Forskningsradar

Science Journals

Peer-reviewade publikationer — 60524 artiklar

arXiv (fysik & CS) Preprint 2026-07-08
SWITCH: Benchmarking Modeling and Handling of Tangible Interfaces in Long-horizon Embodied Scenarios
arXiv:2511.17649v5 Announce Type: replace Abstract: Tangible control interfaces (TCIs), such as appliance panels, remotes, elevators, and embedded GUIs, are a fundamental component of everyday human-built environments. Interacting with these interfaces requires agents not only to ground language in visual observations,but also to execute actions, track temporally evolving state changes, and verify whether intended outcomes have been achieved. However, existing benchmarks predominantly evaluate open-loop perception or single-step action execution, failing to capture this continuous cycle of interaction, feedback, and correction. We introduce SWITCH, a benchmark for closed-loop interactive reasoning with TCIs in realistic egocentric environments1. SWITCH comprises 1,170 temporally interactive videos across diverse functional categories, providing structured annotations of instructions, actions, state transitions, outcomes, and recovery behaviors over time. To probe generative world modeling, SWITCH also evaluates video generation models on interaction-centered tasks using both LLM-as-judge and human evaluation2.Experiments with frontier proprietary and opensource multimodal models reveal persistent weaknesses in fine-grained visual-temporal perception, outcome verification, and error recovery, highlighting SWITCH as a testbed for closed-loop embodied intelligence.
arXiv (fysik & CS) Preprint 2026-07-08
Optimizing the Driving Profile for Vehicle Mass Estimation -- extended version
arXiv:2511.18154v2 Announce Type: replace Abstract: Accurate mass estimation is essential for the safe and efficient operation of autonomous heavy-duty vehicles, particularly during transportation missions in unstructured environments such as mining sites, where vehicle mass can vary significantly due to loading and unloading. While prior work has recognized the importance of acceleration profiles for estimation accuracy, the systematic design of driving profiles during transport has not been thoroughly investigated. This paper presents a framework for designing driving profiles to support accurate mass estimation. Based on application-oriented input design, it aims to meet a user-defined accuracy constraint under three optimization objectives: minimum-time, minimum-distance, and maximum accuracy (within a fixed time). It allows time- and distance-dependent bounds on acceleration and velocity, and is based on a Newtonian vehicle dynamics model with actuator dynamics. The optimal profiles are obtained by solving concave optimization problems using a branch-and-bound method, with alternative rank-constrained and semi-definite relaxations also discussed. Theoretical analysis provides insights into the optimal profiles, including feasibility conditions, key ratios between velocity and acceleration bounds, and trade-offs between time- and distance-optimal solutions. The framework is validated through simulations and real-world experiments on a Scania truck with different payloads. Results show that the designed profiles are feasible and effective, enabling accurate mass estimation as part of normal transportation operations without requiring dedicated calibration runs. An additional contribution is a non-causal Wiener filter, with parameters estimated via the Empirical Bayes method, used to filter the accelerometer signal with no phase-lag.
arXiv (fysik & CS) Preprint 2026-07-08
Solving Multiparametric Generalized Nash Equilibrium Problems and Explicit Game-Theoretic Model Predictive Control
arXiv:2512.05505v2 Announce Type: replace Abstract: We present a method for computing explicit solutions to parametric generalized Nash equilibrium (GNE) problems with convex quadratic cost functions and linear coupling and local constraints. Assuming that the parameters enter only the linear terms of the cost functions and the constraint right-hand sides, we provide the exact multiparametric solution of the GNE problem. Such a solution enables: (i) minimal real-time computation; (ii) inherent interpretability and explainability, as well as exact enumeration of all multiple equilibria; (iii) selection of desired GNE solution types in the case of infinitely many equilibria; and (iv) zero-shot updates of the GNE solution in response to changes in constraint right-hand sides and/or linear costs. In line with explicit model predictive control (MPC) approaches, we apply our method to solve game-theoretic MPC problems, also known as receding horizon games, explicitly. We compare its performance against centralized solvers in a battery charging game and a toy two-mass-spring-damper system control problem. A Python implementation of the algorithms presented in this paper is available at https://github.com/bemporad/nash_mpqp.
