Artificial Intelligence
Papers
Fresh from arXiv, typeset properly. Highlighted words trade on the market.
All of arXiv
Wednesday 9 September
Selecting LoRA rank for diffusion fine-tuning requires balancing quality and compute cost. We present a controlled study on CIFAR-10 using a DDPM U-Net with ranks {2,4,8,16,32}, fixed optimization settings, and a reproducible local-folder pytorch-fid protocol. We report FID, trainable parameters, runtime, and GPU memory, then validate trends with extended-budget DDPM runs (20 epochs; ranks 4/8/16) and a Tiny DiT backbone (10 epochs; ranks 4/8/16). Results show moderate ranks are most efficient: rank 4 achieves the best DDPM FID (124.1380), rank 8 is close (124.2136), and higher ranks provide limited gains despite larger adaptation cost. These findings support small-to-moderate ranks as practical defaults under fixed training budgets.
Quantum Physics
High-fidelity conditional-squeezing gate in superconducting circuits via refocusing
We propose a conditional-squeezing gate for superconducting circuits, which squeezes one or two modes of a resonator along a direction determined by the state of a dispersively coupled qubit. We show that the fidelity of this gate can be substantially improved by a refocusing protocol, in an architecture accessible with current technology whose key element is a SQUID-terminated resonator. The refocusing protocol relies on a symmetry of the Hamiltonian that is absent in the previously proposed controlled-squeezing gate, which squeezes the resonator only for one state of the control qubit and, ideally, acts trivially otherwise. We apply the refocused gate to encode an arbitrary qubit state into superpositions of single- and two-mode squeezed vacuum states, in which single-photon losses are detectable through parity measurements. Including nonlinearities and dissipation, the conditional gate consistently outperforms the controlled-squeezing gate, and, for optimistic but achievable experimental parameters, the encoding reaches fidelities above 0.99 for all qubit states.
Read the full abstract
Artificial Intelligence
A Multi-Stage Rule-Chaining Framework for Compositional and Interpretable Cognitive Reasoning
The Abstraction and Reasoning Corpus (ARC) benchmarks cognitive generalization, the ability to infer and apply abstract rules from limited examples. This paper presents a multi-stage rule-chaining framework that performs compositional reasoning across symbolic, structural, and conceptual levels. The framework integrates three complementary solvers: (1) a deterministic rule discovery module that induces atomic transformations through geometric, color, and object-based analysis; (2) a pattern-composition engine that reconstructs outputs via block merging, repetition, and spatial heuristics; and (3) a structural abstraction layer that infers hierarchical and nested relationships across grids. These solvers operate sequentially within a progressive fallback hierarchy, where each stage reuses prior reasoning traces to enhance interpretability and generalization. Training passed for 995 tasks out of 1000, further evaluated on 105 tasks out of 120 and solved 230 test tasks out of 240 ARC-AGI-2 tasks. The system achieved strong coverage across deterministic, compositional, and abstract categories, demonstrating an overall accuracy exceeding 95 percent. The proposed architecture bridges symbolic reasoning and pattern synthesis, providing interpretable insight into cognitive generalization. The results suggest that rule chaining and hierarchical composition can advance machine reasoning toward transparent, human-aligned abstraction without relying on task-specific tuning.
Read the full abstract
Operator Algebras
Nonamenable groups whose reduced group C*-algebras are not pure
We exhibit nonamenable groups whose reduced group C*-algebras are not pure. More precisely, if $\Gamma$ is any countably infinite discrete group, then the reduced group C*-algebra of the restricted wreath product $(\mathbb{Z}/2\mathbb{Z}) \wr \Gamma$ has an ideal-quotient isomorphic to $\mathcal{K}(\ell^2(\Gamma))$. It therefore fails to be nowhere scattered and, in particular, is not pure. Taking $\Gamma$ to be any nonamenable group, we thereby obtain a negative answer to a question of Thiel.
