Data Structures and Algorithms
Truthful Calibration Measures for Sequential Prediction
Anagha Gokul, Jason Hartline, Lunjia Hu et al. (+2 more)
Calibration requires probabilistic reports to be conditionally unbiased and reliably interpretable as probabilities. A calibration measure assigns numerical error to miscalibrated reports. Haghtalab et al. (2024) proposed an approximately truthful calibration measure for online prediction, leaving open whether exact truthfulness is compatible with completeness and soundness. We resolve this question negatively for sequential binary prediction: exact truthfulness is incompatible with completeness and soundness, even for independent outcomes. We then show that this impossibility is specific to exact truthfulness. On the positive side, we give two general reductions that transform a base calibration measure satisfying a uniform completeness condition into additively and multiplicatively approximately truthful measures, respectively. Applying the multiplicative reduction, for every $0<\varepsilon<1$ we construct a sound and complete calibration measure that is $1+\exp\left(-\Omega(T^{1-\varepsilon})\right)$-multiplicatively truthful. This improves the approximate truthfulness guarantee of Haghtalab et al. (2024). We also prove a general transfer
theorem showing that the uniform completeness of one calibration measure transfers to another when they are quantitatively related. Using this result, we show that every consistent calibration measure (Blasiok et al., 2023) satisfies uniform completeness and can serve as the base calibration measure to achieve $1 + o(1)$ multiplicative truthfulness.
Read the full abstract ▾
All authors Anagha Gokul, Jason Hartline, Lunjia Hu, Jonathan Ullman, Yifan Wu
Machine Learning
Asymmetric Capacity Allocation in Self-Refinement Pipelines
Zhuoyi Yang, Ian G. Harris, Salar Hashemitaheri et al. (+7 more)
Self-refinement, typically structured as generation, critique, and revision, is a widely adopted paradigm for improving LLM generation and serves as a core mechanism in many LLM
agents. While the three stages involve different cognitive demands, most existing approaches conveniently treat the model size as an implementation detail rather than a subject of study, which may lead to a waste of resources. Little work has systematically examined how model size affects each stage or whether effective self-refinement requires equally capable models for generation, critique, and revision. We present the first stage-wise model size study of the self-refinement pipeline on 5
benchmarks from different domains using 6 model sizes of Qwen3 and 4 model sizes of Gemma 3. We conclude that larger generators and refiners generally improve the pipeline, whereas an undersized refiner can even harm performance. Second, performance is highly insensitive to the size of the critic, although including even a small critic consistently outperforms omitting critique altogether. Our findings demonstrate that model capacity should not be allocated uniformly across self-refinement pipelines. Instead, different stages exhibit distinct size scaling characteristics, providing practical guidance for designing more computationally efficient multi-stage
language model systems.
Read the full abstract ▾
All authors Zhuoyi Yang, Ian G. Harris, Salar Hashemitaheri, Cassie Huang, Yuangang Li, Hyunwoo Oh, Paul Dourish, Tony Givargis, Mohsen Imani, Li Zhang
Artificial Intelligence · Cross-listed from eess.AS
TurboBias 2.0: Streaming Context-Biasing for Production-Efficient ASR Systems
Vladimir Bataev, Lilit Grigoryan, Andrei Andrusenko et al. (+3 more)
Contextualization is essential for production automatic
speech recognition (ASR) systems, where user-provided phrases must be recognized accurately under strict latency constraints. Although many context-biasing methods improve recognition accuracy, they often do not address the practical requirements of modern production ASR systems: streaming inference, efficient batched decoding, user-specific context lists, and low runtime overhead. We propose TurboBias 2.0, a production-oriented framework for efficient phrase boosting in Transducer-based ASR systems. The framework extends GPU-accelerated TurboBias with a case-insensitive boosting graph and per-stream batched decoding, allowing each utterance in a batch to use an independent context-biasing configuration. This enables personalized context biasing for multiple simultaneous users without sharing or mixing their context lists. The proposed framework supports both offline and streaming inference and can be used with greedy and beam-search decoding. Experiments show that TurboBias 2.0 improves contextual phrase recognition while preserving low latency and high throughput.
Read the full abstract ▾
All authors Vladimir Bataev, Lilit Grigoryan, Andrei Andrusenko, Nikolay Karpov, Vitaly Lavrukhin, Boris Ginsburg
Software Engineering
Natural-Language Workflows Are Not Software Yet: Artifact-Driven Compilation for Reliable Agent Execution
Xiangzhe Xu, Hanxi Guo, Guangyu Shen et al. (+2 more)
Natural-
language workflows offer a software-like interface for
agents: domain experts can write reusable procedures, and
agents can execute them as instructions. This promise is not yet reliable. Workflow descriptions often leave data dependencies implicit, so the executor must infer which prior results a step should use;
agents can also fail to follow long or branching instructions under context pressure. We propose Artic, an artifact-driven workflow compiler that transforms a natural-
language workflow into an artifact-driven workflow in which each step declares the artifacts it reads and writes, constraints gate produced artifacts, and explicit control transfers route execution. This representation exposes the enforcement burden placed on
agent execution, allowing the compiler to identify steps that depend on too much state or contain difficult control logic and refine them through constrained optimization. To validate the LLM-assisted transformation, Artic decomposes faithfulness checking into local obligations and uses scenario-based dry runs to test whether compiled workflow regions conform to the source workflow. We evaluate Artic on 488 problem instances from 11 real-world domain workflows; it improves task resolve rate by 28 percentage points over the original text workflow. We also show that workflows compiled by Artic are 32 and 56 percentage points more consistent in cross-model and repeated-execution setups, respectively.
