arXiv:2607.17411v1 Announce Type: new
Abstract: Inverse rendering is traditionally solved via differentiable renderers and gradient descent, which requires substantial problem-specific engineering and is prone to getting stuck in local minima due to ambiguities. Derivative-free approaches alleviate engineering requirements, but often heavily depend on a good problem initialization. In this work, we propose Feature-Informed Diffusion Evolution (FIDE), a fully black-box framework that requires no gradients or specific initialization: the renderer is treated as an opaque function whose only requirement is to produce images. Our key insight is feature guiding: rather than reducing each candidate rendering to a scalar loss value, we use a Vision Transformer (ViT) to extract dense visual features from it. We subsequently use these features to train a diffusion-based candidate proposal model, allowing the network to use visual cues to predict parameters that would match the target image. The candidate solutions proposed by this diffusion model are then refined in a closed loop with a CMA evolution strategy, continuously narrowing the proposal region as optimization progresses. We validate across diverse inverse problems from path tracing, vector splines, Voronoi shaders, and robotics, and demonstrate that feature-guiding substantially improves convergence over scalar-loss baselines and reliably escapes local minima where gradient-based methods stall.
Science Journals
arXiv:2607.17978v1 Announce Type: new
Abstract: We present RuNNer 2.0, the "Ruhr University Neural Network energy representation", a highly optimized software suite for training and evaluating high-dimensional neural network potentials (HDNNPs) of the second, third, and fourth generation. Long-range electrostatics and charge equilibration (QEq) for the description of non-local charge transfer in fourth-generation (4G) HDNNPs are accelerated by quasi-linear-scaling plane-wave methods, reducing QEq computational complexity from $\mathcal{O}(N^3)$ to $\mathcal{O}(N\log^2 N)$ such that linear or quasi-linear scaling is achieved across all HDNNP generations. An optimized memory management strategy eliminates the training overhead traditionally associated with long-range interactions, allowing 4G-HDNNPs to be trained with the same efficiency as their local counterparts. Developed in modern Fortran (2003/2008 standards), combined with a hybrid MPI/OpenMP parallelization scheme, RuNNer 2.0 has been designed to run efficiently in any CPU environment, from cost-effective local workstations to massive HPC clusters. Its modular library architecture facilitates straightforward binding to external simulation software; native interfaces to LAMMPS and the Atomic Simulation Environment (ASE) provide full access to all its features, including built-in committee-based uncertainty quantification. The high efficiency and scalability of the RuNNer 2.0 ecosystem are demonstrated through detailed benchmarks.
arXiv:2607.18203v1 Announce Type: new
Abstract: A recent paper [Hazeltine, et al., Phys. Plasmas 33, 072501 (2026)] has questioned whether the standard result, (ultimately) due to Ferraro, that the plasma angular velocity is approximately constant along individual equilibrium magnetic field-lines in a rotating axisymmetric magnetic mirror machine, continues to hold when the rotation becomes sonic or supersonic. In order to resolve this issue, the equilibrium of a rapidly rotating mirror is investigated, starting from first principles, using an ideal two-fluid model with anisotropic pressure. It is found that, as long as the ion gyro-radius is much less than the machine size, and the angular velocity of the plasma is much less than the ion gyro-frequency, the Ferraro result holds good.
arXiv:2607.17152v1 Announce Type: new
Abstract: Jailbreak attacks on large language models are usually evaluated by attacker-centric metrics such as attack success rate (ASR), yet an attack that breaks a model is not necessarily useful for improving its safety. We propose a defender-centric view of jailbreak evaluation, where attacks are evaluated by the downstream safety improvements they enable when used as red-teaming data for safety training. Building on this view, we introduce A-MESS (Minimal Effective Attack-Subset Selection), a setting-agnostic framework for attributing and selecting jailbreak attacks from black-box subset utility observations. A-MESS estimates AttackSHAP, a Shapley-based score that attributes marginal utility to individual attacks and selects compact attack subsets under user-specified budgets via greedy or surrogate-based optimization. Across controlled utility landscapes and real LLM safety settings, we find that ASR rankings are weakly aligned with defender-centric utility, that AttackSHAP can be estimated accurately with limited utility queries, and that directly optimizing subsets yields stronger safety utility than attacker-centric or attribution-only selection. These results suggest evaluating jailbreak attacks as resources for improving safety, not only as tools for breaking models.
