Science & Mathematics

The ability to assemble and reconfigure individual atoms is central to performing quantum science using optical tweezer arrays. Here, we demonstrate parallel omnidirectional atom transport and rearrangement using dynamic holograms generated by a single kilohertz-rate spatial light modulator. By carefully characterizing survival and heating, we show that atoms remain confined even when successive holograms are commanded before the liquid crystal has settled, enabling significantly faster transport. To generate holograms in real time, we develop and implement a warm-start, weighted Gerchberg-Saxton algorithm that achieves performance comparable to neural-network methods without requiring any training. Combining these techniques, we assemble ytterbium atoms into a 2078-site kagome lattice with 99.0% mean filling. In addition, we demonstrate parallel transport between array layers, combining axial and transverse motion to reconfigure atoms in three dimensions. Our approach is readily adaptable to existing tweezer platforms and provides a method for large-scale SLM-based atom assembly as well as programmable three-dimensional control.

Generating human-object interactions (HOIs) with multiple objects is challenging because the resulting motions arise from different underlying causes. Human motion reflects semantic intent, intentionally manipulated objects must remain coordinated with the human, while surrounding objects move primarily as physical consequences of contact and collision. We present BUMP, a text-driven framework for multi-object HOI generation based on a simple principle: learn intentional motion and simulate physical consequences. Our method first generates full-body human motion from text and scene context, then predicts trajectories for a small number of intentionally manipulated driver objects conditioned on the human motion. We then obtain responder object motions through physics simulation, allowing the framework to handle varying numbers and configurations of surrounding objects without expanding the learned trajectory output space. To address the scarcity of paired full-body multi-object HOI data, we combine heterogeneous supervision from human-only motion, full-body HOI, and hand-centric manipulation data augmented with plausible body motion. We further introduce outcome-conditioned interaction editing, which infers a local modification of the causal driver object trajectory from a desired responder object behavior without requiring an explicit target driver object trajectory. Experiments demonstrate strong performance on both single-object and multi-object HOI benchmarks. Qualitative results showcase diverse scenarios involving inter-object relationships and edits, highlighting the advantage of our unified model formulation.

The microphysics of dark matter (DM) is imprinted on cosmological structure across an enormous dynamic range. Nonlinear cosmological observables are most reliably modeled with N-body and hydrodynamic simulations. However, systematic N-body scans over DM theory space are prohibitively expensive. Machine learning emulators provide a possible solution, but are currently limited by either low resolution or only the ability to emulate specific observables. We introduce Astrolabe, a flow-matching-based generative emulator designed to be free of both constraints. Astrolabe acts directly on the 6N phase-space coordinates of N individual particles. With no voxel grid in the output, there is no voxel-resolution ceiling, and in principle any observable can be measured from inference. Astrolabe is computationally tractable via a field-particle hybrid transformer in which every particle cross-attends to multiscale field tokens, so compute and memory are linear in N. Astrolabe incorporates physical flow matching by taking the prior distribution for the flow to be approximate COmoving Lagrangian Acceleration (COLA)-based coordinates rather than unphysical noise, with momentum and mass conservation enforced at all points along flow time. In the Lambda-CDM cases we test, Astrolabe produces realistic fields with accurate power spectra, halo mass functions, velocity profiles, and halo density profiles. We also observe out-of-distribution generalization, with accurate emulation of dark-acoustic oscillation initial conditions. Astrolabe is also capable of being fine-tuned to simulate very large N and we demonstrate this by emulating a TNG50-3-dark simulation at a fraction of the cost. This method represents a first step toward a general and powerful cosmology simulation surrogate, which may eventually permit the inclusion of baryonic physics and a systematic exploration of richer dark sector interactions.

