arXiv:2607.09753v1 Announce Type: new
Abstract: Diffusion models have achieved remarkable success across diverse domains, with performance closely related to the denoising backbones that parameterize the score function. In this paper, we present a systematic, phase-aware analysis of diffusion components and show that abrupt, early-stage fluctuations in deep latents are strongly associated with artifacts. Guided by these findings, we introduce DUNE (Diffusion Unified Network refiNEr), a training-free refinement framework that detects abrupt deviations in deep low-noise internal latents using a shared EMA-based criterion, and applies backbone-specific suppression to the detector-selected entries. Although derived from U-Net, the same detect-suppress principle extends naturally to Transformer-based diffusion models by acting on the latents of deep self-attention blocks. Extensive experiments across multiple backbones indicate that DUNE improves fidelity while reducing hallucinations, offering new insight into where and when diffusion backbones should be controlled.
Science Journals
arXiv:2607.11342v1 Announce Type: new
Abstract: Despite their central role in fault detection, test oracles remain challenging to construct effectively. Recent learning based methods address this challenge by automatically generating test assertions, yet even if syntactically correct, they are often ineffective in revealing bugs. Rather than generating assertions, this study explores a different approach by training a model to directly predict whether a given test prefix passes or fails. We present FOCAL, an emerging code LLM-based discriminative oracle predictor. It learns from labeled pairs of test prefixes and methods under test, employs losses that emphasize failing cases during training, and grounds its predictions in statement level behavioral evidence. Compared with the baseline method SEER, we substantially improve performance on failing cases for unseen projects and provide richer explanations. A preliminary evaluation on fault-detection benchmarks and automated test-generation artifacts shows that our approach is highly accurate within its training distribution and substantially improves failure detection on previously unseen projects where prior discriminative oracles collapse. Moreover, the highlighted statements are supported by behavioral explanation checks. These early results suggest that fail-aware discriminative oracle prediction can complement existing approaches such as fuzzing, search-based testing, and LLM-based test generation. These techniques produce test prefixes at scale but often lack fault oriented oracles. In future work, FOCAL could take generated test prefixes and attach fault-aware predicted oracles to them, turning high-volume input generation into executable tests that are more likely to expose semantic failures.
arXiv:2607.10765v1 Announce Type: cross
Abstract: The Lee-Yang theorem and its quantum extensions state that, for a broad class of Hamiltonians on any graph, the partition function's zeros in the complex magnetic field plane lie only on the imaginary axis. For these Hamiltonians, we prove that under a uniform Z-field of any strength h, the ground state has a spectral gap of at least h/4, independent of the system size and of the coupling strengths. The proof uses the zero-freeness of the partition function as given by Asano and Suzuki-Fisher to show exponential decay of the imaginary-time correlations for any product of Z-operators. Our result gives a polynomial-time quantum algorithm for computing the ground state energy of any Lee-Yang Hamiltonian.
arXiv:2607.10865v1 Announce Type: cross
Abstract: We develop a two-axis interpolation framework for the O$(n)$ universality family, treating the spatial dimension $D$ and the spin-component number $n$ as independent continuous parameters connecting exact limiting solutions. On the spatial axis, anchoring between the Onsager solution at $D=2$ and mean-field theory at $D\to\infty$ yields a closed-form prediction for the 3D Ising critical coupling that agrees well with Monte Carlo benchmarks $K_c = 0.2204$ (benchmark: $0.22165$) with no adjustable parameters. Wilson--Fisher-constrained polynomial interpolation gives $\nu=2/3$, $\beta=31/96$, and $\eta=35/864$ at $D=3$ (benchmarks: $0.6299$, $0.3265$, $0.0362$), and reproduces conformal-bootstrap results across $3 \le D < 4$. On the spin axis, we establish a necessary compatibility criterion: two-anchor interpolation succeeds only for observables that vary monotonically between the anchor values. The critical coupling $K_c(n)$ violates this criterion because the Heisenberg value falls below the spherical limit, whereas the correlation-length exponent $\nu(n)$ satisfies it. A perturbative $1/n^2$ expansion yields $\nu(3) = 0.7493$ (benchmark: $0.7112$), and propagation through exact scaling relations gives $\beta(3) = 0.3797$ (benchmark: $0.3689$) and $\gamma(3) = 1.489$ (benchmark: $1.396$), without introducing additional parameters. The framework naturally extends to non-integer spin, producing the prediction $\nu(2.5) = 0.7143$ for the O$(2.5)$ universality class. These results establish dimensional and spin interpolation as a unified and predictive approach to critical phenomena, while clarifying the structural conditions under which interpolation succeeds.