Read the full abstract ▾
All authors Xiangzhe Xu, Hanxi Guo, Guangyu Shen, Siyuan Cheng, Xiangyu Zhang
Information Theory
The first tight classification of skew-constacyclic codes over finite fields
Monica Nevins and Susanne Pumluen
We parametrize the isometry and equivalence classes of skew constacyclic codes over a finite field by classifying the corresponding classes of their ambient rings, and present
algorithms for the parametrizations. We achieve a tight classification by taking all possible Hamming-weight preserving isomorphisms between their ambient Petit rings into account. We also count the number of equivalence classes of these rings. We present many examples where isometry is a strictly stronger relation than equivalence, that is, codes that are isometric but not equivalent.
Read the full abstract ▾
Machine Learning
Across-Design Uncertainty in Short Pricing Panels: Inference and Identification
Pedro Cadahia Delgado
Short observational pricing panels often contain many data points but very few actual price changes. This paper shows that this sparsity creates a hidden source of error that standard statistical methods miss. When estimating price effects, most of the uncertainty does not come from sample size within a panel, but from the specific history of price movements observed. Standard confidence intervals fail because they only measure variation within the panel, ignoring this broader design-level error. Using simulations, we find that this cross-design variation accounts for most of the estimation error, causing standard methods to significantly understate uncertainty. First, we show that cross-design error decreases predictably as the total volume of price variation increases. Second, adding more data from regions that share the same price trends does not fix the issue; true precision improves only when combining data across units with independent price trajectories. Third, applying a simple variance-component adjustment across independently priced units restores accurate statistical coverage. We confirm these findings in real-world store scanner data, showing that products and pricing zones behave as if they have far fewer independent price movements than their raw counts suggest. Ultimately, reliable inference in passive pricing data requires genuine, independent variation, which can be achieved through controlled regional price testing.
Read the full abstract ▾
Artificial Intelligence
Anatomy-Informed Neural Networks: Encoding Anatomic Priors in Loss and Architecture, with an SE(3) Formulation of Guidewire-Induced Aortoiliac Deformation
David P. Stonko
Deep-learning models of anatomy can be numerically plausible yet anatomically impossible, and they generalize poorly when data are scarce. We introduce Anatomy-Informed Neural Networks (AINN), in which soft anatomic priors enter as penalty terms in the loss (e.g., a branching penalty that treats a renal transplant artery off the iliac instead of the aorta as unexpected rather than impossible), in direct analogy to a physics-informed neural network, and hard anatomic priors (e.g., continuity of the vessel) are built into the architecture and state representation, making such invalid predictions impossible by construction wherever the prior admits architectural enforcement. We develop it on a clinical test case with limited data: how the aortoiliac tree deforms when a stiff wire is introduced endoluminally. This is important to contemporary aortic surgery and will matter to autonomous endovascular
navigation. We lift the vessel centerline and the wire path from R^3 to curves of frames in the Lie group SE(3), and couple a Cosserat-rod wire to a tortuosity-modulated, anatomically anchored vessel through a unilateral lumen-contact inequality. The prediction is a constrained minimizer of the coupled elastic energy, with contact forces as its Lagrange multipliers. Supervision is a Wasserstein-2 optimal-transport loss between the predicted projection through the C-arm geometry and the observed angiogram, so a 2D angiogram can train a 3D prediction. The kinematics, loss and projection are verified against known ground truth; the mechanics solver only against its own optimality conditions, and predicted displacement is not yet mesh-converged. Here, no network is trained. Future work will transfer this in silico model to real CT scans and test whether it improves predictive accuracy and reduces the training data required.
Read the full abstract ▾
Robotics
NeSAM: Neuro-Symbolic Kinodynamics with Soil Adaptation for Off-Road Mobility
Chenhui Pan, Tong Xu, Francesco Cancelliere et al. (+1 more)
Accurate prediction of off-road vehicle motion over deformable terrain remains challenging because sinkage, slip, and traction vary with local soil conditions. Existing learning-based kinodynamic models directly approximate vehicle-terrain interactions from data but do not explicitly represent soil mechanics and offer limited physical interpretability. To address these limitations, we present NeSAM, a neuro-symbolic framework that combines differentiable Bekker-Wong terramechanics with learned terrain representations and a
Transformer-based residual dynamics model for long-horizon, six degree-of-freedom kinodynamic prediction. The terramechanics component models soil-dependent interaction forces, while the residual model corrects discrepancies between the analytical prediction and the observed vehicle dynamics. NeSAM further estimates physically meaningful soil parameters from terrain observations and updates them online using an extended Kalman filter. We evaluate NeSAM in Verti-Bench, a simulator built on the Chrono multiphysics engine, and validate its performance on a physical Verti-4-Wheeler platform. NeSAM improves prediction accuracy by up to 30% in simulation and 29% on real-world data relative to the strongest compared baselines. When integrated with a close-loop
navigation controller, NeSAM further improves traversal success rate through online soil adaptation while reduces Hausdorff distance to the reference trajectory by 69.4%, indicating improved trajectory tracking accuracy.
