Software Engineering
Assessing the Quality of Concrete Scenarios Generated from Scenario Classes for Automated Driving
Lina Putze, Thies de Graaff, Jan Steffen Becker et al. (+1 more)
A scenario-based development approach has become state of the art for automated driving systems. Once a set of scenario classes covering the operational design domain is available, concrete scenarios are generated from these scenario classes, e.g., to be used for hazard analysis and risk assessment or for the system's verification and validation. For such applications, a high diversity among the generated concrete scenarios is paramount. The question arises how the quality of collections of concrete scenarios can be assessed in such a setting. In this work, we aggregate scenario distance measures into so-called quality functions that quantify the diversity of concrete scenario collections in the context of automated driving. In particular, we consider the case where scenario classes are formalized as constraint systems and the associated concrete scenario space is not explicitly known. Based on a set of desired properties we construct candidates for such quality functions and give proofs for their property fulfillment. To show the practical relevance of quality functions, we perform a variety of experiments in the context of an abstract scenario-based virtual verification toolchain for automated driving systems. More specifically, we use quality functions to compare the effectiveness of variation methods for sampling, examine the choice of different scenario distance measures, and investigate the effects of temporal aspects on concrete scenario collection quality.
Read the full abstract ▾
All authors Lina Putze, Thies de Graaff, Jan Steffen Becker, Christian Neurohr
Artificial Intelligence
Back to the Future: Reverse-Consequence Modeling for Long-Horizon Particle Dynamics
Leyan Zhang, Nantian He, Jiajun Ma et al. (+2 more)
Accurate prediction of object trajectories is fundamental to mechanical understanding and robotic control. However, autoregressive models inherently suffer from cumulative errors, leading to severe drift in long-horizon rollouts. Observing that in many physical systems with dissipative or convergence-to-rest behavior, the final states are often more predictable than transient dynamics (e.g., a sliding object resting under friction or a pendulum stabilizing at
equilibrium), we propose ReStep, a cascaded approach that reverses the conventional forward-prediction paradigm. ReStep first employs a
stochastic model to directly predict the final state, then leverages Anchored
Diffusion to generate goal-conditioned intermediate trajectories. By conditioning on both ”boundary states”, it effectively constrains dynamics and reduces uncertainty, producing physically plausible and temporally coherent transitions. We further integrate ReStep dynamics as the physical engine within a 3D Gaussian Splatting pipeline, enabling a unified physical-visual framework for intuitive inference from images. Extensive experiments on diverse rigid and non-rigid collision
datasets show that ReStep achieves superior long-term accuracy, substantially mitigating error accumulation via reverse-consequence modeling.
Read the full abstract ▾
All authors Leyan Zhang, Nantian He, Jiajun Ma, Yanpeng Zhao, Yunbo Wang
Robotics
Beyond Planning: Differentiable Control-in-the-loop End-to-End Autonomous Driving
Zhidong Wang, Zirui Li, Zitong Shan et al. (+6 more)
End-to-end autonomous driving has achieved remarkable progress by planning trajectories directly from raw sensor data. Yet the step from planning to control still relies on hand-tuned tracking controllers that block the gradient path from vehicle motion back to the network. This gap stems from non-differentiable vehicle dynamics, aliased supervision from coarsely sampled expert labels, rate mismatch between high-rate control and low-rate re-planning, and vehicle model discrepancies between training and deployment. This paper proposes DIVA (Differentiable Identified Vehicle dynamics for end-to-end Autonomous driving) to close this gap. An identified actuator-level vehicle model is embedded in the training loop of the end-to-end driving network, forming a differentiable chain from perception to vehicle states. A learning-based control decoder replaces the external controller and is supervised on the rolled-out vehicle states, bypassing the aliased control-input labels. A rate-consistent control-input parameterization reconciles the heterogeneous sampling rates and enables asynchronous execution with a 3.8 ms control-only cycle. On-policy re-identification and an online disturbance observer address training-time distribution shift and deployment-time model mismatch respectively. On the closed-loop Bench2Drive
benchmark, DIVA achieves a driving score of 86.7, surpassing the widely adopted state-of-the-art approach by 3.3 points while nearly halving mean jerk and reducing brake switching by 64%, and maintains superior path tracking accuracy under significant vehicle model mismatch at deployment.
Read the full abstract ▾
All authors Zhidong Wang, Zirui Li, Zitong Shan, Zhiqi Mao, Qihang Huang, Zhan Chen, Jingsong Liang, Han Yu, Chen Lv
Artificial Intelligence · Cross-listed from stat.ML
Lifting Finite-Sample Gaussian Approximation of Wide Neural Networks to Functional Convergence
Luis Wiederhold, Nil Ayday and Debarghya Ghoshdastidar
Randomly initialised neural networks of width $m$ converge to Gaussian processes as $m \to \infty$, with two types of quantitative bounds characterising this convergence. Finite-sample bounds compare the network with its Gaussian limit at a fixed finite sample set, on an arbitrary domain, in the quadratic Wasserstein distance $\mathcal W_2$, achieving a convergence rate of $m^{-1/2}$ with explicit constants. However, these constants grow with the number of samples, so the bound becomes uninformative as the sample set is refined. Functional bounds remove that dependence by comparing the two as random functions, at the price of spherical input, metrics weaker than $\mathcal W_2$, or slower rates. In this paper, we therefore lift a finite-sample bound to a functional one. The lifting refines the sample set until it is dense in the domain, so the bound must first be made independent of the number of samples. We achieve this by measuring the error per sample, which makes the constant uniform. For a
polynomial activation and any compact input set, this gives a rate of $m^{-1/2}$ in the supremum norm, with an explicit constant.
