Computer Vision
Papers
Fresh from arXiv, typeset properly. Highlighted words trade on the market.
All of arXiv
Saturday 29 August
Ground image localization with respect to satellite imagery is a key enabler for metrically-accurate, geo-localized 3D scene reconstruction from unconstrained image collections. Existing cross-view localization methods have strict requirements such as panoramic imagery or known initial locations, limiting their applicability for in-the-wild reconstruction settings. We propose a robust hierarchical cross-view localization framework that leverages geometric constraints from Structure-from-Motion (SfM) models derived from unconstrained ground image collections. Our method generates coarse-to-fine pose hypotheses through a cross-view matching approach and aggregates noisy predictions across SfM model(s) using Kernel Density Estimation to recover consensus alignments while filtering outliers. Experiments demonstrate reliable localization performance from challenging image collections. Empirically we found satellite-referenced alignment enables accurate metric scale estimation, doppelgänger detection, and merging of disjoint SfM reconstructions, resulting in more complete, geo-localized site models than are possible with SfM alone.
Artificial Intelligence
Imag-Eval: a language-grounded framework for interpretable Text-to-Image instruction following evaluation
Text-to-Image (T2I) models have recently achieved impressive visual fidelity, yet their evaluation remains constrained by benchmarks that are often difficult to interpret and insufficiently diagnostic. Existing skill-based evaluations tend to overlook critical failure modes that strongly impact usability but fall outside standard taxonomies, such as global incoherence arising from missing parts or physically implausible configurations (e.g., floating objects). In addition, prompt difficulty is typically controlled along a single dimension; either prompt length or the number of elements to generate. To address these limitations, we introduce Imag-Eval, a controlled benchmark designed to assess how T2I models ground compositional natural-language instructions into visual outputs. Unlike prior work that conflates surface linguistic complexity with compositional difficulty, Imag-Eval explicitly seeks to disentangles these factors by independently varying both the number of instances and the combination of constraints (rules), while avoiding error propagation. This design enables fine-grained and interpretable analysis of where cross-modal instruction following fails. Our benchmark comprises 1,140 prompts and 8,842 combined rules, and we evaluate it on several state-of-the-art models. Complementing this analysis with an additional study of over 2,000 prompts from a concurrent benchmark, our results suggest that, for structured skills, compositional difficulty is primarily governed by the number of grounded rules and their binding to instances,, rather than by prompt length alone.
Read the full abstract
Computation and Language
Toward Cultural Alignment: Human-Centered Evaluation of Multimodal AI Stories Across Five African Communities
In this paper, we examine how well AI-generated multimodal stories align with the lived practices, relationships, language, values, and visual expectations of the communities they represent. We conduct a community-grounded mixed-methods evaluation with 19 culture representatives across five African communities, combining quantitative annotations with qualitative focus group discussions. We find that cultural alignment depends not simply on recognizable cultural markers, but on how those markers fit social, linguistic, procedural, and visual context. From these evaluations, we develop a taxonomy of cultural alignment comprising five broader cultural marker categories and eight recurring mechanisms of misalignment. We additionally evaluate five multimodal LLM judges to examine whether automated evaluation can approximate community-grounded judgments at scale. Judge reliability and score calibration vary substantially across communities, with no single judge performing consistently across all five settings. These findings motivate community-calibrated evaluation pipelines in which automated judges are validated against community judgments to determine where they can be trusted and where human review remains necessary.
Read the full abstract
Robotics
AdaVLA: Adaptive Step Flow Matching for Training-free Acceleration of Vision-Language-Action Models
Vision-Language-Action (VLA) models, built upon Vision-Language Models (VLMs), have significantly enhanced robotic capabilities by leveraging internet-scale knowledge and multimodal reasoning. However, the intensive computational overhead of VLAs constrains on-device deployment, hindering real-time responses to environmental changes. While various acceleration techniques have been proposed, they often rely on fine-tuning or access to training datasets, which are frequently unavailable due to privacy and proprietary concerns. Moreover, although flow-matching-based VLAs have emerged as efficient alternatives to standard diffusion models, current acceleration efforts largely target VLM inference costs, failing to address the iterative ODE solving process inherent in flow matching inference. To address these limitations, we propose AdaVLA, an online, training-free adaptive framework for fast yet accurate flow-matching-based Vision-Language-Action models. We introduce a novel metric derived from the flow matching trajectory curvature to quantify action generation confidence during inference. This metric enables the dynamic reduction of inference steps and the adaptive adjustment of MLP pruning ratios through an efficiently computed importance evaluation, requiring no access to training data. Experimental results on the LIBERO benchmark using a Jetson AGX Orin device demonstrate that our method achieves $1.87\times$ and $2.24\times$ speedups for $\pi_{0.5}$ and X-VLA, respectively, with negligible degradation in success rates. Furthermore, we validate the robustness of our approach on real-world robotic tasks using SmolVLA.