Read the full abstract ▾
All authors Chenhui Pan, Tong Xu, Francesco Cancelliere, Xuesu Xiao
Computers and Society
Invisible Agents, Uninformed Patients: Towards Responsible Deployment Of Autonomous AI Diagnostic Agents In Sub-Saharan Africa
Percy Brown and Kweku Yamoah
Autonomous AI diagnostic
agents, systems that analyse patient-specific clinical data and produce diagnostic outputs or triage decisions without mandatory real-time human review, are increasingly deployed across eHealth platforms in sub-Saharan Africa at a pace that has outrun the governance infrastructure needed to oversee them. While significant bodies of work address AI accountability, transparency and explainability in healthcare, existing frameworks are largely clinician-centered and assume regulatory conditions that do not uniformly exist in low-resource settings. A patient-centered analysis of the disparity in patient awareness regarding autonomous
agents, which results in a structural accountability gap, is mostly missing from the literature. This paper synthesizes existing research on informed consent, algorithmic accountability, and explainable AI to highlight three distinct challenges introduced by deploying AI
agents in the sub-Saharan African context. Drawing on three documented deployment cases, including computer-aided tuberculosis detection in Tanzania, diabetic retinopathy and TB screening in Zambia, and mobile health chat-bot triage in Ghana, it demonstrates that these gaps are already present in active deployments across the region. In response, the paper proposes three foundational principles;
agent-aware informed consent, human override as a structural requirement and contextually adapted explainability. This triad of principles lays a practical minimum standard for developers, health system administrators and policymakers in contexts where formal AI regulation remains nascent.
Read the full abstract ▾
Computation and Language
Move by Move: Measuring and Steering How LLMs Conduct Psychotherapy
Afonso Baldo, Hugo Pitorro, Areti Vassilopoulos et al. (+5 more)
As the demand for mental health services outpaces clinical workforce capacity, users are increasingly turning to large
language models (LLMs) for accessible and private support. Although recent evaluations suggest LLM care can reduce clinical symptoms, how models actually conduct therapeutic interactions compared to trained human clinicians remains largely unknown. Drawing on established clinical psychology frameworks that conceptualize therapy as a sequential decision-making process, we can systematically analyze the conversational interventions used in AI-driven sessions. To achieve this, we introduce an original, clinically validated ontology of ten distinct therapeutic moves to descriptively categorize strategies LLMs favor or neglect. Applying this ontology at scale reveals that models drastically over-use probing, disregard common CBT-favored techniques and are strongly context-anchored. Moreover, we propose exposing these categories as callable tools, bringing models significantly closer to the human clinical baseline.
Read the full abstract ▾
All authors Afonso Baldo, Hugo Pitorro, Areti Vassilopoulos, Anabela C. Areias, Maya D'Eon, Fabíola Costa, Ricardo Rei, Nuno M. Guerreiro
Machine Learning
Time-Aware Tranformer-Based Prediction Model for AECOPD
Weihao Qu, Ling Zheng, Dongyang Wang et al. (+2 more)
The rapid symptom change of Acute exacerbation of chronic obstructive pulmonary disease (AECOPD) makes it critical to have time-sensitive prediction models. However, most current machine learning models studying AECOPD use clinical and laboratory data, which will inevitably cause latency. To ensure timely detection of AECOPD and minimize latency, this paper focuses on home monitoring scenarios where only respiratory data from daily-use ventilators is available. We introduce a Time-Aware
transformer-based AECOPD prediction model, which generates meaningful patient representations using the Time-Aware
transformer to capture the symptoms and their temporal progression in ventilator data. Our experimental results demonstrate that our Time-Aware
transformer-based approach outperforms traditional methods in multiple classification tasks, highlighting its potential to enhance AECOPD prediction accuracy.
Read the full abstract ▾
All authors Weihao Qu, Ling Zheng, Dongyang Wang, Jiacun Wang, Haowen Pan
Artificial Intelligence
Unified Branch-and-Bound Search for the Steiner Traveling Salesman Problem on Graphs of Convex Sets
Jingtao Tang and Hang Ma
We formalize the Steiner Traveling Salesman Problem (Steiner-TSP) on Graphs of Convex Sets (GCS), which seeks a minimum-cost closed trajectory through required convex sets while allowing optional transit vertices and revisits. To explore the resulting infinite solution space, we propose a unified branch-and-bound search over rooted walk prefixes. Additive lower-bound-graph costs bound committed prefixes, while a cut-separated connected-flow relaxation lower-bounds the residual cost of visiting every remaining target and returning to the root. Under a uniform positive-cost assumption, best-first traversal terminates after finitely many expansions on every feasible instance without an initial incumbent, whereas depth-first traversal does so once a finite incumbent is available. For a user-specified factor $\epsilon\geq1$, a global lower bound certifies that either strategy's incumbent cost is at most $\epsilon$ times the global optimum. We further demonstrate joint sensing-mode, visitation-order, and continuous-trajectory selection for a mobile-manipulator inspection task, including action precedences expressed in linear temporal logic over finite traces (LTL$_f$). Both traversal strategies find feasible solutions on all
benchmark instances within 30s with mean certified optimality gaps of 28.1% and 29.7%, respectively, whereas two recent baselines succeed on only about half of the instances
Read the full abstract ▾
Artificial Intelligence
From Regulation to Implementation: A Critical Evaluation of LLM-Assisted Regulatory Compliance in Industry
Adriana Watson, Marco Bücheler and Grant Richards
The European Union (EU) has emerged as a leading regulatory body in the development of sustainability and
privacy regulations. While new regulation requirements vary, many include a documentation artifact to ensure compliance. Notably, the Ecodesign for Sustainable Products Regulation (ESPR) introduces Digital Product Passports (DPPs) for life cycle transparency, while the General Data Protection Regulation (GDPR) mandates Data Protection Impact Assessments (DPIAs) to mitigate