Read the full abstract ▾
Software Engineering
Soliduzz: Finding Miscompilations in Smart-Contract Compilers with Assertion-Driven Equivalence Oracles
Alexander Stromberger, Christoph Hochrainer, Valentin Wüstholz et al. (+1 more)
Smart contracts compiled to the Ethereum Virtual Machine (EVM) can manage high-value assets and are difficult to patch after deployment. Compiler miscompilations are therefore especially dangerous: compilation succeeds, but the generated bytecode does not preserve the semantics of the source program. Existing smart-contract compiler fuzzers either focus on robustness failures or rely on differential oracles that compare compiler configurations. We present Soliduzz, an approach for finding miscompilations in EVM smart-contract compilers using assertion-driven equivalence oracles. Soliduzz uses product programs to encode semantic equivalence as executable assertions. The key shift is to compile these self-checking programs and let off-the-shelf fuzzers and symbolic execution engines explore the resulting bytecode for assertion violations, without requiring compiler-specific support from the analyzer. Soliduzz instantiates this idea with two complementary families of metamorphic transformations. The first starts from equivalent expressions and lifts them into smart-contract code that deliberately exercises EVM-specific features such as storage, data locations, control flow, contract interactions, and inline assembly. The second generalizes Equivalence Modulo Inputs (EMI) from fixed concrete inputs to symbolic input spaces. Our symbolic EMI records coverage in the original computation and uses it to guard mutations in the transformed computation. Together, these transformations produce product programs that assert that paired computations yield the same outputs. We evaluate Soliduzz on three EVM smart-contract compilers (solc, solx, and vyper) using several analysis tools. Overall, Soliduzz exposes 29 bugs. Of these, 24 are compiler bugs: 6 miscompilations, 17 internal compiler errors, and 1 hang. The remaining 5 are bugs in the analysis tools.
Read the full abstract ▾
All authors Alexander Stromberger, Christoph Hochrainer, Valentin Wüstholz, Maria Christakis
Machine Learning
Random Probing for Structural Self-Interactions in Graph Neural Networks
Saku Peltonen, H. Çağr{\i} Bilgi and Kubilay Atasu
Graph Neural Networks (GNNs) fundamentally lack the ability to capture structural properties defined by node self-interactions, such as cycle counts, subgraph centrality, and heat kernel signatures. Existing solutions often rely on computationally expensive higher-order architectures or hand-crafted features. In this work, we introduce DIAL, a message-passing layer that gives nodes access to graph structure through the diagonal of graph-derived operators. Our approach uses randomized probing to provide nodes with learnable, permutation equivariant access to diagonal entries. We prove that our approach can express important structural features, including cycle counts, subgraph centrality, and heat kernel signatures. The proposed method can be integrated directly into message passing layers. Experiments on molecular property prediction, cycle counting, and pattern detection demonstrate the effectiveness of DIAL across a range of graph learning tasks.
Read the full abstract ▾
Robotics
RESTA: Time-Scaling Augmentation for Data-Efficient Robot Imitation Learning
Yunpeng Jiang, Jianshu Hu, Han Fang et al. (+3 more)
Robot manipulation policies trained through imitation learning commonly rely on costly demonstrations recorded in one task direction and at one temporal resolution. We propose Reversal and Stride-based Trajectory Augmentation (RESTA), which exploits time-scaling data augmentation to improve data efficiency through trajectory reversal and temporal resampling. Trajectory-reversal augmentation reverses and relabels demonstrations between paired tasks with opposite objectives, reusing recorded rollouts without additional environment interaction. Because contact-rich manipulation is not exactly reversible, we treat reversal as an approximate paired-task prior rather than a physical
symmetry. Temporal-stride augmentation varies observation and action sampling strides and conditions the policy on the resulting offsets, enabling fixed or randomized temporal schedules. Experiments across 16 paired bimanual manipulation tasks show that trajectory reversal is particularly effective in the low-data regime and that temporal-stride augmentation improves success under randomized temporal schedules. With 25 and 50 independently collected demonstrations per task, RESTA achieves 78.4% and 85.2% average success, respectively, matching or exceeding baselines trained with twice as many demonstrations (77.3% and 85.4%). These results show that RESTA can improve data efficiency in
robot imitation learning.
