inference runtimes. BEVPIPE partitions the model into runtime-managed dense subgraphs and three external operator extensions (voxelizer, sparse encoder, BEV projector), connected through a shared GPU memory space. BEVPIPE achieves a 19.5x end-to-end speedup over conventional deployments while retaining 98.5% of reference mAP. We also showcase that BEVPIPE is portable across different GPU backends.

Large Language Models (LLMs) have shown strong reasoning capabilities when fine-tuned with reinforcement learning (RL), particularly through Group Relative Policy Optimization (GRPO). However, existing GRPO methods assume centralized access to training data, which may not hold in practice due to privacy or regulatory constraints. To this end, we propose Fed-GRPO, a federated GRPO training framework that addresses these privacy constraints by enabling collaborative reasoning training without sharing raw data, which leverages the reward statistics naturally produced during GRPO training as zero-cost signals to guide aggregation, local training, and communication. Fed-GRPO contains three reward-signal-driven mechanisms: (i) signal-weighted aggregation that weights clients by their reward standard deviation, prioritizing clients with stronger learning signals; (ii) global reward calibration that re-weights per-prompt objectives based on the local-global reward gap, steering each client toward its relative weaknesses; and (iii) adaptive sparse communication that allocates bandwidth based on the informativeness of each client's update. Extensive experiments on mathematical reasoning tasks demonstrate that Fed-GRPO achieves the best performance among all federated methods, clearly outperforms FedAvg and approaches centralized training performance, while losslessly reducing communication by $32\times$ and supporting up to $621\times$ compression under tight bandwidth budgets with only graceful accuracy degradation. Our code is available at https://github.com/HKU-HealthAI/Fed-GRPO.

Despite the adoption of large language models (LLMs) in recommendation systems, prevailing approaches mostly model single-type behaviors (e.g., views or purchases). Even when incorporating multiple behaviors, existing methods flatten heterogeneous actions into homogeneous token sequences, ignoring their distinct decision-making roles. This flattening fails to capture semantic hierarchies and contextual nuances in complex decision-making, such as trade-offs between price and quality. Consequently, performance degrades in critical “difficult-choice” scenarios involving highly similar items. To bridge this gap, we propose MARI (Memory-Augmented Recommendation with Interpretability), which grounds predictions in explicit, structured decision evidence. MARI maintains a Decision Memory Bank (DMB) that archives users' past rationales as Structured Decision Memories (SDMs): concise records of goals, constraints, and trade-offs. These SDMs are generated offline via Post-Hoc Decision Distillation from heterogeneous behaviors and user-generated content. By retrieving relevant SDMs to augment LLM reasoning, MARI achieves interpretability and scalability without the prohibitive cost of processing long raw sequences. Extensive experiments show MARI significantly outperforms state-of-the-art baselines on standard next-item prediction and a newly introduced Difficult Choice Prediction task, incurring low latency overhead by decoupling memory construction from online inference. Qualitative analyses reveal actionable, human-readable insights into user decision-making, marking a concrete step toward reasoning-aware recommendation systems.

Bound states in the continuum (BIC) metasurfaces are typically described by geometry-specific parameters, hindering cross-geometry comparison, while ultranarrow qBIC features are easily diluted in full-spectrum learning. Here, 2015 samples from seven dielectric metasurface families are mapped to a shared reciprocal-lattice grid, where two frozen low-order Fourier channels capture resonance shifts with mean within-branch $R^2$ values of 0.871-0.999. Field-level analysis of two representative branches further confirms that these shifts are consistent with the Maxwell-Fourier perturbation picture. A five-channel K-space backbone models the broadband spectrum, while a local complex K-space expert parameterizes the qBIC resonance through a differentiable Fano layer. The expert reduces resonance-position mean absolute error (MAE) from 3.2 to 0.95 nm and the resonance-depth error by 14-fold on a geometry-blocked test set. The same coordinate supports spectrum-to-structure reconstruction.

Video diffusion and flow models require many sequential evaluations, making generation computationally expensive. Few-step distillation reduces this cost but poses a capacity allocation problem: a student must match the teacher's iterative generation with far less sequential computation. Existing trajectory methods ask the student to reproduce teacher transitions that are highly curved at high noise, which can exceed its capacity and degrade fine detail. We introduce Parametric Trajectory Distillation (PTD), which lets the student parameterize teacher trajectory segments as polynomials and learn from teacher guidance along its own predicted path. PTD is designed to let the learned curvature adapt to the backbone's predictive capacity, preserving motion and diversity. The curvature head is used only in training; inference keeps the original backbone architecture. On Wan2.1-14B, four-step PTD sets a new state of the art for trajectory distillation, significantly improving dynamic quality and naturalness over PDD, the best-performing trajectory-only method on this model, under the same training setting. On the 33B audio-video MiniMax-H3, LoRA-trained PTD significantly improves diversity and naturalness over the state-of-the-art LightX2V Turbo. Blinded human votes give PTD 55.1% and 63.4% preference shares against PDD and LightX2V Turbo. Project page: https://alan-lanfeng.github.io/PTD/.