Read the full abstract
Artificial Intelligence
Hyper-Fold: Exploring the Expressive Limit of Sequence-Geometry Learning for Proteins via Hypergraph Modeling
Protein structure modeling rests on a single computational primitive: the interaction between what a residue is (sequence content) and where it sits (three-dimensional geometry). What is the expressive limit of this layer class? We show that the complete bilinear operator over content-geometry outer products–the sufficient statistic of all second-order interactions–is the expressive ceiling, while the additive message passing of mainstream geometric GNNs is provably blind to content-geometry binding. We then introduce Hyper-Fold, a rank-K separable convolutional backbone approaching this ceiling at message-passing cost: each radius neighborhood is organized into a sequence hyperedge and a contact hyperedge, modulated by an edge-conditioned matrix-valued operator factorized into K learned basis operators with geometry-generated coefficients. Across enzyme function prediction, fold classification, and ligand binding site detection, Hyper-Fold and its hierarchical variant Hyper-Fold-Deep achieve the best results among protein-specific structure encoders; Hyper-Fold-Pocket, an anchored set-prediction head, surpasses UniSite-3D on UniSite-DS and two zero-shot benchmarks with no sequence language model features, 68x fewer parameters, and 4.8x lower latency–suggesting that a sufficiently expressive 3D backbone recovers information that fusion architectures previously borrowed from evolution-scale pretraining.
Read the full abstract
Artificial Intelligence
Benevolent Bias in Multi-Turn Human-Agent Dialogue
Bias in human-agent interaction can manifest not only through hostile language but also as benevolent bias, whereby unequal treatment hides behind a warm, positive tone. To make it detectable, we operationalise benevolent bias along two dimensions, tone and treatment, yielding three classes: neutral support, overt bias, and benevolent bias. Building on these definitions, we construct BENEVDIAL, a class-balanced corpus of 362,880 multi-turn support dialogues spanning user and agent demographics, roles, and generators, to support controlled evaluation. We then test two detector families on it: off-the-shelf safety detectors and prompted large language model (LLM) judges. Our findings reveal a notable detection gap: off-the-shelf detectors reliably flag overt bias yet largely fail to identify benevolent bias. LLM judges improve sensitivity when guided by explicit detection criteria, but this comes at the cost of increased misclassification of neutral supportive statements as benevolent bias, a tendency that is further exacerbated by the presence of demographic context. These findings suggest that fair monitoring of human-agent dialogue must look beyond surface cues to whether the agent's treatment is disparate.
Read the full abstract
General Relativity and Quantum Cosmology
Yukawa modification of Traversable Wormholes supported by Holographic Dark Energy
Inspired by holographic dark energy models, we extend the analysis performed in Ref.\cite{RGPC} taking into account Yukawa deformations on the different energy density (HDE) profiles. The profiles we have explored are: the Bekenstein-Hawking HDE, the Moradpour energy density, the Standard Renyi HDE and the Mixed Energy Density. With the help of an inhomogeneous Equation of State of the form $p_{r}(r)=\omega _{r}\left( r\right) \rho (r)$, we have considered Zero Tidal Forces. Differently from the original case, only a particular case of the family of the Bekenstein-Hawking HDE profiles distorted by the Yukawa term has developed a divergent $\omega _{r}\left( r\right) $ for $r\rightarrow \infty $. To cure such a divergence we have introduced a modification at large distances in order to have a finite result. This means that the Zero Tidal Forces can be imposed in a very large but limited region of the spacetime. Such modification did not change the behavior of the Equation of State close to the throat. For every proposal we have computed the components of the Stress-Energy Tensor. What we have found is that the Yukawa-Bekenstein-Hawking HDE is the only positive profile, while the Yukawa-Moradpour and the Yukawa-Renyi have a limited region where the energy density is positive. The remaining Yukawa-Mixed Energy Density is always negative. This means that the Yukawa distortion introduces a kind of negative energy density.