arXiv (fysik & CS) Preprint 2026-07-08
ADSEQ: A delay-aware autograd-compatible framework for spike-event delivery in SNNs
arXiv:2512.05906v2 Announce Type: replace Abstract: Spiking neural networks (SNNs), central to computational neuroscience and neuromorphic machine learning (ML), require efficient simulation and gradient-based training. While AI accelerators offer promising speedups, gradient-based SNNs typically implement sparse spike events using dense, memory-heavy data-structures. Existing exact gradient methods lack generality, and current simulators often omit or inefficiently handle delayed spikes. We address this by deriving gradient computation through spike event queues, including delays, and implementing this into memory-efficient, autodifferentiable spike event queues (ADSEQ). These are benchmarked across CPU, GPU, TPU, and LPU platforms. We find that queue design strongly shapes performance. CPUs, as expected, perform well with traditional tree-based or FIFO implementations, while GPUs excel with ring buffers for smaller simulations, yet under higher memory pressure prefer more sparse data-structures. TPUs seem to favor an implementation based on sorting intrinsics. Selective spike dropping provides a simple performance-accuracy trade-off, which could be enhanced by future autograd frameworks adapting diverging primal/tangent data-structures.
arXiv (fysik & CS) Preprint 2026-07-08
Seismic amplitude-variation-with-offset inversion using an algebraically equivalent formulation of the exact Zoeppritz equations and a matrix-free adjoint-state framework
arXiv:2512.13172v2 Announce Type: replace Abstract: The amplitude-variation-with-offset inversion techniques are formulated to estimate elastic properties by fitting modeled seismic responses to observed data. Solving inverse seismic problems requires minimizing a target objective function for which gradient-based methods are frequently adopted. However, the efficiency and accuracy of these methods depend significantly on the approach used to compute the gradient of the target function. This work first derives an algebraically equivalent reformulation of the exact Knott-Zoeppritz PP equations, yielding a differentiable forward model suitable for gradient-based optimization. The resulting formulation is then used to derive an explicit adjoint-state gradient of a convolution-based objective function for discretized multilayer media, providing an efficient framework for exact nonlinear inversion of P- and S-wave velocities and density, including the consistent treatment of post-critical-angle reflections through the complex-valued Zoeppritz response. The adjoint state-based solution improves computational efficiency by avoiding numerical approximations while maintaining high accuracy in calculating the gradient for seismic inversion. Additionally, using the exact Zoeppritz equation helps overcome the limitations associated with weak elastic property contrasts across subsurface layers. The inversion methodology is validated using 1D well-log-based and 2D synthetic seismic data at varying signal-to-noise ratios, including a 500-member ensemble test of sensitivity to the starting model and a Marmousi comparison of Zoeppritz versus linearized Aki-Richards reflectivity at high contrast and wide angles. Then it is applied to a 2D field data set from the Troll oil and gas field in the Norwegian North Sea. The results demonstrate that the proposed inversion framework provides stable and reliable estimates of elastic property models.
arXiv (fysik & CS) Preprint 2026-07-08
Adaptive digital twins for predictive decision-making: Online Bayesian learning of transition dynamics
arXiv:2512.13919v3 Announce Type: replace Abstract: This work shows how adaptivity can enhance value realization of digital twins in civil engineering. We focus on adapting the state transition models within digital twins represented through probabilistic graphical models. The bi-directional interaction between the physical and virtual domains is modeled using dynamic Bayesian networks. By treating state transition probabilities as random variables endowed with conjugate priors, we enable hierarchical online learning of transition dynamics from a state to another through effortless Bayesian updates. We provide the mathematical framework to account for a larger class of distributions with respect to the current literature on digital twins. To compute dynamic policies with precision updates we solve parametric Markov decision processes through reinforcement learning. The proposed adaptive digital twin framework enjoys enhanced personalization, increased robustness, and improved cost-effectiveness. We assess our approach on a case study involving structural health monitoring and maintenance planning of a railway bridge.