arXiv:2607.09828v1 Announce Type: cross
Abstract: The differentiable shift-variant filtered backprojection (SV-FBP) framework enables data-driven estimation of redundancy weights for cone-beam CT reconstruction under general source trajectories, removing the need for analytically derived weighting schemes. In this work, we present a systematic study of the robustness and adaptability of differentiable SV-FBP under challenging acquisition settings. We show that the framework remains stable across highly irregular and discontinuous trajectories, indicating that reconstruction performance is largely insensitive to trajectory ordering or continuity. Instead, the spatial distribution of sampling points plays a more dominant role. Under sparse-view conditions, differentiable SV-FBP achieves competitive reconstruction quality while providing an order-of-magnitude reduction in computation time compared to iterative reconstruction methods at moderate sampling densities. However, we identify a clear transition regime under severe undersampling, where the absence of iterative data consistency leads to performance degradation. Furthermore, we demonstrate that the framework remains applicable to non-planar multi-isocenter geometries, such as Lissajous-saddle trajectories, without requiring architectural modifications. These findings provide new insights into the behavior and limitations of the differentiable SV-FBP model and highlight it as a flexible and efficient solution for non-standard and robotic CBCT acquisition scenarios.
arXiv:2607.11353v1 Announce Type: new
Abstract: The creation of digital collections involves not only the digitisation of content, but also the creation of catalogue records for it. This often-overlooked task requires slow and costly expert manual work. In this project, we have evaluated the application of AI models to this task, comparing different implementations and models. This work includes a qualitative and quantitative evaluation of the experiments carried out, as well as recommendations on the use of AI models that go beyond the specific use case.
arXiv:2607.09976v1 Announce Type: cross
Abstract: Hybrid programs combine a quantum circuit with a classical host program that consumes measurement outcomes. In such programs, an outcome may be syntactically read by the host but semantically non-contributory: changing the outcome cannot change the returned value. Such outcomes obscure gates that are dead only relative to the host semantics, and are therefore invisible to circuit-local optimizers.
We present a semantics-aware host-side static analysis that identifies non-contributory measurement outcomes by abstract interpretation, and prove its soundness. We implement the analysis and evaluate it on $24$ application-faithful hybrid workloads across quantum chemistry, optimization, quantum machine learning, and quantum finance. Compared with a syntactic liveness baseline, our analysis identifies more than $4\times$ as many non-contributory measurements, and it standalone enables the removal of $37.98\%$ of total gates on average. Even after the state-of-the-art optimizers like Qiskit, t|ket$\rangle$, and PyZX have already optimized the circuits, our analysis still enables removal of more than $30\%$ of the post-optimized gates, showing that the host-semantic opportunities exposed by our analysis are not subsumed by circuit-local optimization. To scale our analysis, we further lower host programs to an SSA-style levelized intermediate representation that exposes level-wise parallelism for GPU execution, and implement a CUDA backend. We prove that this lowering preserves the analysis result, and the evaluation shows speedups of up to $6.53\times$ over a sequential baseline as structural parallelism increases.
arXiv:2607.09963v1 Announce Type: new
Abstract: We prove a near-maximum ($2^n / n$) circuit lower bound for the complexity class $\mathsf{E}^{\mathrm{pr}\mathsf{MA}}/_1$, corresponding to exponential time with access to a promise-$\mathsf{MA}$ oracle and one bit of advice. Our proof incorporates the iterative win-win paradigm (Chen--Lu--Oliveira--Ren--Santhanam, FOCS'23), the reduction from the Range Avoidance problem to circuit lower bounds (Je\v{r}\'abek, Ann. Pure Appl. Log. '04; Korten, FOCS'21), and the PCP theorem. Crucial to our proof is the analysis of the complexity class $\mathsf{P}^\mathsf{NP}[{\textsf{#rounds}}=r, {\textsf{length}}=s]$, which is $\mathsf{P}^\mathsf{NP}$ with $r(n)$ adaptive rounds of $\mathsf{NP}$ queries, where each $\mathsf{NP}$ query has witness length $s(n)$.