Read the full abstract
Machine Learning
Artificial Intelligence Algorithms for the Detection of Pathologies Related to Lung Cancer through Image Analysis using Convolutional Neural Networks and Data Augmentation: a systematic mapping of the literature
Lung cancer is one of the leading causes of death worldwide, and its early diagnosis is crucial to improving patients prognosis and quality of life. However, the process of interpreting medical images for the detection of lung cancer is complex and requires trained experts. In this context, artificial intelligence (AI) and deep learning (DL) emerge as potential tools to automate and optimize image analysis. The objective of this work is to review the most recent and relevant applications of AI and DL in the field of radiology for the detection of lung cancer. To this end, an exhaustive search was carried out in scientific databases such as PubMed,IEEEXPLORE, Scopus and Web of Science, and 96 articles published from 2015 to the present addressing the use of AI and DL in biomedical engineering were selected. Emphasis is placed on the use of convolutional neural networks (CNN) with transfer learning and Data Augmentation as promising techniques to improve the accuracy and efficiency of the image interpretation process. The results show that the use of AI and DL can offer an effective alternative for the early diagnosis of lung cancer, with high sensitivity and specificity. However, current limitations and challenges that must be addressed to guarantee its responsible and safe application in clinical practice are also identified, such as the lack of standardized data, the ex plainability of the models, patient privacy, and the ethical and social implications. It is concluded that the use of AI and DL can have a positive impact on the care of patients with lung cancer, but further research and regulation are required to ensure its quality and reliability.
Read the full abstract
General Physics
Constraining $f(R)$ gravity and evolving dark energy via large-scale structure and phase-space trajectories
We present a joint observational analysis confronting viable $f(R)$ modified gravity theories, specifically the Hu & Sawicki and Starobinsky models, with background and large-scale structure (LSS) data. Utilizing Monte-Carlo Markov chain (MCMC) sampling across datasets including baryon acoustic oscillations (BAO), type Ia supernovae (SNeIa), cosmic microwave background (CMB) distance priors, and linear growth measurements ($f\sigma_8$, $f$, $\sigma_8$), we place tight constraints on the model parameters governing deviations from General Relativity. For the full dataset combination, we obtain $\log_{10} b_\mathrm{HS} = -6.325_{-1.138}^{+1.216}$ for the Hu & Sawicki model and $b_\mathrm{S} = (0.8\pm61.0)\times10^{-4}$ for the Starobinsky model. Model comparison based on the Akaike Information Criterion indicates that these $f(R)$ extensions are statistically favored over flat $\Lambda\text{CDM}$ ($|\Delta\text{AIC}| \ge 3.99$) for the combined data. However, when considering the Bayesian Information Criterion, the evidence for support is significantly reduced. Furthermore, we construct two-dimensional phase-space diagrams in the $(\mu, \gamma)$ and $(\mu, \Sigma)$ planes across several redshifts, establishing a novel diagnostic null-test allowing us to probe for deviations from $\Lambda\text{CDM}$, corresponding to the fixed point $(1,1)$ in both planes, using LSS observables. Should future weak-lensing and galaxy surveys provide data points with $\mu-1<0$ and $\gamma-1>1$ or $\Sigma-1 < 0 $, then the aforementioned models could be directly ruled out.
Read the full abstract
History and Philosophy of Physics
The Crab Nebula progenitor: recovering the 1054 AD supernova event as galactic Gamma-ray burst
In 1054 AD a daytime star appeared in the constellation of Taurus, for three weeks, and it was reported in various sources from Europe to China/Japan: it was one of the few documented galactic supernovae of the last two millenia. This paradigm has been established about sixty years ago, as the comprehension of the physics of supernovae progressed with enough observational data. The Gamma-ray bursts were discovered in the same period, but only in the past few years have their observations become daily and their distances have been fully understood as cosmological. After the explosion, the exponential decay of the luminosity in gamma-rays and X-rays has been followed with telescopes onboard dedicated satellites. Also the exponential decay of the afterglow's optical and radio frequencies have been observed with the largest optical and radio telescopes. Within the binary-driven hypernova framework, successful in explaining all the observed phases of the Gamma-ray bursts, the universal exponential decay can be extended to 1000 years after the burst, to account for the present values of Gamma and X-rays as well as optical and radio frequencies of the Crab Nebula. Both the daytime visibility of the burst, and the simultaneous radiation plagues appeared in Constantinople and Cairo is a strong evidence of the presence of Gamma-rays in the lower atmosphere, coming from the same source originating the Crab nebula. The association to the daytime visibility of that star and the following plague meets exactly the etymology of the word dis-aster, bad star.