We introduce Lookup-table (LUT)-aware Efficient Attention on FPGAs (LEAF). LEAF is the first nanosecond-scale LUT-aware transformer for FPGAs. Deploying transformers that are large enough to perform well, small enough to fit on an FPGA, and fast enough to meet the sub-microsecond timescales required by real-time systems such as those at CERN remains difficult, since the same attention mechanism that gives these models their high expressivity also makes them costly to implement in hardware. LEAF converts the query, key, and value projections and the final output feed-forward layer into FPGA-native layers with LUT-aware training, creating the first LUT-aware transformer and linformer layers built for ultra-low-latency inference. Attention resists LUT-aware training because it multiplies two activations where the method needs a constant weight to tabulate; LEAF tabulates the attention block's constant matrix-vector multiplications and leaves the rest in arithmetic form. On a standard jet tagging benchmark this improves accuracy by 1.8 percentage points while using 2.9 times fewer LUTs than the previous state-of-the-art sub-microsecond transformer at the same input size. Our most efficient LEAF-based linformer with 64 input particles reaches 80.4% classification accuracy at 28k LUTs and 35ns.

Large language model (LLM) watermarking operates on the next-token distribution that evolves during generation. In streaming generation, each sampled token becomes part of the text and affects subsequent distributions. Embedding therefore proceeds with a changing host, unlike classical watermarking, which adapts embedding to an existing host signal. To accommodate this difference, we formulate a novel embedding framework for LLM watermarking based on two principles. We first argue that embedding should be adaptive, retaining the state of previous sampling decisions to better guide subsequent choices. We further argue that embedding should be host-dependent, using the current LLM distribution to constrain both token selection and state updates. Guided by these principles and the two requirements of distribution preservation and robustness to text editing, we propose ClockMark. It combines adaptive and host-dependent embedding with message spreading, embedding the full payload in each watermarked token. Theoretically, we prove that ClockMark preserves the original LLM sequence distribution, thereby maintaining generation quality. Empirically, extensive experiments against 12 baselines demonstrate improved robustness to text editing, with gains of up to 16.08 percentage points in bit accuracy over the strongest tested baseline.

We show that the problem of deciding whether a simplicial complex is partitionable is NP-complete in dimensions two and higher, answering a 1995 question of Kleinschmidt and Onn.

Scientific method figures often provide researchers with a clear visual entry point into a paper's methodology and core contributions, yet their dense layouts and complex information flows are difficult for Vision Language Models (VLMs) to interpret. Existing benchmarks predominantly focus on coarse-grained tasks, such as context-based retrieval or holistic comprehension, overlooking the fine-grained content and inter-module relations within complex methodological diagrams. To address these overlooked aspects, we introduce Scientific Figure Benchmark (SciFigBench), a benchmark for evaluating core capabilities required for scientific method figure understanding. Specifically, it targets three capabilities: dense content perception, visual relation comprehension, and information-flow tracing. For this purpose, we curate 4,352 scientific method figures from real-world research papers and construct 40,670 QA pairs to operationalize these capabilities as measurable question-answering tasks. All annotations are generated by dedicated pipelines and then manually verified. We further propose SciFigCue, a training-free visual prompting module that externalizes arrow-induced dependencies as visual relational evidence for method-figure reasoning. Experiments across open-source and proprietary VLMs show that even strong models remain substantially limited in visual relation comprehension, while SciFigCue improves performance across several tasks without parameter updates. The code and related resources are available at https://github.com/LeNcyer/SciFigBench.

Ternary layered transition metal dichalcogenides (TMDCs) offer high tunability and access to correlated and topological states. However, like the broader family of layered TMDCs, their properties are sensitive to stacking order and interlayer coupling. Here, we investigate the electronic structure and surface states of NbIrTe$_4$, a candidate Weyl semimetal in bulk, and predicted to be a quantum spin Hall insulator in monolayer form. Using scanning tunnelling microscopy (STM), angle-resolved photoemission spectroscopy (ARPES), and density-functional theory (DFT), we attribute the strong resonance in tunnelling spectroscopy located approximately 8 meV below the Fermi level to a surface-derived flat band which shows sensitivity to the local stacking order at the surface. In regions where the top layer is laterally shifted, tunnelling spectroscopy instead shows a gap-like feature with intensity near the Fermi level shifted in energy. Combining DFT and quasiparticle interference (QPI) measurements, we find this dramatic change in the spectroscopy in the vicinity of the quasi-1D flat-band surface state becomes dispersive due to a change in the spatial distribution of the density of states in the top two layers. These findings establish NbIrTe$_4$ as a platform where surface band dispersion can be effectively tuned through local structural manipulation.