On exponential convergence of Chebyshev polynomial approximation for multivariate analytic functions
arXiv:2607.10209v1 Announce Type: new
Abstract: This paper presents a new analysis of the Chebyshev projection for multivariate analytic functions, drawing on pluripotential theory. It is proved that in any downward closed convex polynomial space, the Chebyshev projection achieves the same exponential convergence rate as the best polynomial approximation. This result enables a precise quantification of the exponential convergence rate of the Chebyshev projection. The analysis is then extended to several related topics, including tensorized Chebyshev interpolation, tensor product Gauss--Legendre quadrature, Padua interpolation and cubature, and Chebyshev-Galerkin method, with the corresponding exponential convergence rate established in each case. Supporting numerical experiments are provided to validate the theoretical results.
arXiv:2607.10566v1 Announce Type: new
Abstract: Objective. Existing quadratic unconstrained binary optimization (QUBO)-based sparse-view computed tomography (CT) reconstruction neglects photon-counting statistics and anatomical heterogeneity. We address both limitations within the QUBO framework.Approach. We propose a quantum compressed-sensing CT method combining penalized weighted least squares (PWLS) and guided total variation (GTV). PWLS weights projection residuals by photon-count reliability, whereas GTV uses gradients from a prior image reconstructed by the simultaneous algebraic reconstruction technique (SART) to preserve edges and suppress noise in homogeneous regions. After binary encoding, both terms form a unified QUBO model. Experiments used four 40 times 40 CT images under a 10-view fan-beam geometry with Poisson noise. Comparisons included conventional reconstruction methods, QUBO variants, gradient descent, simulated annealing, and a D-Wave hybrid quantum-classical solver.Main results. PWLS-GTV achieved the best reconstruction quality across all cases. In the representative chest case, it reached a peak signal-to-noise ratio (PSNR) of 36.64 dB, compared with 22.48 dB for SART, the best conventional baseline. GTV consistently outperformed conventional total variation. Simulated annealing and the D-Wave hybrid solver produced similar reconstructions, whereas gradient descent was ineffective. Repeated hybrid-solver runs showed stable performance.Significance. The framework incorporates photon-statistical weighting and structure-guided regularization into QUBO-based CT reconstruction without changing its quadratic form, providing a proof of concept for quantum-assisted sparse-view CT reconstruction.
arXiv:2511.09331v3 Announce Type: replace
Abstract: Decentralized collision avoidance is a core challenge for scalable multi-robot systems. A promising approach to this problem is Model Predictive Path Integral (MPPI) control - a framework that naturally handles arbitrary motion models and provides strong theoretical guarantees. Still, in practice an MPPI-based controller may produce suboptimal trajectories because its performance relies heavily on uninformed random sampling. We introduce CoRL-MPPI, a fusion of Cooperative Reinforcement Learning and MPPI that addresses this limitation. We train an action policy, approximated by a deep neural network, in simulation to learn local cooperative collision-avoidance behaviors. This learned policy is then embedded into the MPPI framework to guide its sampling distribution, biasing it toward more intelligent and cooperative actions in scenarios that may differ substantially from those used during training. Moreover, CoRL-MPPI preserves the theoretical guarantees of regular MPPI. We evaluate our approach in dense, dynamic setups against classical and learning-based state-of-the-art baselines. Our results demonstrate that CoRL-MPPI outperforms competing methods and significantly improves navigation efficiency, measured by success rate and delay, as well as safety, enabling agile and robust multi-robot navigation.
arXiv:2511.18850v4 Announce Type: replace
Abstract: Discovering effective predictive signals, or "alphas," from financial data with high dimensionality and extremely low signal-to-noise ratio remains a difficult open problem. Despite progress in deep learning, genetic programming, and, more recently, large language model (LLM)-based factor generation, existing approaches still explore only a narrow region of the vast alpha search space. Neural models tend to produce opaque and fragile patterns, while symbolic or formula-based methods often yield redundant or economically ungrounded expressions that generalize poorly. Although different in form, these paradigms share a key limitation: none can conduct broad, structured, and human-like exploration that balances logical consistency with creative leaps. To address this gap, we introduce the Cognitive Alpha Mining Framework (CogAlpha), which combines code-level alpha representation with LLM-driven reasoning and evolutionary search. Treating LLMs as adaptive cognitive agents, our framework iteratively refines, mutates, and recombines alpha candidates through multi-stage prompts and financial feedback. This synergistic design enables deeper thinking, richer structural diversity, and economically interpretable alpha discovery, while greatly expanding the effective search space. Experiments on 5 stock datasets from 3 stock markets demonstrate that CogAlpha consistently discovers alphas with superior predictive accuracy, robustness, and generalization over existing methods. Our results highlight the promise of aligning evolutionary optimization with LLM-based reasoning for automated and explainable alpha discovery.
