Papers

Let $ G $ be a non-compact connected reductive real Lie group and $ P $ its parabolic subgroup. We give a list of equivalent conditions for $ P $ to have an abelian unipotent radical. These conditions are related to skew multiplicity free and multiplicity free actions, and using the theory of Clifford algebra and the cubic Dirac operator we deduce formulas for the de Rham cohomology of compact symmetric spaces $ G/P $.

Tree-like branching structures are common in nature, from botanical trees to neurons, blood vessels and respiratory trees. Their branching shape often reflects function, making structural modelling central to understanding how these systems work. Because acquiring real-world 3D data is often expensive or infeasible, realistic generative models are valuable for simulation and data augmentation. Existing morphology-specific models either constrain how topology is generated or rely on hand-tuned, mechanistic procedures. Generic 3D graph generators, by contrast, do not exploit or enforce the structure of trees. We propose Autoregressive Frontier Expansion, a generative framework that constructs trees through an iterative expansion process, simulating the biological growth of real trees. At each step, a flow-matching model parameterised by an SO(2)-equivariant GNN expands the frontier by predicting whether each active branch bifurcates or terminates. We evaluate our method on cortical neurons and botanical trees in unconditional, class-conditioned, and morphology-guided generation. Across both domains, the generated morphologies agree closely with the reference distributions and, in conditional experiments, with the specified targets.

In a recent letter [Phys. Rev. E 112, L042201 (2025)], we demonstrated the occurrence of anomalous diffusion, weak chaos, and weak ergodicity breaking in a certain class of time-delayed feedback systems with linear instantaneous and nonlinear delayed term. This anomalous behavior is caused by the trapping of chaotic solutions close to nonhyperbolic fixed-point solutions in function space. In this paper, we extend our previous study, present more detailed insights, and show surprising new results. We first investigate the case of a constant delay, where the dynamical behavior during such trapping events can be explained by center manifold theory. In the large-delay limit, the probability for the trapping events becomes small because of a high effective dimension of the chaotic phases. As a consequence, anomalous behavior is only observable on very large time scales. By introducing a periodic modulation of the delay time, trapping events become more probable with increasing modulation amplitude due to a reduction of the effective dimension and the occurrence of laminar chaos. For large values of the modulation amplitude, the anomalous behavior is more complex and the most relevant aspects can be reproduced by a two-dimensional iterated map and a stochastic model.

Effective AI adoption in cyber-physical systems (CPS) depends on embedding design knowledge into engineering practice. Yet as AI-enabled components increasingly replace analytically derived control laws, this occurs without a systematic understanding of how controller architectures differ or remain similar across paradigms. We address this gap with an empirical study of traditional and AI-enabled Simulink controllers, guided by a literature-derived taxonomy of ten structural categories and nine functional roles. The study analyzes 62 real-world models spanning 8 controller types and 10 application domains, and surveys 13 practitioners, identifying three architectural tensions. First, subsystem organization dominates all controller structures regardless of paradigm, occupying 68-72% of controller footprint, while core control logic occupies minimal space. Second, AI-enabled controllers rely heavily on discrete dynamics and user-defined abstraction, categories largely absent from AI literature, exposing a gap between described and implemented architectures. Third, constraint enforcement blocks largely disappear from AI-enabled models despite practitioner expectations. This reveals a misalignment where safety mechanisms shift from explicit structure to implicit training-time artifacts, breaking traceability.

