arXiv:2607.04655v1 Announce Type: new Abstract: The veracious semantic alignment in autoformalization is significant for formal mathematical reasoning. However, existing evaluations provide only opaque binary verdicts or scalar scores, offering no interpretable insight into where or why translations fail. This opacity severely limits both human understanding and automated system improvement. To bridge this gap, we introduce FormalRx, a comprehensive diagnostic evaluation framework that transforms autoformalization assessment from black-box judgments into actionable feedback. At its core is SCI Error Taxonomy, a hierarchical classification scheme decomposing autoformalization errors into 28 distinct categories with strict priority ordering. Building on this taxonomy, FormalRx provides four critical diagnostic capabilities: alignment verdicts, error categorization, error localization, and correction. We instantiate the framework with a diagnostic model FormalRx-8B, trained on 56,287 NL-FL pairs with fine-grained diagnostic annotations, and release FormalRx-Test as the first fine-grained diagnostic benchmark. FormalRx-8B achieves F1-scores of 0.88 (verdict) and 0.71 (categorization), along with accuracies of 0.75 (localization) and 0.73 (correction), substantially outperforming both general-purpose LLMs and specialized baselines. By connecting evaluation with actionable insights, FormalRx enables systematic diagnosis and improvement of autoformalization systems.
Science Journals
arXiv:2607.04755v1 Announce Type: new Abstract: Model merging offers a practical alternative to conventional continual learning by integrating independently fine-tuned models without retaining previous training data. Recent state-of-the-art model merging methods employ test-time adaptation (TTA-guided merging) to address distribution shifts by adjusting merging-related variables using unlabeled target data. However, these methods have primarily been studied in multi-task or single-target settings, and their behavior under sequential continual learning remains insufficiently understood. We present a benchmark study that maps this family of methods to rehearsal-free continual Whole Slide Image classification and evaluates them against traditional continual-learning approaches. Experiments on six TCGA cancer-subtyping cohorts cover CLASS-IL and TASK-IL scenarios, in-domain and out-of-domain evaluation, and different task orders. The results show that adapting model merging at test time can provide strong task-specific performance and improve retention of previously acquired knowledge without storing historical WSIs. Nevertheless, performance remains sensitive to task order and to the interaction between adaptation on the current distribution and accumulated knowledge. This benchmark identifies model merging with test-time adaptation as a promising direction for continual computational pathology and motivates future methods that balance adaptation to domain shift with explicit preservation of historical knowledge.
arXiv:2607.04761v1 Announce Type: new Abstract: Real-time, text-promptable 4D reconstruction is indispensable for autonomous surgical interaction. Severe misalignment between semantic meaning and physical anatomy still persists, largely because existing solutions integrate Vision-Language Models into deformable fields via a rigid coupling scheme that tightly binds semantic features to geometric warping. In this paper, we propose DeGenseGS, Geometrically and Semantically Decoupled Surgical Scene Understanding in 4D Gaussian Splatting, a novel framework that independently models semantic evolution and geometric deformation. Specifically, we propose a HexPlane-based spatiotemporal entanglement module that uses shared kinematic latents to synchronize semantic mutations with scene dynamics, while explicitly disentangling semantic updates from geometric deformation. To further ensure robustness against reconstruction artifacts, we devise a Rasterization-Native Semantic Extraction mechanism that infers semantics from topologically continuous feature maps. Additionally, we incorporate an angular-aligned optimization strategy that conforms to the native hyperspherical latent space, thereby preventing semantic distortion. Extensive evaluations on the CholecSeg8k and EndoVis18 datasets demonstrate that DeGenseGS achieves state-of-the-art performance. Our framework yields enhanced geometric completeness and robust semantic-anatomic alignment, enabling spatially continuous segmentation despite drastic tissue deformation and topological transitions.