Clinical care, health-system operations and biomedical discovery depend on extracting reliable answers from patient records, yet applying diverse clinical questions across entire populations remains computationally expensive. Here we present Clinical System One (CS1), a 4-billion-parameter decision model for fast, typed judgments from clinical records, along with two agentic harnesses in which CS1 works together with a generalist language model. CS1-4B is trained on 12.4 million questions and answers from 3.6 million patients across a multi-hospital health system, combining structured-data keys, observed outcomes and language-model supervision across clinical notes, diagnostic reports and patient timelines. It answers multiple typed questions in a single forward pass without task-specific retraining, and each answer carries a probability that can route uncertain cases to human review. On 963,408 held-out patients, CS1-4B outperformed open decision models up to 27 billion parameters on 13 of 14 tasks scored independently of its training teacher. Across 21 external data sets, its aggregate performance approached that of the strongest tested 27-billion-parameter model, with 6.4-fold faster inference in a matched workload. Integrating CS1-4B with a language model agent produced cited answers over complete multi-year patient histories and screened a patient population against the eligibility criteria of 17 registered clinical trials, placing four times as many judged-eligible patients among its top 20 candidates as keyword search while using 690-fold fewer language-model tokens than a language model reading every record. These findings support a division of labour in clinical AI, with small decision models performing repeated record reading and generative models formulating questions and synthesizing evidence.

Realistic market impact is essential for generative limit order book (LOB) models to serve as world models for trading agents. If a simulator misprices this mechanism, agents learn policies that are optimal in the simulator but wrong in the real market. Existing metrics assess generated streams purely through distributional comparison with historical data, leaving counterfactual response to an agent's own orders largely unexplored. Here, we show that market impact emerges naturally in neural LOB generators trained solely for next-token prediction. We inject execution-style metaorders into the generated message flow of five autoregressive models across five NASDAQ stocks and measure the response using estimators from the empirical impact literature, validated on baselines whose physics is known in advance. Our benchmark evaluates any message-level generator in three phases: free generation (reading order-flow persistence against price diffusivity); metaorder injection (yielding the impact law's exponent and amplitude); and a post-execution window (measuring relaxation). In addition, comparing each visible injection to a blind, simulator-only control isolates the model's learned response from standard book mechanics. For every model on every stock, we measure how the impact responds as a function of cumulative metaorder participation. Interestingly, we observe emergent impact in generative models of market data, with exponents in the range of 0.63 to 1.21, slightly above the traditional 0.5 of the square-root law of impact. In the family of models tested, we are so far unable to observe impact decay during relaxation. We tentatively attribute this to the still-short context windows of the models. We release the benchmark as open source to validate future generative models for market data.

Recent Omni LLMs (OLLMs) have demonstrated remarkable capabilities in joint audio-visual reasoning, yet cross-modal hallucinations remain a major challenge. Existing mitigation methods mainly operate at the output level, overlooking the entanglement of grounding evidence and learned priors within fused representations. To address this issue, we propose Omni-Woodpecker (OWP), a training-free framework that shifts hallucination mitigation from output-level manipulation to conflict-conditioned representation rebalancing. Specifically, OWP first identifies cross-modal conflicts from complementary modality-specific views. It then employs orthogonal decomposition to separate grounding evidence from competing priors within the hidden states. Finally, OWP adaptively rebalances fused representations by suppressing misleading priors and restoring ignored evidence before decoding. Extensive experiments demonstrate the effectiveness and strong cross-architecture generalization of OWP. Using Qwen2.5-Omni-7B as the backbone, OWP improves performance by 12.9 and 8.6 points on CMM and AVHBench, respectively, while consistently achieving gains of 0.67-2.25 points across four general OLLM reasoning benchmarks. Furthermore, OWP generalizes effectively to VideoLLaMA2-AV, improving performance by 4.6 and 3.9 points on the same benchmarks. Code is publicly available at https://github.com/ZiiJianLiu/Omni-Woodpecker