Read the full abstract
Fluid Dynamics
An exposition on some viscosity dependent issues in refined potential flow theory
This study evaluates viscosity-dependent limitations in refined potential flow theory (RPT) applied to incompressible flow over an impulsively started circular cylinder. RPT adapts classical potential flow theory to account for viscous effects, flow fluctuations, and three-dimensional variables. By utilizing a quasi-irrotational viscous stream function, the framework solves the continuity and Navier-Stokes equations while satisfying both free-stream and no-slip boundary conditions.
Read the full abstract
Software Engineering
Governed Human-AI Prioritization Under Uncertainty: Adaptive Estimation and Dependency-Constrained Portfolio Selection
AI-native software engineering increasingly combines human judgment, historical analogy, parametric estimation, and AI-generated forecasts in the same prioritization decision. We study five quantitative operators used in the D-POAF decision practice and stress-test them under uncertainty, estimator dependence, drift, shared context error, dependencies, and limited capacity. Robustness extensions use 30 deterministic seeds. Moderate scoring changes preserve broad order while materially changing funded-set membership. Under independent estimator errors, inverse-MSE aggregation reaches MAE 0.615, a 42.2% reduction relative to the best individual estimator. The gain falls to 4.4% at error correlation rho = 0.6 and becomes negative at rho = 0.8, while calibration-set regression remains slightly better than the best individual and approaches a test-set oracle. Under estimator drift, four-Wave adaptive weighting reduces average RMSE by 4.64% (95% CI [4.50%, 4.78%]). Model-collective divergence achieves ROC-AUC 0.908 with independent channels and remains above 0.878 with shared noise up to 0.8. We separate budget-constrained portfolio selection from ODP's native dependency-aware sequencing role. When ODP distance is repurposed as a portfolio-selection heuristic, value-to-effort is stronger in the tested generator. For a fixed selected set, ascending ERS/BVS is optimal for an ERS-weighted completion objective without dependencies and reaches mean efficiency 0.982 versus the exact precedence-constrained optimum at dependency intensity lambda = 1.0. Scaling to 100 Feature Blocks, ODP has the highest mean early-value AUC among the transparent sequencing baselines, with a small margin over PVS ordering. These results characterize both the validity region of evidence-weighted human-AI prioritization and the boundary between selection and sequencing objectives.
Read the full abstract
Machine Learning
Byzantine-Robust Federated Fire Detection with a Rotating Coordinator
We study the application of federated learning (FL) to indoor fire detection. Such fire-detection systems use edge cameras that record sensitive footage which cannot easily be collected at a central server. Existing federated solutions leave three practical obstacles unaddressed: limited uplink bandwidth, Byzantine (malicious or faulty) clients, and unconditional trust in a single, permanently fixed aggregation server. Our main contributions address all three. In particular, we provide (i) a curated indoor fire-detection dataset assembled from eight public sources; (ii) an edge-deployable detector whose model updates are compressed up to 10 times with only a small loss in balanced accuracy; and (iii) a semi-decentralized Byzantine-robust FL method that combines history-aware aggregation with a rotating coordinator, evicting stealthy attacks that per-round filters miss while removing the fixed-server single point of failure. On the held-out test set the rotating-coordinator method matches its fixed-server counterpart in accuracy and detection speed, and a physically distributed six-node cloud deployment confirms feasibility.