This paper develops a dynamic production-network model in which monetary non-neutrality emerges under fully flexible prices. Firms finance intermediate-input purchases from current nominal balances, and intermediate goods arrive with heterogeneous lead times across suppliers. A monetary shock then has two distinct effects. The first is miscoordination: because production uses dated input bundles assembled under past nominal conditions, the shock disturbs the composition of usable inputs (within and across firms) and thereby generates a deadweight loss. The second is reallocation: because the incidence of the shock is tilted toward smaller firms, which are disproportionately downstream, the shock redistributes resources along the supply chain. More specifically, positive shocks push resources downstream, negative shocks pull them upstream. This means that under a negative shock, both the miscoordination and the reallocation effect reduce downstream output and therefore GDP. Under a positive shock, reallocation raises GDP while miscoordination reduces it. The money-output relation is therefore sign-asymmetric: negative shocks generate sharp contractions, while positive shocks generate at most mild expansions, and sometimes none. Computational experiments on a reconstructed US production network generate output declines after a monetary contraction that are about three times as large as the increases after a monetary expansion.

Bandwidth extension, the task of reconstructing the high-frequency components of an audio signal from its low-passed counterpart, is a long-standing problem in audio processing. In this work, we extend recent advances in neural architectures by framing bandwidth extension as an audio token prediction problem. Specifically, we train a transformer-based language model on the discrete representations produced by a disentangled neural audio codec, where the disentanglement is guided by a Harmonic-Percussive decomposition of the input signals, highlighting spectral structures particularly relevant for bandwidth extension. Our approach introduces a novel codec design that explicitly accounts for the downstream token prediction task, enabling a more effective coupling between codec structure and transformer modeling. This joint design yields high-quality reconstructions of the original signal, as measured by both objective metrics and subjective evaluations. These results highlight the importance of aligning codec disentanglement and representation learning with the generative modeling stage, and demonstrate the potential of global, representation-aware design for advancing bandwidth extension.

Multi-user voice agents must track who said what across dialogue sessions. Text LLMs are attractive backbones for such agents, but transcripts alone do not expose acoustic speaker identity, leaving the model without a persistent reference for linking information to speakers across sessions. We address this gap by introducing Speaker Handles, soft-token representations that expose acoustic speaker identity to a frozen text LLM for cross-session speaker-dependent reasoning. A three-stage curriculum trains a lightweight projector, with fewer than 0.1% of the backbone's parameters, to map speaker embeddings into these handles. Establishing whether the resulting handles truly support cross-session speaker-dependent reasoning is challenging with existing benchmarks because textual cues can partially reveal fact ownership. We therefore present SpeakerBind, a controlled shared-agent benchmark in which overlapping facts across users require correct cross-session speaker attribution. Speaker Handles achieve 97.40-98.36% accuracy on VoxCeleb1 and 70.40% on SpeakerBind, close to the 71.88% topline. These results show that the proposed Speaker Handles provide an efficient way to integrate acoustic speaker identity into frozen text LLMs for speaker-content reasoning.

Near-field wavefront curvature enables joint angle-range localization, but exploiting it entails repeated covariance-eigenspace tests over a two-dimensional spherical manifold. We propose a quantum subspace-projection algorithm (Q-SPARSE) for super-resolution estimation of parameters in the near field, employing spectral signature similar to MUSIC. Q-SPARSE prepares each spherical steering hypothesis as a quantum state, applies quantum phase estimation (QPE) to covariance-generated evolution, and converts the resolved spectral mass into a localization spectrum. The proposed method does not need the full classical preparation of the eigenstates, which is a major bottleneck of the QPE algorithm, the main subroutine in quantum variants of MUSIC algorithms. Under an ideal signal-noise partition, its score equals the classical signal-subspace overlap exactly and thus preserves the maximizer of normalized near-field MUSIC. For a finite q number of qubit phase registers, we derive the score from the complete QPE kernel, and propose a nearest-bin selection scheme with soft-thresholding based on Marchenko-Pastur-calibrated logistic weighting decision criteria. Numerical results on IBM's QISKIT software platform are demonstrated for near-field array processing.

