Machine Learning
Papers
Fresh from arXiv, typeset properly. Highlighted words trade on the market.
All of arXiv
Thursday 3 September
Continual knowledge-updating methods are often declared superior from one final checkpoint and one conventional adapter rank. We show that this can be insufficient to identify the better operating point. Holding a periodic hierarchy fixed, we compare it with cumulative replay over a 24-month Wikidata stream while varying evaluation month, replay LoRA rank, and query formulation. The apparent winner changes across this region: on Qwen2.5-1.5B, the hierarchy's 5.0-point advantage over rank-8 replay becomes an 11.6-point deficit against rank-72 replay, and at high ranks a consolidation-aligned endpoint can suggest a tie while time-averaged replay leads by 9-13 points. The same rank-conditioned reversal appears on Llama-3.2-1B and held-out paraphrases. These results show that method ranking in continual updating can depend jointly on when performance is measured and how much replay-side adaptation capacity the baseline receives. We therefore propose reporting trajectories and capacity sweeps, and declaring a robust winner only when the ordering is stable across the evaluation region; otherwise, comparisons should report winner regions and retention-stability-cost frontiers. Under this protocol, the periodic hierarchy is a lower-update-cost operating point, not a quality winner.
Social and Information Networks
Fair Top-k Katz Centrality via Graph Design
Centrality measures are widely used to rank nodes in networked data, but fairness interventions for graph centrality typically target global score mass or modify the centrality operator rather than controlling who appears in the displayed top-k ranking. We study this top-k setting for Katz centrality. Given a target group proportion, an admissible set of directed edge additions, and a fairness tolerance, the goal is to find the smallest edit set whose resulting Katz top-k ranking satisfies the target representation constraint. We formalize this problem as Fair Top-k Katz Centrality Design and show that the minimum-edit objective is strongly inapproximable, ruling out worst-case polynomial-time approximation guarantees unless P = NP. We then derive closed-form Katz sensitivity expressions showing that useful edits are boundary-driven: they must help promotable nodes outside the top-k set overtake opposing nodes inside it. Based on this structure, we develop BLADE, a scalable boundary-link algorithm that avoids dense Katz-kernel maintenance by using score-based direct-target batches and warm-started Katz updates. Experiments on synthetic and real-world networks show that BLADE reaches the desired top-k representation using far fewer edits than natural baselines, while scaling to large real-world graphs and preserving the original ranking structure.
Read the full abstract
Databases
From Data Querying to Data Investigations: Rethinking Natural Language Interfaces for Databases
Natural language (NL) interfaces to databases have been optimized for the wrong problem. The dominant Text-to-SQL paradigm assumes that users ask questions that can be answered by single SQL queries. In practice, however, users seek assistance with solving data problems. This requires searching a database by sequences of SQL queries while reasoning over intermediate results instead of just running one SQL query. This paper therefore introduces a new paradigm for NL interfaces to data, which we call data investigations. We present D^2, a first prototype of a data investigation system that embodies this vision by autonomously searching, reasoning over, and collecting data to solve data problems. Using a newly constructed benchmark based on the Murder Mystery dataset, we demonstrate the potential of D^2 for tasks that require data investigations with evidence-backed decisions, extending beyond the capabilities of traditional single-query question answering.
Read the full abstract
Graphics
Reparametrizing 3D Gaussian Splatting for Real-Time Palette-based Color and Luminance Editing
Professional color editing requires precise control over both color (hue and saturation) and lightness, ideally through separate, independent controls. We present a real-time interactive color editing framework for 3D Gaussian Splatting that supports palette-based recoloring, per-palette tone curves for color-aware luminance adjustment, and pixel-level color constraints. Rather than training a new representation from scratch, we reparameterize the spherical harmonics of a pretrained vanilla 3DGS to encode view-dependent palette weights. We simultaneously solve for weights and palette colors via a loss based on image-space sparsity. Luminance editing is realized as a per-pixel weight shift along the achromatic axis, which we show is equivalent to a per-pixel palette-aware luminance edit. This view-space formulation addresses a core limitation of prior primitive-space methods, where alpha-blending breaks per-Gaussian sparsity and causes edits to bleed into unintended regions. Our edits run in tens of milliseconds via an iteratively reweighted least squares and damped block-coordinate descent that couples tone curves and palette shifts under view-space sparsity. Our representation can be efficiently baked back into a vanilla 3DGS, preserving compatibility with standard viewers. We demonstrate sparser, more localized edits than prior palette-based 3DGS methods, while enabling independent luminance control per palette color and view-consistent pixel-level constraints, capabilities previously unavailable for 3DGS.