Read the full abstract
Statistical Methodology
Mixture-based Nonparametric Estimation of Spatial Covariance Functions with Applications to HIV Key Population Size Estimation across Sub-Saharan Africa
Consistent data on the sizes of key populations, such as female sex workers (FSWs), are often scarce, particularly at the sub-national level. Accurate size estimates are critical to effectively allocate resources and achieve HIV targets. Since FSW population sizes may be spatially correlated across areas, models that account for spatial dependence can improve estimation. An important component of such models is the covariance function, which characterizes the spatial dependence structure of the underlying process. In this work, we study spatial covariance functions to estimate FSW population sizes in Sub-Saharan Africa (SSA). Many spatial models rely on parametric covariance functions. However, parametric estimation can suffer from model mis-specification, potentially leading to inefficient or biased predictions. We therefore develop a robust non-parametric approach for estimating the covariance function of a stationary isotropic process in $\mathbb{R}^d$. We focus on a class of covariance functions that are valid in all dimensions, which includes popular kernels such as the exponential and Matérn kernels. Leveraging the fact that such covariance functions can be represented as infinite mixtures of scaled Gaussian kernels, we propose two estimation methods: weighted least squares and nonparametric maximum likelihood estimation to estimate the mixing measure of scaled Gaussian kernels. We also develop computationally efficient methods to solve these optimization problems using non-negative least squares and second-order descent updates. We evaluate the proposed methods through simulations and apply them to estimate the FSW population sizes at the sub-national level in SSA.
Read the full abstract
Sound
Sparse Weight and Edge Circuit Discovery in Transformer-based Acoustic Models
Transformer-based foundation models are powerful but opaque, motivating Mechanistic Interpretation methods to uncover the black-box by identifying small computation subgraphs responsible for a task. DiscoGP is a joint weight-and-edge circuit discovery framework originally developed for text decoders. We extend DiscoGP to speech encoders and present, to our knowledge, the first circuit discovery study for modern speech foundation models. Across HuBERT and Wav2Vec 2.0 on several speech classification tasks, we find that the discovered circuits are extremely compact, yet often match or even exceed the performance of the full pretrained encoder with the same downstream head. Through ablations, we show that these circuits reflect pretrained computation rather than random structure or task-head artifacts. We also introduce a memory-efficient DiscoGP variant that reduces the GPU memory cost of edge-circuit discovery at runtime from quartic to cubic. Overall, our results broaden Mechanistic Interpretation beyond text decoders and show that circuit-level analysis can reveal both explanatory structure and unexpected functional behavior in speech encoders.
Read the full abstract
Biomolecules
Sequence-Informed Geometric Evaluation of RNA 3D Structures
Computational RNA structure pipelines generate many candidate conformations for the same sequence. Reliable evaluation therefore requires more than recognising plausible geometry, it requires determining whether that geometry is compatible with the sequence. We introduce SIRGE, a sequence-informed geometric evaluator that conditions structural representations on nucleotide embeddings from a pretrained RNA language model. Early results show that SIRGE outperforms established evaluators in Kendall–$\tau$ alignment, Top-1 selection, and Top-3 ranking. Controlled comparisons further show that sequence conditioning corrects errors made by an otherwise matched geometric model and improves target-level rank structure. These findings provide initial evidence that pretrained sequence representations supply ranking information that complements geometric reasoning.
Read the full abstract
Machine Learning
Zero-shot rib design: merging training-free generative prior with topology optimization
Natural load-bearing patterns such as leaf venation, trabecular bone, and spider webs achieve high stiffness per unit mass, yet classical topology optimizers rarely reach such geometries, and few let engineers express structural design intent through natural language. This work treats a frozen text-to-image diffusion model as a training-free source of design knowledge and distills it into the physics loop of density-based topology optimization via score distillation sampling, so that a text prompt becomes an explicit, machine-interpretable representation of engineer intent. The prompt-induced generative gradient and the finite element sensitivity are combined at every iteration, letting physics decide which prompt-induced features survive. In 245 primary SDS runs spanning four geometric domains and two physics regimes, 38 of 49 prompt–domain combinations achieved statistically significant compliance reductions (up to $-31.5\%$ mechanical and $-23.0\%$ thermoelastic), outperforming gradient-based baselines. Cross-domain morphological analysis identifies a recurring structural signature of improvement: in most domains the generative prior suppresses dead-end branches in the rib skeleton, with endpoint–compliance correlation $r = +0.56$ to $+0.99$. A Heaviside projection with $\beta$-continuation resolves a pronounced intermediate-density tendency in this diffusion–physics coupling ($42.6\%$ to $<3\%$), and an automated skeleton-based pipeline converts optimized density fields into \rev{candidate geometry ready for computer-aided design. By retargeting the generative prior across domains, loading conditions, and physics objectives through a change of text prompt, with each new problem's physics setup specified separately, the framework uses a pretrained generative model as a reusable, training-free prior for engineering design.