arXiv:2607.17094v1 Announce Type: new
Abstract: The results of a survey on the use of GenAI by the design students of the Politecnico di Milano raises major questions around the role of AI in the Design Process. A domain specific set of questions alongside the more general purpose probes about GenAI usage, delivers insights into the particular practices that are emerging in Design. The very high frequency of use of GenAI tools is concentrated in the initial stages of projects and does not affect the perception of project ownership or creativity. An analysis of AI journals kept by a class during research assignments confirms the range of AI supported activities but also the limited trust in GenAI leading to systematic individual and collective verification and augmentation of outputs. Together, the findings suggest that design students are reflectively experimenting with how GenAI tools can contribute to their creative design process.
arXiv:2507.21987v2 Announce Type: replace
Abstract: Graph modification problems aim to find a small set of modifications to a graph so that it satisfies a desired property. The literature is rather rich in NP-completeness results and polynomial time solvable cases for special graph classes. However, no exact algorithm has been proposed for perfect graph modification problems. In this work, we propose the first exact solution methods based on integer programming for three variants: minimum perfect editing, minimum perfect completion, and the perfect sandwich problems. The minimum perfect editing problem inquires about the smallest number of edge additions and deletions needed to make a graph perfect, while the completion problem allows only for edge additions. The perfect sandwich problem is a decision problem that asks whether a perfect graph can be formed by adding edges from a restricted subset. To solve these problems, we formulate an integer programming model based on the Strong Perfect Graph Theorem. To address the resulting exponential number of constraints, we propose a branch-and-cut algorithm that dynamically generates them on demand. At the core of this approach is an efficient separation routine for enumerating odd holes and odd antiholes. We also release this underlying routine as "is_perfect"-a standalone open-source perfect graph recognizer and odd hole enumerator designed for broader community reuse. To enhance the practical efficiency of the branch-and-cut algorithm, we calculate the expected number of odd holes and odd antiholes in random Erdos Renyi graphs. In addition, we propose "IterativeModificationHeuristic", the first heuristic for the editing and completion problems, which provides upper bounds. Finally, we demonstrate the empirical effectiveness of the proposed methods through computational experiments on a wide range of instance types; all benchmark instances are publicly available.
arXiv:2607.17517v1 Announce Type: new
Abstract: Determining the true minimum distance of an alternant code remains a notoriously difficult problem in coding theory. In this paper, we study the minimum distances of generalized BCH codes and twisted Goppa codes through their parity-check matrices. We first give a necessary and sufficient condition for an alternant code to attain its designed distance and apply it to generalized BCH codes. As applications, we prove that broad classes of generalized BCH codes have minimum distances equal to their designed distances. These classes provide explicit infinite families rather than isolated examples. We characterize when a twisted Goppa code $\Gamma(L,g,\eta)$ with $\operatorname{deg} g=t$ satisfies $d(\Gamma(L,g,\eta))=t+1$, and derive structured classes and infinite families attaining this distance.
arXiv:2607.17713v1 Announce Type: new
Abstract: Span-level rationales are often assumed to improve controllability in text detoxification, but it remains unclear when such guidance helps and when it introduces trade-offs. We present Awareness-Enhanced Guidance for Iterative Safeguard (AEGIS) as an exploratory framework for studying span-guided multilingual detoxification across English, Mandarin Chinese, and Korean. AEGIS combines span-level detector outputs with frozen generator backbones, allowing harmful spans, intensity labels, and target attributes to be provided as structured guidance during rewriting. Rather than claiming state-of-the-art detoxification performance, we analyze how span guidance affects the balance between toxicity reduction and meaning preservation across generator families, model scales, and languages. Our results suggest that span-guided detoxification is conditionally useful: explicit rationales change the trade-off between toxicity reduction and meaning preservation, but their effects depend strongly on the generator backbone and the linguistic context. These findings highlight both the promise and the limitations of span-level control signals for multilingual detoxification.