arXiv:2607.04812v1 Announce Type: new Abstract: Bird's-Eye View (BEV) end-to-end instance prediction has emerged as a robust paradigm for autonomous driving perception, effectively mitigating the error propagation inherent in traditional modular pipelines. However, current state-of-the-art approaches rely predominantly on geometric supervision, such as occupancy regression and optical flow, effectively treating scene agents as generic moving obstacles. This absence of explicit semantic awareness imposes limitations on the capacity of the model to solve ambiguities in complex scenarios, particularly those where object-specific behavior is essential for accurate forecasting (e.g. overtaking, intersections). In this paper, we introduce Text-Guided Representation for Instance Prediction (TGRIP), a novel framework that bridges this gap by injecting rich semantic priors into the instance prediction loop. The proposed teacher-student pipeline employs Vision-Language Foundation Models to generate dense, semantic-enhanced BEV maps from multi-camera images. These maps serve as auxiliary supervision during training, guiding the network to learn spatio-temporal representations that are not only geometrically consistent but also semantically discriminative. To the best of our knowledge, this represents the first attempt to unify semantic guidance with the temporal task of future instance prediction. The experimental results demonstrate that TGRIP surpasses existing state-of-the-art models in nuScenes, validating the hypothesis that semantic enrichment is a fundamental element for robust, end-to-end motion prediction. Code is available on https://github.com/miguelag99/TGRIP.
arXiv:2607.05168v1 Announce Type: new Abstract: Why do intelligent systems need to perform explicit symbolic reasoning? Computer science has traditionally regarded symbolic reasoning as a defining component of intelligence. Yet the remarkable success of modern foundation models raises a fundamental question: if increasingly capable AI systems can operate with little explicit symbolic reasoning, what role do symbolic methods actually play? This article argues that explicit symbolic reasoning is not a fundamental property of intelligence, but a computational consequence of operating on simplified models of reality. We propose the Compression Principle: every computational model is a simplified representation of reality, and explicit symbolic reasoning compensates for information omitted during model construction. From this principle, we derive the Modeling--Reasoning Trade-off: as computational models preserve richer representations of the world, the need for explicit symbolic reasoning correspondingly decreases. This perspective provides a unified explanation for both the historical success of symbolic methods and the remarkable effectiveness of modern foundation models. Paradoxically, the same development makes symbolic methods increasingly important for humans. As intelligent systems become more capable and more opaque, symbolic representations increasingly serve as interfaces through which humans specify requirements, verify behavior, regulate autonomous systems, and establish trust. We therefore argue that the future of symbolic methods lies not primarily as the computational engine of intelligent systems, but as the symbolic interface between increasingly capable AI systems and the humans who build, govern, and depend upon them.
arXiv:2605.23462v2 Announce Type: replace Abstract: Cyclic animation is widely used in computer graphics and interactive content.It supports seamless playback in games, VR, and interactive simulation,where short clips must repeat smoothly over long durations. Achievingphysically plausible cyclic synthesis from an input sequence is challengingbecause the endpoint states of the observed sequence rarely match exactly,and the governing equations of the underlying system are often unavailable.We therefore propose an equation-free framework that identiffes a Koopmansurrogate from the observed trajectory and computes a cyclic trajectory byapplying a Fourier-parameterized, time-varying control force under a hardtemporal periodicity constraint. The resulting formulation reduces cyclicsynthesis to a linearly constrained quadratic program that can be solvedefffciently through a structured KKT system. Our method is applicable toa diverse range of examples, including N-body systems, cloth, deformableobjects, shallow water, etc.
arXiv:2607.04992v1 Announce Type: new Abstract: Nanosecond-pulsed dielectric barrier discharge (DBD) plasma actuators were investigated on a generic blunt body in a Mach 6 Ludwieg tube to characterize the pressure and thermal perturbations relevant to hypersonic boundary-layer transition control. Complementary quiescent experiments were also conducted over ambient pressures representative of those predicted in the model nose region to isolate the influence of local thermodynamic conditions on actuator operation. Pulse-energy measurements and schlieren imaging showed that decreasing pressure reduced the deposited electrical energy per pulse, weakened the actuator-generated shock, and increased the spatial extent of the residual heated region owing to energy deposition over a larger plasma volume. Under Mach 6 Ludwieg-tube conditions, the actuator-generated shock interacted with and reflected from the detached bow shock, temporarily increasing the bow-shock stand-off distance by approximately 11%, while the residual heated region was advected downstream along the body. The schlieren images further permitted the evolution of the thermal disturbance to be distinguished from that of the actuator-generated shock. The results demonstrate two distinct perturbation mechanisms -- a short-duration compression wave and a longer-lived thermal disturbance -- whose relative importance is governed by the local thermodynamic conditions and which may independently promote hypersonic boundary-layer transition.