privacy risks. Creating these compliance artifacts, however, is challenging. Industrial data, which often exists in heterogeneous formats and is scattered across company and supplier systems, is required for DPPs and can be difficult to extract into compliant DPP formatting. Furthermore, DPIA documents require interdisciplinary expertise and follow no standardized format, making development difficult for novel systems. To address the particular complexity of compliance artifact creation for both regulations, researchers have proposed the use of LLMs in the generation process; however, the impact of the aforementioned problems on the output of these systems is largely unaddressed. This work investigates the existing research gap by exploring how data extraction instructions and regulatory vagueness impact the quality and consistency of LLM-produced compliance artifacts. The resulting artifacts are evaluated by benchmarking different models against manually created ground-truth schemas. The results reveal that less strict guidelines, such as DPIA formatting, require higher context
prompts to maintain consistency and completeness. Stricter guidelines, such as formatting for Digital Battery Passports (DBP), result in consistent results regardless of
prompt context, but may lead to more
hallucinations in the output
Read the full abstract ▾
Computation and Language
Prompt-Model Interaction Reaches the Fixed Points: A deterministic, task-free structural readout – and the factorizations of it that failed
Nicolás Vera Zúñiga
That a
prompt's effect is not a property of the
prompt is established:
prompts optimised for one model degrade on another, and rankings reorder under neutral reformatting. That evidence is about task accuracy, which cannot say whether the interaction is a fact about task machinery or about the conditional distribution itself. We ask on a readout with no task in it: the fixed-point structure of the short-window argmax map x_{t+1} = argmax_x p(x | x_{t-1}, x_t), censused from 96 starts. It is deterministic, so nothing can be helped or hurt, and it exists only at short windows – four of six models lose it entirely by window 16 – so everything here concerns how a model reads a fragment. Two results. First, the interaction reaches this readout at full magnitude: nine
tokens of conditioning move the fixed-point fraction across most of its range, change a four-way structural class, and reorder models, while instruction tuning worth 60.5 IFEval points moves the class by zero. Second, nothing we proposed carries it. Prefix length fails: the effect is not monotone. Four phenomenological factors – prose-versus-markup, a universal direction, bidirectionality, instruct-resistance – were each withdrawn within one run of being proposed, dissolved by widening the sample. And the nearest mechanistic account,
attention-sink dominance of early
tokens, predicts the sign of the shift on 2 of 5 models – chance – while a length-by-content cross shows it holds on real text and fails on our probe's uniformly random input, so we are outside its regime, not against it. One fixed nine-
token prefix drives four models toward 0 and two toward 1; the bidirectionality survives in-distribution starts. On this readout the unit of explanation is the
prompt-model pair. The recurring error it caught in us has a name: a criterion with a shape applied to a quantity with no room to vary.
Read the full abstract ▾
Databases
VTRQ: Enabling Verifiable Trajectory Range Queries in Hybrid-Storage Blockchains
Zhongming Yao, Junchang Xin, Yumeng Song et al. (+7 more)
Due to their increasingly large volumes, outsourcing of trajectory storage and querying to third-party service providers has become attractive. However, in such outsourced environments, service providers may return incorrect, e.g., incomplete, tampered, or invalid query results, making verifiability of query results an important consideration. Existing hybrid-storage blockchains offer limited support for trajectory data, lacking authenticated data structures (ADS) that enable efficient verification. For example, ADSs designed for queries on one-dimensional data are unsuitable for queries on multidimensional trajectory data, while ADSs tailored for discrete data may yield incomplete results when applied to continuous trajectory data. We propose the first framework for verifiable trajectory range queries in hybrid-storage blockchains, called VTRQ. It features two efficient ADSs: (i) a spatial ADS for road networks that leverages hierarchical organization to aggregate trajectory, edge, and node hashes, thus reducing redundant computations and improving spatial verification efficiency; and (ii) a temporal ADS based on interval trees, which indexes only the start and end times of trajectories, thereby enabling
pruning and efficient temporal verification. By separating spatial and temporal indexing, the method reduces the need for data comparison, enhancing both query and verification efficiency. To aggregate spatial and temporal query results, VTRQ provides a spatio-temporal edge aggregation mechanism that combines temporal verification of spatial nodes, spatial intersection computation, and temporal intersection analysis to achieve spatio-temporal filtering.
Read the full abstract ▾
All authors Zhongming Yao, Junchang Xin, Yumeng Song, Yusen Mao, Kristian Torp, Yuemin Ding, Divesh Srivastava, Yushuai Li, Christian S. Jensen, Tianyi Li
Software Engineering
AI-to-AI Code Reviews of GitHub Pull Requests
Niruthiha Selvanayagam and Taher A. Ghaleb
AI coding
agents are increasingly integrated into software development workflows, operating on both sides of the pull-request (PR) process: AI authoring
agents create or modify PRs, while AI reviewers evaluate them. This creates a closed loop in which one AI coding
agent reviews a contribution attributed to another. We construct a large-scale
dataset of AI-to-AI code review by linking AI-attributed PRs with AI-attributed review events from CodAGE, a public
dataset of coding-
agent-generated GitHub events. Our
dataset contains 248,641 unique AI-attributed PRs that received at least one AI-attributed review. Of these, 45,269 received cross-product review and 208,145 received same-product review; 4,773 PRs received both. Cross-product AI-to-AI review occurred in approximately 1.6% of identified
agent-authored PRs but was substantial in absolute terms, and its volume increased by more than two orders of magnitude from 2025-Q1 to 2025-Q3. Reviewer output varied across author-reviewer configurations. CodeRabbit labeled 35.0% of its comments on Claude Code-authored PRs as refactor comments, compared with 10.5% on Copilot-authored PRs, although this difference may reflect characteristics of the PRs rather than the reviewer. For three of four dual-role reviewers, mean comments per PR were 58-65% higher in the same-product group, although effect sizes were small or negligible and the difference was concentrated in the upper tail. Among pairs with complete, nonnegative timestamps, the observed median latency was 1.2 minutes for cross-product pairs and 4.7 minutes for same-product pairs; differential timestamp availability and reviewer composition limit this comparison. Overall, closed-loop AI-to-AI review is increasing but remains a minority of identified
agent activity, with review output varying across authoring-
agent groups and product configurations.