arXiv:2606.25715v2 Announce Type: replace
Abstract: We extend the cone-ray exact-NMF pipeline of Ramteke (arXiv:2606.22451) from the uniform-support regime r_+ = r to the gap regime r_+ > r, and classify recoverable nonnegative factorisations by the rank of the W-factor into a three-regime taxonomy. Regime A (rank(W) = r_+, full column rank): a two-sided SVD-gauge cone-ray pipeline W = U_{r+}(G) Q, H = P V_{r+}(K)^T with G, K on Stiefel manifolds and square consistency Q P = diag(S_r, 0). On 10x10 dense random gap matrices it gives 100/100 recovery at r_+ = 5 and 6. We explain this via two geometric facts: slack enclosure (the data cone has codimension r_+ - r in the outer cone) and NRF-variety thickness (valid gauges form a positive-measure set, so the blind SVD lands on one with probability one). Regime B (rank(W) = r, W a column subset of M): a rank-deficient-W branch enumerating r_+-subsets of M's columns with per-column LP tests. On the block-diagonal family diag(C, J_k), where additivity of nonnegative rank collapses the valid gauges to a single point and the blind SVD pipeline fails, the column-subset branch restores recovery in milliseconds. Regime C (r < rank(W) < r_+, W not a column subset): exposed by the regular octagon's slack matrix. An exact size-6 NRF exists and is reachable by the symmetric formulation at an oracle gauge derived from a known factorisation (residual 1.5e-10), but the blind problem is open: 50 Haar random restarts and Riemannian gradient descent on the Stiefel/Grassmann gauge all stall, because the alt-LP residual is piecewise constant on cells of gauge-space, so local descent cannot cross cell walls. A combined toolkit (Regime B then A) covers regimes A and B with no regression on dense draws; Regime C remains open, with the regular octagon as the cleanest unsolved test case.
arXiv:2607.06186v2 Announce Type: replace
Abstract: Most quadruped loco-manipulation designs trade manipulation capability against stance. A trunk-mounted arm sits high and usually carries a single arm; using the legs as manipulators lifts the manipulating leg off the ground; and even leg-mounted grippers reach two-handed tasks only by rearing onto the hind legs. This paper integrates a manipulator with a prismatic slider, two revolute joints, and a gripper into each front calf of a Unitree Go2. The two arms grasp objects at ground level and manipulate with both hands while all four feet stay planted, without rearing. With one arm carrying, the base stays free to walk. A vision-language model sequences skills from a predefined library at each skill boundary, conditioned on the head-camera image and task state, for long-horizon autonomy. In simulation, the design performs three bimanual tasks: a long-horizon cabinet task under autonomous skill selection, a cooperative two-handed lift, and an inter-arm handover.
arXiv:2607.10252v1 Announce Type: new
Abstract: Large language models (LLMs) are increasingly consumed through opaque serving chains - API aggregators, resellers, and inference providers - in which the client has no technical means to confirm that the model answering is the model advertised, and recent audits show that a substantial fraction of commercial endpoints deviate from the vendor's reference weights. Existing identification techniques require long generated texts, token-level log-probabilities, adversarially crafted prompts, or the model owner's cooperation. We show that far weaker evidence suffices. We define a behavioral fingerprint of an LLM as the empirical distribution of its answers to trivial one-word prompts - "name a random number between 1 and 100" - collected across four languages at a cost of one output token per query. Measuring 165 models served via a large commercial aggregator (OpenRouter), we find that (i) these distributions are highly non-uniform (median cell entropy 1.0 bit) and model-specific: split halves of the same model's samples lie an order of magnitude closer than samples of different models; (ii) Jensen-Shannon divergence between fingerprints recovers model lineage, assigning a model to its documented family with 59.5% leave-one-out accuracy against an 18.4% chance rate; and (iii) a biometric-style verification protocol achieves a 7.3% equal error rate with the full 40-cell battery, and below 11% with eight probe cells - roughly a hundred single-token queries per audit. We further report ecosystem anomalies, including a proprietary-branded flagship endpoint distributionally indistinguishable from an open-weight Qwen model. The protocol, prompts, raw data, and analysis code are released for reproduction and operational use.