Reliable language-guided perception is essential for embodied agents, where questions about objects, events, and their locations must be answered from long observation histories. Unlike conventional visual question answering, realistic robotic settings require reasoning over extended temporal horizons and persistent scene memory. We build upon the view-based memory paradigm and propose an efficient framework for long-horizon spatio-temporal question answering. While existing view-based systems typically store either text captions or visual embeddings, we extend the ReMEmbR-style memory with image paths that link each stored description to its original visual content. To make the stored annotations more informative, we replace uniform sampling during memory construction with an optical-flow-based frame selection strategy (PSFR) and store an object-centric descriptive scene representation instead of plain captions. We further equip a VLM-based agent with an adaptive retrieval tool that returns the image associated with the most relevant caption, allowing the agent to decide when visual inspection is needed. On the OC-NaVQA benchmark, LVMem outperforms ReMEmbR on binary and positional metrics with both backbones and matches or exceeds a 4D scene graph DAAAM baseline on binary accuracy, while operating on a lightweight view-based memory that requires no explicit graph construction.

We present the Sparse Particle-interaction Transformer Architecture (SParTA), a modular, scalable, and efficient transformer for jet tagging at high-energy colliders. Exploiting the steeply falling transverse-momentum ($p_T$) spectrum of jet constituents, SParTA factorizes constituent processing into two topological regimes. Ordered by descending $p_T$, only the $N$ hardest constituents (the hard core) participate in quadratic $O(N^2)$ differential interaction attention driven by pairwise physical and learned relational features. The remaining soft constituents are integrated via linear $O(lN)$ cross-attention to the hard core, preserving soft-radiation and color-flow information while avoiding the $O(l^2)$ memory bottleneck of full self-attention over all $l$ constituents. Pairwise interaction features are computed dynamically in-network from constituent four-momenta and conditioned on global jet-level substructure observables. On the quark-gluon tagging benchmark, a 300K-parameter SParTA model achieves an AUC of 92.02% and accuracy of 85.06%, matching the 2.14M-parameter Particle Transformer (ParT) at one-seventh the parameter count and outperforming IAFormer by 21% in background rejection at 30% signal efficiency. On the boosted top-tagging benchmark, SParTA outperforms ParT and ParticleNet with 511K parameters, while we analyze the architectural trade-offs relative to fully dense multi-layer interaction networks.

Credit card fraud is a significant problem for financial institutions, resulting in substantial monetary loss each year. Fraud patterns and methods are continuously evolving, with fraudulent transactions often disguised to resemble legitimate activity. Real world transaction data for fraud detection is severely affected by class imbalance making it harder to detect these transactions. Fraudulent transactions are rarely isolated. Accounts, devices, and other common attributes are often shared across fraudulent activity, revealing relational patterns that individual transaction features alone cannot capture. Graph-based approaches model these relationships explicitly. However, standard graph neural networks aggregate neighbourhood information uniformly, which can dilute the already scarce signal carried by fraudulent transactions and limit detection of camouflaged fraud. This work presents RGH-GNN, a heterogeneous graph neural network that addresses this limitation by incorporating transaction-level risk scores, generated by a supervised LightGBM classifier as edge weights. It is used to guide message passing through a learnable risk-gated aggregation mechanism. It ensures that high-risk transactions contribute more strongly to learned representations rather than being diluted by uniform neighbourhood aggregation. RGH-GNN is evaluated on the IEEE-CIS fraud detection dataset against standard graph neural network models that are trained on the same heterogeneous graph. RGH-GNN outperforms all baselines across every evaluation metric, with an AUC-ROC score of 0.914. This demonstrates that incorporating engineered risk signals into graph-based message passing substantially improves fraud detection performance.

3D Gaussian Splatting grows a dense scene from a sparse initialization through adaptive density control (ADC), driven by per-Gaussian statistics gathered during rendering. These statistics pass a frustum-based visibility filter blind to occlusion: fully occluded Gaussians still count as observed, diluting the averaged gradient signal and suppressing densification precisely where geometry is occluded. We correct this with an occlusion-aware filter that restricts densification statistics to views in which a Gaussian actually contributes, and replace the criterion's raw view-count normalizer with an average-contribution one. Integrated into a StopThePop-based framework, the two together improve every image-quality metric over the occlusion-blind baseline on occlusion-heavy scenes (up to 6% lower LPIPS), at less than half the Gaussians a view-count normalizer requires. Under the longer schedules these large scenes require, the count stays bounded where most criteria diverge, and on smaller benchmarks our method stays competitive at a fraction of the competitors' count.