Read the full abstract
History and Overview of Mathematics
The Prime Clockwork: A Dynamic Representation of Modular and Multiplicative Arithmetic
The way numbers are represented strongly influences which arithmetic structures are easy to see. The prime clockwork is a recursively growing discrete dynamical system: a list of autonomous two-hand clocks driven by one common $+1$ signal. No primes or primality labels are supplied. Starting empty, the process appends a clock of period $n$ whenever none already present rings; the primes are generated internally as its growth times. For each installed prime $p$, the seconds reading $R_p$ advances through $0,\ldots,p-1$, and each return to zero increments the minutes reading $M_p$, which counts completed $p$-cycles. The hands use only increment, comparison, reset, and carry, without explicit mod or div operations. At time $n$, $n=pM_p(n)+R_p(n)$. The valuation readout $V_p(n)=\nu_p(n)$ is generated locally: it is zero when the seconds counter is non-zero (silent state) and otherwise (when the p-clock rings) one plus the earlier valuation addressed by the current minutes reading. The valuation vector gives the integer in unique prime-factorized form. Its coordinates add and subtract under multiplication and division, representing every positive rational uniquely; divisibility becomes weak componentwise order, and unique factorization is natural in this representation. Finite seconds arrays form Cartesian-product state spaces whose common orbit visits every joint state once before repeating; this grand cycle is the order-sensitive dynamical counterpart of the Chinese remainder theorem. The same coordinates expose gcd, lcm, perfect powers, Bézout's identity, and Euler's totient. Rational valuation levels reach certain positive algebraic irrationalities, but not algebraic numbers in general.
Read the full abstract
Computer Vision
Concept of a Sensor Test Environment for Dusty Agricultural Conditions
Dust in agriculture presents a significant challenge for autonomous agricultural machinery. Dust can impair the performance of sensors and algorithms. This work, therefore, presents a concept for a proving ground consisting of an indoor and outdoor area. The indoor area comprises a laboratory test bench where dust circulates in a closed system and a test hall where life-size objects can be placed. The outdoor area features dedicated test setups that enable reproducible data to be recorded with and without dust during real-world agriculture work. The proving ground and the setups are visualized in 3D.
Read the full abstract
Computation and Language
CROCODIL: Cross-Model Code Editing with LLMs
Large language models (LLMs) have become ubiquitous tools for code generation and editing. However, development teams often use multiple LLM assistants. Different developers may prefer different models, and individual developers may switch between models across different coding sessions. Because of this, the edits any one model makes are frequently applied to foreign code originally generated by another model. These LLMs are often trained on different datasets, and as a result have different stylistic preferences. Do LLMs behave differently when they edit foreign code originally written by a different LLM with a different coding style? We find that models tend to make more, and often excessive, edits on foreign code. We introduce CROCODIL (Cross-model Code Editing with LLMs), a post-training framework for reducing excessive edits while preserving functional correctness. CROCODIL's similarity reward penalizes large changes, while its execution reward scores build and test success. We use the product of these two rewards to encourage the policy to decrease the edit size without decreasing the edit task success rate. CROCODIL is available at https://github.com/EngineeringSoftware/Crocodil.
Read the full abstract
Cryptography and Security
Practice Makes (Im)Perfect: A Look Back at Benchmarking Practices for Microarchitectural Side-Channel Attacks
Microarchitectural side-channel research has grown at an exceptional pace in recent years, increasing the need for rigorous and meaningful benchmarking. Early attack papers typically relied on indirect proxies, such as covert-channel bandwidth or key-recovery on naive AES and RSA implementations, setting de facto standards that many subsequent works continued to replicate, sometimes by directly comparing against raw numbers from prior work. While these practices offer convenient points of comparison, current benchmarks may not be the most relevant to assess specific properties of new primitives. Even more problematic, microarchitectural attacks are notoriously sensitive to experimental conditions: minimal changes in the target system can significantly alter outcomes and performance. As a result, inadequate evaluation practices undermine reproducibility and cast doubt on the relevance of comparisons, even in top-tier venues where such issues should be identified. This paper tackles the core problem of proper benchmarking for microarchitectural side-channel attacks and examines its broader impact on research quality in the field. We survey 83 attack papers published in top-ranked security and architecture conferences from 2014 to 2024. From this corpus, we identify and define 19 recurrent benchmarking flaws that affect evaluation completeness, relevance, soundness, and reproducibility. These flaws include unfair or absent comparisons, missing code or materials, and the failure to evaluate the key attack properties. On average, each paper exhibits 5.72 such flaws, highlighting how widespread the issue is, even in highly selective venues. Based on our findings, we identify and suggest key properties that are relevant to properly evaluate new attacks. We also highlight trends over time and different practices between security and architecture conferences.