arXiv:2607.10389v1 Announce Type: new
Abstract: A persistent interactive world model keeps its running state resident on the GPU that serves it: a multi-gigabyte attention cache, almost all of it rewritten at every generation step. That state cannot be recomputed in interactive time or approximated without changing the world, so a live session pins its device. The pin is a scheduling problem. WorldMove moves a live session under one guarantee: the destination is bit-identical to the source, or nothing is installed. It relocates the cache in 18.8 ms same-node, 101x faster than save/load. It holds a checksum-verified 92.1-94.8 Gb/s on a 100 Gb fabric. At that rate the cache fits inside one interactive block. Migrating an actively generating session, it converges at a block boundary and the destination continues the world bit for bit. An admissibility condition decides each move. The move must complete inside the readout horizon, over bandwidth that covers the state plus its dirty rate. Lifted to a fleet schedulability test, it governed a consolidation loop that executed 48 of 48 migrations bit-identical across two providers. Two constraints are structural. Bit-exactness survives only inside a controlled configuration of one GPU architecture, so moving the state is the only way to preserve it exactly in interactive time. Verification cannot hide inside the wire on this fabric. Receive-path checksums stall the transport at protocol timescales under fan-in, and unscheduled incast silently collapses a receiver while every delivered byte stays correct. An incast-aware admission controller holds zero misses to 1.4x offered load and sheds overload as rejects. A lossless GPU codec widens the admission gate to fabrics raw motion cannot use. We exercise the serving loop and the mover separately, each end to end. Their composition on one fabric is unbuilt. Exact-state elasticity is a joint scheduling problem over transport and verification.
arXiv:2607.10262v1 Announce Type: cross
Abstract: Hybrid Schrodinger-Feynman (HSF) simulation offers an attractive memory-path tradeoff for exact quantum-circuit emulation, but its practical runtime is often dominated by exponential path growth from cross-boundary two-qubit gates. Existing GPU and FPGA quantum simulators are largely optimized for full-state Schrodinger execution and therefore do not align well with HSF's path-centric workflow. This paper presents HSF-S, a compiler-accelerator co-designed framework for exact HSF-based quantum circuit emulation. HSF-S lowers input circuits to an HSF-compatible basis, formulates a rank-aware effective path-cost model, and applies dependency-preserving reordering together with discounted-gain SWAP insertion to suppress recurring cross-boundary interactions while preserving exact circuit semantics. A regression-free selector guarantees that the compiled circuit never increases effective path cost relative to the naive lowered baseline. We further design a dedicated HSF-S accelerator and execution flow, and integrate them into a stand-alone processor for efficient per-path dual-slice evaluation and final accumulation without materializing the full state vector. Across 56 benchmark circuits, HSF-S matches reference amplitudes to within floating-point precision, reduces effective path cost by up to 90.0%, and substantially improves practical tractability, including representative timeout-to-sub-second reductions under a 1-hour budget. On the resulting compiled workloads, the HSF-S processor prototype delivers up to 4.34x additional speedup.
arXiv:2606.07804v2 Announce Type: replace
Abstract: Smart cities rely on Internet of Vehicles (IoV) networks for critical services. However, this vast connectivity enlarges the attack surface, exposing vehicular systems to evolving cyber threats. Conventional static defenses struggle to autonomously adapt to these dynamic, multi-stage intrusions. To address this, we propose the Communication Aware Quantum Inspired Reinforcement Learning (CA-QIRL) framework, built on a lightweight deep Q-Network architecture for autonomous cyber defense.
V2X defense is formulated as a communication-aware Markov Decision Process (MDP). The agent observes intrusion, mobility, Road Side Unit (RSU), and communication metrics to select optimal mitigation actions. CA-QIRL integrates quantum-inspired encoding, rotation exploration, and an interference reward, combined with a cost function penalizing false negatives, false positives, delay, packet loss, and RSU overload.