Read the full abstract ▾
All authors Yunpeng Jiang, Jianshu Hu, Han Fang, Zhiyuan Guan, Paul Weng, Yutong Ban
Machine Learning
LoOPSD: On-Policy Self-Distillation with Looped Teachers
Wei Fan, Qingyue Yang, Chao-Hong Tan et al. (+6 more)
On-policy self-
distillation improves a model using guidance from the model itself, but it still requires a teacher that provides more useful predictions than the student. Existing approaches obtain such guidance by giving the teacher additional context or instructions. We ask whether it can instead come from changing how the same model uses its own layers. To address this, we introduce LoOPSD, which searches over a model's layer execution order to construct a stronger teacher and distills its advantage into the student. Specifically, we keep the model weights unchanged while searching for layers to repeat, and use the resulting teacher's predictions on student-generated prefixes to guide learning. We then alternate search and
distillation, rebuilding the teacher from the updated student in each round so that the student progressively absorbs the gains from alternative layer execution. Experiments on mathematical and logical
reasoning tasks show improvements across models. On GSM8K with a 400-
token thinking budget, LoOPSD raises Qwen3-1.7B accuracy from 42.30% to 79.61%; subsequent
distillation rounds yield further gains beyond the first round. Together, these findings show that by searching layer execution to construct teachers, we can turn a model's additional computation into improved student capabilities, opening a new avenue for self-
distillation to derive supervision from the model's own computation.
Read the full abstract ▾
All authors Wei Fan, Qingyue Yang, Chao-Hong Tan, Qian Chen, Wen Wang, Xiangang Li, Kejiang Chen, Weiming Zhang, Nenghai Yu
Computer Vision
Benchmarking Multimodal Large Language Models for Zero-Shot Human-Object Interaction
Oriol Rabasseda, Zenjie Li, Kamal Nasrollahi et al. (+1 more)
Multimodal Large
Language Models (MLLMs) have demonstrated strong zero-shot
reasoning across diverse domains, but their potential for Human-Object Interaction (HOI) remains unexplored. Unlike traditional HOI approaches that rely on extensive labeled
datasets and fixed interaction taxonomies, MLLMs offer flexible
reasoning, making them well-suited for dynamic, real-world scenarios. In this work, we present the first systematic study of MLLMs for HOI recognition and detection within a pipeline that combines an external object detector for entity localization with zero-shot
multimodal reasoning for interaction understanding. We introduce a
benchmark framework for principled zero-shot evaluation, consisting of (i) a unified taxonomy of zero-shot prompting strategies, and (ii) a methodology that transforms
token-level logits into interaction scores, enabling fair comparison with specialized HOI models under standard metrics. To support
prompt-based evaluation, we additionally construct new question-answer annotations for standard HOI
datasets, transforming them into a format suitable for MLLMs. Building on this, we propose two visual prompting strategies, namely Objects as Context and Black Other Objects, that guide MLLMs to target objects while preserving scene context. Experiments on HICO, HICO-DET, and V-COCO demonstrate state-of-the-art zero-shot performance, achieving 53.50 mAP on HICO and 23.69 mAP on HICO-DET. Overall, our work provides a principled benchmarking framework for MLLMs in HOI, opening the door to new models and
datasets and enabling flexible open-world interaction understanding.
Read the full abstract ▾
All authors Oriol Rabasseda, Zenjie Li, Kamal Nasrollahi, Sergio Escalera
Robotics
Force Sensor-Based Picking Pattern Analysis and Two-Jaw Gripper Design for Cherry Tomato Robotic Harvesting in Greenhouse
Rajmeet Singh, Manveen Kaur, Yusuf Adamu et al. (+1 more)
Robotic harvesting of cherry tomatoes in greenhouses remains challenging due to dense fruit clusters, delicate skin, and the need for gentle yet effective detachment. This study presents a force-data-driven framework for autonomous cherry tomato harvesting that integrates six-axis force/torque sensing, a bio-inspired compliant gripper, and deployment on a mobile robotic platform. A wearable ATI six-axis force/torque sensor mounted on the index finger captured kinetic signatures of three elemental picking patterns (gripping, twisting, and lateral pulling) across 150 manual trials (50 per pattern) in laboratory and greenhouse settings. Gripping required a peak normal force of 3.92 and 0.45 N, twisting produced a dominant torque of 19.20 2.10 Nmm about the pedicel axis with minimal truss disturbance, and pure lateral pulling alone was insufficient for reliable detachment. These measurements informed a novel fingertip-localised Finray gripper, with compliant TPU 95A Finray geometry restricted to the fruit-contact region and a rigid PLA body, emulating the dual-stiffness human fingertip. A Dynamixel XM430-W350 servomotor under current-based torque control applies the measured gripping force and a 45{\deg} wrist twist. Finite element analysis showed a maximum von Mises stress of 0.24 MPa under the 3.92 N load, below the bruising threshold of cherry tomato skin. Integrated on a UR5 arm and Clearpath Husky A200 base, the system was validated in NVIDIA Isaac Sim and in 50 autonomous trials at the SILAL commercial greenhouse, Abu Dhabi, UAE. The hybrid grip-twist-pull strategy achieved an 86.0% picking success rate, 2.3% fruit damage, and an 8.4 s cycle time per fruit, outperforming
vacuum (66.3%), rotating (70.1%), clamping (76.0%), and suction-cutting (52.4%) end-effectors. These results demonstrate preliminary feasibility under the tested facility and cultivar conditions
Read the full abstract ▾
All authors Rajmeet Singh, Manveen Kaur, Yusuf Adamu, Irfan Hussain
Computation and Language
Claim Framing as a Driver of LLM Sycophancy
Jean Seo, Pepa Atanasova and Isabelle Augenstein
Sycophancy in Large
Language Models, the tendency to align responses with user preferences over ground truth, has largely been studied using questions without a clear ground truth and with a single, plainly stated user opinion. We present a systematic study of which framings of a user's claim trigger sycophancy in factual question answering, where every answer is objectively verifiable without extra context. Evaluating 14 models on 42K questions, we express correct and incorrect user opinions through various claim framings over 8 dimensions and 21 types, and measure a broad set of opinion-driven behaviors. We find that all models are susceptible to user opinions regardless of size. Models are most influenced by how confidently an opinion is stated, whom it is attributed to, and whether it is presented as shared between the user and the model. In contrast to prior work, framings based on the user's expertise, authority over the model, or social pressure matter little in our factual setting. Larger models resist incorrect opinions better but are no better at recovering from their own errors. Our results call for sycophancy evaluations that probe how an opinion is framed, not merely whether one is present, and that measure multiple behaviors beyond a single flip rate.