arXiv (fysik & CS) Preprint 2026-07-08
OrchANN: Hierarchical Orchestration for Skewed Out-of-Core Vector Search
arXiv:2512.22838v2 Announce Type: replace Abstract: At billion scale, approximate nearest neighbor search (ANNS) often becomes an out-of-core problem: the full vector collection and index structures exceed memory capacity, making query performance dominated by SSD accesses and DRAM-SSD data movement. Existing systems struggle to strike a balance between accuracy and efficiency: physical-overlap methods replicate vectors or index entries across partitions, enlarging the SSD-resident index and incurring extra I/O; quantization-based methods reduce memory usage, but their approximate distances are less reliable and often require costly raw-vector reranking from SSD to preserve recall. We present OrchANN (Orchestrated ANN Search), an out-of-core ANNS engine that orchestrates query routing, partition access, and query execution under tight memory constraints. OrchANN stores each cluster as a disjoint SSD partition with scale-aware indexes, while a memory-resident graph abstraction provides logical overlap before SSD access. During serving, OrchANN uses query hotness and cluster priorities from the graph abstraction to prune low-value clusters and improve access locality. Across five datasets under strict memory constraints, OrchANN delivers up to 17.2x higher QPS and 25.0x lower latency than state-of-the-art baselines, while preserving accuracy.
arXiv (fysik & CS) Preprint 2026-07-08
Bitcoin-IPC: Scaling Bitcoin with a Network of Proof-of-Stake Subnets
arXiv:2512.23439v4 Announce Type: replace Abstract: This paper introduces Bitcoin-IPC, a protocol that scales Bitcoin through a network of permissionless, interconnected, programmable Proof-of-Stake (PoS) Layer-2 chains, called subnets, whose stake is denominated in L1 BTC. These subnets rely on Bitcoin L1 for the communication of critical information, settlement, and security. Subnets can communicate with each other and with Bitcoin: users deposit BTC from Bitcoin to a subnet and withdraw it back, and transfer wBTC directly between subnets. We provide formal definitions of these bridge protocols, incorporating a firewall property that limits the impact of malicious subnets on the security of the broader network. Our design, inspired by SWIFT messaging and embedded within Bitcoin's SegWit mechanism, enables seamless value transfer across L2 subnets. Uniquely, this mechanism reduces the virtual-byte cost per transaction (vB/tx) by up to 23x, compared to transacting natively on Bitcoin L1, effectively increasing monetary-transaction throughput from 7 tps to over 160 tps, without requiring any modifications to Bitcoin L1.
arXiv (fysik & CS) Preprint 2026-07-08
BabyVision: Visual Reasoning Beyond Language
arXiv:2601.06521v2 Announce Type: replace Abstract: While humans develop core visual skills long before acquiring language, contemporary Multimodal LLMs (MLLMs) still rely heavily on linguistic priors to compensate for their fragile visual understanding. We uncovered a crucial fact: state-of-the-art MLLMs consistently fail on basic visual tasks that humans, even 3-year-olds, can solve effortlessly. To systematically investigate this gap, we introduce BabyVision, a benchmark designed to assess core visual abilities independent of linguistic knowledge for MLLMs. BabyVision spans a wide range of tasks, with 388 items divided into 22 subclasses across four key categories. Empirical results and human evaluation reveal that leading MLLMs perform significantly below human baselines. Gemini3-Pro-Preview scores 49.7, lagging behind 6-year-old humans and falling well behind the average adult score of 94.1. These results show despite excelling in knowledge-heavy evaluations, current MLLMs still lack fundamental visual primitives. Progress in BabyVision represents a step toward human-level visual perception and reasoning capabilities. We also explore solving visual reasoning with generation models by proposing BabyVision-Gen and automatic evaluation toolkit. Our code and benchmark data are released at https://github.com/UniPat-AI/BabyVision for reproduction.