Read the full abstract
General Mathematics
A Four-Genus Kronecker-Limit Evaluation of the Alternating Rogers-Ramanujan Continued Fraction
We evaluate the six odd class-number-four cases left unevaluated in Ramanathan's treatment of the Rogers-Ramanujan continued fraction. Let \[ S(q)=-R(-q),\qquad R(q)=\cfrac{q^{1/5}}{1+\cfrac{q}{1+\cfrac{q^2}{1+\cfrac{q^3}{1+\cdots}}}}. \] For $n=39,87,111,119,159,287$ we determine the value of $S(e^{-\pi/\sqrt{5n}})$ by applying the genus-character form of the Kronecker limit formula to the four ideal classes of $\mathbb Q(\sqrt{-5n})$. The resulting expressions are given in terms of fundamental units of real quadratic fields. In particular, if $X_n=S(e^{-\pi/\sqrt{5n}})$, then \[ X_n^{-5}+11-X_n^5=\frac{5\sqrt5}{U_n}, \] where the six quantities $U_n$ are displayed explicitly below. We also give the corresponding radical expressions and quartic algebraic certificates, together with numerical checks of the six evaluations.
Read the full abstract
Data Structures and Algorithms
A Deadline-Driven Algorithm for Polyamorous Scheduling
In Polyamorous Scheduling Problem, we are given an edge-weighted graph and must find a periodic schedule of matchings in this graph which minimizes the maximal weighted waiting time between consecutive occurrences of the same edge. This NP-hard problem generalises Bamboo Garden Trimming and is motivated by the need to find schedules of pairwise meetings in a complex social group. We present a $4 G^*$ algorithm for Polyamorous Scheduling, improving the previously known bound of $3 + \sqrt{5} \approx 5.236$. Our algorithm is inspired by the Deadline-Driven Heuristic which is optimal for Bamboo Garden Trimming (BGT).
Read the full abstract
Quantum Physics
A nonrecursive method for computing the off-diagonal small-time heat kernel expansion
We present a nonrecursive method to compute the small-time heat kernel asymptotic expansion. The Klein-Gordon operator in flat space-time coupled to electromagnetic fields is considered. We obtain closed-form expressions for the expansion coefficients up to the second order in time. Our approach can be regarded as an extension of the method described in [R.I. Nepomechie, Phys. Rev. D 31, 3291 (1985)], but unlike that work, our approach is valid for arbitrary space-time dependence of the heat kernel, i.e. for off-diagonal values. We verify that our results correctly reproduce small-time asymptotics of the heat kernel for a plane wave field and for constant electromagnetic fields.
Read the full abstract
Numerical Analysis
Vectorial lattice Boltzmann solver for compressible inviscid flows with generic equation of state
We develop a vectorial lattice Boltzmann model with a space-time adaptive relaxation coefficient, combined with adaptive time-stepping for compressible Euler dynamics with generic equation of state. The model, due to special form of the relaxation coefficient devised here is shown to converge to the Euler limit with second-order accuracy in the absence of shocks under acoustic scaling. The solver is robust and able to properly capture compressible gas dynamics for both ideal and non-ideal equations of state, including the Bethe–Zel'dovich–Thompson regime. This is verified and demonstrated through a variety of configurations of increasing complexity.