Experimental evaluations on vehicular intrusion datasets and a mobility-aware V2X simulation demonstrate robust performance. CA-QIRL achieves competitive detection accuracies of 97.89% on CICIDS2017 and 80.31% on CAN-MIRGU, outperforming state-of-the-art ensemble methods in inference latency. Furthermore, end-to-end delay and Channel Busy Ratio (CBR) drop by up to 95.7% and 90%. Statistical significance is confirmed on ROAD and VeReMi. These findings establish CA-QIRL as a highly practical and resilient defense mechanism for next-generation V2X and IoV networks.
arXiv:2606.07820v2 Announce Type: replace
Abstract: We describe new dependent-rounding algorithms for bipartite graphs. Given a fractional matching $x$ of graph $G = (U \cup V, E)$, the algorithms return an integral solution $X$ such that each right-node $v \in V$ has at most one edge, and where the variables $X_e$ also satisfy broad non-positive correlation properties. In particular, for any edges $e_1, e_2$ sharing a left-node $u \in U$, the variables $X_{e_1}, X_{e_2}$ have \emph{strong} negative-correlation, i.e. the expectation of $X_{e_1} X_{e_2}$ is significantly below $x_{e_1} x_{e_2}$.
Dependent rounding schemes with these properties have been used for a approximation algorithms for job-scheduling on unrelated machines to minimize weighted completion times, among other applications. Our new algorithm achieves simpler and qualitatively stronger bounds compared to prior algorithms. In particular, we achieve a negative-correlation property $$ \E[X_{e_1} X_{e_2}] \leq 0.79751 \ x_{e_1} x_{e_2}, $$ which is a significant constant-factor improvement over Baveja, Qu & Srinivasan (2023).
arXiv:2607.09827v1 Announce Type: new
Abstract: Visual saliency aims to predict the regions of an image most likely to attract human visual attention. While most saliency models assume free-viewing conditions, human attention is often shaped by explicit task goals. In this work, we address task-driven saliency prediction by proposing a model that conditions visual attention on natural-language task descriptions. The model produces task-dependent saliency maps that reflect how attention shifts under different viewing intents. Through quantitative and qualitative analysis, we show that incorporating explicit task semantics enables more faithful modeling of goal-directed visual attention.
arXiv:2607.11498v1 Announce Type: new
Abstract: Vision-language-action (VLA) models predict robot actions from visual observations and language instructions. These actions are defined in the robot's own 3D coordinate frame, yet most VLAs observe the scene in the camera frame, creating a frame mismatch between where the scene is observed and where actions are defined. The mismatch is benign under a fixed viewpoint, where the policy can memorize a single observation-to-action mapping, but grows harder as large-scale datasets aggregate demonstrations across diverse camera setups and the policy must generalize this mapping across viewpoints. We address this mismatch with robot-centric pointmaps, images whose pixels store the 3D coordinates of scene points in the robot frame. Pointmaps provide robot-frame 3D geometry while preserving the dense H x W grid expected by pretrained 2D VLAs, so they integrate into existing VLAs with minimal architectural change. On RoboCasa, pointmaps improve both pi0.5 and SmolVLA and outperform representative camera-viewpoint and 3D-aware baselines. In real-robot experiments, their advantage over an RGB-only policy widens when the camera is moved to a placement unseen during training.
arXiv:2607.10277v1 Announce Type: new
Abstract: The oracle problem (determining the correct expected outcome for a test) remains a major bottleneck in automated testing, and is increasingly relevant as non-experts rely on AI-generated code they cannot reliably validate. We study whether large language models (LLMs) can generate generalizable test oracles directly from natural-language business requirements, without access to source code or example input-output pairs. We propose a reproducible, requirement-driven pipeline grounded in Defects4J. For each of 10 real bugs from Defects4J Lang (Bugs 1 and 3-11), we (i) extract behavioral changes via buggy/fixed diffs, (ii) manually translate the change into a business requirement, (iii) construct a requirement-derived oracle (REQ) as a gold standard, and (iv) prompt five LLMs (DeepSeek-V3, Gemma-3n, Llama-3, Mistral-7B, and Qwen-3) to generate Java oracle code. We evaluate oracle correctness and generalization under two targets: agreement with REQ and agreement with the system under test (SUT), reporting macro-averaged accuracy, precision, recall, and F1. LLMs achieve non-trivial generalization but with substantial bug- and model-level variance. Generated oracles align more closely with REQ than with SUT, and correlations between requirement technicality/ambiguity ratings and oracle accuracy are weak with wide confidence intervals. No detectable linear relationship exists between requirement properties and oracle accuracy in this dataset, suggesting that pretraining coverage and the semantic specificity of the required behavior dominate oracle correctness. As a pilot proof of concept, these findings are preliminary and are intended to establish feasibility and motivate larger-scale empirical investigation.