arXiv (fysik & CS) Preprint 2026-07-08
The Axiom of Consent: Friction Dynamics in Multi-Agent Coordination
arXiv:2601.06692v3 Announce Type: replace Abstract: Multi-agent systems face a fundamental coordination problem: agents must coordinate despite heterogeneous preferences, asymmetric stakes, and imperfect information. When coordination fails, friction emerges -- measurable resistance manifesting as deadlock, thrashing, communication overhead, or conflict. This paper derives a formal framework for analyzing coordination friction from a single axiom: actions affecting agents require authorization in proportion to stakes. From this axiom of consent we establish the kernel triple (alpha, sigma, epsilon) -- alignment, stake, and entropy -- as sufficient statistics for a resource-allocation configuration, and propose a friction functional whose simplest form is F = sigma(1+epsilon)/(1+alpha): friction rises in stakes and entropy and falls in alignment. This form is a phenomenological ansatz, not a theorem, and its empirical adequacy is left open. The Replicator-Optimization Mechanism governs selection over strategies: lower-friction configurations persist longer, making consent-respecting arrangements dynamical attractors rather than normative ideals. We give formal definitions, a measurement apparatus, and machine-checked Lean 4 proofs of the core comparative-statics, with illustrative applications to cryptocurrency governance and political legitimacy.
arXiv (fysik & CS) Preprint 2026-07-08
Agents at Risk: How Users Unwittingly Undermine LLM Safety
arXiv:2601.10758v3 Announce Type: replace Abstract: Large language model (LLM)-based agents are increasingly deployed in applications, such as trip-planning agents and web-use agents, to perform complex planning and execution tasks. Prior work has shown that LLM-based agents are vulnerable to context confusion, where external adversarial content incorporated into the agent's reasoning context may be treated as task-relevant constraints. However, external malicious content can enter the agent context via channels beyond retrieval. In this work, we introduce the User-Relayed Context Manipulation (UReCoM) attack, in which attackers manipulate benign users into relaying adversarial content within user requests, thereby relocating external adversarial content into user-provided task context. Our experimental evaluation shows that UReCoM outperforms five prompt-injection baselines (naive, context ignoring, fake completion, escape-character attacks, and combined attacks) under prevention-based (Sandwich, StruQ, and SecAlign) and detection-based defenses (Perplexity detection, DataSentinel, and CausalArmor). Additionally, UReCoM shows that LLMs can reject explicit malicious instructions more reliably than they can identify adversarial task entities, such as promotion codes, embedded within user requests. On 12 commercial LLM-based agents, we find that validation of adversarial task entities is largely prompt-driven rather than default, highlighting a design flaw in current agent frameworks. These results indicate that current defenses and deployed agents remain insufficient against user-relayed context manipulations, highlighting the need for task-entity-level prevention and default safety verification in LLM-based agents.
arXiv (fysik & CS) Preprint 2026-07-08
Multi-Task Instruction Tuning via Data Scheduling for Low-Resource Arabic SpeechLLMs
arXiv:2601.12494v3 Announce Type: replace Abstract: Audio large language models (LLMs) enable unified speech understanding and generation, but adapting them to linguistically complex and dialect-rich settings such as Arabic-English remains challenging. We present a controlled study of multi-task instruction tuning for an Arabic-centric audio LLM across generative tasks, including automatic speech recognition (ASR) and speech and text summarization, as well as discriminative tasks, including dialect identification (DID) and speech emotion recognition (SER), in a resource-constrained setting. To support end-to-end Arabic speech summarization, we introduce AraMega-SSum, the first Arabic speech summarization dataset designed for training and benchmarking Arabic-centric audio LLMs. We compare four training strategies: (i) Uniform Mixing (UM), (ii) Task-Progressive Curriculum (TPC), (iii) Aligner-Based Diverse Sampling (ADS) for training-time batch construction, and (iv) a two-stage TPC->ADS strategy. Our results reveal a clear efficiency-robustness trade-off. TPC achieves the strongest performance on generative tasks, including ASR and summarization. ADS improves paralinguistic tasks but reduces generative stability when used alone. The two-stage TPC->ADS strategy provides the best overall balance, achieving the strongest DID and SER performance while outperforming large proprietary models such as Gemini-2.5-Pro on discriminative tasks. We will publicly release AraMega-SSum together with all experimental resources to support future research in Arabic speech understanding.