Read the full abstract
Image and Video Processing
Physics-Informed Sequential Reconstruction of Scanning Probe Microscopy Images with Calibrated Uncertainty
Scanning probe microscopy (SPM) has become a mainstay of materials science, condensed matter physics, nanotechnology, biology, and semiconductor metrology. An SPM image, however, is acquired sequentially as the probe traverses the surface under closed-loop feedback while thermal drift, feedback dynamics, tip state, mechanical disturbances, and electronic noise evolve during the experiment. Image correction has therefore become an intrinsic part of SPM data analysis. Classical approaches rely on line leveling, filtering, registration, interpolation, and explicit models of tip or scanner distortions, whereas recent work increasingly uses neural networks trained to map corrupted images to corrected ones. Here, we explore a different formulation in which SPM reconstruction is treated as a Bayesian sequential state-estimation problem. The hidden state represents surface height and its low-order evolution along the slow-scan direction; an artifact-dependent quality score continuously modifies the observation likelihood; a forward Kalman pass provides the causal estimate and innovation diagnostics; and a Rauch-Tung-Striebel (RTS) backward pass incorporates information from the complete acquisition. The result is a reconstructed surface, a relative posterior uncertainty map, and line-resolved diagnostic signals. The recursion is computationally lightweight and scales linearly with the number of image pixels, while the physical parameters can in principle be initialized from instrument characterization or from priors learned during previous operation, an extension not exercised here. We organize the method as a multi-axis family of models so that the contributions of state representation, artifact recognition, lateral coupling, and acquisition redundancy can be evaluated independently.
Read the full abstract
High Energy Physics (Theory)
Metric Source Completions and Integrability in Celestial Gravity
Opposite-helicity soft limits carry ordering information, whereas a metric generating functional has a symmetric mixed response. We compare these structures for Einstein tree amplitudes, retaining the full momentum-conservation distribution. Normalized transport of primary sources fixes the metric response, and a common Banerjee–Pasterski prescription fixes the two ordered scattering currents on compact regular hard packets. At the reference metric, both ordered scattering responses agree with the transport Hessian, so the two matching defects vanish separately. A four-hard-particle example displays the same nonzero recoil coefficient in both orders and in the transport Hessian; an overlapping-source packet resolves the lower-order spin-frame contact. Six-graviton factorization independently tests the recoil and undifferentiated coefficients of the ordering difference. At seven points, angular integration produces a nonzero relative current whose connector variation contains a conjugate bulk term. We derive this term and its cancellation under coherent transport; global cubic scattering matching still requires the physical Ward comparison on the remaining singular overlaps. Transporting the states, kernels and inverse chains together gives a finite Ward identity and coefficientwise reconstruction for that chosen family. Regular packets leave extensions supported solely on excluded hard–hard coincidences undetermined, and the vacuum functional, independent of the hard sources, remains free.
Read the full abstract
High Energy Physics (Phenomenology)
Solar Capture and Suppressed Annihilation of Inelastic Scalar Dark Matter
Solar capture of inelastic dark matter does not by itself fix the annihilation signal. We consider a scalar low-energy model with a $B-3L_\tau$ current, motivated by the high-energy recoil candidate reported by LUX-ZEPLIN. For $m_\chi=840~\mathrm{GeV}$, a $320~\mathrm{keV}$ splitting, and a $160~\mathrm{MeV}$ mediator, the public LZ response gives 0.965 expected accepted events, while micrOMEGAs gives $\Omega h^2=0.11999834$. Finite-momentum nuclear responses leave iron as an important solar-capture target despite the near cancellation of its coherent charge at zero momentum transfer. The captured population remains extended after thermal nuclear motion, elastic and inelastic scattering, and excited-state decay are included, with $r_{\rm eff}=0.215R_\odot$. Same-state annihilation is open, but the inferred rate is only 0.0404 of the public IceCube $W^+W^-$-calibrated proxy and is smaller than the corresponding isothermal estimate by a factor of 628. For this benchmark, substantial solar capture is accompanied by a weak neutrino signal.