arXiv:2607.04993v1 Announce Type: new Abstract: Many phenomena of deep learning are dynamical: they concern not only which minima exist, but how gradient descent reaches, avoids, or selects among them. Edge-of-stability behavior, sharpness oscillations, catapult phases, balancing, and movement toward flatter representations are effects of the training map itself, and are poorly captured by the small-step gradient-flow limit. This paper studies fixed-step gradient descent as a discrete dynamical system in a hierarchy of exactly solvable models retaining basic structures of deep learning: depth, factorization, width, data coupling, activation, and stochasticity. The starting point is the balanced scalar reduction of a deep linear chain, giving a quartic loss and a cubic gradient map whose post-edge behavior is explicit. Under the natural large-depth scaling, this dynamics converges to a universal Ricker-type map. The edge of stability is therefore not a breakdown of optimization, but the first bifurcation of the training map. Embedding the scalar dynamics back into factored models turns these regimes into learning phenomena. Finite steps break conservation laws of gradient flow and contract factorization imbalance; residual oscillations move parameters toward flatter, more balanced representations. Wider linear networks produce a ladder of spectral edges, so the optimal learning rate can lie beyond the first edge. Data coupling, nonlinear activations, and stochastic targets preserve the same organizing principle: finite-step oscillations drive alignment, balancing, and representation selection. Thus the learning rate is not merely a numerical stability parameter. It is a structural parameter of the training dynamics, determining its attractors and shaping the representations gradient descent selects.
arXiv:2607.04995v1 Announce Type: new Abstract: Mass lumping enables explicit time stepping for finite-element discretisation of the wave equation if the resulting quadrature weights are positive and accuracy is preserved. New rules for elements of degree two and three on the 4-simplex are presented. Numerical quadrature for the stiffness matrix can be more efficient than exact evaluation if it requires fewer nodes. For the latter, new rules in two and four space dimensions for the lower-degree elements on the simplex were found, as well as some additional results for three dimensions.
arXiv:2607.04996v1 Announce Type: new Abstract: Timely communication of information in Internet of Things (IoT) networks is critical to enhancing system performance and energy efficiency by minimizing the transmission of outdated or redundant data. Although timeliness metrics such as the Age of Information (AoI) effectively quantify information freshness, they do not account for content evolution. The Version Age of Information (VAoI) addresses this gap by tracking version lag at the receiver, thereby providing a practical content-aware metric. However, prior research has primarily focused on first-moment analyses in single-hop settings, leaving the distributional properties of VAoI in multi-hop networks, as well as the impact of feedback mechanisms, unexplored. In this study, we provide a comprehensive characterization of VAoI in multi-hop networks with transmission constraints and acknowledgment-based feedback. A bi-level optimization framework is formulated to jointly optimize the update policy of a rate-constrained source and the feedback-aware forwarding policies of the intermediate nodes, aiming to minimize communication overhead while maintaining VAoI performance at the destination. We show that the optimal source policy follows a threshold-based update strategy and derive the optimal threshold in closed form. For both the optimal threshold policy and a randomized baseline, we obtain closed-form expressions for the stationary distribution and average VAoI, along with the corresponding update rates across network nodes under feedback-aware forwarding. Numerical results corroborate the analytical findings and illustrate the advantages of utilizing VAoI and feedback to reduce redundant transmissions while preserving data freshness and informativeness in multi-hop systems.
arXiv:2607.05004v1 Announce Type: new Abstract: We revisit the online construction of \emph{smallest suffixient sets} and the online computation of the \emph{longest repeating suffix} (LRS). We give the first compressed-space online construction of smallest suffixient sets, and present two space-time trade-offs for both problems: $O(r\log n+n)$ bits of working space and $O(\log^2 n/\log \log n)$ worst-case time per character, and $O(r\log n+n \log \log n)$ bits of working space and $O((\log n/\log \log n)^2)$ worst-case time per character. Here, $r$ is the number of runs in the Burrows-Wheeler transform of the reverse of $T[1..n]$. In particular, for highly repetitive texts satisfying $r=O(n/\log n)$, the first trade-off uses $O(n)$ bits of working space, while the second uses $O(n\log\log n)$ bits. We also prove that any deterministic online algorithm for computing LRS requires \(\Omega(n)\) bits of peak working space in the worst case, even over a constant-size alphabet. Through reductions from online LRS computation, we extend this lower bound to deterministic online algorithms maintaining either an arbitrary smallest suffixient set augmented with the length of the supermaximal right extension represented by each selected position, or the position-only smallest suffixient set obtained by selecting the rightmost occurrence of every such extension. For constructing smallest suffixient sets, our algorithms are the first online solutions using compressed working space, improving the $O(n)$-word space required by previous online constructions. For compressed-space online LRS computation, compared with the algorithm of Prezza and Rosone~[CiE 2020], our bounds improve their $O(\log^2 n)$ amortized time per character by factors of $\Theta(\log\log n)$ and $\Theta((\log\log n)^2)$, respectively, while also providing worst-case guarantees.