Read the full abstract ▾
Artificial Intelligence
When and How Order-Flow Interventions Work After Dealers Adapt
Guanghao Wu, Zhuo Cai, Zhidong Li et al. (+1 more)
Learning dealer
agents can settle on persistently wide quotes, raising what customers pay to trade. A venue can route more orders to tighter quotes, but a wide quote may reflect real risk, and once dealers adapt, an end-to-end comparison cannot show why a rule works: because it intervenes more, picks better moments, is announced in advance, or rewards the right quote. We propose SOFA (State-aligned Order-Flow Allocation), a public order-flow contract with three parts: a State Scorer ranks
market states from lagged public data, a Commitment Scheduler announces a budgeted intervention strength before dealers quote, and a Quote-to-Flow Allocator shifts a bounded share of fixed customer demand toward tighter quotes. Our paired re-adaptation protocol gives competing rules the same intervention strengths, only re-timed, lets dealers re-adapt to each rule from a shared checkpoint, and breaks one link at a time. Across 20 independently trained dealer populations in an ABIDES
market, concentrating the intervention in states that the score ranks as low-risk lowers low-risk execution cost relative to random timing in every population, by 0.040 tick on average, while a stale announcement or a broken quote-to-flow link raises this cost. In the primary setting the gain comes at the expense of high-risk trades, whose costs rise, so state
alignment pays off only for a venue that weighs low-risk costs more than 1.38 times as heavily as high-risk costs. Our code is available at https://github.com/moderncavemann/state-aligned-order-flow-contracts.
Read the full abstract ▾
All authors Guanghao Wu, Zhuo Cai, Zhidong Li, Shoujin Wang
Computation and Language
Fan Out or Relay? Ownership-Aware Routing for LLM Agents Querying Siloed Tables
Zhexi Feng
Inside organizations, a question's data often sit with several owners, who query their own tables but never hand them over. An LLM coordinator can fan out, sending the question to every owner at once, or relay, querying owners in turn and reusing earlier rows. We show that neither is safe in advance, because data ownership decides which one is right. Across thirteen settings on PeopleJoin and BIRD with three model families, fan-out is more accurate in twelve on questions one owner can answer, yet answers at most 4.3% of those joining two owners' tables. The flip follows ownership, not query shape: on BIRD two-table joins, fan-out answers 32-44% when one owner holds both and 1-4% otherwise. Relay wins by conditioning on returned rows, and hiding them from an otherwise identical planner cuts two-owner accuracy from 23.2% to 5.1%. Since ownership is public, we propose Ownership-Aware Routing (OAR), which drafts a schema-only SQL query, counts the owners whose tables it joins, and routes the question before any private row is read. In ten settings OAR is never significantly less accurate than the best fixed protocol, and beats every fixed protocol on PeopleJoin, by up to 9.3 points on held-out organizations.
Read the full abstract ▾
Machine Learning
When Do Learned Blocks Preserve Causal Meaning? Block Identifiability and Constructive Causal Abstractions
Karim Zaghw, Wieland Brendel and Patrik Reizinger
Standard identifiability theory assumes one-dimensional latent factors, but learned representations can encode latent information across multidimensional blocks whose internal coordinates are not uniquely determined. When such blocks are used to represent and intervene on higher-level concepts, it is unclear whether their causal meaning survives the transformations left unresolved by block-identifiability. We connect block-identifiability to constructive causal abstraction. We show that block-diagonal invertible transformations followed by compatible coordinate permutations form the maximal class that preserves every constructive causal abstraction built on those blocks. This guarantee assumes that the correspondence between blocks and concepts is updated with the representation. A small synthetic experiment with additive autoencoders shows that high-level interventions can be transferred between independently learned representations while approximately preserving their intended effects.