Read the full abstract
Software Engineering
AgentLogs: A Dataset for Opening the Black Box of Cloud Agents on GitHub
Coding agents now complete development tasks in real software projects and submit the results as pull requests. A pull request shows the outcome of an agent's execution but not the process behind it, such as which files the agent read, which commands it ran, or how many tokens it consumed. This is recorded only in the agent's log, and no existing dataset provides such logs of real-world use, including failed agent executions, across a wide range of repositories. We introduce AgentLogs, a large-scale dataset of activity on GitHub's cloud agent platform on public repositories, collected without requiring opt-in from developers. AgentLogs comprises 363,624 tasks and 635,886 sessions, each session an execution of the agent, in 40,565 popular public repositories, with 75,631,136 log entries that record each session step by step, including prompts, intermediate reasoning, tool calls (e.g., file edits, git operations, and GitHub interactions), and token usage, together with the linked pull requests, timelines, workflow runs, and users. Sessions are included regardless of their outcome. AgentLogs enables research on agent behavior and failure modes, efficiency and cost, task formulation, human-agent collaboration, and repository adoption. A demonstration video can be found at: https://youtu.be/ZnFpjuK9Azg
Read the full abstract
Combinatorics
Interval Deformations of Coxeter Arrangements
We study integer-interval deformations of Coxeter arrangements of types $A$, $B$, and $D$, assigning an integer interval to each root direction through an integer-valued root function. We introduce a local compatibility condition on this root function. For compatible root functions, subject to additional local bounds in types $B$ and $D$, we derive characteristic-polynomial formulas as binomial sums over cyclic orders in type $A$ and signed permutations in types $B$ and $D$. The proofs use the finite-field method and reduce hyperplane avoidance to consecutive-gap inequalities. These formulas recover, in particular, the classical Catalan and Shi characteristic polynomials.
Read the full abstract
Physics and Society
A Generalized Gompertz Law for the Rise and Fall of Empires
Human societies tend to evolve toward increasing scale and complexity, often followed by stagnation and decline. Historical empires provide a prominent example of this trajectory; however, existing quantitative models capture expansion and rarely account for decline within a unified framework. Here, we introduce a minimal dynamical model that describes both growth and collapse in a single formalism. The model yields a closed-form trajectory that generalizes the Gompertz law and admits equivalent maximum-entropy and optimal-control formulations. We show that this functional form captures the rise and fall of fourteen diverse empires across widely varying time scales. Despite substantial historical heterogeneity, all cases follow a common asymmetric rise-peak-decline pattern governed by a small number of interpretable parameters. We identify the mechanism driving this behavior as the chronophage: an exponentially accumulating burden of administrative and coordination costs that offsets the gains of expansion. These findings uncover a general quantitative principle, yielding new insights into how increasing complexity constrains large-scale social systems.
Read the full abstract
Combinatorics
Resolvable quantum Latin squares with maximal cardinality
In this paper we investigate the existence of a maximal-cardinality \(t\)-resolvable quantum Latin square (\(t\text{-}\text{RQLS}(v)\)). We determine the existence of a maximal-cardinality \(1\text{-}\text{RQLS}(v)\) with \(17\) possible exceptions based on the established research result and construct a maximal-cardinality \(2\text{-}\text{RQLS}(n^2)\) for every integer \(n\geq6\).
Read the full abstract
Quantum Physics
Fully integrated continuous-variable quantum key distribution with composable security over 100 km
Quantum key distribution (QKD) guarantees information-theoretic security by the laws of physics, but deployment at scale requires compact, manufacturable photonic terminals. Continuous-variable QKD (CV-QKD) is well suited for this transition through telecom-compatible, room-temperature coherent detection. However, unifying full on-chip core terminal integration, room-temperature operation, high loss tolerance, and composable end-to-end security in long-distance QKD remains a key bottleneck. Here we report a fully integrated CV-QKD platform in which two hybrid III–V/Si$_3$N$_4$ integrated lasers, a silicon transmitter, and a silicon coherent receiver implement the core terminal functions, operating with a local local oscillator (LLO) over fibre links of 25–150 km. A Bayesian machine-learning algorithm maintains robust phase lock throughout the long records required for composable security, consistently outperforming the conventional unscented Kalman filter, while rate-matched multidimensional reconciliation approaches the Shannon limit. The system certifies a composable finite-size secret-key rate of 29.3 kbps at 100 km from a 140-billion-symbol block, with 12.9 kbps at 125 km under finite-size analysis and 9.17 kbps at 150 km under asymptotic analysis. By establishing the longest finite-size and asymptotic reaches and the highest secret-key rate per symbol reported for integrated CV-QKD, this work advances the development of practical chip-based quantum networks.