arXiv:2607.05026v1 Announce Type: new Abstract: This paper investigates the existence, construction and classification of limited magnitude error-correcting codes, with a focus on splitter sets and their connections to group splittings. We establish new nonexistence results for quasi-perfect splitter sets and provide a complete classification of quasi-perfect $B[0,3](n)$ splitter sets in both singular and nonsingular cases. Furthermore, we derive improved lower bounds for the size of maximal $B[0,3](q)$ sets by investigating Cayley graphs, where $q$ is a prime. We also provide existence criteria for perfect $B[0,6](q)$ splitter sets and quasi-perfect $B[-4,4](2p)$ sets for prime $p$. For perfect burst-correcting codes, we develop a general construction framework, and prove the existence of infinite families of $(k_2,k_1)$-limited-magnitude cyclic $b$-burst-correcting codes for $k_1+k_2\le 4$ and arbitrary burst length $b$. We further provide sufficient existence conditions for general parameters $k_1$ and $k_2$. Our results combine algebraic, combinatorial, and number-theoretic methods to advance the understanding of codes tailored for flash memory and related storage systems.
arXiv:2607.05029v1 Announce Type: new Abstract: Persistent memory has enabled large language model (LLM) agents to store factual knowledge, prior decisions, reasoning histories, tool usage information, and context. While this has improved the agent's functionality and continuity across tasks, it has also introduced a new attack surface: the agent's own reasoning history. In this paper, we introduce the Forged Amplifying Rationale Memory Attack (FARMA), which poisons an agent's remembered reasoning rather than its factual knowledge. It inserts forged reasoning traces using evasive language that bypasses keyword-based defenses, then amplifies them through self-referential reinforcement that defeats consensus-based defenses. To address FARMA, we introduce SENTINEL, a layered defense pipeline to detect forged reasoning entries. Its central component is the Reasoning Guard that structurally analyzes candidate entries for forgery using five weighted signals. We evaluate FARMA and SENTINEL across multiple agents and different LLM models with 50 trials and show that FARMA achieves an attack success rate of up to 100% under baseline conditions and is capable of defeating defense mechanisms like keyword filter and A-MemGuard. Our evaluation also shows that SENTINEL reduces FARMA's attack success rate to as low as 0% with no false positives observed across 326 benign agent traces. Our work demonstrates the need to protect not only an agent's retrieved content but also the integrity of its reasoning history.
arXiv:2607.05030v1 Announce Type: new Abstract: Ensemble Kalman filters (EnKFs) are widely used for data assimilation in geophysical systems. Among various implementations, the local ensemble transform Kalman filter (LETKF) has gained popularity because of its computational efficiency. However, the deterministic EnKF such as the LETKF is known to be less robust than the stochastic EnKF in strongly nonlinear regimes. We generalize the LETKF such that it contains a stochastic term and includes the stochastic EnKF within it. We adaptively optimize the parameter that determines the weight of the stochastic term based on an identity of mutual information, which is satisfied by the Kalman filter in linear Gaussian systems. As the analysis perturbation equations of EnKFs are decomposed into a system of equations for modes that are uncorrelated with each other, the application of mutual information is easily achieved. The generalized LETKF thus optimized is named the mutual information-based ensemble Kalman filter (MI-EnKF). The MI-EnKF indirectly uses the third- and fourth-order moments of the forecast ensemble through entropy. To speed up calculations of entropy, we create a lookup table based on maximum entropy distributions. We conduct data assimilation experiments using the Lorenz-96 model to confirm the validity of the optimization method of MI-EnKF. When the observation operator is linear, the ML-EnKF shows the same analysis accuracy as the LETKF. When the observation operator is strongly nonlinear, the MI-EnKF is more accurate than both LETKF and stochastic EnKF regardless of ensemble size. Optimizing just the first mode can lead to significant improvements, but positive impacts of increasing the number of optimized modes are not observed unless the ensemble size is large. The optimized parameter values indicate that the optimal EnKF lies between the deterministic EnKF and the stochastic EnKF.