Read the full abstract ▾
Software Engineering
Beyond Fault Localization: A Trajectory-Level Study of LLM Agents for Microservice Root Cause Analysis
Qisheng Lu, Aoyang Fang, Junjielong Xu et al. (+5 more)
Automated microservice root cause analysis (RCA) is usually evaluated by whether the final answer localizes the fault source. This is insufficient for on-call site reliability engineers (SREs), who need the evidence and
reasoning behind the fault. Evaluating the process is hard because
agent trajectories interleave free-form
reasoning with heterogeneous tool calls and telemetry results, making it difficult to assess
agents' behavior. To assess how well RCA
agents ground diagnoses in evidence, we propose a trajectory-analysis methodology that characterizes evidence
navigation and utilization, SRE-oriented diagnostic intent, and recurring failure modes. Our key insight is that fault-propagation paths give a causal reference for where
agents investigate and how they use evidence. With it, we analyze 5,000 trajectories from ten framework-backbone configurations (nine
agent frameworks, two backbone models) on RCABench. Correct localization can hide deficiencies in causal reconstruction. Successful investigations use retrieved evidence to guide later queries, consider broader diagnostic intents, and investigate more deeply. Failed investigations drift toward unrelated services or repeat shallow queries; recurring failures arise from omitted evidence, misinterpreted evidence, and unsupported
reasoning. To test whether these patterns can guide improvement, we build a lightweight defense prototype targeting the three failure families. Across two held-out
benchmarks, two backbone models, and six diagnostic frameworks, the failure patterns recur; the prototype reduces their prevalence and raises Acc@1 by 3.5-35.1 percentage points in every configuration, suggesting that the mined failures generalize and are actionable. Our findings motivate a process-level view of automated RCA, in which trajectories reveal whether a diagnosis is correct, well-grounded, and open to improvement.
Read the full abstract ▾
All authors Qisheng Lu, Aoyang Fang, Junjielong Xu, Jin'ao Shang, Songhan Zhang, Yifan Yang, Xiaochuan Yan, Pinjia He
Machine Learning
Rethinking Expressivity and Efficiency in Test-Time Training
Zeyun Zhong, Joya Chen, Manuel Martin et al. (+3 more)
Test-Time Training (TTT) enables long-context processing via continuous weight updates during inference, but current methods struggle to balance the expressivity of per-
token update dynamics with the hardware efficiency of chunk-wise approximations. We propose E$^2$-TTT (Expressive and Efficient TTT) to bridge this gap. Under the standard approximation of taking gradients at the chunk-start weights, we derive a closed-form state transition that exactly reproduces the chunk-end fast-weight and momentum states of the per-
token recurrence. This enables fully parallelized chunk-level training while preserving the temporal structure of the update rule that prior chunk-wise methods discard. We validate E$^2$-TTT by training models up to 1.3B parameters from scratch. It performs on par with previous TTT and hybrid
attention baselines in
language modeling while outperforming them on in-context
retrieval. Its advantage is most pronounced in length extrapolation: on the standard “Needle in a Haystack” passkey test, it retains over 90% accuracy at $8\times$ the training context length. Meanwhile, E$^2$-TTT can match the training throughput of efficient chunk-wise methods, demonstrating that it effectively reconciles expressivity with efficiency. The code is available at https://github.com/zeyun-zhong/E2-TTT.
Read the full abstract ▾
All authors Zeyun Zhong, Joya Chen, Manuel Martin, Frederik Diederichs, Juergen Gall, Juergen Beyerer
Machine Learning
SPARCL: Spectral Partitioned Analytic Continual Learning
James Hartley, Zeropy Surio, Daniel Whitmore et al. (+2 more)
Analytic continual learning has emerged as a strong exemplar-free alternative to gradient-based class-incremental learning because it replaces iterative optimization with closed-form ridge updates. Yet the usual forgetting narrative, centered on
stochastic gradient overwriting, does not explain why analytic methods still drift on old classes despite exact recursive solvers. We identify the culprit as spectral interference: the joint ridge classifier for all tasks shares the inverse autocorrelation operator $(R+\lambda I)^{-1}$, so incoming task samples that load onto old dominant eigendirections dilute the spectrum and perturb old-class logits even when old labels are never revisited. Based on this view, we propose SPARCL, a spectral partitioned analytic continual learner that decomposes the running autocorrelation into a high-energy core and a residual complement, freezes old-class classifier components in the core subspace, and updates only the residual block through recursive least squares with an optional residual random-projection expansion. This yields a simple closed-form update with a provable invariance guarantee for the core contribution of old logits. Across CIFAR-100, CUB-200, ImageNet-R, and ImageNet-A under a frozen ViT-B/16 protocol, SPARCL closes most of the gap from classical analytic learners to strong representation matchers, while remaining complementary to sparse feature-decorrelation approaches such as Fly-CL.