Read the full abstract
Statistics Applications
Burn-in-Free Simulation of VARMA Time Series
Varmapack is a software package for efficient, exact simulation of VARMA time series without a burn-in period. For stationary models, Varmapack can generate initial states and innovations from their joint stationary distribution. Alternatively, the user can supply initial states, in which case innovations are generated from their conditional distribution. The core C library has interfaces to R, Python, and Matlab. It also offers VARMAX simulation and computation of autocovariances and correlations, spectral radii, and impulse responses. Varmapack automatically selects between vector-Yule-Walker and state-space methods for covariance computation and uses level-3 BLAS operations for efficient generation of multiple replicates. Benchmarks on several platforms show substantial performance gains over existing simulation software, ranging from several-fold to more than three orders of magnitude for the packages and models considered. Varmapack is open source and publicly available through GitHub.
Read the full abstract
Artificial Intelligence
How Identity and Opinion Shape Political Sycophancy in LLMs
As Large Language Models (LLMs) increasingly encourage users to disclose personal profiles for tailored assistance, measuring their political alignment becomes increasingly important. However, many existing benchmarks for assessing political behavior rely on closed-ended questions and do not fully capture how a model's stance may adapt to user-provided context during interaction. We introduce a framework that disentangles two distinct triggers of political sycophancy: opinion (aligning with explicit narratives) and identity (stereotyping based on demographic labels). Using 450 manually-checked political dilemmas as controlled probes, we evaluate 13 instruction-tuned LLMs. We uncover a dissociation: a model's susceptibility to explicit opinions does not necessarily predict its susceptibility to identity cues, and vice versa. When both signals are present, their effects are generally sub-additive rather than simply additive. Additionally, system-level personas primarily shift a model's baseline stance while having limited effect on the stance shift caused by user opinion or identity. Ultimately, our results suggest that LLM political stance is interactively and steerably vulnerable rather than being a fixed trait, highlighting how personalization may amplify identity- or opinion-conditioned shifts in the model's behaviors.
Read the full abstract
Machine Learning
PokaiTrainer: Scaling Equilibrium Search to Competitive Pokémon VGC
Decision-time equilibrium search carried poker to superhuman play, but it has so far relied on tractable subgames: a handful of actions per decision, chance confined to card deals, one player moving at a time. Competitive Pokémon in its official doubles format (VGC) breaks all three assumptions at once. Both players act simultaneously from joint menus in the hundreds, each joint action resolves to hundreds of stochastic outcomes, and the opponent's reserves and stat allocations are hidden. No prior Pokémon agent performs equilibrium search, and whether it scales to this regime was open; we show that it does, and report what it took. PokaiEngine, our Rust battle engine, enumerates a joint action's full weighted outcome distribution in one pass, at ${\sim}99\%$ parity with Pokémon Showdown and a fraction of the cost of sampling it. PokaiTrainer adapts Student of Games to this scale and trains it by self-play over hundreds of human teams. Each decision is solved by counterfactual regret minimization as a Bayesian matrix game over public belief states, subgames grow under an explicit compute budget, and value targets are harvested from the interior of every solve and grounded by realized outcomes. The strength is in the search. The network's policy alone loses even to a shallow heuristic search. PokaiTrainer is, to our knowledge, the first VGC agent rated on the live Showdown ladder. Under open team sheets it wins 59% of 150 best-of-three sets against a human field averaging ${\sim}1320$ Elo, holds a 1350-1400 Elo band, and at its peak reached 1492 Elo, entering the format's top 500.
Read the full abstract
Computer Vision
Bayesian-Optimized Superpixel-GrabCut for Traceable Optic Disc Segmentation
Optic disc (OD) segmentation is essential for diagnosing ophthalmic pathologies from retinal fundus images. However, prevailing deep learning approaches operate as opaque black boxes, lacking the inference-stage mathematical traceability–a critical requirement for algorithmic auditing and failure analysis in clinical workflows. This paper presents a fully algorithmically traceable and trainable segmentation pipeline that jointly combines superpixel decomposition, hybrid brightness-proximity superpixel scoring, morphological regularization, iterative GrabCut refinement, and elliptical shape fitting. The hyperparameter optimization is formulated as an objective function and solved via Bayesian optimization to eliminate manual parameter tuning. A quantitative evaluation on the Drishti-GS dataset demonstrates that our method achieves a Dice coefficient of 0.9536, matching state-of-the-art performance. By maintaining explicit mathematical transparency across all processing stages, our framework offers a deterministic, traceable alternative to black-box architectures for medical review and debugging.