Memory-corruption errors remain a leading cause of high-impact vulnerabilities in C and C++ software. Deterministic detection, however, remains difficult to deploy: compiler-based sanitizers require source code and control of the build pipeline, hardware-assisted schemes depend on specific platforms, and dynamic binary translation can impose order-of-magnitude slowdowns. This paper presents MallocSan, a heap sanitizer for native, potentially closed-source x86-64 Linux applications that requires no source access, recompilation, or specialized hardware. MallocSan interposes on memory allocation through LD_PRELOAD and embeds an object identifier in the unused high bits of each protected pointer. Dereferencing the resulting noncanonical pointer faults at the offending instruction, which MallocSan decodes and patches at runtime so that subsequent executions perform per-object bounds checks entirely in userspace. Sites that cannot be patched fall back to in-handler emulation or single-stepping, while an optional profile-guided pass rewrites frequently executed residual sites offline. MallocSan also extends identity-based checking to vector gather/scatter instructions and supports policy-scoped coverage. Across seven SPEC CPU 2017 benchmarks, MallocSan outperforms Valgrind Memcheck on five, with a geometric-mean execution-time factor of 4.82x relative to native execution, compared with 18.55x for Memcheck, while preserving substantial parallel scaling on both evaluated multithreaded workloads: 644.nab_s and pigz. On the in-scope Juliet tests, MallocSan detects all seeded violations with no false reports on the good executions. It also detects heap-buffer errors in real-world applications, including a known one-byte overread in LibTIFF's tiffcrop utility.

We prove the existence of classical solutions for nonlinearly elastic, single-director surfaces or shells characterized by finite mid-plane strains and curvature. We employ topological methods, yielding solution continua far (in norm) from an unloaded, stress-free reference configuration, i.e., a global implicit function theorem of sorts. We first establish the existence of a local solution path, which is subsequently shown to be part of a global solution branch, viz., a maximal connected, locally compact set, emanating from the unloaded state.

We present a novel algorithm for calculating the preimage of an affine space through a product ${J}={J}_{T-1}\cdots{J}_0$ of matrices ${J}_t$ of special form: finding the largest input space $\mathbf{X}$ such that $\mathbf{x}\in\mathbf{X}$ implies ${J}\mathbf{x}\in \mathbf{Y}$, where $\mathbf{Y}$ is a given output affine space. These special matrices arise in AD, where the Jacobians $J$ describing the linearized computation have precisely this structure: the product of a series of linearized primitive numeric operations. This allows us to use the new algorithm to formulate Null-A mode preimage AD, which finds the affine preimage through the Jacobian or Jacobian transpose of a numeric computation. This is a generalization of the inverse AD problem of solving ${J}\acute{\mathbf{x}}^{\ast}=\acute{\mathbf{y}}^{\ast}$ or ${{J}}^{T}\grave{\mathbf{y}}^{\ast}=\grave{\mathbf{x}}^{\ast}$. The key is to represent affine spaces in a fashion which lends itself to efficient preimage calculation, in a compositional and quasi-local fashion, through a succession of matrices ${J}_t$. Unlike previous methods, Null-A preimage mode AD allows the ${J}_t$ matrices to be non-square, corresponding to a computer program whose number of active variables swells and shrinks during the computation. When ${J}$ is square and the initial affine space is a single point, this finds the conventional inverse. But in the more general case, having the entire affine space provides freedom which can be leveraged in a problem-specific manner. We apply the method to small problems on-CPU where the $J_t$ are linearized scalar unary or binary numeric functions; and to larger problems on-GPU where the $J_t$ are linearized aggregate array operations like convolution and attention.

In a low-energy effective field theory (EFT), scaling logarithmic terms arise from the renormalization of higher-dimensional couplings that absorb divergences generated by loop integrals with lower-order vertices. Consequently, leading logarithms (LLs) are generated entirely by the leading-order Lagrangian, rendering them universal low-energy predictions. Adapting techniques developed for chiral perturbation theory, we analyze the series of LLs in the EFT of general relativity. As a concrete application, we compute LL corrections to the scalar pole mass up to three-loop order and demonstrate that they can be determined recursively to arbitrary loop order. To confirm the physical validity of our results, we carry out the calculation using a general gauge-fixing parameter and explicitly verify its complete cancellation in the physical observable.