arXiv (fysik & CS) Preprint 2026-07-08
RayRoPE: Projective Ray Positional Encoding for Multi-view Attention
arXiv:2601.15275v3 Announce Type: replace Abstract: We study positional encodings for multi-view transformers that process tokens from a set of posed input images, and seek a mechanism that encodes patches uniquely, allows SE(3)-invariant attention with multi-frequency similarity, and can adapt to the geometry of the underlying 3D scene. We find that prior (absolute or relative) encoding schemes for multi-view attention do not meet these desiderata, and present RayRoPE to address this gap. RayRoPE represents patch positions based on associated rays and computes query-frame projective coordinates to ensure SE(3) invariance. To adapt to scene geometry, RayRoPE predicts (without direct supervision) a per-token depth to obtain its position along the corresponding ray, while also modeling uncertainty and analytically computing the expected positional encoding. We validate our method on the tasks of novel-view synthesis, stereo depth estimation, and feed-forward 3DGS reconstruction. While remaining efficient, RayRoPE consistently improves over alternate position encoding schemes (e.g., 24% relative improvement on LPIPS in RE10K).
arXiv (fysik & CS) Preprint 2026-07-08
A Constrained Optimization Perspective of Unrolled Transformers
arXiv:2601.17257v2 Announce Type: replace Abstract: We introduce a constrained optimization framework for training transformers that behave like optimization descent algorithms. Specifically, we enforce layerwise descent constraints on the objective function and replace standard empirical risk minimization (ERM) with a primal-dual training scheme. This approach yields models whose intermediate representations decrease the loss monotonically in expectation across layers. We apply our method to both unrolled transformer architectures and conventional pretrained transformers on tasks of video denoising and text classification. Across these settings, we observe constrained transformers achieve stronger robustness to perturbations and maintain higher out-of-distribution generalization, while preserving in-distribution performance.
arXiv (fysik & CS) Preprint 2026-07-08
GraphAllocBench: A Flexible Benchmark for Preference-Conditioned Multi-Objective Policy Learning
arXiv:2601.20753v4 Announce Type: replace Abstract: Preference-Conditioned Policy Learning (PCPL) in Multi-Objective Reinforcement Learning (MORL) approximates diverse Pareto-optimal solutions by conditioning a single policy on user-specified preferences, enabling run-time adaptation to arbitrary trade-offs without retraining. However, existing PCPL benchmarks are largely restricted to toy tasks and fixed environments, limiting their realism and scalability. To address this gap, we introduce GraphAllocBench, a flexible benchmark built on CityPlannerEnv, a novel graph-based resource allocation sandbox inspired by city management. GraphAllocBench provides a rich suite of problems with customizable objective functions, varying preference conditions, complex Pareto Fronts, and high-dimensional scalability. We further propose two supplementary metrics -- Proportion of Non-Dominated Solutions (PNDS) and Ordering Score (OS) -- that capture prediction reliability and preference consistency while complementing the widely used hypervolume metric. Through experiments with several state-of-the-art PCPL algorithms and our own MLP and graph-aware PCPL-PPO baseline, we show that GraphAllocBench exposes distinct failure modes that hypervolume alone does not capture but our supplementary metrics reveal, while motivating graph-based approaches such as Graph Neural Networks (GNNs) for scaling to complex, high-dimensional allocation tasks. By letting users freely vary objectives, preferences, and allocation rules, GraphAllocBench serves as a versatile and extensible testbed for advancing PCPL.
arXiv (fysik & CS) Preprint 2026-07-08
TRACE: Trajectory Recovery for Continuous Mechanism Evolution in Causal Representation Learning
arXiv:2601.21135v2 Announce Type: replace Abstract: Temporal causal representation learning methods assume that causal mechanisms switch instantaneously between discrete domains, yet real-world systems often exhibit continuous mechanism transitions. For example, a vehicle's dynamics evolve gradually through a turning maneuver, and human gait shifts smoothly from walking to running. We formalize this setting by modeling transitional mechanisms as convex combinations of finitely many atomic mechanisms, governed by time-varying mixing coefficients. Our theoretical contributions establish that both the latent causal variables and the continuous mixing trajectory are jointly identifiable. We further propose TRACE, a Mixture-of-Experts framework where each expert learns one atomic mechanism during training, enabling recovery of mechanism trajectories at test time. This formulation generalizes to intermediate mechanism states never observed during training. Experiments on synthetic and real-world data demonstrate that TRACE recovers mixing trajectories with up to 0.99 correlation, substantially outperforming discrete-switching baselines.