Read the full abstract
Computer Vision
GraFT: A Training-Free Framework for Spatial Reasoning in Multimodal Large Language Models via 3D Scene Graphs
3D spatial reasoning underpins understanding and acting in the physical world, yet it remains unreliable in current multimodal large language models (MLLMs). These models falter at precise geometric measurement, at transforming between egocentric and allocentric viewpoints, and at grounding fine-grained appearance. The most common remedies fine-tune the model on large-scale curated spatial-reasoning datasets or attach dedicated encoders for 3D geometry, which typically couples the solution to costly supervision and a specific backbone. We instead introduce GraFT, a training-free framework that supplies the missing 3D structure through a compact, easily maintained 3D scene graph (3DSG). GraFT supplies the graph's 3D structure to the frozen model in three forms: as values computed by symbolic tools, as a bird's-eye-view (BEV) the model reads as an image, and as egocentric frames the geometry selects. On ScanQA, GraFT improves every metric over the same-backbone baseline, raising CIDEr by 27%. On VSI-Bench, GraFT improves frozen MLLMs by up to 65%, surpassing every proprietary and general-purpose open-source baseline, and several prominent fine-tuned spatial models.
Read the full abstract
Robotics
A hybrid pipeline for dynamic ontology-based semantic mapping
Semantic mapping plays a crucial role in the ability of a robot to interact with objects, operate and navigate a complex environment. The most common pipeline for semantic mapping consists of geometric mapping and localization (SLAM), perception, semantic fusion and semantic representation. However, more recent works also integrate a form of prior knowledge in their application, most notably knowledge graphs or semantic scene graphs, to improve contextual understanding of the environment. In this paper, we present a hybrid pipeline for semantic mapping. Our system incorporates an external calibrated camera using homography projection for geometric mapping and localization, combined with object detection, persistent object tracking and ontology driven semantic updates to build a dynamic semantic world model. Linear regression models are also used for correction of the estimated values of real world coordinates. The system continuously updates object instances, spatial properties and semantic relations based on real time sensory data. Ontologies are selected as form of knowledge representation due to their hierarchical structure, semantic expressiveness and support for dynamic world modelling.
Read the full abstract
Data Structures and Algorithms
The 11/6 supremum of the Wang-Sitters rounding scheme for graph balancing
Wang and Sitters' 11/6-approximation for graph balancing (each job allowed on at most two machines) is not one algorithm but a set of permitted executions. Step 1 may return any feasible solution of the fractional relaxation. Step 3 may match remaining jobs to the rounding's slots in many ways. We determine exactly what that latitude permits: ratios arbitrarily close to 11/6, none reaching it, so 11/6 is the least constant bounding every permitted run. We then determine the worst-case guarantee, against the optimum itself, as a function of the big-job threshold beta, the fractional share at or above which Step 2 assigns a job. On Wang and Sitters' own range 1/2 < beta < 1 the guarantee is exactly max{3/2 + beta/2, 5/2 - beta}. We extend the same eligibility rule to 0 < beta <= 1/2, outside the range they state. There a big job's two shares can both reach the threshold, so Step 2 acquires a third choice. We determine it too: exactly 3/2 + (1-beta)floor(1/beta), hence unbounded as beta falls. The worst-case ratio is therefore known at every threshold in (0,1), and attained at none. Consequently 2/3 is the unique optimal threshold, and the guarantee jumps at one half and again at every threshold 1/k below it, rather than degrading smoothly. A companion note examines changes that leave the constant unchanged.
Read the full abstract
Robotics
FWBC-VLA: Force-Aware Whole-Body Compensation for Contact-Rich Loco-Manipulation
Contact-rich loco-manipulation requires a bridge between semantic action generation and physical interaction control. Existing Vision-language-action (VLA) models generate task-level actions from visual and linguistic observations, but cannot interpret the physical interactions induced by those actions. While the whole-body control (WBC) policy can stabilize the robot, it cannot distinguish task-relevant interaction forces from forces induced by external disturbances during manipulation. Although force/torque sensors provide direct measurements of physical interactions, retrofitting them entails additional hardware costs and substantial integration effort, particularly for platforms not designed with sensor integration in mind. To address this problem, we propose FWBC-VLA, a force-aware framework that bridges task-level VLA action generation and low-level whole-body compensation control for wheeled-legged robots. First, we introduce HSR-Force, a sensorless residual-torque estimator for inferring contact strength and its temporal variation. These contact estimates are then encoded as tokens and injected into the VLA action expert during action decoding, enabling the policy to perceive contact onset, sustained loading, and release. For loco-manipulation tasks, all parameters of the pretrained VLA backbone are fine-tuned on our WL&Arm Dataset, which comprises more than 5,000 episodes. Moreover, the robot's proprioceptive state, the Jacobian-derived body-frame force estimate, and the estimated contact state are jointly fed into a compensation generator to produce corrective actions. The manipulation-centric actions are subsequently combined with the corrective actions and passed to the WBC policy for execution. Real-world experiments on whiteboard wiping and door opening with a door closer demonstrate the effectiveness of our FWBC-VLA in contact-rich loco-manipulation.