arXiv:2607.02576v1 Announce Type: cross Abstract: Parameter-Efficient Fine-Tuning (PEFT) commonly adapts pretrained weights through low-rank updates, and recent methods further exploit the singular value decomposition (SVD) of the base weight for initialization or subspace selection. However, these methods do not explicitly preserve the coupled geometry between the pretrained left and right singular bases. Motivated by recent minimum-perturbation theory, which shows that stable finetuning follows a coherent SVD rotation in which a single orthogonal $Q$ acts on both the left singular basis $U_0$ and the right singular basis $V_0$, we prove a per-slice analogue: each row slice of $W_0$ can be adapted by a shared orthogonal rotation $Q_i$ on its left basis $U_i$ and right basis $V_i$ together with a diagonal spectrum shift. We implement this form as CORA (Coherent Orthogonal Rotation Adaptation), which applies per-slice orthogonal rotations and a per-layer diagonal scale to the rank-$r$ SVD truncation of $W_0$. CORA uses $\tfrac{1}{2}m(r{-}1)$ trainable parameters per linear layer, about $4{\times}$ fewer than LoRA at the same rank. CORA outperforms LoRA, DoRA, PiSSA, and MiLoRA on commonsense reasoning and code generation while using about $8{\times}$ fewer parameters.
arXiv:2607.02622v1 Announce Type: cross Abstract: Multiplex CRISPR-Cas9 gene editing requires selecting one guide RNA per target gene subject to cross-gene interactions: a constrained combinatorial problem that can be formulated as a Quadratic Unconstrained Binary Optimization (QUBO) and solved via the Quantum Approximate Optimization Algorithm (QAOA). The one-hot per-gene constraint is conventionally enforced by adding quadratic penalty terms to the cost Hamiltonian, but penalty coefficient selection is heuristic and penalties amplify hardware noise. An alternative is to enforce the constraint structurally via the XY-mixer, which preserves feasibility by construction. We present COMET, a systematic comparison of penalty-based and XY-mixer QAOA on a three-gene, twelve-qubit multiplex editing instance targeting the immune-checkpoint genes PDCD1, LAG3, and HAVCR2. In simulation, the XY-mixer exceeds 95% probability of the optimum by QAOA depth p=3, while three penalty variants spanning an order of magnitude in penalty coefficient remain below 6% at every depth. On IBM's ibm_kingston (Heron r2) processor, the XY-mixer's simulator-hardware energy gap stays within |0.8| across all depths, while the worst-tuned penalty variant's gap reaches +53.9. We provide an honest account of where the structural guarantee partially breaks under gate-level noise. The twelve-qubit instance is classically trivial; our contribution is a methodological comparison of constraint-enforcement strategies in a biologically motivated domain, with real-hardware validation.
arXiv:2607.05101v1 Announce Type: new Abstract: The application of machine learning-based predictive algorithms to Anti-Money Laundering (AML) has grown rapidly, driven by the vast volume of financial transaction data available to banks. These algorithms are typically trained not only on transactional data but also on sensitive client information, which may raise fairness concerns. Despite this, AML detection systems remain largely underexplored from a fairness perspective, even though deeper analytical methods based on counterfactuals are now available. Such techniques enable the decomposition of the direct and indirect effects of potentially sensitive features on model predictions, thereby supporting the evaluation of whether their influence is acceptable from a fairness perspective. Closing this gap, we consider the synthetic IBM AMLSim transaction dataset and construct additional features of the country of an account and its average behaviour. This improves the predictive performance of diverse machine learning models, ranging from baseline decision trees to state-of-the-art graph neural networks. We assess the potential unfairness associated with these features through a counterfactual, path-specific effect analysis. This reveals that fairness violations tend to be more pronounced for models whose predictive performance benefits the most from the extended features. Such a finding highlights a concrete instance of the trade-off between predictive accuracy and fairness in AML applications, thus underscoring the urgency of a systematic fairness analysis in such critical domains.