Read the full abstract ▾
Machine Learning
Learning a Group Element from Its Action: Galois Correspondences, Optimal ERM, and an Abel-Ruffini Threshold for the Natarajan Dimension
Xiaoyu Li
We study learning an unknown element $g$ of a group acting with finite orbits on a countable set, from examples $(x,gx)$. After a sample on $A$, the consistent hypotheses are the restrictions of a coset of the pointwise stabilizer of $A$, and the sample forces exactly the labels of points this stabilizer fixes; for a Galois group, they form the field generated by the data. Through this correspondence, multiclass dimensions become group invariants: the DS and graph dimensions equal the height, the Littlestone dimension the longest irredundant sequence length, and the teaching dimension the minimal base size. Among deterministic learners,
symmetry makes empirical risk minimization optimal: on nondegenerate classes every consistent learner, whatever its tie-breaking, attains the optimal realizable sample complexity up to constant factors, and every consistent online learner the optimal mistake bound. Agnostically, every ERM's sample complexity exceeds the known lower bound by at most the DS-to-Natarajan ratio, up to constants. For finite Galois extensions of countable fields, the Natarajan dimension equals a group invariant, the coset-cube number, which is at least the number of complemented factors of any chief series of any subgroup. Hence the dimensions coincide for every solvable Galois group, while for the groups $A_5$ and $S_5$ of the quintic they are $2<3$ and $3<4$, and $A_5$ is the smallest group with a gap. For the Galois group $S_n$ of the general equation of degree $n$, they coincide exactly when $n\le 4$, i.e., in characteristic zero, exactly when the equation is solvable by radicals; the implication is one-way, as the simple group $\mathrm{PSL}(2,8)$ has no gap. For linear, orthogonal and
unitary maps of $\mathbb R^d$ or $\mathbb C^d$ both dimensions equal $d$, and a two-fibered pseudo-cube is a disjoint union of Natarajan cubes exactly when its monodromy group is abelian.
Read the full abstract ▾
Computation and Language
Attacking Only Half the Data: Two-Sided Evaluation of AI-Text Detector Robustness
Apu Chandraw Shill and Rabindra Nath Nandi
We investigate the robustness of widely used AI text detectors under attacks applied to both human and machine text. Prior evaluations typically attack only machine text, overlooking that human text is routinely modified through grammar correction, formatting, or invisible character insertion. We introduce a two-sided protocol that applies attacks to both classes. Across two corpora and nine detectors, reported AUROC gains vanish under two-sided evaluation, with attacked human text misclassified at rates up to 100%. While simple preprocessing reverses character-level attacks, word-level synonym substitution remains effective meaning-preserving, and irreversible.
Read the full abstract ▾
Robotics
Walk Like This: Stylized Locomotion for Physics-Based Characters through Example-Driven Diffusion
Joshua Aurand, Agon Serifi, Sammy Christen et al. (+4 more)
We present a real-time framework for expressive legged locomotion that enables independent control over velocity and style. Unlike existing data-driven controllers, which typically require explicit style labels for training, our method efficiently leverages large-scale unstructured motion data and generalizes to unseen styles. We disentangle style from functional locomotion through a tailored sampling strategy used to train a style encoder whose output conditions an autoregressive
diffusion model. During inference, the user provides a style reference clip and a joystick velocity command. The
diffusion model then generates a kinematic reference tracked by a reinforcement learning controller. To improve user controllability and stylization strength, we propose an
attention biasing mechanism and a velocity command correction strategy. We demonstrate our method on a physics-based humanoid character in simulation and on a physical
robot, showcasing robust real-time control and continuous style modulation. The results can be best seen in our supplemental
video (https://www.youtube.com/watch?v=roFnS6N7nJ4).
Read the full abstract ▾
All authors Joshua Aurand, Agon Serifi, Sammy Christen, Ruben Grandia, Espen Knoop, Marco Hutter, Moritz Bächer
Data Structures and Algorithms
A Better-than-2 Approximation for k-Median Part I: The Power of General Bids
Rares-Darius Buhai, Vincent Cohen-Addad, Henry Fleischmann et al. (+6 more)
The Greedy
algorithm of Jain, Mahdian, Markakis, Saberi, and Vazirani was the best Lagrangian-multiplier-preserving (LMP) approximation
algorithm for facility location, and thus the engine behind the best $k$-median guarantees, for over two decades. Its $2$-LMP guarantee was only recently broken by a small amount ($(2-\eta)$-LMP where $\eta \approx 2.25 \cdot 10^{-7}$) through local-search post-processing. That analysis relies on the structure of worst-case instances, and it appears difficult to obtain substantial further improvements along those lines. We take a different approach and change the
algorithm itself: we introduce general bidding functions, replacing the classical linear bid with a richer family of rules and prove a
polynomial-time $1.95$-LMP approximation, a large improvement over the long-standing factor of~$2$. We further show that these generalized bidding rules can be made smooth (log-adaptive), building on the recent framework developed for the classical Greedy
algorithm by Cohen-Addad, Grandoni, Lee, Schwiegelshohn, and Svensson. This yields a
polynomial-time $1.95$-approximation that opens $k+O_\varepsilon(\log n)$ centers. Combined with a $(1.89+\varepsilon)$-approximation
algorithm for stable instances in our companion work \cite{stableinstances}, we achieve a better-than-2 approximation for the $k$-median problem in
polynomial-time.