arXiv:2607.18221v1 Announce Type: cross
Abstract: Magnetometers are among the most widely used probes in science and technology. Comagnetometers increase sensitivity by self-cancellation of magnetic noise, but only at low frequencies. We suggest a correlated measurement of two alkali species in one cell to cancel the magnetic background also at high frequencies. The inter-species phase difference of the light-matter interaction response function is found to be calibration free and insensitive to common-mode intensity noise. Utilizing a dark-matter signal as a testcase, the method achieves a thirtyfold background suppression, raising the signal-to-noise ratio by an order of magnitude or more, depending on the coupling to the different subatomic particles. We show that the method also provides model differentiation. The higher sensitivity and model differentiation open a path to novel probes for precision measurements in general and exotic fields in particular.
arXiv:2607.17171v1 Announce Type: new
Abstract: Monocular foundation models provide dense geometry but usually lack a stable metric scale. This paper presents VIDAR, a visual-inertial dense reconstruction framework that couples SVO+IMU odometry with Depth Anything 3. VIDAR uses the visual-inertial front end as a metric anchor: it provides camera poses, scale, and a consistent world frame for aligning dense foundation-model predictions across time. The foundation model then contributes detailed local geometry that is fused into a global reconstruction. We study both pose-conditioned DA3 and a decoupled alignment strategy. On EuRoC, pose injection reduces scale error to about 1\% and reaches 0.463 mean F@0.10; the decoupled hybrid improves this to 0.676 without ground-truth poses. Results on EuRoC and TUM RGB-D show that VIDAR is a practical route to metric dense monocular reconstruction.
arXiv:2607.17671v1 Announce Type: new
Abstract: Large-scale single-cell perturbation atlases make it possible to ask an inverse question: given an observed transcriptional response, which annotated targets and compounds in a fixed library are most consistent with that response? We present \model, a Transformer retrieval model for this closed-library setting. Each input is a cell-level perturbation signature formed by contrasting one treated cell with a cell-line-specific mean DMSO reference. The encoder maps the signature to a target-retrieval vector and a molecular-embedding vector, trained jointly with supervised target losses and structure--transcriptome alignment. We evaluate on Tahoe-100M conditions with mapped target annotations using a within-compound stratified 90/10 condition-pair split of 10,505 training and 1,168 validation drug--cell-line pairs. Because compounds and cell lines can occur in both partitions, the experiment measures held-out condition-pair retrieval rather than generalization to unseen compounds or cellular contexts. In a Monte Carlo evaluation over 38,400 sampled validation cells, \model\ achieved target Recall@10 of 0.408 and Recall@20 of 0.544, together with compound Hit@1 of 0.129, Hit@10 of 0.343, and mean reciprocal rank of 0.205 over a 379-compound bank. A separate diagnostic evaluation produced nearly identical values for the main model and large gains over a random-vector control and post-hoc bag-of-genes controls. These results demonstrate that a single multi-task model can recover both mapped target annotations and recorded compound identities from observed cell-level responses in the evaluated Tahoe-100M closed-library setting. Generalization to unseen compounds and cellular contexts remains to be established.