arXiv:2506.23281v3 Announce Type: replace Abstract: Random testing has proven to be an effective technique for compiler validation. However, the debugging of bugs identified through random testing presents a significant challenge due to the frequent occurrence of duplicate test programs that expose identical compiler bugs. The process to identify duplicates is a practical research problem known as bug deduplication. Prior methodologies for compiler bug deduplication primarily rely on program analysis to extract bug-related features for duplicate identification, which can result in substantial computational overhead and limited generalizability. This paper investigates the feasibility of employing bisection, a standard debugging procedure largely overlooked in prior research on compiler bug deduplication, for this purpose. Our study demonstrates that the utilization of bisection to locate failure-inducing commits provides a valuable criterion for deduplication, albeit one that requires supplementary techniques for more accurate identification. Building on these results, we introduce BugLens, a novel deduplication method that primarily uses bisection, enhanced by the identification of bug-triggering optimizations to minimize false negatives. Empirical evaluations conducted on five real-world datasets demonstrate that BugLens significantly outperforms the state-of-the-art analysis-based methodologies Tamer and D3 by saving an average of 33.56% and 10.68% human effort to identify the same number of distinct bugs. Given the inherent simplicity and generalizability of bisection, it presents a highly practical solution for compiler bug deduplication in real-world applications.
arXiv:2607.02537v1 Announce Type: cross Abstract: Jamming and spoofing pose significant threats to wireless and satellite navigation by disrupting radio-frequency (RF) signals and compromising availability and integrity. Robust RF interference direction finding through angle-of-arrival (AoA) estimation is therefore essential for detecting and localizing anomalous signals. Although data-driven methods perform well under line-of-sight (LoS) conditions, their performance degrades in practical environments due to non-line-of-sight (NLoS) multipath propagation. In this work, we propose a hybrid learning framework that incorporates physics-informed constraints into deep neural networks to improve the robustness of AoA estimation. A neural network is trained to estimate the azimuth and elevation of incoming signals received by a four-element antenna array, while a physics-informed loss enforces consistency between the predicted angles and inter-antenna phase differences under a plane-wave model. We further introduce a latent-space classifier to distinguish LoS from NLoS samples. Since inter-antenna phase differences under LoS propagation exhibit domain-invariant structure across environments, the physics-based loss is applied only to LoS samples, promoting physically consistent and domain-invariant representations without over-constraining the model in NLoS scenarios. In addition, domain-incremental learning (DIL) across NLoS environments with varying scatterer distributions improves cross-domain generalization. Evaluations on real-world datasets show that the proposed method reduces AoA estimation error by up to 6{\deg} in low-exemplar settings compared with DIL baselines.
arXiv:2607.05120v1 Announce Type: new Abstract: AI agents act on behalf of user prompts, consuming external data and taking actions based on the agent context. Prior research on AI agent security has primarily focused on indirect prompt injection (IPI). Its most well-studied category is instruction injection, where attacker-controlled untrusted data is interpreted as an instruction. In response, many mitigations have been proposed to prevent instruction injection attacks. In this paper, we introduce a new category of IPI, agent data injection attacks (ADI). ADI injects malicious data disguised as trusted data, such as security-critical metadata (e.g., resource identifiers or data origins) or agent context data (e.g., tool call and response formats). As a result, agents unknowingly execute unintended actions based on attacker-controlled data. ADI has similar attack impacts as instruction injection attacks, because it causes agents to misbehave and execute unintended actions. Despite the similar impact, ADI remains underexplored and easily bypasses existing IPI defenses. We found several critical vulnerabilities in real-world agents that allow an attacker to launch various attacks: arbitrary click attacks on web agents (Claude in Chrome, Antigravity, and Nanobrowser), and remote code execution and supply-chain attacks on coding agents (Claude Code, Codex, and Gemini CLI). We evaluate ADI vulnerabilities across off-the-shelf models and AI agents, and find that ADI is effective in both standalone LLMs and AI agent settings. ADI exposes a critical gap in agent security, signifying that current AI agents do not employ a fundamental security principle: current agents do not isolate trusted data from untrusted data.