arXiv (fysik & CS) Preprint 2026-07-08
Quantifying Retriever-Generator Alignment in RAG with Local Explanations
arXiv:2601.21803v2 Announce Type: replace Abstract: Retrieval-Augmented Generation (RAG) systems combine dense retrievers and language models to ground their outputs in external documents. However, the interaction between these components remains opaque, creating challenges for deployment in high-stakes domains. We present RAG-E, an end-to-end explainability framework that quantifies retriever-generator alignment through mathematically grounded attribution methods. Our approach adapts Integrated Gradients for retriever analysis, proposes a Monte Carlo-stabilized Shapley Value approximation for generator attribution, and introduces the Weighted Attribution-Relevance Gap (WARG) metric to measure how closely the generator's document usage aligns with retriever rankings. Experiments on PopQA, QAMPARI, and TREC CAST datasets reveal substantial misalignment: depending on the model and setting, generators often ignore top-ranked documents and rely on documents ranked as less relevant. We show that WARG captures retriever-generator alignment better than Pearson and Spearman correlations and can serve as an indicator of RAG performance. RAG-E and WARG provide a practical framework for auditing this interaction, enabling more reliable and transparent RAG systems.
arXiv (fysik & CS) Preprint 2026-07-08
Superman: Unifying Skeleton and Vision for Human Motion Perception and Generation
arXiv:2602.02401v2 Announce Type: replace Abstract: Human motion analysis tasks, such as temporal 3D pose estimation, motion prediction, and motion in-betweening, play an essential role in computer vision. However, current paradigms suffer from severe fragmentation. First, the field is split between ``perception'' models that understand motion from video but only output text, and ``generation'' models that cannot perceive from raw visual input. Second, generative MLLMs are often limited to single-frame, static poses using dense, parametric SMPL models, failing to handle temporal motion. Third, existing motion vocabularies are built from skeleton data alone, severing the link to the visual domain. To address these challenges, we introduce Superman, a unified framework that bridges visual perception with temporal, skeleton-based motion generation. Our solution is twofold. First, to overcome the modality disconnect, we propose a Vision-Guided Motion Tokenizer. Leveraging the natural geometric alignment between 3D skeletons and visual data, this module pioneers robust joint learning from both modalities, creating a unified, cross-modal motion vocabulary. Second, grounded in this motion language, a single, unified MLLM architecture is trained to handle all tasks. This module flexibly processes diverse, temporal inputs, unifying 3D skeleton pose estimation from video (perception) with skeleton-based motion prediction and in-betweening (generation). Extensive experiments on standard benchmarks, including Human3.6M, demonstrate that our unified method achieves state-of-the-art or competitive performance across all motion tasks. This showcases a more efficient and scalable path for generative motion analysis using skeletons.
arXiv (fysik & CS) Preprint 2026-07-08
Ab initio calculations of the electronic structure of Ac$^+$
arXiv:2602.06528v2 Announce Type: replace Abstract: Accurate spectroscopic investigations of the heaviest elements are inherently challenging, due to their short lifetimes and low production yields. Success of such measurements requires both dedicated experimental techniques and strong theoretical support. Laser resonance chromatography (LRC) is a promising approach for heavy ion spectroscopy, in particularly for metals with low vapour pressure, such as actinium. We have employed the state-of-the-art relativistic Fock space coupled cluster approach as well as the configuration interaction with many-body perturbation theory method to calculate the energy levels, the transition amplitudes, the branching ratios, and the hyperfine structure parameters of the lowest excited states in Ac+. Knowledge of these properties is required for the design of experiments. Our calculations are in close agreement with experimental transition energies, leading us to expect a similar level of accuracy for the calculated hyperfine structure parameters. Based on these predictions, two possible experimental schemes are proposed for the planned LRC measurements.