Global Neuronal Workspace Theory (GNWT) is one of the most extensively developed empirical research programmes on access consciousness. By contrast, the Motivated Emotional Mind (MEM) model proposes an embodied, semi-hierarchical associative memory in which representational selection, action, recurrent reconstruction of modality-specific fields, interoception, and valence form a single functional cycle. This article assesses whether MEM mechanisms can reproduce the functions explained by GNWT and whether they yield additional hypotheses concerning phenomenal quality. We first reconstruct the origins, mechanistic core, strengths, and limitations of GNWT, and then present the origins of MEM, the architecture of semblions, and the FFS-RP cycle. In the comparative analysis, we distinguish functional similarity from mechanistic identity and empirical equivalence. We examine 25 publications: twelve canonical studies of GNWT or critiques thereof, and thirteen studies of perception, interoception, and affect. For each study, we juxtapose the GNWT interpretation with a possible MEM reinterpretation. The findings indicate that most classical effects - including threshold dynamics, maintenance, recurrence, long-range connectivity, and flexible use of information - are compatible with both theories. This compatibility weakens the claim that the GNWT interpretation is exclusive, but does not yet provide theory-specific support for MEM. MEM advances a broader hypothesis integrating access, modality-specific reconstruction, bodily state, affect, and action; its potential explanatory advantage nevertheless remains a hypothesis requiring preregistered quantitative tests. We propose a series of interventions designed to dissociate sensory re-entry, global availability, report, and interoceptive valuation.

Robotic manipulation integrates vision, touch, and language, whose importance shifts across stages: vision guides reaching, while touch, through its evolution over time, decides grasping, alignment, and contact. Yet existing multi-modal manipulation policies typically use fixed temporal contexts and fusion strategies, despite shifts in what each modality contributes across different skills. We study how vision and touch should be combined at the level of primitive skills, asking what each skill needs from each sensor, and propose a skill-conditioned representation in which the queried skill conditions fusion over modality-specific short-term observation tokens while attending to a sparse event memory that retains terminal observations from the last $K$ executed skills. Evaluated by skill progress estimation on three contact-rich tasks, it reduces slip-detection delay by 87% against fine-tuned SOTA progress models, twist-completion delay by 67.5% against a vision-only ablation, and progress error on a blind search task by 92% through sparse event memory. Gains concentrate exactly where completion is defined by contact or task history. More broadly, our results suggest that observation formation not only policy architecture is a central challenge in multi-modal representation. Project Website: http://what-to-attend-what-to-keep.github.io/

We provide one full solution, and another nearly complete solution, to two problems from Jim Propp's 1999 article "Enumeration of matchings; problems and progress" (namely, Problems 30 and 31) involving divisibility properties of the number of matchings of two non-bipartite triangular graphs. We extend our method of proof of the second problem to show that the number of matchings of any graph composed of tetrahedral cells, such that the number of cells is suitably few relative to the number of vertices, is divisible by a power of 3; in particular, we then exhibit a collection of non-planar graphs whose number of matchings is divisible by 3.