From Failing to Passing: Evolving Natural Language Prompt Optimization Rules for LLM Code Generation
arXiv:2607.05121v1 Announce Type: new Abstract: Large language models are known to be sensitive to prompt formulation. Even minor variations in wording can substantially degrade performance. This sensitivity reveals an opportunity: if prompt phrasing can harm performance, can it be used to improve it? To investigate this question, we introduce a search-based approach that identifies and evolves a set of natural language transformation rules with strong downstream effects on coding performance. We then propose DUALFIX, a staged repair pipeline that combines the evolved transformation rules with execution-feedback repair, addressing both specification-level and implementation-level failures. A key strength of our approach lies in its generality: the evolved rules are error-agnostic, reusable across problems, and transferable across models. We evaluate DUALFIX against execution-feedback repair baselines across three models on two challenging benchmarks, LiveCodeBench and APPS. Our results show that the evolved transformations fix from 10-30% of failing cases, including 12-17% of failures that execution-based repair alone cannot resolve. Overall, DualFix recovers up to 30% of baseline failures and fixes 3-5 times more failing cases than Self-Fix across all evaluated settings. Furthermore, we also show that rules evolved on one model transfer zero-shot to other models, outperforming execution-feedback repair without any re-optimization.
arXiv:2607.05122v1 Announce Type: new Abstract: Vision-language-action (VLA) models enable robot navigation from natural language and visual goals, but remain susceptible to perceptual distractions and ambiguous scene interpretations. This paper presents the first empirical evaluation of visual grounding for VLA navigation policies. We propose a real-time segmentation-based grounding method that highlights traversable areas in green and non-traversable areas in red using SegFormer. Two variants are evaluated: observation-only segmentation and joint observation-goal augmentation. Using OmniVLA on the Grand Tour dataset, we show that visual grounding reduces the mean waypoint error by 27-44% at the farthest waypoint, depending on the instruction length. The benefits are greater for long instructions than for short instructions, and grounding provides little improvement for image goals. Normalized error analysis indicates that grounding primarily acts as a trajectory length regularizer, reducing the predicted path length by 30% without improving per-unit-distance reasoning. Our results indicate that visual grounding offers a simple, computationally inexpensive method to improve VLA navigation without model retraining, although it cannot compensate for missing training signals in out-of-distribution instructions.
arXiv:2607.05126v1 Announce Type: new Abstract: We investigate quantifier-free interpolation properties for several fragments generalising the extensional theory of arrays. Our results include the (general) quantifier-free interpolation properties of combinatory array logic with iterated diffs and the uniform interpolation property of the simple flat array fragment. To our knowledge, these are the first positive quantifier-free interpolation results obtained for theories of arrays featuring expressive specifications over unbounded domains.
arXiv:2512.04073v2 Announce Type: replace Abstract: Stop-and-go waves in traffic flow are captivating collective phenomena with important safety and environmental consequences. While classical theories attribute these oscillations to linear instabilities caused by reaction delays and inertia, this study explores an alternative stochastic perspective. Using a linearly stable car-following model, we show that white Gaussian noise in the measurement of the inter-vehicle distance can destabilise the flow, inducing a phase transition to periodic stop-and-go dynamics via a nonlinear instability mechanism. Furthermore, we demonstrate that a simple linear transformation of the model, which amplifies the system response while introducing a positive acceleration bias, can counteract noise-induced effects and restores the stability of uniform traffic flow. These findings, supported by numerical simulations, aim to provide new insights into the modelling and control of oscillatory traffic dynamics.
arXiv:2607.05129v1 Announce Type: new Abstract: Two-photon fluorescence imaging achieves deep-tissue penetration through long excitation wavelengths and nonlinear excitation confinement. The 1700 nm transparency window is particularly attractive, as it optimally balances tissue scattering and absorption. However, efficient fluorophores for two-photon excitation in this window remain limited. Moreover the weak near-infrared emission of individual emitters, and the low photon detection efficiency, has so far precluded single-particle imaging. Here, we characterize the two-photon excitation properties of chirality-sorted pristine and quantum color center-functionalized single-walled carbon nanotubes under 1700 nm excitation. By measuring and comparing their two-photon action cross-sections, we identify quantum color center-functionalized (6,5) nanotubes emitting at 1140 nm, as the most promising emitter, with an exceptionally large cross-section of (57 \pm 2).103 GM. Leveraging these favorable photophysical properties, we image individual nanotubes under 1700 nm excitation, to our knowledge the first demonstration of single-particle imaging at this wavelength. These results establish quantum color center-functionalized (6,5) nanotube as a strong candidate for long-wavelength two-photon imaging and lay the groundwork for deep-tissue single-particle imaging.