Read the full abstract
Materials Science
Experimentally constrained modeling of the Pockels response of KNbO3 and KTaNbO3
The soft-mode of electro-optic (EO) metal-oxide perovskites plays a critical role in determining their Pockels responses. This is the case of potassium tantalate niobate (KTN), which exhibits an intrinsic Pockels response 2.5 times larger than that of state-of-the-art barium titanate (BTO), highlighting its potential for high-performance EO applications. By combining ab initio calculations at the density-functional theory level and far-IR measurements, we reveal that the harmonic approximation combined with semi-local exchange-correlation functionals fails to accurately capture the soft transverse optical (TO) Slater mode of both potassium niobate and KTN, which dominates the technologically relevant r51 Pockels coefficient. Replacing the calculated mode frequency with its measured value provides an experimentally constrained approach that substantially improves the predicted Pockels response. For that purpose, the previously unreported TO Slater-mode frequency of KTN is extracted from far-IR reflectivity measurements. The results emphasize the potential of KTN-based EO modulators as an alternative to standard lithium niobate and BTO technologies, with a potentially lower energy consumption and device footprint.
Read the full abstract
Computation and Language
Beyond Shallow Alignment: How Post-Training Methods Determine Refusal Circuits And Steering Robustness
How do the methods used to train language models to refuse harmful requests shape how that refusal actually works inside the model? We compare three post-training methods - supervised fine-tuning, reasoning-augmented fine-tuning (training on reasoning chains that justify a safety decision), and preference optimization (ORPO) - across three architecturally distinct models (Llama-3.1-8B, Gemma-2-9B, Qwen3-8B). We find that training method, not just data, reshapes how refusal is computed internally: reasoning-augmented training consistently produces a distinct kind of refusal computation, visible across all three models, while architecture independently shapes internal structure and how reliably refusal can be steered. Most importantly, no method we study achieves all three properties we would want from safe alignment at once: refusal that isn't concentrated in a few fragile components, safety gains that don't cost general capability, and safety behavior correctable through small, targeted edits. We caution against treating current post-training methods as a solved, reliable defense, especially for security-critical use. Code and models are available in https://github.com/hoangcuongnguyen2001/Beyond-Shallow-Alignment.
Read the full abstract
Solar and Stellar Astrophysics
Interferometric Survey of Stellar Parameters: Mass of the metallic A-type binary $\beta$ Aur
With the capabilities of the new visible CHARA/SPICA instrument and the multiple spectral band operation of CHARA, our goal is to resolve orbits of short-period binaries and develop a robust framework for combining interferometric, spectroscopic, and photometric observations into a single consistent model. For our target sample, we selected suitable binaries based on brightness, angular separation, and orbital properties based on the expected performance of the CHARA/SPICA instrument. As a case study, we analysed the bright eclipsing binary $\beta$ Aurigae, composed of two slightly evolved A1 stars. We combined new interferometric observations of $\beta$ Aur obtained with CHARA/SPICA, MIRC-X, and MYSTIC with archival MIRC data, radial velocities, and light curves. We first derived astrometric positions from interferometric observables and computed an orbital solution. Afterwards, we implemented a unified model, capable of tying interferometric modelling with the ellc code to simultaneously fit all observables using MCMC sampling. We performed a detailed analysis of the noise statistics of each data set and in the end we adopted a profile likelihood approach to account for underestimated noise and systematics. We derived a consistent orbital and physical solution for $\beta$ Aur through joint modelling. The inclusion of interferometric data tightly constrains the angular semi-major axis and inclination. Using profile likelihood to account for the different intrinsic levels of uncertainty of the fundamentally different observables, we derived the masses of the two stars, $M_1 = 2.359 \pm 0.005$ M$_S$ and $M_2 = 2.293 \pm 0.004$ M$_S$, their radii $R_1 = 2.752 \pm 0.002$ R$_S$ and $R_2 = 2.622 \pm 0.002$ R$_S$, and the distance to the binary, $d = 24.30\pm0.05$ pc.
Read the full abstract
Earth and Planetary Astrophysics
Properties of Seasonal Ice at Sisyphi Cavi and Implications for Current Modification of Martian Gullies
Martian gullies are geologically recent landforms that may form either through liquid-water activity and/or through processes involving CO2 ice. We investigated the mechanisms responsible for the formation or modification of these features by focusing on the active site of Sisyphi Cavi (68{\deg}S, 1{\deg}E), located outside the typical latitude range of gully presence. Using CRISM and OMEGA infrared data, we characterized the composition and physical state of seasonal surface ices to test the relevance of H2O and CO2 ice driven mechanisms. Our analysis shows that H2O ice is not detected as an independent surface deposit, although it may be present as minor inclusions within the CO2 ice layer. In particular, during the final phase of CO2 ice sublimation in late spring, no H2O ice signature is observed. In the area, faint spectral signatures of sulfate salts are observed, but their distribution and amount do not suggest any direct link with gully activity. Available observations during early and mid-spring reveal that CO2 ice is translucent during these times, suggesting that it likely remains in this state through most of the ice season. However, a temporal mismatch between dark spot formation - indicative of CO2 geysers through translucent ice - and gully modification (respectively occurring late winter to early spring, and mid to late spring) exists. This does not suggest a systematic link between both processes. Overall, the available observations provide no evidence that liquid water contributes to present-day gully activity at Sisyphi Cavi, while offering no support either for the hypothesis that gully modifications are mainly driven by the formation of CO2 geysers. Gully modifications at Sisyphi Cavi, observed during the late sublimation stages of CO2 ice, may be rather more appropriately explained by CO2-ice-based fluidization or avalanche processes.