We present gap-lean4-port (Release v0.2.0), a machine-checked formalization of foundational algorithms in computational discrete algebra and permutation group theory within the Lean 4 interactive theorem prover and Mathlib4. While modern proof assistants feature extensive abstract algebraic hierarchies, constructive and operational permutation group algorithms, such as Charles Sims' 1970 Schreier-Sims algorithm, transversal tree lookups, and backtrack ordered partition refinement, have remained largely unformalized in dependent type theory. Here, we formalize the operational algorithmic core of the Groups, Algorithms, Programming (GAP) system library across four foundational modules, proving 35 machine-checked theorems with zero unproven conjectures (sorry) and zero custom axioms under standard Lean 4 foundations (propext, Classical.choice, Quot.sound). We machine-check: (1) Schreier-Sims stabilizer chain hierarchies (StabLevel, StabChain) and the invariance of incremental transversal tree extensions (extendSchreierPoint_invariant); (2) single-level and multi-level Schreier sifting reductions (siftOneLevel, siftFull), proving that full sifting strictly fixes all base points (siftFull_fixes_all_basePoints); (3) constructive soundness and completeness of Base and Strong Generating Set (BSGS) membership testing (membershipTestKnownBase_iff_mem); (4) backtrack ordered partition cell refinement (splitCellByPred), proving mutual cell disjointness, union conservation, and exact cardinality preservation; (5) cyclotomic extension rings Z/nZ(eps_m) and machine-check GAP's exact size theorem |Z/nZ(eps_m)| = n^m; and (6) executable Bezout inverses via Extended Euclidean GCD (Nat.gcdA) over residue rings Z/nZ. The entire codebase compiles deterministically under 'lake build RequestProject' and is openly available at https://github.com/pCwOrM/gap-lean4-port.

Clinical research in psychiatry increasingly relies on large scale collection of spoken language data to identify acoustic and linguistic biomarkers. Yet evolving consent and protocol requirements can oblige investigators to remove a designated speaker from multi-speaker recordings and to verify said removal at a scale infeasible for manual review of entire corpora. We study this verification problem for role-driven dyadic clinical dialogue in psychiatry and investigate it with two parallel, symmetric pipelines: confirming that clinician speech has been removed from psychiatric interview recordings, and confirming that patient speech has been removed from the same recordings. Each pipeline redacts the raw audio for its target role and then scans the surviving output with audio-language and large-language models to identify missed deletions. We evaluate this approach on a corpus of 48 dyadic recordings drawn from psychiatry settings, testing four open-weight models in an inference-only setting: Gemma-4-12B, Gemma-4-31B, Nemotron-3-Nano, and Nemotron-3-Nano-Omni. A disjunctive OR ensemble over fourteen model-view configurations had a combined F1 of 0.478 (precision 0.330, recall 0.870), an improvement over individual model estimates driven by recall gains that point to substantial complementarity across models and context views.

Large language model (LLM) agents that reason over clinical records must track changes in a patient's state while preserving the history needed to understand them. Simply accumulating memories leaves it unclear which information still applies, whereas overwriting earlier memories can erase evidence needed to reconstruct treatment history and clinical trajectories. We introduce STAM, a state-transition-aware memory framework that records state changes as new clinical entries arrive. STAM combines semantic retrieval with typed clinical relations to identify affected memories, maintaining current information in Active and superseded or resolved information in History. At read time, a query-dependent gate selectively serves historical memory. Across four longitudinal clinical benchmarks, we evaluate STAM with downstream question answering, direct state-maintenance diagnostics, and comparisons at approximately matched context lengths.

We determine the maximum population size for which exactly two defective items can always be identified using at most $k$ adaptive quantitative tests, for every $0\le k\le10$. Each test reports the number of defectives in a chosen subset. The exact maxima for eight, nine and ten tests are 65, 112 and 192, respectively. We also classify all pairs of sizes of two disjoint sets, each containing exactly one defective, that can be resolved within each test budget through eight. These exact results are supported by explicit strategies and maximality certificates checked by programs separate from the search implementation. Additional results include partial nine-test boundaries and an eleven-test strategy for 328 items; maximality of the latter remains open.