Read the full abstract ▾
All authors James Hartley, Zeropy Surio, Daniel Whitmore, Hannah Clarke, Thomas Reed
Computer Vision
Re$^3$Cap: Retrieval-Guided Refinement for Image Captioning Enhancement via Reinforcement Learning
Haonan Jia, Shichao Dong, Zenghui Sun et al. (+7 more)
Reinforcement Learning (RL) has demonstrated significant gains in image captioning, yet it is still limited in encouraging Large
Vision-
Language Models (LVLMs) to explore novel
reasoning strategies. This limitation leads to a performance gap between RL and Supervised Fine-Tuning (SFT). In this paper, we argue that multi-modal
retrieval can serve as an effective
reasoning signal for caption refinement. Based on this insight, we present the
Retrieval-Guided Refinement for Image Captioning (Re$^3$Cap), a
retrieval-guided
reasoning strategy that enhances image captioning without requiring additional annotations. Instantiated by Caption Refinement Suggester (CRS) and Caption Quality Assessor (CQA), this strategy identifies
hallucinations and omissions in image captions, leading to more accurate and detailed descriptions. Extensive experiments demonstrate the superiority of our method in image captioning, even compared with Supervised Fine-Tuning. Especially, Re$^3$Cap outperforms GRPO with an average improvement of 8.64% in relation
reasoning on the COCO-LN500
benchmark.
Read the full abstract ▾
All authors Haonan Jia, Shichao Dong, Zenghui Sun, Jiawen Zheng, Ziqi Miao, Gege Shi, Qiuyu Zhao, Jinsong Lan, Xiaoyong Zhu, Bo Zheng
Numerical Analysis
Basin-Preserving Discretizations of Modern Hopfield Retrieval Dynamics: Energy Cells, Dissipation, and the Attention Limit
Francisco R. Villatoro
The
retrieval dynamics of a modern Hopfield network is the gradient flow of a log-sum-exp energy, while the
attention update is its exact difference-of-convex minimization step. We study which time discretizations preserve not only energy decay and
equilibria but also basins of attraction. We introduce energy
cells, connected components of sublevel sets containing one attractor and no other critical point. Our main
theorem shows that every finite energy
cell below the escape energy is contained simultaneously in the basin of the continuous flow, every relaxed
attention map $\Psi_\theta=(1-\theta)\,\mathrm{id}+\theta\,\mathrm{attention}$ for $0<\theta<2$, and implicit Euler throughout its uniqueness regime. A parameter-uniform unit-
curvature majorant yields unconditional dissipation and a monotone interpolation of each discrete step. We also derive explicit local contraction bounds near well-separated patterns, with a certified optimal slight overrelaxation; characterize proximal tunneling and overshoot beyond the preservation regimes; compare first-order error constants; establish an order barrier for scalar reparametrizations of the relaxed family; construct a second-order scalar-auxiliary-variable scheme; and extend
cell preservation to damped difference-of-convex iterations in Bregman geometry, including a certified overrelaxed window under bounded asymmetry. Nine numerical campaigns test the bounds and failure mechanisms. In two-dimensional basin experiments, all observed disagreements between continuous and discrete
retrieval occur above the attractor-specific numerically inferred escape level.
Read the full abstract ▾
Systems and Control
Fault Diagnosis of Dynamic Systems Under Unknown Operating Conditions: A Condition-Guided Selective Adaptation Approach
Jiaming Liu, Zeyi Liu, Hongshuo Zhao et al. (+2 more)
Fault diagnosis under unknown operating conditions remains challenging for dynamic industrial systems, as the distribution shift caused by changing operating conditions can significantly degrade the performance of diagnostic models in real-world applications. To address the problem, a condition-guided selective adaptation approach is proposed. Specifically, a novel continuous operating-condition
adversarial learning strategy with progressive training is developed in the offline stage to enhance the generalization ability of the diagnostic model. During online deployment, residual operating-condition responses are exploited to identify reliable unlabeled samples from streaming data, which are then used to update the diagnostic model. Extensive experiments on real-world gearbox and motor
datasets have demonstrated that the proposed framework outperforms state-of-the-art methods in diagnostic accuracy while maintaining relatively low test-time, showing its potential for practical industrial applications.