Read the full abstract
Analysis of PDEs
Two-point estimates for the logarithmic p-flux of the first Dirichlet p-eigenfunction
Let \(u>0\) be the first Dirichlet \(p\)-eigenfunction on a bounded convex domain \(\Omega\subset\mathbb R^N\), and set \[ X_\Omega:=|\nabla\log u|^{p-2}\nabla\log u. \] We study whether the sharp one-dimensional two-point modulus for \(X_\Omega\), which for \(p=2\) reduces to the logarithmic-gradient estimate of Andrews and Clutterbuck, persists for \(p\neq2\). We prove sharp estimates on intervals and balls for every \(p>1\), with the radial modulus on balls strictly larger than the one-dimensional one. In dimensions \(N\ge2\), however, the one-dimensional modulus fails on general convex domains for every \(p\neq2\). For \(1<p<2\), at every sufficiently small fixed scale there are smooth uniformly convex domains for which the corresponding two-point flux tends to zero. For \(p>2\), on thin domains \(\Omega_\varepsilon=D\times(-\varepsilon,\varepsilon)\), with \(D\subset\mathbb R^{N-1}\) bounded and convex, the normalized first eigenfunctions satisfy \[ u_\varepsilon(x,\varepsilon z)\longrightarrow \frac{\phi(z)}{\phi(0)} \left(\frac{G_D(x)}{G_D(x_0)}\right)^{2/p} \] locally uniformly in \(D\times(-1,1)\), where \(\phi\) and \(G_D\) are the first Dirichlet eigenfunctions of the \(p\)-Laplacian on \((-1,1)\) and of the Laplacian on \(D\), respectively. This yields the failure for \(p>2\). Finally, for arbitrary \(C^2\) functions, positivity of the symmetric differential of the \(p\)-gradient implies convexity, and the converse holds universally if and only if \(p=2\).
Read the full abstract
Computer Vision
RLG-TPV: Radar- and LiDAR-Guided Tri-Perspective View Fusion for Camera-Radar 3D Object Detection
Tri-Perspective View (TPV) representations describe 3D scene structure through top, side, and front feature planes, but existing TPV lifting is primarily camera-based, leaving the depth of sampled image evidence ambiguous along projected camera rays. We propose RLG-TPV, a multimodal TPV framework for camera-radar 3D object detection in which radar and training-time LiDAR provide complementary geometric guidance during representation construction. A ray-guided deformable-attention lift weights sampled image features using LiDAR-supervised camera depth probabilities and radar frustum occupancy, while radar additionally refines the depth distribution before lifting. Because conventional radar provides limited elevation information, LiDAR-derived class-occupancy targets supervise the side and front planes during training; the corresponding heads are removed at inference, so deployment requires only cameras and radar. For temporal aggregation, Doppler-guided temporal fusion aligns past features using a motion field anchored by measured radar radial velocity, with gating that limits warping in regions without supported motion. An RCS-aware radar scatter further allows radar evidence to spread over spatial neighborhoods conditioned on radar cross section. On the nuScenes validation set, RLG-TPV achieves 0.4981 mAP and 0.5959 NDS, reducing orientation and velocity error by 31.9% and 30.7% relative to the published CRN baseline. Ablation studies show that ray-level geometric guidance is a major contributor to the final performance.
Read the full abstract
Machine Learning
HalluPrism: When Multimodal Uncertainty Should Diagnose, Not Decide
Multimodal Large Language Models (MLLMs) can assign similar confidence to answers that fail for different reasons. We propose HalluPrism, a behavioral diagnostic that re-runs an answer after visual degradation, blank-image replacement, and grounding or relation checks. These targeted probes yield a signature over visual-perturbation sensitivity (V ), image-removal confidence retention (L), and grounding/relation-probe instability (A). Across 58K+ examples from four benchmarks and four MLLMs, image-removal confidence retention is most prevalent, while grounding/relation-probe instability better separates failure families. Only 18 of 48 source-target checks are diagonally aligned, so the coordinates should be interpreted jointly rather than as independent causal sources. With the dataset fixed, the joint signature improves failure-family AUROC from 0.634 to 0.769 on HallusionBench and from 0.707 to 0.817 on VizWiz, with smaller gains on POPE and VSR. In pooled XGBoost analysis, AUROC rises from 0.78 with scalar confidence to 0.95 with (V, L, A) and 0.97 when confidence is added. The same signature does not automatically improve correctness ranking. The three tested direct scalarizations can harm it. These results separate failure diagnosis from abstention scoring: multimodal uncertainty should characterize failure structure before it is used to decide whether to abstain or correct.