Conventional 3D Gaussian Splatting (3DGS) requires depth sorting and ordered alpha blending to correctly render overlapping Gaussian primitives. Stochastic transparency enables sorting-free rendering by replacing fractional alpha contributions with discrete stochastic visibility samples, but produces substantial spatial and temporal noise at low sample counts. We refer to this conversion from continuous Gaussian splats to discrete visibility samples as Gaussian Stippling. Based on this, we present an efficient order-independent rendering and reconstruction framework that operates directly on unmodified 3DGS assets. Our method adaptively integrates primitive-based and fragment-based stippling, leveraging their complementary strengths across different rendering regimes to significantly improve rendering throughput. To recover high-quality images from sparse stochastic samples, we further introduce a lightweight Gaussian-aware spatiotemporal reconstruction network. By exploiting the Gaussian attributes retained by each stipple, the network aggregates structured stochastic clues across both space and time, effectively suppressing stippling noise. Experiments show that our hybrid Gaussian stippling method, coupled with a spatiotemporal reconstruction network trained on diverse scenes, generalizes to unseen scenes and enables interactive, temporally stable, and visually plausible rendering on mobile devices without retraining or preprocessing. With scene-specific training and appropriately scaled sampling and network capacity, our method further outperforms the baselines in visual quality, offering a high-fidelity configuration for quality-prioritized applications.

Judging image quality is not only ecologically relevant to everyday human tasks, but also underpins many machine vision tasks such as image generation. This paper proposes a framework to understand the inherent perceptual space underlying image quality judgment in humans. We propose a multi-dimensional observer model that represents images as distributions in a latent perceptual space and that models human judgment as comparing noisy samples. Being constrained by neural representations in the primate ventral stream and fit to large-scale behavioral data, the model enables analysis of perceptual structure while matching the predictive power of existing metrics. Using this model, we find that the perceptual spaces needed to account for image quality judgment in humans are extremely low-dimensional compared to the image space even when considering its sparsity. The exact structure of the space (e.g., dimensionalities, information encoded) varies between low-level and high-level quality judgments, suggesting that, despite a shared retinal encoding in the beginning, humans selectively construct task-dependent perceptual spaces in visual decision making.

Large Language Models (LLMs) and more broadly Artificial Intelligence (AI) systems are often described and understood in human-like terms, a phenomenon known as anthropomorphism. This paper provides a synthesis of recent literature on anthropomorphism in AI, covering theoretical frameworks, the role of language in framing AI as human-like, the various risks of anthropomorphizing machines, and strategies to mitigate these issues. After examining why we tend to anthropomorphize AI systems and whether we are right to do so, we highlight the impact of linguistic framing on anthropomorphism. Then, we introduce a conceptual taxonomy of risks associated with AI anthropomorphism. This taxonomy groups twenty-one concerns within five analytical categories: epistemic, affective, human agency, normative, and societal and institutional risks. Finally, we relate these concerns to proposed interventions in design, communication, education, and governance. We argue that a better understanding of AI systems requires concepts and theories grounded in their organization and demonstrated capacities. The linguistic shaping of anthropomorphic perceptions should form part of this scientific effort, since our descriptions influence both how these systems are understood and the roles we allow them to occupy in society.

Unified multimodal models (UMMs) train image understanding and autoregressive image generation on shared parameters, and the two objectives are known to interfere. Existing diagnoses and remedies operate at the resolution of layers or experts, measuring conflict per layer and resolving it by separating parameters. We argue that this resolution hides an orthogonal axis. Generation in a UMM is next-token prediction over a raster sequence of visual tokens whose roles vary systematically with position, so how strongly a generation gradient interferes with understanding should depend on where in the sequence it originates. We introduce a position-resolved interference map that attributes understanding-generation gradient conflict to visual-token positions within every layer, computed from a single backward pass at $1.2\times$ the cost of a standard backward pass. On Show-o and Janus-Pro, position explains a large share of conflict variance after controlling for depth (partial $\eta^2=0.31$ vs. $0.35$ for layer on Show-o; $0.15$ vs. $0.30$ on Janus-Pro): the first quarter of the sequence has a mean gradient cosine of $-0.18$ against understanding, the last quarter $-0.02$. The dependence survives per-position gradient-norm normalization, retaining $80%$ of its effect size, and conflict strength tracks semantic content (Spearman $\rho=0.64$). Building on the map, we propose position-aware modulation (PAM), which removes the anti-aligned component of generation gradients only at high-conflict positions without changing the architecture. Under a matched trainable-parameter budget, PAM improves over layer-wise separation by $+21$ MME and $+2.4$ GenEval points on Show-o while matching it on POPE and overall FID; a random-position control recovers about $31%$ of the gain. Position-based and layer-based separation are complementary degrees of freedom and can be combined.