arXiv:2508.01438v3 Announce Type: replace
Abstract: The capacity of bandlimited direct-detection channels is challenging to compute or approach due to the receiver non-linearity. A generalized vector approximate message passing (GVAMP) detector is designed to achieve high rates at a reasonable level of complexity. The rates increase by using multi-level coding and successive interference cancellation. The methods are applied to fiber-optic channels with intersymbol interference caused by spectrally efficient pulse shapes, chromatic dispersion, and receiver sampling at twice the baud rate. Bipolar modulation operates within 0.26 bits per channel use (bpcu) of the real-alphabet coherent capacity for optically amplified links, reducing the best-known theoretical gap of 1 bpcu. Remarkably, bipolar modulation achieves 6 dB and 3 dB of power gain over unipolar modulation with and without optical amplification, respectively. Simulations with polar-coded modulation confirm the gains. The GVAMP complexity, measured in multiplications per information bit (mpib), is proportional to the number of iterations and to the logarithm of the block length, and is substantially lower than that of other equalizers. For example, a system with 64-ary bipolar modulation and a root-raised cosine pulse with a 1% roll-off factor was simulated over 4 km of optically amplified standard single-mode fiber in the C-band. The GVAMP receiver requires 93 mpib to achieve 5 bpcu at 300 gigabaud.
arXiv:2607.17527v1 Announce Type: new
Abstract: Computer Use Agents (CUAs) can autonomously execute complex, multi-step tasks within GUIs, enhancing efficiency through parallel multitasking. However, our formative studies with CUA experts and GenAI users indicated that current feedback is primarily text-based, requiring sustained attention to monitor progress and offering limited visibility to trace past GUI interactions. Based on the findings, we developed a prototype system, Sidekick, for communicating CUAs' status with multimodal feedback across different stages of interaction: (i) When CUAs run in the background, Sidekick signals its execution state through ambient cues. (ii) Upon resuming interaction with CUAs, Sidekick provides multimodal summaries of completed actions to support rapid context resumption. (iii) When CUAs operate in the foreground, Sidekick enhances transparency by verbalizing and visualizing the agent's reasoning. A study with 30 participants demonstrated that Sidekick significantly improved multitasking performance with CUAs compared to baseline systems that presented textual feedback either in a typical chat or in an ambient display. Sidekick supported progress awareness, and error and action traceability more effectively. Finally, we demonstrate the promise of Sidekick through several example applications, and discuss implications for long-horizon human-agent collaboration.
arXiv:2607.17120v1 Announce Type: new
Abstract: Multi-object tracking (MOT) is conventionally decomposed into detection followed by association, with object identity maintained as external state: track buffers, motion models, and appearance embeddings. We ask whether a video generator can maintain that state in pixels. We fine-tune a 22B text-to-video diffusion model (LTX-2.3) with a lightweight in-context LoRA to translate an RGB clip into an ID-map clip, a video in which every person is painted a flat, distinct color that persists over time: same color, same identity. Long videos are generated as chained windows, where each window is conditioned on the cleaned tail of the previous one. A brief continuation fine-tune teaches the model to extend a given coloring, after which identity flows through the chain with no tracker, no motion model, and no re-identification module. On the DanceTrack test server, our system, to our knowledge the first generative tracker evaluated there and the only entry with no detector and no tracking stack, reaches 40.3 HOTA. This is well below today's specialist state of the art (>=70 HOTA), but with a unique, inverted error profile: its association score (AssA 44.1) exceeds every tracker of the original benchmark suite while detection remains the sole deficit. Controlled comparisons show the mechanism matters: the same generated windows linked by classical post-hoc association score 2x worse (18.2 HOTA), and frame-to-frame IoU association fragments tracks that the generator's colors keep whole. On 383 mined occlusion events, the generator re-acquires identities after gaps at a 42% conditional rate where appearance-embedding baselines score zero, including gaps longer than its temporal context, evidence that the generator's color assignment functions as an emergent re-identification signal. We release code, checkpoints, and the full pre-registered experimental log.