Read the full abstract
Human-Computer Interaction
An Eye-Tracking Dataset for Viewing Distance Categories in Real-World Scenarios
Estimating viewing distance from gaze behavior is essential for understanding user intent and enabling distance-aware interactive systems. However, most existing eye-tracking datasets have been collected in constrained settings, such as laboratory environments or static tasks. Consequently, they only partially capture viewing behaviors in real-world situations where viewing distance changes with natural head and body movements. We introduce GazeDepth, an eye-tracking dataset collected from 19 participants using a wearable tracker during tasks reflecting real-world scenarios. GazeDepth includes fixed-distance viewing scenarios with constant observer-target distances at near (33 cm), middle (50 cm), and far (300 cm), as well as variable-distance viewing scenarios in which participants shift gaze among targets at different depths in indoor and outdoor environments. The dataset provides synchronized gaze data, pupil size, 3D eye-vectors, and head-motion signals, along with distance labels. Statistical analyses showed that distance-related gaze features, such as vergence angle and estimated viewing distance, differed consistently across viewing-distance categories. In addition, classification models trained on GazeDepth further demonstrated that the dataset captures gaze characteristics that distinguish the three viewing-distance categories, supporting gaze-based distance inference and distance-aware interaction in realistic scenarios.
Read the full abstract
Cryptography and Security
A Broadcast Authenticated Encryption with Keyword Search in the Standard Model: Tightly Secure in Multi-User, Multi-Challenge Settings
However, no known work considered the functionality requirement in its most realistic setting. We propose a new security definition of BAEKS in the multi-user (with adaptive corruptions) and multi-challenge (both in terms of ciphertext and trapdoor in an interleaved manner) settings. We also study the question of the unforgeability of BAEKS. In fact, our strong hiding requirement already implies a significant amount of unforgeability. We then propose a new BAEKS construction in the bilinear pairing groups. We prove this scheme achieves adaptive tight full-hiding security under (almost) standard MDDH assumptions. Restricting our BAEKS construction for a single receiver also gives an efficient and tightly secure PAEKS construction. We further run experiments to implement and evaluate our scheme.
Read the full abstract
Astrophysics of Galaxies
The Cygnus Allscale Survey of Chemistry and Dynamical Environments: CASCADE: VI. Molecular outflows in the DR21 ridge
Star formation takes place in varied environments, from isolated clumps to massive molecular cloud complexes. However, whether the environment in which a star forms has any effect on the formation process remains a matter of debate. The molecular outflows, launched during the formation of protostars present a more easily accessible way to study star formation in different environments. The DR21 ridge, in the Cygnus-X high-mass star-forming complex, hosts a high number of massive dense cores and embedded protostars with outflows. We aim to identify molecular outflows associated with dense molecular cores along the DR21 ridge, and investigate whether the extended environment impacts the formation process of stars within it. We identified molecular outflows along the DR21 ridge using HCO+ J=1-0, H13CO+ J=1-0, and SiO J=2-1 observations obtained with the IRAM 30m telescope and NOEMA as part of the CASCADE program. We calculated outflow properties and performed statistical comparisons between the DR21 ridge sources and a literature sample of low- to high-mass outflow sources. Based on the morphology of HCO+, H13CO+ and SiO, we identify molecular outflows in 14 out of 34 dense cores (41%) along the DR21 ridge. Despite the high density of star formation in DR21, the resulting outflow properties are found to be in good agreement with the established correlations between outflow and source properties. Little variation is seen in the outflow properties of sources along the ridge, with the exception of sources located at the intersection of the DR21 ridge with large-scale (~1 pc) accretion filaments. These sources are found to drive the most powerful outflows. Overall, our results indicate that protostellar outflow properties, even when driven by sources forming in an extreme and clustered region, such as the DR21 ridge, remain largely unaffected.
Read the full abstract
Neurons and Cognition
Rate-Coding Bundle Memory: A Unified Model of Memory and Control for Symbolic Computation in the Brain
We propose a neurobiologically plausible model of cognition that combines the advantages of connectionist and symbolic systems, and that can explain a wide range of cognitive phenomena. This model, called Rate-Coding Bundle Memory (RCBM), is based on the Symbolic Subsystem Hypothesis, which posits that the brain implements a symbolic subsystem within its fundamentally connectionist nature. RCBM is a hybrid model that uses rate coding to represent symbols in a continuous space, and it uses a bundle memory system to store and retrieve these symbols. The model is capable of solving a wide range of cognitive phenomena, including one-shot learning, pattern separation, and the binding problem. We argue that RCBM provides a promising framework for understanding the nature of cognition, and that it can be used to develop more sophisticated models of cognition in the future.