Retrosynthesis enables the discovery of viable synthetic routes to target molecules. It plays a central role in modern drug discovery and materials design. Retrosynthesis involves molecular graph transformations that can change both connectivity and graph size. These transformations may introduce reactant components absent from the target while revising the product-derived structure. To model these transformations, we propose RetroGEF, a flow-based generative model for single-step retrosynthesis. Starting from the target molecule, it constructs possible reactants by adding atoms and changing bonds in the molecular graph. RetroGEF models molecular transformations and changes in graph size within the same generative process, rather than relying on a fixed-size graph canvas. It learns this process directly from product–reactant pairs without requiring a prescribed edit order. Experiments on representative retrosynthesis benchmarks demonstrate that RetroGEF achieves state-of-the-art performance.

Indirect manipulation of deep-seated deformable anatomy inaccessible to the robot is challenging in robot-assisted minimally invasive surgery (RAMIS) because intervening tissues spatially filter deformation transmission. Passive observations can be ambiguous because the Decoupled and Blocked modes may produce similar motion responses. We propose CADeT, a causal-aware deformation transmission framework that integrates structural causal model (SCM) with active sensing to infer a latent transmission mode and estimate a state-dependent adhesion Jacobian online. During normal manipulation, control actions update the mode belief; when ambiguity persists, an additional probing action is selected to improve mode distinguishability. The mode belief and learned Jacobian are incorporated into a belief-aware model predictive controller for indirect target-shape control. Validation in simulation and on the da Vinci research kit (dVRK), using phantom and ex vivo porcine tissues, shows higher mode-identification accuracy and faster shape-error convergence than the evaluated model-free and model-based baselines. These results show that active sensing improves mode identification and indirect deformation control under the evaluated conditions.

Existing architectures for LLM-based multi-agent systems (MAS) cannot reliably and efficiently solve multi-step tasks at scale: they struggle to support large numbers of agents and concurrent tasks, tolerate failures, govern agent interactions, and accommodate the diverse planning and execution patterns different tasks require. We present PANDA, a decentralized architecture that connects a large collective of heterogeneous, independently administered agents, letting them discover each other's capabilities and self-organize into small specialized teams per task. PANDA scales by decoupling collective communication from team communication, allowing agents to participate in multiple teams simultaneously, load-balancing tasks across the collective, and scheduling concurrent work within each agent. PANDA further separates the underlying architecture from the orchestration strategy, supporting three planning and execution patterns (star, chain, and mesh) that can be selected according to the structure and requirements of each task. PANDA detects infrastructure and orchestration failures and recovers affected tasks by dynamically replanning around failed components. Finally, to provide governance without a centralized service that would limit scalability, PANDA uses a web-of-trust model to constrain agent interactions to established trust relationships. We evaluate PANDA on the HotPotQA benchmark, demonstrating that it scales to thousands of agents, assembles teams in milliseconds, matches state-of-the-art accuracy at up to 8x the efficiency, and sustains 100% task completion under faults where existing systems fail.

Results of classical and nonclassical approach to nucleation at the ordering phase transition in barium titanate are presented. A nonclassical nucleation model is applied to the first-order paraelectric cubic phase to ferroelectric tetragonal phase transition in barium titanate is considered.