arXiv:2607.18016v1 Announce Type: new
Abstract: Vision-language-action policies are a promising foundation for general robot control, but long-horizon humanoid loco-manipulation requires the robot to treat task objects as persistent physical entities across movement, contact, occlusion, and recovery. We study this problem as object-state divergence: the object state used to condition a whole-body action can differ from the state used to decide whether the action achieved the intended physical relation. We propose \emph{Persistent Object Tokenization} (POT), which maintains role-indexed 3D object records from RGB-D observations and converts them into object tokens for a whole-body action expert. Instantiated as \emph{POT-VLA}, the same object records condition action generation and support geometric predicate checks, yielding a closed-loop execution system in which object state is both actionable and verifiable. On a Unitree G1, POT-VLA improves a matched direct GR00T-N1.7 baseline from 39/80 to 71/80 successes over eight real-world task families. In an external Being-0-aligned reference, POT-VLA achieves 44/50 successes on aligned service tasks, compared with the 37/50 success reported by the Being-0 paper. The largest gains occur on tasks requiring maintained 3D relations, suggesting that persistent object-centered state is a useful abstraction for verifiable humanoid VLA execution.
arXiv:2607.17032v1 Announce Type: new
Abstract: This paper investigates the optimal safety control problem of nonlinear control systems by proposing novel high-order control barrier functions (HOCBFs). Different from zeroing HOCBFs, two novel HOCBFs are derived and the safety controllers are designed in an explicit way. Next, we implement vector Lyapunov function approach to propose a novel high-order control Lyapunov function (HOCLF) for the stabilization control problem. The relations between the proposed and existing HOCBFs are discussed. Afterwards, the compatibility of the proposed HOCLF and HOCBF is addressed to guarantee the stabilization and safety control objectives simultaneously, and thus the optimal controller is established. Finally, a numerical example from the navigation problem of quadrotors is presented to illustrate the efficacy of the derived results.
arXiv:2508.05298v3 Announce Type: replace
Abstract: We present GhostShell, a novel approach that leverages Large Language Models (LLMs) for streaming and concurrent behavioral programming in embodied systems. In contrast to predefined behavioral structures and plan-then-execute paradigms, GhostShell enables reasoning-while-acting by incrementally invoking functions during LLM streaming generation. We define function tokens as an XML-based function-call representation that GhostShell parses from the LLM generation stream and maps to callable functions. A multi-channel scheduling algorithm further orchestrates these calls with intra-channel synchronous and inter-channel asynchronous dispatch, coordinating sequential-parallel behavior execution across multiple robotic components. We evaluate GhostShell on our robotic prototype CoCo across 33 real-world tasks with LLMs from nine providers. On 30 grounded Human-Robot Interaction (HRI) tasks, our approach achieves the highest Directed Structured Behavior Correctness (DSBC) score of 0.83 with Claude-Sonnet-4, while on three long-horizon multimodal tasks, GPT-4.1 attains a top human evaluation score of 7.0/10. Compared to native LLM function calling, our function token schema achieves a 15/15 task completion rate versus 6/15, particularly in coordinating concurrent linguistic and physical actions. Supplementary materials, including videos, are available at https://coco-robot.github.io/GhostShell.
arXiv:2607.17205v1 Announce Type: new
Abstract: Supervised fine-tuning (SFT) of open-weight LLMs on expert agent trajectories has emerged as a prominent approach to building capable code agents without reliance on proprietary models. A central yet underexplored question is how trajectory quality and quantity jointly shape model performance. We present a systematic empirical study of trajectory data filtering for LoRA fine-tuning of Qwen2.5-Coder-7B-Instruct on the SWE-trajectory dataset (67,074 trajectories, of which 32,161 are resolved). We propose a two-axis quality scoring framework -- Efficiency and Style -- and evaluate it through 16 controlled experiments spanning strategy, scale, and ablation analyses. Since 7B-scale models attain near-zero SWE-bench resolve rates, we adopt cross-entropy (CE) loss on held-out trajectories as the primary metric, validated via first-action generation: CE loss and ROUGE-L are perfectly rank-correlated (Spearman $\rho$ = -1.00), with limited-sample evidence supporting but not conclusively establishing this proxy. Our results reveal a scale-dependent quality-quantity trade-off: at small scales, doubling the dataset (500 to 1,000) yields ~12.7% CE-loss reduction whereas the TopQ-Random gap stays <1% (Mann-Whitney p > 0.10); at 2,000 trajectories this same gap widens to 3.6% (p = 0.016). Ablation further identifies error-retry rate as the dominant sub-dimension, performing comparably to the full composite ($\Delta$ < 0.2%). Together, these findings establish trajectory-level quality scoring as a viable but scale-sensitive lever for code-agent SFT and offer a proxy-validated evaluation protocol for the regime where end-to-end resolve rate is statistically infeasible.