Read the full abstract
Machine Learning
Locked at the Entrance, Open Inside: Where RLVR Narrows the Solution Space
Reinforcement learning with verifiable rewards (RLVR) substantially improves single-sample accuracy (pass@1) but causes the policy's solution space to contract, diminishing the returns of test-time scaling. In this work, we investigate where inside a reasoning trajectory this breadth is lost: does the policy fail to access a valid solution family, or does it fail to execute computation once initiated? To disentangle access from execution, we analyze the Countdown task, whose solution space can be exhaustively enumerated into discrete entrance families defined by the first operand and operator, across PPO on Qwen2.5-3B and GRPO on Qwen2.5-3B-Instruct. Across both training setups, solution coverage falls by up to 67%, halving even on problems solved across all checkpoints. We show that this contraction is heavily concentrated at the entrance: per-token likelihood shifts are 11x–16x larger prior to the first arithmetic operation than during downstream reasoning. Supplying only an unselected entrance prefix restores completion rates in low-access families by over an order of magnitude (0.018 -> 0.212 under PPO), demonstrating that alternative solutions remain executable but are no longer initiated. Guided by this localization, we find that while surface prompting fails to recover diversity, entrance-targeted interventions succeed: late-layer parameter interpolation with early checkpoints increases solution coverage by 37% at no loss in pass@1. Finally, we show that early-step entropy collapse recurs across six math benchmarks with 7B and 14B models, but is not an inevitable byproduct of reasoning optimization: an SFT baseline preserves more than double the coverage, and staged SFT–DPO–RLVR pipelines retain early-step entropy. In summary, reasoning breadth is lost at the door, not inside the room. Code: https://github.com/ershiyidian/early-branch-locking.
Read the full abstract
Computer Vision
OPUS-V2: Bridging the Gap between Sparse Points and Dense Voxels
The point-based occupancy prediction paradigm has achieved an attractive trade-off between accuracy and efficiency by modeling 3D space sparsely. However, its predictions inherently mismatch the dense voxel-based occupancy required by self-driving systems, necessitating hand-crafted heuristics during training and inference that limit final performance. To overcome these limitations, we propose OPUS-V2, a novel framework built upon the pioneering OPUS (occupancy prediction using a sparse set) point-based approach. OPUS-V2 incorporates a lightweight point-voxel transformation (PVT) module behind the decoder to adaptively map sparse predictions into the dense voxel space, eliminating the need for suboptimal operations and improving model accuracy. Furthermore, our architecture decouples feature and occupancy generation processes, allowing OPUS-V2 to adapt to arbitrary occupancy resolutions. OPUS-V2 achieves a state-of-the-art rayIoU of 44.0 on the Occ3D dataset. On the more challenging OpenOccupancy dataset, it attains a competitive 16.4 mIoU while running in real time at 20.6 FPS.
Read the full abstract
Computer Vision
Multi-Scale Temporal Domain Alignment for Federated Video Domain Adaptation
Federated Video Domain Adaptation (FVDA) enables collaborative learning across distributed and non-IID video datasets while preserving privacy, but is under-explored due to challenges in aligning temporal information. We propose Multi-scalE Temporal domAin aLignment (METAL), a novel framework that leverages temporal information at multiple resolutions to improve cross-domain video action recognition with only model parameter transfers. METAL trains per-scale transformer encoders on source-clients, then performs independent knowledge voting at each temporal scale to generate robust pseudo-labels on the target-server. A novel $L_2$ variance penalty enforces cross-scale consistency during scale-based knowledge distillation, preventing a singular dominant scale. The late fusion aggregates features across different scales, where the fusion head is trained via knowledge distillation using confidence-weighted aggregation of scale-wise predictions, enabling the model to effectively exploit complementary temporal information for final predictions. Experiments on Epic-Kitchens-55 and Daily-DA demonstrate state-of-the-art performances, with gains up to 28.47% over current FDA methods. Ablation studies prove that multi-scale distillation and scale coordination are critical for effective temporal knowledge transfer.
Read the full abstract
Quantum Physics
Second harmonic phase locking and synchronization blockade in quadratically coupled driven quantum van der Pol oscillators
We investigate the dynamics of a quadratically coupled system under the influence of an external drive applied to the second oscillator, where the coupling facilitates a high-order synchronization with phase-locking emerging in the form of 2:1 between the oscillators. Our analysis reveals a synchronization blockade in the first oscillator, characterized by the complete suppression of conventional 1:1 phase-locking with the drive. In contrast, we observe that the directly driven second oscillator synchronizes with the drive, showing 1:1 phase-locking but notably at second harmonic frequency. A classical mean-field analysis of the corresponding equations of motion reproduces this asymmetric phase-locking geometry which demonstrates that the phase-locking structure itself can be understood from the nonlinear classical dynamics. The quantum analysis, however, reveals the microscopic origin of the synchronization blockade. Furthermore, we show that the system exhibits mutual synchronization when both the oscillators satisfies the resonance condition, enabling coherent energy exchange facilitated by nonlinear quadratic coupling. The mutual synchronization shows synchronized regimes and also subtle suppression of synchronized regimes near resonance occurring due to spectral splitting of the energy states. Using perturbation analysis of the master equation within the low excitation subspace, we analyze steady-state phase distribution and synchronization measures, supported by population statistics and spectral responses. We also propose possible experimental realizations in trapped-ions and optomechanical setups. These findings highlight the crucial role of quadratic coupling in enabling nonclassical synchronization phenomena, offering deeper insights for quantum control strategies and the development of quantum information platforms.