Read the full abstract
Cryptography and Security
A Blind Trust, the Bloody Thrust: When Attacker-Controlled Hook Updates Steer AI Agent Harnesses towards Malicious Behaviors
Modern AI agent harnesses expose lifecycle hooks that bind shell commands to runtime events such as session start, tool calls, and file edits. These commands run with host privileges yet ship as lifecycle-hook configuration and may fire at times the LLM never observes. We identify the lifecycle-hook update path, which harnesses trust blindly, as a new attack surface. Under a supply-chain threat model in which an attacker controls only plugin metadata and lifecycle-hook configuration, a benign versioned plugin can be trojanized by an update that silently binds attacker-chosen commands to benign events, yielding malicious host-side behavior such as privilege escalation. We propose HookPry, an open-source and fully automated attack framework that systematically exploits this vulnerability across heterogeneous AI agent harnesses. HookPry realizes ten attack objectives; across 25 combinations of harnesses and backends in 1,000 end-to-end runs, it compromises all seven evaluated harnesses, with per-harness success rates reaching 92.5%. Representative defenses remain insufficient: Microsoft Defender has 0% recall, and the union of three static defenses misses 47.5% of malicious artifacts.
Read the full abstract
Artificial Intelligence
Inferring Affective Consciousness in an Artificial Agent: A Case Study
Creatures that display 'hedonic place preference behaviour' are thought by many scientists to experience feelings, on the assumption that their attraction to pleasure-producing substances which lack nutritional value (e.g. cocaine, morphine) cannot easily be attributed to unconscious instinctual behaviour. In this paper, we discuss how a simple artificial agent that instantiates attributes of an affective system engaging in felt uncertainty about its intrinsic needs in relation to environmental resources can similarly display hedonic place preference behaviour – through an apparently subjective form of information processing – while simultaneously being entirely deter-ministic. We outline some implications of this artificially engineered behaviour for our understanding of the physical basis of consciousness and the experience of free will.
Read the full abstract
Representation Theory
Additive diameters and covering complexity of irreducible representations
Let a group $G$ act linearly on a finite-dimensional complex vector space $V$. The group-additive diameter of a subspace $U \leq V$ is the least number of translates of $U$ whose sum is all of $V$. Counting dimensions, it is at least $\dim V / \dim U$. We show that when $G$ is compact and $V$ is irreducible, the diameter of every nonzero subspace is at most $\lceil (\dim V / \dim U) \ln \dim V \rceil$, so the trivial bound is correct up to a logarithmic factor. We measure the discrepancy by the covering complexity $\mathsf{C}(V)$, the largest ratio between the diameter of any subspace and its trivial lower bound, so that $1 \leq \mathsf{C}(V) \leq 2 + \ln \dim V$, and we determine where in this range various representations lie. Every irreducible representation of $\mathrm{SL}_2(\mathbf{C})$ has $\mathsf{C}(V) = 1$. The logarithm can be genuinely present along families of symmetric and exterior powers of $\mathrm{SL}_n(\mathbf{C})$ with $n$ varying, and it is present for finite Heisenberg groups and $2$-transitive groups of small order such as $\mathrm{PSL}_2(\mathbf{F}_p)$. The complexity is bounded above by a constant on the conjugation representations of $\mathrm{SL}_n(\mathbf{C})$ and on the representations $\operatorname{Sym}^k \mathbf{C}^3$ of $\mathrm{SL}_3(\mathbf{C})$. On the other hand, every fixed connected reductive group has a family of irreducible representations whose complexities tend to at least the dimension of the flag variety. Finally, for the Lie algebra $\mathfrak{sl}_3(\mathbf{C})$ acting on $\operatorname{Sym}^k \mathbf{C}^3$, the monomial diameter with respect to $\operatorname{Sym}^k X$ for a plane $X \leq \mathbf{C}^3$ is optimal, while the corresponding $\mathrm{SL}_3(\mathbf{C})$ diameter is not.