Read the full abstract ▾
All authors Jiaming Liu, Zeyi Liu, Hongshuo Zhao, Pengyu Han, Xiao He
Computer Vision
When Adaptation Hurts: Connecting Representational Drift to OOD Failures in MedSAM Fine-Tuning
Marko Haralović, Sounic Akkaraju, Carlo Baretta et al. (+2 more)
Foundation models for medical image
segmentation, like
prompt-based MedSAM, generalize well across domains and modalities, often in zero or few-shot setups. However, their performance depends on the quality of
prompts and the adaptation of the models to custom
datasets. This work systematically examines how MedSAM generalizes across diverse medical imaging
benchmarks, with six adaptation strategies: full-model and encoder-only LoRA, shallow and deep visual
prompt tuning (VPT), and decoder-only and full fine-tuning. Models are trained on the International Skin Imaging Collaboration Challenge (ISIC 2018)
dataset and evaluated under clean and increasingly noisy
prompts on IN and Out-of-Distribution (OOD)
datasets: close-OOD PH2 (dermoscopy), far-OOD BUSI (Breast Ultrasound Images
Dataset) and CBIS-DDSM (Curated Breast Imaging Subset of the Digital Database for Screening Mammography). We show that adaptation improves performance on IN and close-OOD data but often reduces performance on far-OOD data. Full fine-tuning provides the best tradeoff, while encoder-only LoRA is the strongest parameter-efficient alternative, outperforming standard LoRA and VPT under far-OOD shifts. Using Centered Kernel
Alignment (CKA), we show that far-OOD degradation is strongly associated with drift in decoder representations, whereas encoder similarity alone does not explain robustness. This suggests encoder-only LoRA provides stronger robustness than standard LoRA by adapting the encoder to distribution shift in visual features, while preserving the decoder pathway. We further show that random 0-100 pixel jitter on
prompts produces more robust and better performing models. We thus conclude that robust MedSAM adaptation requires the combined consideration of
prompt noise exposure, domain shift, and representation preservation. We release our code: https://github.com/ImSounic/medsam-vpt
Read the full abstract ▾
All authors Marko Haralović, Sounic Akkaraju, Carlo Baretta, Vasil Zapryanov, Alexia Briassouli
Software Engineering
Human-AI Collaboration in Requirements Engineering: Evidence of the Negative Effect of LLMs on Requirements Inspection
Giovanna Broccia, Julian Frattini, Chetan Arora et al. (+4 more)
Background. Requirements inspection (RI) is a well-established practice for detecting potential defects in requirements artifacts early in the software lifecycle. Recent advances in large
language models (LLMs) have stimulated interest in their potential to support requirements engineering (RE) tasks. However, empirical evidence on the effects of LLMs when used as collaborative assistants in human-performed RI remains scarce. Aims. We aim to investigate the impact of LLM support on human-performed RI, considering inspection effectiveness in terms of smell identification and severity classification (i.e., nocuous vs innocuous), as well as inspection duration. Method. We conducted a controlled crossover design experiment with 34 participants, who inspected textual specifications with and without LLM support, identifying and classifying requirements smells while recording inspection time. We analyzed the data using one
Bayesian regression model per outcome variable, accounting for validity threats induced by the crossover design as well as
covariates and mediators. Results. Results show that LLM support negatively affects smell detection accuracy but has no significant effect on smell classification or task duration. A learning effect is present across experimental periods, but reduced when RI is first performed with LLM support. Conclusions. Our findings provide empirical evidence that LLM support does not necessarily improve performance and may, instead, hinder it for novice inspectors. Moreover, the results suggest that learning RI with LLM-support from the beginning may slow down the skill acquisition process, implying threats for LLM-supported learning.
Read the full abstract ▾
All authors Giovanna Broccia, Julian Frattini, Chetan Arora, Maurice H. ter Beek, Alessandro Fantechi, Andreas Vogelsang, Alessio Ferrari
Multiagent Systems
Level-k Distinguishable Mechanisms for Evaluating Bounded Rationality in LLMs
Binchi Zhang and Atrisha Sarkar
Strategic depth of
reasoning is essential for human interaction of Large
Language Models (LLMs) operating in boundedly rational environments. However, existing evaluations are primarily based on canonical games prevalent in pretraining corpora, making it difficult to disentangle true strategic
reasoning from memorisation. To address this, we formalise a necessary level-K distinguishability condition for strategic depth inference and construct a suite of novel game structures that meet this standard. Using these games, we evaluate strategic depth in LLMs from both the Chain-of-Thought
tokens and actual actions under recursive
reasoning and an inductive trace of opponent game-play data. Across experimental trials spanning four LLMs, four game structures, and ten levels of iterated
reasoning, we find that model models maintain accurate strategic depth under recursive
reasoning, with strong internal consistency between stated
reasoning and actions at every level. Errors arise from using the wrong number of iterated depth of
reasoning steps, not from computing best responses incorrectly. However, inductive inference from opponent play degrades accuracy sharply and unevenly across games, and explicit strategic mentalizing in the chain of thought substantially improves overall performance.
Read the full abstract ▾
Human-Computer Interaction
Event-Time Confounding Under Bursty Human Dynamics
Michael Iannelli and Alan Ai
Studies of digital behavior often align users at moments they choose, such as opening an AI assistant, clicking a recommendation, or visiting a product page, and interpret higher activity afterward as an event effect. We show how this creates an endogenous time zero: the event occurs during an ongoing task episode, so the aligned curve can trace episode continuation rather than a response to the event. In same-user, cross-surface web logs, AI, shopping, news, coding, and reference events are all preceded by broad activity increases that peak before time zero. Our strongest test uses known-null timestamps that cause nothing. Among the 5.8% of AI responses meeting strict pre-event activity and washout criteria, these timestamps show 3.42 times the post-event search activity of a within-user placebo, compared with 4.32 times for real events. The fraction of excess reproduced by the known null falls from 0.56 at detectably active moments to -0.04 at quiet moments, where the design detects none. We formalize this episode-selection bias, prove that a single-surface event window cannot separate it from a genuine effect without additional assumptions, and show in zero-effect simulations why user fixed effects and coarse activity matching can fail: the confound is within-user and time-varying. We provide a diagnostic protocol, public-data
benchmarks, and burstcheck, a lightweight audit tool. User-timed events may have real effects, but post-event volume does not identify them by default; studies should compare similar episodes with and without the event.
Read the full abstract ▾
Artificial Intelligence
AUSO: Action-Level Unified Skill Optimization from Internalization to Utilization
Huizu Lin, Chengkai Huang, Tianqi Gao et al. (+5 more)
Skills play different roles as an
agent's policy evolves: they should first provide learnable knowledge, then support capability formation, and finally be invoked only when they improve individual decisions. Existing methods rarely model this lifecycle. They either keep skills outside the model, fully internalize them, or select among internalization and utilization objectives through noisy task-level success rates. Such designs fragment training and assign uniform importance to actions within the same trajectory, even though skill guidance may help some decisions while distracting others. To solve these problems, we introduce AUSO (Action-level Unified Skill Optimization), which unifies skill learning and skill use through a progressive, action-aware optimization process. At the beginning of training, AUSO jointly learns from teacher guidance and environmental outcomes, enabling the policy to acquire foundational skills without losing task-oriented feedback. It subsequently emphasizes outcome-based policy optimization to consolidate autonomous problem-solving ability. As the policy matures, AUSO evaluates each sampled action under both skill-conditioned and skill-free contexts. The resulting action-level information signal is coupled with the trajectory outcome advantage, allowing beneficial skill-sensitive actions to receive stronger updates and harmful ones to be suppressed. Therefore, skills gradually transition from an external source of supervision into decision knowledge whose utilization is adapted to its action-level benefit, while reinforcement learning remains the shared backbone across all stages. Experiments on ALFWorld, WebShop, and SearchQA show that AUSO consistently improves
agent performance and out-of-distribution generalization over competitive baselines.