Read the full abstract ▾
All authors Rares-Darius Buhai, Vincent Cohen-Addad, Henry Fleischmann, Anupam Gupta, Antoni Jubés Monforte, Madhusudhan Reddy Pittu, Miltiadis Stouras, Ola Svensson, Lukas Vogl
Artificial Intelligence
RadixFlow: Filling Execution Bubbles in Prefix-Cached Hybrid LLM Serving
Huihao Jing, Wenbin Hu, Shaojin Chen et al. (+6 more)
Prefix reuse improves LLM inference by sharing the KV cache of common
prompt prefixes. This helps only if the scheduler can execute a matched branch in the current worker round. Schedulers assume that a matched prefix is ready, which holds for full
attention but fails for hybrid LLMs. There, a branch (the consumer) may need recurrent state that another branch (the producer) is still computing in the same round. The native scheduler stops its waiting-queue scan at this blocked consumer, and the resulting head-of-line blocking creates execution bubbles: idle batch slots that independent work could fill. We introduce RadixFlow, which decouples prefix matching from state readiness. Its state-ready traversal turns a global scan barrier into a local wait: only the consumer is blocked until the next round, while feasible independent branches keep filling the batch under the native state check. On five hybrid models, RadixFlow lowers mean latency by 3% to 24% and worker rounds by 11% to 44% at almost unchanged reuse. On the larger Qwen3.5 models, ablations show that most of the gain comes from traversal, which adds little on a full-
attention control. Heavier rounds limit how much of this saving reaches latency.
Read the full abstract ▾
All authors Huihao Jing, Wenbin Hu, Shaojin Chen, Haochen Shi, Haoyu Huang, Sirui Zhang, Yuxuan Liu, Zhongwei Xie, Haoran Li
Computer Vision
FlowCred: Hierarchical Credit Assignment and Functional Coordination for Joint Text-to-Video Post-Training
Mingxuan Zhu, Yanwei Huang, Shiyuan Liu et al. (+4 more)
Prompt optimization and generator post-training are two complementary ways to improve text-to-
video (T2V) generation, and recent systems increasingly train both modules jointly. However, optimizing a
prompt policy and a
video generator from the same terminal reward creates two ambiguities:
prompt quality is confounded with
stochastic generation, and the two modules may realize overlapping local improvements. We introduce FlowCred, a two-level framework for joint T2V post-training. First, hierarchical credit assignment uses matched
prompt-seed rollout grids to separate
prompt-level credit from
prompt-conditioned generator residuals under shared randomness. Second, functional coordination probes alternative current-policy
prompts at states from actual generation trajectories, providing generator-aware evidence for
prompt optimization while selectively attenuating generator updates already supported by conditioning changes. Across multiple T2V
benchmarks, FlowCred consistently improves matched joint-training baselines. On LTX-2.5, FlowCred improves a strictly matched joint-training control from 61.47 to 64.46 on T2V-CompBench and from 55.58 to 58.08 on StoryEval, without increasing decoded-
video or reward-model-call budgets. Factorial and realized-update analyses show complementary contributions from
prompt shaping and generator routing, while human evaluation and transfer across backbones and updater families provide further supporting evidence.
Read the full abstract ▾
All authors Mingxuan Zhu, Yanwei Huang, Shiyuan Liu, Xinzhi Zhang, Shujie Li, Yuanxing Zhang, Lu Chen
Artificial Intelligence
Mitigating Cross-Modal Hallucinations by Correcting Modality Misrouting in Omnimodal LLMs
Yu Mi, Licheng Pan, Jiayi Zhu et al. (+7 more)
Omnimodal large
language models (OLLMs) are susceptible to cross-modal
hallucinations under audio-visual interference. We find that, for a substantial fraction of incorrect joint predictions, the question-relevant unimodal branch can still recover the correct answer, suggesting that task-relevant evidence can remain recoverable from the relevant modality even when it is not reflected in the joint prediction. We refer to this recoverable failure pattern as modality misrouting. We propose CaRe (Conflict-aware Residual Intervention), a training-free inference-time method for identifying and correcting modality misrouting. CaRe uses a text-only modality probe to identify a candidate modality-specific counterfactual branch, and flags potential misrouting when its prediction conflicts with the original joint prediction. For such cases, CaRe computes a residual direction from the discrepancy between the counterfactual and fused representations and uses it to steer the original
multimodal state toward the relevant modality evidence, while retaining both modalities in the original audio-visual inference path. The intervention strength is adapted based on the confidence of the counterfactual branch. Experiments on CMM and AVHBench across two OLLMs show that CaRe consistently outperforms the evaluated inference-time
hallucination-mitigation baselines.
Read the full abstract ▾
All authors Yu Mi, Licheng Pan, Jiayi Zhu, Qirui Liu, Yitong Qiao, Changtong Xu, Tao Wei, Fei Sun, Zhixuan Chu, Kui Ren
Computation and Language
SparPR: Runtime KV-Cache Recovery for Resource-Constrained LLM Inference on Consumer Hardware
Jieyun Deng, Keming Hu, Yanghe Liu et al. (+1 more)
Running large
language models (LLMs) on consumer hardware is bottlenecked by the key-value (KV) cache, which grows with every decoded
token and quickly exhausts the
memory and bandwidth of a single consumer GPU. We present SparPR (Sparse Partial Recomputation), a runtime KV-cache recovery mechanism that does not modify model weights and requires no offline training. SparPR couples three online decisions: Query-Aware key-block selection scores historical blocks against the current query; dynamic budget control estimates per-block copy and recompute costs to decide how many blocks to retain; and on-demand local recomputation with cost-aware fallback merges retained blocks into continuous spans and switches between sparse recomputation and full recovery according to a runtime cost comparison. Implemented on FlexGen/FlexLLMGen, SparPR is evaluated on a single consumer laptop (RTX 3070 Ti 8GB, Ryzen 9 6900HX, 16GB RAM). It cuts end-to-end latency by 31%, 52%, and 63% at 1k, 2k, and 4k
prompt tokens respectively, with 1.4x-2.7x higher throughput, and completes an 8k-
token workload in 44 minutes where the FlexGen baseline times out. Under a fixed cache budget, SparPR retains a Needle Accuracy that stays well above StreamingLLM and H2O at every keep ratio, and ablation studies confirm that the three mechanisms act jointly.