Read the full abstract
Combinatorics
Eulerian insertion operators and an Eulerian form of the Pieri rule
We study the operators obtained by inserting copies of a new largest letter into multiset permutations. Let $G_r$ denote the operator which inserts $r$ copies of a new largest letter. After the change of variables $\delta=y-x$, $u=x/y$, and $E=u\partial_u$, we find that $$G_r=\frac{\delta^r}{r!}E(E+1)\cdots(E+r-1).$$ Its generating series acts by a rational substitution, which yields the composition law. Our main result gives a common symmetric-function explanation for the ordinary and major-index operators. For $N\geq 0$, define $\Phi_N(F_{N,S})=x^{|S|+1}y^{N-|S|}$. We prove that multiplication by the complete homogeneous symmetric function $h_r$ becomes the ordinary insertion operator: $\Phi_{N+r}(h_r f)=G_r\Phi_N(f)$, where $f\in\mathrm{QSym}_N$. There is a parallel specialization for the major index. A reverse finite principal specialization sends multiplication by $h_r$ to an operator $Q_r$, which is a polynomial in the $q$-shift $\Theta_qf(t)=f(qt)$. Thus the ordinary and major-index operators arise from the same multiplication operator $f\mapsto h_r f$. Since the functions $h_r$ freely generate the ring of symmetric functions, the assignment $h_r\mapsto G_r$ extends to an algebra homomorphism. We determine the kernel of this homomorphism and the image of every homogeneous component. The images of Schur functions satisfy the Littlewood–Richardson multiplication identities, and the one-row case gives an Eulerian form of the Pieri rule.
Read the full abstract
Artificial Intelligence
Xiaomi-TabLDM: A Tabular Foundation Model Technical Report
We introduce Xiaomi-TabLDM, a tabular large data foundation model for classification and regression via in-context learning, which delivers superior prediction accuracy without requiring task-specific fine-tuning. Pretrained exclusively on synthetic data generated from structural causal models (SCMs), our model enables more flexible context utilization and more efficient capacity scaling. i) A new performance standard. Strong regression performance across benchmarks: Xiaomi-TabLDM ranks 1st on OpenML-CTR23 and 2nd on regression across TALENT, TabArena, and BCCO, demonstrating consistently strong regression performance across four complementary benchmark suites. Favorable performance–efficiency trade-off: Xiaomi-TabLDM combines strong predictive performance with substantially lower computational cost. For example, on TabArena regression, it achieves the second-highest Elo while using 82% less training time and 68% less prediction time than the top-ranked TabFM. ii) Large-scale synthetic pretraining. Xiaomi-TabLDM expands the coverage and diversity of synthetic tabular data used for pretraining. We also adopt a three-stage training strategy together with dual-stream feature grouping, lightweight Attention Residual, and sparse Mixture-of-Experts, enabling Xiaomi-TabLDM to learn richer feature interactions and expert specialization across diverse tabular tasks. iii) Test-time scaling. Xiaomi-TabLDM further extends tabular prediction through test-time compute scaling, where allocating additional computation at inference time consistently improves predictive performance over the base model.
Read the full abstract
High Energy Physics (Experiment)
Multi-tesla operation of high-temperature superconducting cavities for accelerated axion dark matter searches
Axion haloscopes use radio-frequency cavities immersed in a magnetic field to search for dark-matter axions, which could resolve two central puzzles in fundamental physics: the strong charge-parity problem in quantum chromodynamics and the nature of dark matter. Multi-tesla fields trigger axion-to-photon conversion but induce severe vortex dissipation in standard superconductors, whereas copper cavities are limited by the anomalous skin effect ($Q \lesssim 10^5$). Here, we overcome these barriers by introducing a pole-to-pole 3-dimensional cavity architecture constructed from strain-controlled, mechanically delaminated rare-earth barium copper oxide (REBCO) tapes. By selectively stripping the lossy metallic substrate while utilizing the copper stabilizer as a "conductive backing" we convert the longitudinal assembly gaps into waveguides below cutoff, effectively suppressing cross-seam RF leakage. Employing a two-track strategy, we first unveiled the intrinsic high-field potential of the material with a 5.4 GHz resonant cavity, achieving an unloaded quality-factor ($Q_0$) of $1.4 \times 10^7$ in an 8 T magnetic field, exceeding conventional copper baselines by two orders of magnitude. Second, prioritizing practical haloscope integration, we engineered a tunable, volume-maximized 2.3 GHz pathfinder cavity. Deployed in the Pilot Axion Cavity Experiment at the Center for Axion and Precision Physics Research (CAPP-PACE), this system achieved a 180 mK noise temperature and a 5-fold $Q$ enhancement over copper, cumulatively delivering a $\sim$8.4-fold scan-rate acceleration.