Read the full abstract ▾
All authors Huizu Lin, Chengkai Huang, Tianqi Gao, Tao Huang, Daijiao Liu, Tongxin Li, Xiaoyan Sun, Lina Yao
Robotics
VT-MUSE: Multimodal Unified Sequential Visuotactile Representation Learning for Manipulation
Congsheng Xu, Qiaochu Yang, Fangyuan Shi et al. (+9 more)
We propose VT-MUSE, a
Multimodal Unified SEquential representation learning framework for visuotactilemanipulation. Existing approaches often encode visual and tactile observations independently before fusion, limiting their ability to capture fine-grained cross-modal dependencies. Moreover, most methods focus on observations at the current time step and overlook the temporal
evolution of contact. VT-MUSE addresses both limitations through a two-stage representation learning framework. In Stage I, modality specific encoders are jointly adapted via cross-modal temporal
alignment and masked-view consistency. In Stage II, a conditional variational latent model processes masked visual sequences together with full tactile histories. Auxiliary decoders reconstruct the masked recent visual observations and predict tactile depth changes, encouraging the latent representation to retain both global visual context and local contact dynamics. The learned representation is subsequently integrated into a lightweight
Transformer policy through gated cross-
attention. On the simulation
benchmark, VT-MUSE outperforms the strongest baseline evaluated on all tasks by 11 percentage points and also achieves substantial improvements in real-world experiments.
Read the full abstract ▾
All authors Congsheng Xu, Qiaochu Yang, Fangyuan Shi, Yifan Han, Baijun Chen, Yiming Wang, Haonan Zhao, Zhe Liu, Yao Mu, Daolin Ma, Xiaokang Yang, Hesheng Wang
Human-Computer Interaction
Supporting The Many Lives of Personal Data with Rebite: LLM-Powered Goal-Directed Framing in Food Journaling
Weijun Li and Daniel A. Epstein
People's health and tracking goals frequently change, but most personal informatics systems struggle to adapt, leading people to abandon their data and start over. We propose goal-directed framing, an approach that repositions goals within personal informatics systems. Instead of fixing the meaning of data at capture time, the approach frames the collected data through the current goal and reframes it whenever the goal changes. We realize this in Rebite, a photo-based food journaling system that uses LLMs to read unstructured meal photos and produce goal-directed feedback. In a one-week deployment with 21 participants managing multiple dietary goals, we find that goal-directed framing shaped how participants engaged with their goals. Translating a goal into metrics helped them see what it meant in practice, confirming existing priorities, surfacing what they overlooked, and revealing where the metrics fell short. When goals changed, seeing past meals reframed under the new goal exposed overlaps and conflicts, prompting participants to negotiate trade-offs and refine priorities. We discuss how goal-directed framing both supports and complicates reflection as goals change, and offer design implications for personal informatics systems to support evolving goals.
Read the full abstract ▾
Hardware Architecture
Assessing Triple Modular Redundancy for Wide-Link, Low-Latency NoC Routers: Reliability and Physical Design Challenges
Chen Wu, Michael Rogenmoser, Luca Benini et al. (+1 more)
Protecting the Network-on-Chip (NoC) of physical-AI tile-based accelerators deployed in harsh environments against single-event effects (SEEs) is paramount for preventing NoC failures that can lead to deadlocks and silent data corruption (SDC). Prior work on reliable NoCs has largely focused on narrow links (e.g., 32-bit), deeply pipelined routers, and single-event upsets (SEUs). However, the state of the art has evolved toward low-latency NoC routers with ultra-wide links, implemented on advanced technology nodes and operating at frequencies above 1 GHz. We evaluate the cost and reliability trade-offs of implementing Triple Modular Redundancy (TMR) at three granularities (coarse, state-only, and full) for a 2-cycle-latency NoC router with 512-bit wide links. We carry out RTL-to-GDSII physical design in TSMC 7nm technology, as well as both RTL- and netlist-level SEU and SET fault injection campaigns. We evaluate the three TMR approaches in terms of reliability, cost, and physical design strategies, further extending the assessment from a standalone router to a full AI acceleration tile. Our results show that state-only and coarse-grained TMR do not provide sufficient protection against SEEs, whereas full TMR eliminates all observed failures across more than one million injected faults per experiment. Although the standalone full-TMR router incurs a 7.04x area overhead, this cost is drastically amortized once integrated into a complete AI accelerator tile with processors and local L1
memories: the same design adds only 16.8% area and 15.2% power consumption under a GEMM
benchmark at the system level, with the critical path of the tile entirely unaffected. These results demonstrate that advanced technology nodes provide sufficient routing capacity to make full TMR a practical and deployable solution for protecting NoCs in Physical AI systems operating in harsh environments.
Read the full abstract ▾
All authors Chen Wu, Michael Rogenmoser, Luca Benini, Angelo Garofalo