arXiv:2607.17852v1 Announce Type: new
Abstract: Localization is a key task in robot navigation, and many techniques exist for it. In many plausible scenarios, a robot might face unforeseen, dynamic obstacles, rendering any pre-determined map inaccurate for localization. In this work, we propose a robust lifelong localization framework in dynamic planar indoor environments, using the robot's odometry and sparse distance sampling. We demonstrate how distance samples can be used to provide a robust prior on the robot's location. This technique can solve the kidnapped robot problem in real time, up to symmetries. Based on insights from real-world recorded data, we also account for dynamic obstacles. We then fuse this prior, over time, with the odometry to converge to the robot's location. A central property of our method is that it provably converges to the robot's ground truth pose even in large indoor environments when the environment is static. We further show that this guarantee also holds in dynamic environments, as long as the nature of those changes has been correctly learned. We demonstrate the effectiveness of our approach in different real-world indoor environments. In particular, we achieve a localization comparable to SLAM with merely a few (sixteen) distance samples, as opposed to the full LiDAR range. Sufficing with only sparse distance sampling is advantageous in terms of sensor cost, privacy, storage space, and transmission bandwidth.
arXiv:2607.17861v1 Announce Type: new
Abstract: Establishing interpretable decision-making processes in long-horizon robotic manipulation is critical for enabling reliable human oversight and intervention. However, existing approaches to robotic manipulation largely treat skill selection as opaque mappings from observations to actions, offering limited transparency into how decisions are formed. In this work, we propose ConceptTree, a framework that reframes high-level manipulation skill selection as reasoning over human-interpretable concepts, representing high-level policies as a sequence of concept-level predicates over visual observations. Rather than relying on implicit latent representations, our method learns a normalized concept space grounded in visual inputs, over which a decision tree is trained to predict high-level skills. This formulation yields a transparent decision process that is both traceable and intervenable, enabling direct inspection and modification of policy behavior. We evaluate our approach on a set of real-world robotic manipulation tasks with increasing complexity. Experimental results show that ConceptTree consistently outperforms existing concept-based baselines, particularly in complex, long-horizon scenarios. Furthermore, we provide qualitative case studies showing that our model supports fine-grained intervention by modifying individual concepts, enabling targeted correction of decision errors without retraining.
arXiv:2607.17559v1 Announce Type: new
Abstract: The Contrastive Olfaction-Language-Image Pre-training 2 (COLIP-2) model is a multimodal embeddings space that places olfaction as a first-class citizen among vision and language. Molecular structure, gas-sensor readings, odor-descriptor language, and images are all trained into a single shared representation space, so that a robot can localize a detected aroma to objects in a scene probabilistically. No ImageNet-scale datasets of paired image-scent examples exists which warrants the need for their collection. Our intent with the release of COLIP-2 is to demonstrate the limit of what can be built for robotics with open-sourced olfactory data in order to ground the argument for why new methodologies and datasets are necessary in order to enable advanced olfactory-oriented perception capabilities. We enumerate results from internal testing of the COLIP-2 architecture and make necessary optimizations to run the model at the edge for real-time robotics applications. While developed with robotics in mind, the design of COLIP-2 has been influenced by experts across many disciplines of science in academia and industry, and we hope that the model can be useful in any multimodal domain requiring olfactory intelligence.