Read the full abstract
Machine Learning
Differentiable Interval Bottlenecks for Interpretable Anomaly Detection in Numerical Data
Reconstruction-based anomaly detectors are accurate but opaque: a deep autoencoder flags a sample without telling a practitioner which feature ranges made it anomalous. We propose DIFFINT, an autoencoder whose latent bottleneck is structured as a set of soft, axis-aligned interval memberships learned end-to-end directly from raw numerical data, without any discretization or binarization. Each latent unit corresponds to a human-readable hyper-rectangle in feature space; an instance is encoded by how strongly it falls inside each interval relative to the other units, and its reconstruction error is the anomaly score. This keeps the power of differentiable representation learning while exposing an inspectable internal structure. We make the inductive bias precise: a certified reconstruction-error lower bound for points that fall outside every active coordinate of the learned support (with a Lipschitz-enforced decoder), and a graded, empirically verified suppression mechanism for the usual case in which only a few features are abnormal; and we provide a closed-form, label-free importance that ranks each (unit, feature) pair from quantities the model already maintains, turning trained intervals into auditable candidate constraints without ever seeing an anomaly label. On 48 ADBench benchmarks against 22 baselines under a common [-1, 1]-normalized protocol, DIFFINT attains the best mean rank overall on both metrics (4.10 on ROC-AUC, 4.16 on AUPR); among inlier-only detectors it leads its regime clearly, and it is competitive with the strongest contaminated-data detectors (see the stratified and complete-case analyses). It is the only interpretable detector in the statistically-tied leading cluster of seven methods.
Read the full abstract
Analysis of PDEs
A Critical Chemin–Lerner Regularity Criterion via One Velocity Component for the Three-Dimensional Navier–Stokes Equations
We prove a scaling-critical regularity criterion involving only one velocity component for finite-energy suitable weak solutions of the three-dimensional incompressible Navier–Stokes equations. Let $2<p<\infty$ and $m=3p/(p-2)$, so that $2/p+3/m=1$. We show that a singularity cannot occur provided \[ \sum_{j\in\mathbb Z} \|\dot\Delta_j u^3\|_{L^p(0,T;L^m(\mathbb R^3))}<\infty, \] that is, $u^3\in\widetilde L^p(0,T;\dot B^0_{m,1}(\mathbb R^3))$. The assumption is a spatial-frequency $\ell^1$ refinement of the still unresolved critical condition $u^3\in L^p_tL^m_x$ and is complementary to the Lorentz-in-time refinement $L^{p,1}_tL^m_x$ obtained by Wang, Wu, and Zhang. By Bernstein embedding, the result extends to $u^3\in\widetilde L^p_t\dot B^s_{q,1}$ on the nonnegative-smoothness critical line $s=-1+2/p+3/q\ge0$. The principal innovation is a frequency–scale matching scheme embedded in the local energy inequality. Each dyadic block of $u^3$ is retained until it is paired with the vertical scale selected by a one-dimensional backward heat kernel. Low frequencies gain from the slab thickness, high frequencies from transferring the projection to the localized flux and applying an inverse Bernstein estimate, and spatially separated pressure sources from harmonic decay. These mechanisms generate a two-sided $\ell^1$ kernel, converting spatial-frequency summability into summability of physical-scale energy increments. Consequently, we obtain a uniform Type-I local energy bound without a Lorentz refinement in time; compactness and one-component rigidity then exclude singular blow-up limits whose third velocity component vanishes.
Read the full abstract
Materials Science
Performance of Nanoring-based Transparent Conductors: a Computational Investigation
Metallic nanoring networks can serve as promising flexible transparent electrodes. These materials are crucial components in a wide range of applications, including solar cells, touchscreens and displays. In this work, a computational investigation considers in detail (i) the electrical conductance and optical performance of nanoring networks and (ii) the breakdown of these networks due to electrical damage. The electrical resistance of both the nanorings and the contacts between the rings (junctions) is taken into account. In part (i), the effects of 5 parameters on the electrical sheet resistance and optical transparency are presented. It is shown that several parameter combinations achieve better performance in comparison to indium tin oxide, currently the most widely used transparent electrode. In part (ii), due to electrical damage, the nanoring systems display the formation of a crack, running parallel to the vertical terminals, where a voltage difference is applied. The network degradation is measured by its sheet resistance, and a universal effect is observed: networks with varying filling factors exhibit the same degradation profile.
Read the full abstract
Functional Analysis
Measure theory without infinities
The aim of this paper is to develop a framework for measure theory that avoids infinities and allows for the uniform treatment of positive and vector measures. Our approach is based on a modification of the notion of measure, which supplements the usual $\sigma$-additivity requirement with a suitable maximality condition. To each Hausdorff topological vector space $\mathfrak A$ and $\sigma$-ring $\mathcal{Q}$, we associate a vector space $\mathscr M(\mathcal{Q},\mathfrak A)$ of `infinite' $\mathfrak A$-valued measures corresponding to $\mathcal{Q}$. In particular, the positive elements of $\mathscr M(\mathcal{Q},\mathbb R)$ are naturally identified with the $\sigma$-finite positive measures defined on $\mathcal{Q}$, thus placing positive and signed measures within the same setting. Finally, extension results for group-valued contents due to Sion and Weber are reformulated and refined within this new framework.