arXiv:2507.12138v2 Announce Type: replace
Abstract: We introduce a principled, data-driven approach for modeling a neural prior over human body poses using normalizing flows. Unlike heuristic or low-expressivity alternatives, our method leverages RealNVP to learn a flexible density over poses represented in the 6D rotation format. We address the challenge of modeling distributions on the manifold of valid 6D rotations by inverting the Gram-Schmidt process during training, enabling stable learning while preserving downstream compatibility with rotation-based frameworks. Our architecture and training pipeline are framework-agnostic and easily reproducible. We demonstrate the effectiveness of the learned prior through both qualitative and quantitative evaluations, and we analyze its impact via ablation studies. This work provides a sound probabilistic foundation for integrating pose priors into human motion capture and reconstruction pipelines.
arXiv:2607.11119v1 Announce Type: new
Abstract: Robot manipulation is a complex task that requires visual understanding, physical reasoning, planning, and closed-loop control. General-purpose foundation models (FMs) have grown remarkably capable of some of these, especially vision and reasoning. To leverage this for generalist robot policies, current methods typically involve converting existing FMs into vision-language-action (VLA) models by fine-tuning on robot data to output low-level actions. However, VLAs are often orders of magnitude smaller than frontier FMs given the limited data and compute available for fine-tuning, which in turn limits their general capability. Inspired by the growing ability of FMs to operate software through visual interfaces, we ask whether that same competence suffices to control a robot. We present VIA (Visual Interface Agent for robot control), a framework that recasts robot control as an agentic task: an off-the-shelf FM-powered agent drives a manipulator through a browser-based 3D interface by taking screenshots, issuing intuitive commands, observing the outcome, and adjusting. The agent receives no robot-specific fine-tuning and no access to privileged state information: it perceives visual input and acts through a small set of general tools. VIA inherits the agent's general reasoning, closed-loop error recovery, and ability to plan and re-plan from what it observes. It solves a diverse suite of tabletop manipulation tasks zero-shot with both Claude Code and Codex. With the strongest model (Fable 5) it achieves 96.7% success on three LIBERO-Goal tasks and 100% on a long-horizon rainbow assembly task. Performance improves with the scale and strength of the underlying model. These results suggest that frontier agents already possess skills that transfer directly to robot control given the right interface: your coding or computer-use agent is, in a sense, secretly a robot-control agent.
arXiv:2603.28921v3 Announce Type: replace
Abstract: The critical damping condition of the damped harmonic oscillator model of SGD with momentum (Qian, 1999) yields a momentum schedule with no tuned hyperparameters: mu(t) = 1 - 2*sqrt(alpha(t)). Across five seeds on ResNet-18/CIFAR-10 (200-epoch cosine schedule) it reaches 90% test accuracy 2.34x faster than constant mu=0.9 (range 1.71-2.86x, 5/5 seeds, one-sided paired t-test p=4e-4), at the cost of a real final-accuracy deficit of 0.46 pp (5/5 seeds, p=0.009). A short-schedule control rules out a schedule-length artifact: compressed baselines either pay 0.5-0.9 pp of accuracy or stay slower to 90% at equal accuracy. A hybrid recipe -- critical-damping momentum until 90%, then constant mu=0.9 -- removes the deficit and keeps the speedup: 95.45 +/- 0.05% final accuracy at 2.4x faster progress to 90% (n=5). The speedup generalizes across architectures (VGG-16 without skip connections: 1.72x, n=3); on CIFAR-100 early gains persist (2-4x to mid-training thresholds) but the accuracy cost grows (-1.7 pp), compressing the accuracy-matched gain to 1.14x. We also report an exhaustive negative result on surgical layer selection. Version 2 of this paper claimed that gradient attribution on misclassified images selects which layers to retrain; running the identical correction protocol on all 35 combinations of 3-of-7 layer groups ranks the selected triple 11th of 35 (exact p=0.31) -- no better than random. What survives is weaker but real: combinations containing the top-ranked layer outperform the rest (+6.2 vs -1.4 mean net error reduction), and the bottom of the gradient-norm ranking reliably predicts the most harmful interventions (down to -20 net errors). Gradient attribution on errors is a harm-avoidance signal, not a selector of repair targets. We release the full 35-combination landscape as a baseline for layer-selection claims.