Read the full abstract ▾
All authors Jieyun Deng, Keming Hu, Yanghe Liu, Li Feng
Machine Learning
Revisiting Error Feedback for Model Quantization
Wai-Yiu Keung, Chi Cui, Wing-Kin Ma et al. (+2 more)
We study causal error-feedback coefficients as design variables for model quantization. Modeling instantaneous quantization errors as independent, zero-mean variables with a common fixed variance, we formulated a metric-dependent trace problem for designing feedback coefficients. The formulation resembles fixed-order GPTQ/LDLQ feedback and supports Jacobian-aware designs that account for downstream error propagation. For a positive-definite metric $\boldsymbol{H}$, we prove that the joint rotation-feedback trace minimum is $n\det(\boldsymbol{H})^{1/n}$ and present an attaining pair using the
geometric mean decomposition. Experiments on synthetic networks, trained MLPs, and
language models show benefits from downstream-sensitive feedback, but also reveal that trace optimality need not improve finite-level performance. Small QAT experiments further illustrate direct loss-driven feedback tuning with fixed or trainable network weights.
Read the full abstract ▾
All authors Wai-Yiu Keung, Chi Cui, Wing-Kin Ma, Jesús Cid-Sueiro, Hei Victor Cheng
Systems and Control
Preliminary Exploration of Large Models Based on Self-Supervised Learning for Health Monitoring of Power Electronic Converters
Xinyuan Liao, Xiangtian Deng, Yang Liu et al. (+3 more)
Current artificial intelligence (AI) methods for power electronics converters make limited use of the abundant unlabeled data generated during real-world operation. This underexplored resource holds significant potential for advancing model performance. To address this gap, this paper presents PE-TSFM, a time-series foundation model tailored for power module health monitoring. The model adopts the PatchTST architecture combined with self-supervised learning and is trained on more than 143 million timestamps, where over 98% of this training data is unlabeled and collected directly from an operating traction converter. In degradation classification tasks, PE-TSFM achieves an accuracy of 87%, considerably surpassing five conventional time-series AI models whose performance ranges from 27% to 45%. These results demonstrate the model's strong few-shot learning capability and generalization across different operating conditions. This work promotes a shift from narrowly focused models in power electronics toward unified, scalable foundation models that can unlock value from large-scale unlabeled operational data.
Read the full abstract ▾
All authors Xinyuan Liao, Xiangtian Deng, Yang Liu, Zhaobing Cai, Haoran Wang, Yi Zhang
Hardware Architecture
FPGA Based Design And Implementation of Neural Network Compensators for Nanocrystalline MEMS Pressure Sensor
Somsubhra Talapatra, Sudip Ghosh, C Roy Chaudhuri et al. (+2 more)
In this paper the effects of nonlinearity and temperature of nanocrystalline MEMS pressure sensor have been compensated using neural network
algorithm which has been implemented on a FPGA chip. The neural network employs back propagation
algorithm to train the multilayer perceptron feed forward network that causes an error reduction due to temperature upto the level of 98%. It is realized on FPGA with pipelined architecture for parallel and fast sensing operation. The pressure sensor along with the FPGA based compensation results in an intelligent sensor system capable of high speed and accurate pressure measurement which has been tested for a typical biomedical application.
Read the full abstract ▾
All authors Somsubhra Talapatra, Sudip Ghosh, C Roy Chaudhuri, Biplab K Sikdar, Hafizur Rahaman
Robotics
SkipWorld: Latent Skip Planning with Distilled World Models for Efficient Visual Navigation
Yifei Yang, Zehua Fan, Zike Yan et al. (+2 more)
World-model-based visual navigation enables action planning by predicting future states under candidate trajectories. However, long-horizon rollout remains challenging: models are often trained with ground-truth visual histories but deployed in closed-loop prediction, and evaluating many candidates with step-by-step rollout is computationally expensive. We propose SkipWorld, a latent skip planning framework for efficient visual navigation. SkipWorld introduces three components: (i) Latent AutoRoll, which trains a world model to recursively roll out future states in latent space, reducing the mismatch between training and closed-loop inference; (ii) SkipDistill, which distills the latent autoregressive model into a skip world model that directly predicts long-stride future latent states; and (iii) Skip-then-Refine Planning, a dual-model coarse-to-fine planner that first filters candidate trajectories with the skip model and then reranks selected candidates with the stronger autoregressive model. This design preserves accurate trajectory evaluation while reducing redundant rollout computation. Experiments on RECON, HuRoN, SCAND, and TartanDrive show that, compared with NWM, SkipWorld reduces the average ATE by 4.2% and RPE by 2.7%, while decreasing planning time by 62.4% with a 2.7-fold speedup.