Read the full abstract
Cryptography and Security
GhostSplat: Input-Triggered Backdoors for Multi-View-Consistent 3D Content Manipulation in Feed-Forward Gaussian Splatting
Feed-forward 3D Gaussian Splatting (3DGS) reconstructs a 3D scene from sparse images in one forward pass. Its shared pretrained weights also expose a supply-chain attack surface. Existing Neural Radiance Field and 3DGS backdoors modify individual scenes and activate at selected viewpoints; they do not install persistent behavior in shared generator weights. We introduce GhostSplat, an input-triggered backdoor that installs such behavior in feed-forward 3DGS. A low-amplitude pattern added to the input images causes the poisoned generator to render an attacker-chosen payload on unseen victim scenes. Anchoring the payload to a 3D point and reprojecting it into each target view makes the payload multi-view consistent. Exact projection onto the generator's representation-specific consistency set leaves a realized payload unchanged because the output already belongs to that set. The GhostSplat training framework succeeds across three architectures (MVSplat, pixelSplat, DepthSplat) and two datasets (RealEstate10K, ACID). Its strongest evaluated injection and deletion settings reach 96% and 100% ASR, respectively, with zero observed false positives while surviving JPEG, blur, and resampling. Defenses that use only that exact projection are therefore insufficient; effective mitigation requires information or intervention beyond same-set consistency projection.
Read the full abstract
Computer Vision
Foundational feature fusion for conditional flow matching in 6D pose estimation
Conditional flow matching has enabled a step forward in object 6D pose estimation, achieving state-of-the-art performance by progressively denoising and registering object representations to observed scenes. Existing methods require training task-specific encoders supervised on object-scene overlap and rely on trivial feature fusion strategies to resolve pose ambiguities. We present FunFlow6D, a novel flow matching-based formulation that leverages features from geometric and appearance foundation models for pose estimation, eliminating the need for task-specific encoder training. We also introduce a cross attention-based fusion mechanism that dynamically combines geometric and appearance features to provide richer conditioning for the flow matching module. Experiments on four datasets from the BOP benchmark show that FunFlow6D outperforms the previous state of the art while reducing supervision requirements and memory overhead. Extensive ablations validate the contribution of each proposed component. Project website: https://tev-fbk.github.io/FunFlow6D/.
Read the full abstract
Signal Processing
Movable Antenna Arrays with Imperfect Channel State Information in Rich Scattering Environments
The growing demand for high spectral efficiency in 6G and beyond has driven research into adaptive antenna architectures capable of exploiting the spatial structure of multipath propagation channels. Conventional base station arrays are deployed with fixed element positions and cannot adapt to the instantaneous spatial structure of the propagation channel. In contrast, movable antenna (MA) systems enable dynamic reconfiguration of antenna positions, allowing the array geometry to track the channel characteristics of the current user set. While prior MA studies have demonstrated significant gains under perfect or statistical channel state information (CSI) assumptions, the interplay between imperfect instantaneous CSI, antenna placement optimization, and precoder design has received limited attention. This paper addresses this gap by proposing a practical end-to-end framework encompassing uplink pilot transmission, MMSE channel estimation under a clustered multipath model, and downlink ZF precoding designed from estimated CSI. Antenna positions are optimized via particle swarm optimization under two objectives: sum-rate maximization and max-min fairness. Simulation results show that MA gains are most pronounced under sparse, near-LoS propagation and diminish as channel richness increases. Furthermore, we reveal a fundamental coupling between array geometry and precoder design: a fairness-oriented antenna geometry encodes spatial fairness information that is only recoverable when evaluated with a compatible power allocation strategy. A mismatched precoder can completely mask the geometric advantage, leading to misleading conclusions about the robustness of antenna placement to the choice of optimization objective. These findings provide practically relevant guidance for the design and evaluation of movable antenna systems under realistic operating conditions.