arXiv:2607.16925v1 Announce Type: cross
Abstract: This paper presents a novel intelligent chaotic-based code-division multiple access (CDMA) scheme for underwater optical wireless communication (UOWC), addressing critical performance degradation caused by severe scattering and multipath dispersion in underwater environments. Unlike conventional modulation techniques such as on-off keying, which depend on precise pulse timing and show high sensitivity to channel distortions, the proposed approach leverages unpredictable deterministic chaotic sequences generated by the logistic map to enhance robustness against scattering-induced impairments. A Multi-Agent Reinforcement Learning (MARL) framework enables distributed agents to dynamically adapt chaotic map parameters, including initial conditions and bifurcation parameters, based on real-time environmental feedback, optimizing sequence generation to maintain low cross-correlation properties and improve resilience to multipath effects. Experimental validation using a 2-meter water tank testbed with controlled turbidity demonstrates superior performance compared to conventional schemes. The adaptive framework exhibits rapid convergence and relaxed synchronization requirements, making it highly suitable for distributed underwater networks where centralized coordination is impractical.
arXiv:2508.08368v2 Announce Type: replace
Abstract: Molecules with heavy, radioactive nuclei promise extreme sensitivity to fundamental nuclear and particle physics. However, these nuclei are available in limited quantities, which challenges their use in precision measurements. Here we demonstrate the gas-phase synthesis, cryogenic cooling, and high-resolution laser spectroscopy of radium monohydroxide, monodeuteroxide, and monofluoride molecules ($^{226}$RaOH, $^{226}$RaOD, and $^{226}$RaF) in a tabletop apparatus by combining novel radioactive target production protocols, optically driven chemistry in a cryogenic buffer gas, and low-background spectroscopic detection methods. The molecules are cooled in the lab frame, creating conditions that are the same starting points as many current molecular precision measurement and quantum information experiments. This approach is readily applied to a wide range of species and establishes key capabilities for molecular quantum sensing of exotic nuclei.
arXiv:2607.18075v1 Announce Type: new
Abstract: The Duke Robotics Club presents Oogway and Crush, our AUVs for RoboSub 2026. This year's strategy expands on our previously narrowed scope, targeting all four of RoboSub's design goals for the first time: movement, vision, manipulation, and acoustic tracking. This expansion is based on sustained reliability investment across all three subsystems. Mechanically, Crush gained two additional thrusters and a CFD-optimized case, providing pitch stability. Electrically, we addressed accumulated failure points by repairing unreliable connections and upgraded thruster control hardware. We also redesigned our acoustics system, adding a new custom PCB with higher-order filters, significantly improving pinger detection reliability. On the software side, improvements to state estimation, sonar-based object detection, vision-driven task planning, and IVC enable more capable and coordinated autonomous runs. Paired with investments in our testing infrastructure to maximize our limited pool time, we can now attempt a broader set of tasks while maintaining the reliability our competition strategy demands.
arXiv:2607.18078v1 Announce Type: new
Abstract: Vision-only occupancy prediction requires recovering a semantic 3D occupancy field from calibrated surround-view images, where each view provides observations with ambiguous depth along camera rays. Existing methods have progressed from dense structured representations to sparse Gaussian primitives, improving the efficiency of 3D scene representation. However, Gaussian learning still relies primarily on image domain features, which provide limited explicit geometric information for volumetric reasoning. Our key observation is that effective Gaussian occupancy modeling requires not only sparse primitives, but also richer geometric and semantic learning cues. In this paper, we propose VGOcc, which learns visual and geometric cues from foundation models for Gaussian modeling. VGOcc incorporates these cues into primitive initialization and refinement, yielding a representation termed Visual-Geometric Gaussians tailored to semantic occupancy prediction. Specifically, we propose Visual-Geometric Gaussian Birth to form spatially balanced Gaussian centers from ray depth hypotheses, while visual semantic features initialize primitive attributes. Next, we design Pose-Aware Feature Learning to combine foundation tokens with camera embeddings and calibrated ray information. Features from neighboring views are then aggregated at projected 3D locations for each Gaussian refinement stage. Finally, Gaussian decoder refines birth Gaussians with pose-aware features and renders them into semantic occupancy. Experiments on nuScenes demonstrate that VGOcc achieves state-of-the-art performance in vision-only 3D occupancy prediction. Codes will be available at https://github.com/JHLin42in/VGOcc.