Instrumentation and Detectors
Papers
Fresh from arXiv, typeset properly. Highlighted words trade on the market.
All of arXiv
Saturday 5 September
The Cr$_2$O$_3$ solid target for low-energy $^{16}$O+$^{16}$O fusion experiments was fabricated using filtered cathodic vacuum arc (FCVA) deposition. Its composition, oxygen areal density and impurity content were characterized by elastic backscattering spectrometry (EBS), and the impurity-induced background contributions were investigated by in-beam $\gamma$-ray spectroscopy. EBS results indicate that the Cr$_2$O$_3$ film exhibits good stoichiometry and uniformity, with $^{16}$O areal densities ranging from $(3.24-3.25)\times10^{17}$ atoms/cm$^2$. The EBS analysis reveals a carbon atomic fraction of approximately 1.25-1.29% in the Cr$_2$O$_3$ layer, while a large amount of carbon impurities are also identified on the surface of Cr substrate. In-beam $\gamma$-ray spectra reveal prominent transitions associated with $^{27}$Al and $^{24}$Mg at 844, 1015, and 1369 keV, mainly originating from $^{12}$C+$^{16}$O fusion reactions induced by carbon impurities under $^{16}$O irradiation. Meanwhile, characteristic $\gamma$-rays emissions from evaporation channels of the $^{16}$O+$^{16}$O reaction, including $^{31}$S, $^{31}$P, and $^{28}$Si, were also observed and can be used to extract the $^{16}$O+$^{16}$O fusion cross sections. This work provides an experimental basis for the development of high-purity oxide targets and the optimization of target configurations for future low-background $^{16}$O+$^{16}$O fusion cross section measurements.
Computation and Language
What the Window Does Not Contain: Auditing Provenance in a Document-Grounded Instability Benchmark
Ask a language model the same question about the same document twenty times, and it sometimes returns two different answers. We built Probity, a benchmark of 60 tasks and 470 items from real venture-financing filings, to measure how often this happens. Then we audited our own corpus and found a defect any excerpt-built benchmark can carry: items whose evidence is missing from the window of text the model is shown. The audit flags 36 items and separates two failures a single flag would conflate: evidence genuinely absent from the window and answers that must be computed from numbers the window does supply. Flagged items change their answers far more often, wobbling at 0.255 against 0.087 on the 427 clean items, and excluding them cuts apparent cross-model agreement by about a fifth. Before testing whether the missing evidence explains the instability, we registered a prediction: re-cut each window to hold its evidence, and instability should fall below a set threshold. It failed: the repair moved wobble by 0.058, with an interval containing zero. We report the association as correlational. Almost all measurements sit where instability cannot show, which bounds what a corpus built for accuracy can say about stability. We release the corpus, all 112,800 raw responses, and the audit as a runnable check for any document benchmark.
Read the full abstract
Optimization and Control
Low-rank matrix recovery landscapes beyond RIP with application to rank-one measurements
We study the problem of low-rank matrix recovery from linear measurements via the global nonconvex landscape of a low-rank factored formulation of the matrix LASSO (nuclear-norm–regularized least-squares). If the landscape is benign, that is, has no bad local optima, then practical and scalable algorithms can compute good statistical estimates. Previous state-of-the-art landscape guarantees have typically assumed that the linear measurement operator has the restricted isometry property, that is, the operator is approximately an isometry over all low-rank matrices. This is an unrealistic assumption for many applications; in particular, when the individual measurement matrices are themselves low-rank, we typically have poor upper isometry constants. To overcome this, we establish new guarantees of a benign landscape under a weaker isometry condition: rather than requiring upper isometry over all low-rank matrices, we only require it over the linear low-rank tangent space to the low-rank ground truth matrix. To illustrate the utility of this result, we apply it to the problem of matrix recovery from random rank-one linear measurements; via high-probability concentration bounds on the random measurement operator, we prove a novel landscape guarantee with statistically near-optimal sample complexity and recovery error.
Read the full abstract
Combinatorics
Bounded Ratios of Lorentzian Polynomials I: The Ternary Theory and Optimal Bounding Constants
We study bounded ratios and optimal bounding constants among the normalized coefficients of ternary Lorentzian polynomials. For every fixed $M$-convex support and in arbitrary degree, we give an explicit presentation of the bounded-ratio cone in terms of quadratic Hessian slices. We then express the optimal bounding constants through a variational formula combining local support functions with linear compatibility constraints between slices. For full support, we determine all compatibility relations in arbitrary degree; in degree three, this yields explicit optimal constants for every two-generator section. Finally, we compare the resulting Lorentzian bounds with those for volume polynomials and rank-three matroid basis profiles.