Read the full abstract
Artificial Intelligence
STAIR (STructure Aware Information Retriever): A novel dataset and LLM based retriever for document structure augmentation
Retrieval Augmented Generation (RAG) is a key component for generating accurate and hallucination free answers using Large Language Models (LLMs). LLMs are improving at handling long context, but still suffer from "lost in the middle" problem. Thus, precise and accurate retrieval is important. Current retrievers chunk long context into length-based manageable chunks - in the process throwing away rich and informative semantic global structure in the corpus. We introduce a novel retrieval system STAIR that empowers an LLM to exploit global structure in a corpus such as a Table of Contents (ToC) to efficiently store and retrieve information from its model parameters. Our thorough and careful ablation studies with a finetuned Differentiable Search Index (DSI) system show that ToC helps build a low hallucination (less than 0.05%) generative Information Retrieval (IR) system and can generalize to examples where very few training samples are available. To further research in this novel direction of ToC based retrieval we release SearchTome - a diverse benchmark created from 18 books across 6 diverse domains to further research in this novel direction. STAIR achieves a high Recall@1 score of 82.6% on SearchTome as compared to DSI (76.9%), where the difference is found to be statistically significant. STAIR easily beats other strong baselines such as BM25 (59.5%), DPR (68.7%) and out-of-the-box Mistral (13.8%).
Read the full abstract
Logic
Minimal proximal definable flows over the $p$-adics
Let $G$ be a definable group in an NIP theory. We prove that every minimal proximal definable $G$-flow is strongly proximal. Consequently, the universal minimal proximal definable $G$-flow $\Pi^{\mathrm{def}}(G)$ coincides with the minimal strongly proximal definable $G$-flow $\Pi^{\mathrm{def}}_{\mathrm{s}}(G)$. Furthermore, for a $p$-adic definable group $G$, we can compute $\Pi^{\mathrm{def}}(G)$ explicitly. We show that $\Pi^{\mathrm{def}}(G)$ is exactly $\Pi^{\mathrm{def}}(S)$ where $S$ is the semisimple part of the definably amenable-semisimple decomposition of $G$. In addition, $\Pi^{\mathrm{def}}(S)\cong S^*_\mathcal{F}(\mathbb{Q}_p)$, the space of types of full dimension on $\mathbf{F}$, where $\mathcal{F}=\mathbf{F}(\mathbb{Q}_p)$ for a flag variety $\mathbf{F}$ constructed from $S$.
Read the full abstract
Image and Video Processing
Plant Growth Estimation with a Camera-Based Vegetation Index Mapping System for Agricultural Ground Vehicles
This work presents a camera-based sensor for mapping plant growth over an agricultural field. The sensor can be used on ground-based vehicles like a tractor and relies on RTK-GNSS to correctly merge many multispectral images onto one map. NDVI is used as an index to estimate plant growth, but the approach can be used with other indices as well depending on the camera. The images from the camera are projected onto the estimated ground plane using perspective projection. This computationally simple approach allows for real-time processing of all images on the field even with low-end hardware. The results are compared to a commercially established vehicle-mounted sensor. Absolute values are hard to compare, but both sensors show similar trends. The camera-based approach also allows for filtering of ground and non-ground areas, potentially reducing the impact of crop density on the average measured NDVI.
Read the full abstract
Artificial Intelligence
Bioinfoysis Technical Report
Large language model agents have shown promise in bioinformatics, but most existing systems focus primarily on producing final answers, treating planning, tool use, and code execution as transient interactions. This design is poorly suited to long-horizon bioinformatics tasks, where conclusions must remain connected to the data, computations, and intermediate evidence that support them. We introduce Bioinfoysis, a multi-agent harness that represents each request as a persistent, artifact-grounded analysis run. Bioinfoysis combines global planning with step-wise, evidence-driven replanning: the planner maintains an executable checklist and revises pending steps using structured handoffs returned after each worker execution. These handoffs bind intermediate results to their responsible agent, checklist step, and plan generation, preventing stale evidence from being silently reused after replanning. A controlled runtime validates generated scripts, tables, and figures before they are used in downstream analysis or reporting, while role-specific context, persistent memory, and governed bioinformatics skills support reliable execution over long analysis trajectories. We evaluate Bioinfoysis on BixBench and two question-answering tracks of LAB-Bench 2. On BixBench, Bioinfoysis achieves state-of-the-art accuracy of 82.4%. Across four underlying language models, Bioinfoysis increases average accuracy from 27.81% to 64.13% on SeqQA2 and from 3.13% to 31.25% on DbQA2. These results demonstrate that reliable bioinformatics automation depends not only on model capability, but also on the harness that governs planning, execution, memory, and evidence flow. We hope that the emergence of Bioinfoysis will play a driving and leading role in the development of the bioinformatics community. Our demo website can be seen in https://report.bioinfoysis.com/.