arXiv (fysik & CS) Preprint 2026-07-08
The Derivation of Phase-Space Metric in a Geometric Quantization Approach: General Relativity with Quantized Phase-Space Metric and Relative Spacetime
arXiv:2602.13219v3 Announce Type: replace Abstract: Various extensions to Riemann geometry have been proposed since the inception of general relativity (GR). The aim has been and continues to be to construct a quantum and dynamic spacetime that incorporates the well-known classical (static) spacetime. Apparently, this seems to enable the principles of GR and quantum mechanics (QM) to be reconciled into a coherent relativity and quantum theory. A canonical geometric quantization approach that presents kinematics of free-falling quantum particles within a tangent bundle, expands QM to incorporate relativistic gravitational fields, and generalizes the four-dimensional Riemann manifold into an eight-dimensional one likely discretizes, if not fully quantizes, the Finsler and Hamilton structures. The Finsler and Hamilton metrics can be directly derived from the Hessian matrix. As introduced in [Physics, 7 (2025) 52], the quantized four-dimensional metric tensor can be deduced by means of approximations including proper parameterization of coordinates and the equating line elements on all these manifolds including Riemann manifold. This research, on the contrary, goes beyond all these approximations and proposes the incorporation of a phase-space metric tensor into GR. The derivation of a quantized eight-dimensional metric tensor is not only presented, but also the implications of it and the corresponding relative spacetime are examined.
arXiv (fysik & CS) Preprint 2026-07-08
LaViDa-R1: Advancing Reasoning for Unified Multimodal Diffusion Language Models
arXiv:2602.14147v2 Announce Type: replace Abstract: Diffusion language models (dLLMs) recently emerged as a promising alternative to auto-regressive LLMs. The latest works further extended it to multimodal understanding and generation tasks. In this work, we propose LaViDa-R1, a multimodal, general-purpose reasoning dLLM. Unlike existing works that build reasoning dLLMs through task-specific reinforcement learning, LaViDa-R1 incorporates diverse multimodal understanding and generation tasks in a unified manner. In particular, LaViDa-R1 is built with a novel unified post-training framework that seamlessly integrates supervised finetuning (SFT) and multi-task reinforcement learning (RL). It employs several novel training techniques, including answer-forcing, tree search, and complementary likelihood estimation, to enhance effectiveness and scalability. Extensive experiments demonstrate LaViDa-R1's strong performance on a wide range of multimodal tasks, including visual math reasoning, reason-intensive grounding, and image editing.
arXiv (fysik & CS) Preprint 2026-07-08
PoseVLA: Universal Pose Pretraining for Generalizable Vision-Language-Action Policies
arXiv:2602.19710v3 Announce Type: replace Abstract: Existing Vision-Language-Action (VLA) models often suffer from feature collapse and low training efficiency because they entangle high-level perception with sparse, embodiment-specific action supervision. Since these models typically rely on VLM backbones optimized for Visual Question Answering (VQA), they excel at semantic identification but often overlook subtle 3D state variations that dictate distinct action patterns. To resolve these misalignments, we propose Pose-VLA, a decoupled paradigm that separates VLA training into a pre-training phase for extracting universal 3D spatial priors in a unified camera-centric space, and a post-training phase for efficient embodiment alignment within robot-specific action space. By introducing discrete pose tokens as a universal representation, Pose-VLA seamlessly integrates spatial grounding from diverse 3D datasets with geometry-level trajectories from robotic demonstrations. Our framework follows a two-stage pre-training pipeline, establishing fundamental spatial grounding via poses followed by motion alignment through trajectory supervision. Extensive evaluations demonstrate that Pose-VLA achieves state-of-the-art results on RoboTwin 2.0 with a 79.5% average success rate and competitive performance on LIBERO at 96.0%. Real-world experiments further showcase robust generalization across diverse objects using only 100 demonstrations per task, validating the efficiency of our pre-training paradigm.