Read the full abstract
Functional Analysis
Spectral and numerical radius preservers on pairs of unitarily similar operators
Let $\mathcal{H}$ be a complex Hilbert space. We characterize the bijective linear maps $\Phi:\mathcal{B}(\mathcal{H})\to\mathcal{B}(\mathcal{H})$ that preserve the spectral radius, respectively the numerical radius, on pairs of unitarily similar operators. We also characterize bijective linear maps that transform unitary similarity into similarity. Our results cover both finite and infinite-dimensional Hilbert spaces, with no separability assumption in the latter case.
Read the full abstract
Cryptography and Security
HermiCache: Enclave-Aware Cache Replacement for Trusted Execution Environments
Trusted Execution Environments (TEEs) protect enclave memory from untrusted software but remain vulnerable to cache-based side-channel attacks due to shared microarchitectural resources. Existing countermeasures use techniques such as cache partitioning or randomization: these solutions are not ideal if a designer wants fine-grained configurations and a deterministic protection. In this paper, we introduce HermiCache which is an answer to these requirements. HermiCache is designed for RISC-V cores and has been implemented in the OpenHwGroup CVA6 core with a Keystone TEE for the software layer. The solution has an area overhead of 6% on the processor core.
Read the full abstract
General Mathematics
A Proof of Liu's Conjecture on the Fundamental Triangle Inequality
Let $a,b,c$ be the side lengths of a triangle, and let $R$ and $r$ denote its circumradius and inradius, respectively. Liu proposed the conjecture \[ \sum_{\text{cyc}} \left(\frac{a(b+c-a)}{bc}\right)^k \ge 2+\left(\frac{2r}{R}\right)^k,\qquad k>1, \] with the reverse inequality for $k<1$. We prove this conjecture by reducing it to an algebraic inequality for three positive variables with prescribed sum and product. We also determine the equality cases.
Read the full abstract
Software Engineering
Numbat: Building and Verifying a Self-Contained Machine-Learning Stack
Machine-learning systems are built almost exclusively on a few large Python-orchestrated frameworks, and they inherit those stacks' engineering costs: environments of hundreds of version-coupled packages, separate export toolchains for deployment, and the split between the language research is written in and the language products ship in. We report on the construction and verification of numbat, a machine-learning stack written in one general-purpose language (Zig) with no third-party runtime dependencies. The stack spans tensor computation, automatic differentiation, neural-network modules, mixed precision, multi-GPU training, data loading and monitoring; an SDK exposes it behind a stable, additively versioned C ABI of over 1,400 entry points, with bindings for six languages; and its clinical domain planes encode regulatory requirements as executable acceptance gates rather than documentation. Verifying such a stack is the harder half of building it: a defective training run rarely fails, it converges quietly to a slightly worse model. We treat a widely used reference implementation as an executable specification and verify against it at five levels, from operator gradient checks to an automated trajectory gate against a same-machine reference run - the arrangement our companion study formalizes as a trajectory-level differential oracle. The protocol surfaced ten silent recipe divergences, which we catalog with mechanisms and symptoms. As the acceptance test, we train a 25.9M-parameter detector of the YOLOv8m class from random initialization on COCO 2017 for the full 500-epoch schedule: the exported weights score 0.4956 mAP50-95 under the official protocol, scored by the reference stack's own validator (published endpoint 0.502), with single-GPU step time at parity on identical hardware. Weights, per-epoch metrics and the full run manifest are released.