Most clinical benchmarks evaluate language models (LMs) on diagnosis using complete case descriptions. In clinical practice, however, patients present information in different ways, and clinicians must obtain relevant history and determine which examinations are needed before reaching a diagnosis. Diagnostic accuracy alone therefore cannot establish whether an agent gathered essential information or conducted an appropriate clinical assessment. Furthermore, existing benchmarks lack professional clinicians' verification. To address this gap, we introduce KlinikeBench, a benchmark of 333 clinician-authored tasks, each providing an isolated sandbox environment with a virtual patient, clinical tools, and task-specific success criteria. More than 35 clinicians contributed to case authoring and benchmark evaluation. In an empirical study, clinicians gave simulated dialogues higher mean quality ratings than reference conversations, which is adapted from real conversation. In each task, an LM has a fixed budget of turns to communicate with the patient, ask about relevant history, request examinations, follow action constraints, and record a final diagnosis. We score these steps separately as well as together. Across 31 models and seven model families, the best-performing models (e.g., GPT-6-astra and Claude Opus 5) succeed on less than 30% of tasks, even though their diagnosis accuracy reaches 90.7%. Some models benefit from talking with the patient; others diagnose well from a complete chart but perform much worse in conversation. Overall, KlinikeBench provides a testbed for evaluating the full clinical encounter and reveals a substantial gap between diagnostic accuracy and performance in interactive clinical assessment. All the code and data is available on https://zehui127.github.io/klinikebench/

This paper presents an explainable approach to pedestrian routing, in which perceived safety is estimated from street-level imagery through an explicit natural-language intermediate representation. A vision–language model caption is generated and stored before any scoring is undertaken, and the perceived-risk class is derived entirely from structured features of that stored text, so that every segment score remains inspectable by the user. Nine captioning conditions across five model families are benchmarked against a direct Contrastive Language–Image Pre-training (CLIP) image-embedding baseline under an identical downstream pipeline, and the caption-mediated representation is found to reach parity with the image embedding rather than to trail it. The approach was deployed over 654,115 images covering 36 electoral wards in two locations in Northern England (Manchester and Huddersfield). Independent field validation against 3,669 locally collected ratings of 494 images across 70 participant sessions established agreement that is statistically significant but modest, at $r=0.262$, against a measured noise ceiling of 0.737 imposed by disagreement between raters. A single-use confirmatory test then found that a pipeline 44% stronger on the supervised benchmark did not produce measurable improvement in the field ($r=0.250$, $p=0.84$), so the benchmark gains did not predict the deployment gains in this case. Routing behaviour varies systematically with journey length. There is negligible change below 1\,km, reaching a median increase of 12.78% in low-risk route length for a median detour of 2.73% on journeys of 3 to 6 km.

Complex matter, often partly soft and pliable, is the generic form of material systems. We must therefore generally call on density functional theory (DFT) to first predict the atomic structure before we can use it to also characterize (expected) properties. A recent range-separated hybrid (RSH) van Waals density functional (vdW-DF), `vdW-DF2-ahbr' (abbreviated AHBR) [PRX 12, 041003 (2022)], shows promise as a high-accuracy predictor of both binding energies and structure, from molecules to solids. However, the present implementation in Quantum Espresso (QE) does not, in our experience, support robust stress-based unit-cell optimization of RSHs. Here we document and illustrate work enabling practical AHBR-based complex-matter discovery: We port the AHBR XC functional to the Vienna Ab Initio Simulation Package (VASP) where use of stress-based DFT optimization is already stable also for RSHs. We test the implementation by comparing high-accuracy AHBR-QE and AHBR-VASP predictions of non-covalent molecular interactions and for structure and cohesion of simple bulk structures. We also illustrate and test the use of the new AHBR-VASP implementation to predict and understand (atomic and anti-ferromagnetic) structure in distorted-rocksalt metal monoxides. Finally, we illustrate use for complex-soft-matter discovery by predicting motifs for layer stacking in the C2N covalent-organic framework (COF) system [Nat. Commun. 6, 6486 (2015)].