Read the full abstract ▾
All authors Yifei Yang, Zehua Fan, Zike Yan, Zhongyu Li, Yan Wang
Computer Vision
Human-Inspired Asymmetric Collaboration for Dual-Student Semi-Supervised Medical Image Segmentation
Huiyi Ren, Ziming Fang, Jiaming Cai et al. (+2 more)
Dual-student learning enriches predictive diversity in SSMIS by introducing a strongly augmented view alongside a weakly augmented view. Insufficient strong augmentation may homogenize the students, whereas excessive augmentation can corrupt anatomical evidence and destabilize collaboration. This challenge echoes feedback-regulated human instruction, where learners receive differentiated tasks and the task difficulty is continually adjusted according to their evolving responses. Inspired by these, we propose Human-Inspired Asymmetric Collaboration (HAC), a closed-loop dual-student framework that maintains weak–strong learning through adaptive difficulty calibration and geometry preservation. Specifically, Reinforcement-guided Difficulty Calibration (RDC) formulates strong-augmentation control as a window-level reinforcement learning process, using augmentation history, realized perturbations, and training responses to dynamically adjust the augmentation magnitude, operation number, and magnitude range. Meanwhile, the Gramian Geometry Preservation Constraint (G2PC) preserves the early-stage three-branch geometry through global volume and pairwise area constraints, preventing shared supervision from homogenizing the two students. Comprehensive evaluations across LA, Pancreas, and BraTS-2019 validate both the segmentation effectiveness and capability-aware difficulty calibration of HAC.
Read the full abstract ▾
All authors Huiyi Ren, Ziming Fang, Jiaming Cai, Rui Xu, Haojie Li
Machine Learning · Cross-listed from quant-ph
Distance Prediction for Quantum Codes
Nadav Schneider and Eliya Nachmani
The distance of a
quantum code is a fundamental metric that dictates its error-correction capability. While
quantum low-density parity-check (qLDPC) families such as Bivariate Bicycle (BB) codes offer a promising pathway to low-overhead fault-tolerant
quantum computing, computing their distance is challenging. Current exact solvers and even search heuristics that provide a distance upper bound suffer from severe computational bottlenecks, restricting the discovery of optimal codes. In this paper, we construct a
dataset of BB codes with corresponding distance upper bounds obtained through high-budget heuristic searches. We introduce BB-Net, an
attention-based neural network architecture for predicting the distance of BB codes. It achieves 96.39% accuracy on the reference distances in the test set, compared with 94.86% for state-of-the-art heuristics with a budget of up to $10^{4}$ times more computational runtime. On a subset with certified exact distances, BB-Net achieves 98.94%. By relieving the distance-calculation bottleneck, BB-Net enables faster screening of large code spaces, particularly in ML-driven search pipelines, and thereby can accelerate the discovery of novel
quantum error-correcting codes, an important capability for advancing scalable fault-tolerant
quantum computing. Our code is publicly available: https://github.com/NadavSc/BBNet
Read the full abstract ▾
Artificial Intelligence
Monitoring latent reasoning in looped language models
Pawel Olszowiec, Paulina Tomaszewska, Maciej Żelaszczyk et al. (+2 more)
Looped
language models (LMs) reuse one block of layers several times per
token generation, promising more latent computation at no additional storage cost. Despite their rising popularity, the analysis of looped LMs is still in its infancy: little is known about whether the iterative transformations actually improve their capabilities, or whether these models remain stable under
prompt perturbation. We construct an inspection protocol rooted in Mechanistic Interpretability comprising five diverse lenses. We apply it to examine eight looped LMs from four families. Models differ in many aspects but the key one is whether the
prompt is transmitted only within the looped state (carry loops: Ouro, LoopUS) or re-supplied at every iteration (re-read loops: Parcae, and Rec-OLMo & Huginn from the Raven family). The results of our study provide valuable insights concerning safety and efficient use of looped
language models.
Read the full abstract ▾
All authors Pawel Olszowiec, Paulina Tomaszewska, Maciej Żelaszczyk, Michal Byra, Neo Christopher Chung
Artificial Intelligence
Fresh at Decision, Stale at Action: Routing New Information to LLM Decisions Already Made
Ayoub Merimi, Souhail Bakkali and Abdeljalil Sakat
Freshness shapes an LLM decision at the moment it is made; the action it commits to is latent, carried out later, in a changing world. We therefore introduce a timestamp theory of decision validity and a decision ledger that realises it. Six timestamps define a decision; a freshness ceiling bounds what up-to-date evidence guarantees at action time, placing the remaining risk in information published afterwards. Accuracy then factorises into two capabilities: routing, which delivers that information to affected decisions before their revision deadlines, and temporal judgment, which reads it for the time of action. The ledger is persistent state holding each decision's times, evidence and background knowledge, indexed by watch conditions written at decision time. Controlled decision streams with exact labels confirm four predictions across five LLMs, from 8B to frontier scale: freshness alone yields exactly the accuracy the ceiling predicts; models act on decisive news once it arrives, so accuracy follows routing; the ledger's index sets what routing reaches, and watch conditions reach indirect dependencies at low exposure; and the remaining errors concern when a change takes effect. Freshness makes a decision right when it is made; routing and temporal judgment keep it right when it is carried out.
Read the full abstract ▾