Read the full abstract
Cryptography and Security
From Cycle Space to Cycle Manifold: Limits and Achievability of Blind False Data Injection Attacks
A false data injection attack (FDIA) can change the estimated grid state while evading a residual-based bad data detector (BDD). Existing blind attacks learn a low-rank measurement subspace, but this algebraic view does not state the physical grid constraints that make an attack stealthy or the minimum information needed to recover the complete attack space. Under the connected direct-current (DC) branch-flow model, we show that the residual-sensitive subspace of the noiseless orthogonal test is exactly the weighted cycle space. Its orthogonal complement is therefore the complete stealthy attack space, making weighted cycle-space knowledge both necessary and sufficient for complete blind FDIA. This space identifies the topology only up to 2-isomorphism and the relative cycle-edge parameters only up to one scale per biconnected component; bridge parameters are neither identified nor required. We then formulate a computationally unconstrained benchmark and a tractable measurement-only reconstruction method. Experiments on IEEE systems compare BDD bypass rate at a 95% nominal-acceptance threshold against state impact. As a compact alternating-current (AC) extension, we characterize feasible branch P/Q measurements by a cycle manifold and demonstrate topology-assisted manifold fitting and measurement generation on a graphics processing unit (GPU). In the lossless fixed-voltage small-angle limit, the normal space of the active-power slice reduces to the DC weighted cycle space.
Read the full abstract
Software Engineering
AI Safety: Not Optional, Not Later
Incidents show that AI safety failures often arise across multiple layers. We present a safety-by-design assurance architecture combining model-level supervision, such as Scientist AI, with system-level controls over scaffolds and harnesses, independent verification, monitoring, and evidence infrastructure, supported by governance for accountability and evidence interoperability.
Read the full abstract
Artificial Intelligence
QuantumQUBO Agent: Automating Quadratic Unconstrained Binary Optimization (QUBO) Formulation Generation from Natural Language
Quadratic Unconstrained Binary Optimization (QUBO) is a central formulation for combinatorial optimization and has gained increasing attention due to its compatibility with quantum, hybrid quantum-classical, and quantum-inspired solvers. However, translating natural-language problem descriptions into correct QUBO formulations remains difficult, requiring the identification of binary variables, constraints, objective functions, penalty terms, and suitable penalty weights. This process is time-consuming and often demands substantial domain expertise. To address this challenge, we propose an end-to-end multi-agent framework that automatically generates QUBO formulations from natural-language problem descriptions, supported by structured or unstructured test cases. To evaluate its performance, We also introduce QUBOBench, a benchmark containing 100 combinatorial optimization problems across 12 application domains, curated from peer-reviewed literature, competitions, and canonical NP-hard problems. Experimental results show that our framework achieves 68% accuracy on QUBOBench, outperforming a direct single-call baseline by 22%. Further analysis identifies iterative self-repair as the most important component contributing to improved performance. The data and code are open-sourced at https://quitttcat.github.io/QuantumQUBOAgent.
Read the full abstract
General Mathematics
Convexity of a ratio function involving polygamma functions and monotonicity of the sectional curvature of the Fisher–Rao manifold
This paper proves that the sectional curvature of the Fisher–Rao manifold of beta distributions, restricted to the diagonal, is strictly decreasing. The proof is based on a general convexity lemma for ratios involving positive decreasing functions and is applied to a quotient of polygamma functions.
Read the full abstract
Cryptography and Security
SoK: Privacy Attacks on Machine Learning via Explainable AI
Machine learning explanations reveal model behavior beyond predictions, creating attack surfaces for model confidentiality and data privacy. We systematize 25 studies that exploit explanations for model extraction, membership inference, and model inversion, treating attribute inference as partial inversion. Existing work is often labeled only black- or white-box, obscuring substantial differences in what explanation signal reaches an adversary. We therefore separate model knowledge from explanation acquisition and identify five paths: target-released, attacker-derived, secondary disclosure, privileged access, and released global artifacts. Across these paths, explanations reduce extraction cost, expose membership signals through explanation statistics, recourse distance, and explanation-guided robustness, and support spatial or algebraic reconstruction of private inputs. We compare system and threat models, explanation signals, auxiliary knowledge, target models, modalities, query budgets, evaluation metrics, reported performance, and defenses. Our analysis shows that no explanation family is uniformly unsafe and no defense is uniformly effective. Risk depends on which signal is exposed, how it is acquired, which asset is targeted, and what the attacker already knows. We argue that explanation privacy should therefore be evaluated as an end-to-end disclosure problem, with defenses matched to the acquisition path and protected asset.