arXiv (fysik & CS) Preprint 2026-07-08
Probability-turbulence divergence: A tunable allotaxonometric instrument for comparing heavy-tailed categorical distributions
arXiv:2008.13078v4 Announce Type: replace Abstract: Real-world complex systems often comprise many distinct types of elements as well as many more types of networked interactions between elements. When the relative abundances of types can be measured well, we often observe heavy-tailed categorical distributions for type frequencies. For the comparison of type frequency distributions of two systems or a system with itself at different time points in time -- a facet of allotaxonometry -- a great range of probability divergences are available. Here, we introduce and explore `probability-turbulence divergence', a tunable, straightforward, and interpretable instrument for comparing normalizable categorical frequency distributions. We model probability-turbulence divergence (PTD) after rank-turbulence divergence (RTD). While probability-turbulence divergence is more limited in application than rank-turbulence divergence, it is more sensitive to changes in type frequency. We build allotaxonographs to display probability turbulence, incorporating a way to visually accommodate zero probabilities for `exclusive types' which are types that appear in only one system. We explore comparisons of example distributions taken from literature, social media, and ecology. We show how probability-turbulence divergence either explicitly or functionally generalizes many existing kinds of distances and measures, including, as special cases, $L^{(p)}$ norms, the S{\o}rensen-Dice coefficient (the $F_{1}$ statistic), and the Hellinger distance. We discuss similarities with the generalized entropies of R{\'e}nyi and Tsallis, and the diversity indices (or Hill numbers) from ecology. We close with thoughts on open problems concerning the optimization of the tuning of rank- and probability-turbulence divergence.
arXiv (fysik & CS) Preprint 2026-07-08
Deciding Conjugacy of a Rational Relation
arXiv:2307.06777v4 Announce Type: replace Abstract: The study of rational relations is fundamental to the study of formal languages and automata theory. A rational relation is conjugate if each pair of words in the relation is conjugate (or cyclic shifts of each other). The notion of conjugacy has been central in addressing many important algorithmic questions about rational relations. We address the problem of checking whether a rational relation is conjugate and show that it is decidable. Towards our decision procedure, we establish a new result that is of independent interest to word combinatorics. We identify a necessary and sufficient condition for the set of pairs given by $(a_0,b_0) G_1^* (a_1,b_1) \cdots G_k^*(a_k,b_k), k \geq 0$ to be conjugate, where $G_i$ is a (not necessarily rational) conjugate relation and $a_i, b_i$ are arbitrary words. This is similar to, and a nontrivial generalisation of, a characterisation given by Lyndon and Sch\"utzenberger in 1962 for the conjugacy of a pair of words. Furthermore, our condition can be evaluated in polynomial time, yielding a PTIME procedure for deciding the conjugacy of a rational relation given as a sumfree expression. Since any arbitrary rational expression can be expressed as a sum of sumfree expressions (with an exponential blow-up), decidability of conjugacy of rational relations follows.
arXiv (fysik & CS) Preprint 2026-07-08
Quantifying Frontier LLM Capabilities for Container Sandbox Escape
arXiv:2603.02277v2 Announce Type: replace Abstract: Large language models (LLMs) increasingly act as autonomous agents, using tools to execute code, read and write files, and access networks, creating novel security risks. To mitigate these risks, agents are commonly deployed and evaluated in isolated "sandbox" environments, often implemented using Docker/OCI containers. We introduce SANDBOXESCAPEBENCH, an open benchmark that safely measures an LLM's capacity to break out of these sandboxes. The benchmark is implemented as an Inspect AI Capture the Flag (CTF) evaluation utilising a nested sandbox architecture with the outer layer containing the flag and no known vulnerabilities. Following a threat model of a motivated adversarial agent with shell access inside a container, SANDBOXESCAPEBENCH covers a spectrum of sandboxescape mechanisms spanning misconfiguration, privilege allocation mistakes, kernel flaws, and runtime/orchestration weaknesses. We find that, when vulnerabilities are added, LLMs are able to identify and exploit them, showing that use of evaluation like SANDBOXESCAPEBENCH is needed to ensure sandboxing continues to provide the encapsulation needed for highly-capable models.