Read the full abstract
Computer Vision
TBDub: Production-Oriented Visual Dubbing
Visual dubbing must synchronize mouth motion with replacement speech while preserving identity, appearance, and temporal consistency. Although X-Dub provides a strong mask-free video-editing baseline, its application to livestream and generated-video content reveals limitations in production-domain robustness, temporal and motion stability, identity and oral-detail preservation, and inference efficiency. We present TBDub, a production-oriented extension of X-Dub that combines task-adaptive post-training with task-aware few-step distillation. Post-training adapts the video DiT using production-domain data, production-specific conditioning and filtering, and enhanced audio features to obtain a 30-step Teacher. Distillation adapts DMD/DMD2 to conditional video editing and compresses the Teacher into a two-step Student. On 38 TalkVid clips, the Teacher improves all eight reported reconstruction, perceptual, identity, and synchronization metrics over X-Dub. In the MOS evaluation, it improves lip-sync consistency, identity consistency, and visual quality over X-Dub by 0.14, 0.95, and 0.90 points, while the Student achieves the highest lip-sync and visual-quality scores and remains close to the Teacher in identity consistency. In paired end-to-end generation timing from the first VAE encode through the final VAE decode on a single NVIDIA H20 GPU at $512\times512$, the Student reaches 7.13 effective FPS and reduces total latency by $13.93\times$; the DiT stage alone is accelerated by $42.49\times$. The Student largely retains the Teacher's generation quality and audiovisual synchronization. The code is available on GitHub at \https://github.com/TaoLiveAIGC/TBDub, and the 30-step Teacher and two-step Student weights are available on Hugging Face at https://huggingface.co/TaoLiveAIGC/TBDub.
Read the full abstract
Analysis of PDEs
Modified scattering for semiclassical Bose-Fermi mixtures
The Vlasov-Hartree system models a mixture of bosons and fermions interacting via Coulomb forces in which the bosons are described quantum-mechanically, while the fermions are described classically. For small initial data, we prove that the system exhibits modified scattering, i.e. the bosonic and fermionic subsystems each converge to solutions of the free Schrödinger and Vlasov equations respectively, but with coupled logarithmic corrections owing to the long-range nature of Coulomb interactions. We use a mixed Lagrangian and Fourier-based approach to derive explicit descriptions of the asymptotic dynamics of the solution and associated fields and Lagrangian trajectories. Notably, due to our purely Lagrangian analysis on the Vlasov side, we require no derivatives on the initial fermion density, the asymptotic fermionic profile is as regular as the initial data, and convergence occurs in almost the same space as the initial data.
Read the full abstract
Fluid Dynamics
Hydrodynamic mechanism and suppression of cavity breathing oscillations in a twin-cavity supersonic combustor
Low-frequency cavity breathing oscillations can strongly affect flame stabilization and operability in cavity-stabilized supersonic combustors, yet their hydrodynamic origin and suppression remain insufficiently understood in opposed twin-cavity configurations. We investigate the breathing dynamics of an opposed twin-cavity supersonic combustor under low- and high-enthalpy conditions using synchronized high-speed Schlieren imaging, wall-pressure measurements, numerical simulations, nonlinear phase-space reconstruction, Hilbert-transform analysis, and reduced-order modelling. Under low-enthalpy conditions, a self-sustained breathing mode is observed at a dominant frequency of approximately 156 Hz, with large phase differences between cavity-floor and ramp pressures. Measurements and simulations show that periodic mass exchange between the cavity and freestream drives cyclic cavity depressurization and repressurization, establishing a pressure-shear-layer feedback loop that sustains the oscillation. A two-pressure-state reduced-order model, coupled with a shear-layer displacement coordinate, reproduces the measured frequency, pressure phase relationship, and stability characteristics, demonstrating that both bulk-cavity and reattachment/shock-foot pressure states are required to capture the dynamics. Upstream nitrogen injection produces only transient attenuation, with the oscillation recovering as the injected gas convects downstream. In contrast, ethylene injection followed by ignition permanently suppresses the oscillation, with the nonlinear attractor collapsing to a stable equilibrium. Heat release reorganizes the cavity pressure field and alters density, compressibility, and pressure-response timescales, thereby weakening the feedback mechanism responsible for sustaining the breathing mode.
Read the full abstract
Fluid Dynamics
Sloshing reduction in a swaying tank with porous baffles using scaled boundary finite element method
Sloshing is an inevitable phenomenon in an ocean-going vessel that can have adverse effects. In this work, the mitigation of sloshing is investigated using multiple thin porous baffles of various configurations in a partially filled swaying tank. The boundary value problem is solved within the framework of a linearized potential flow theory using the scaled boundary finite element method (SBFEM). The flow through the thin porous baffles is assumed to follow Darcy's law. The computational domain is divided into a minimum number of subdomains due to the presence of porous baffles and to ensure star convexity. Higher-order polynomials are used along each subdomain edge to represent the unknown field, i.e., velocity potential. The developed numerical model is validated with the known results in the literature. Subsequently, various results, such as the amplification factor and the forces of the tank wall, are presented and discussed for the effect of configuration, porosity, slosh tank width, depth of baffle submergence and the space between adjacent baffles. From the parametric study, it is observed that top-mounted baffles enhance sloshing suppression by $50\%$ compared to bottom-mounted vertical baffles, considering all sloshing modes. Assessing the overall effectiveness, top-mounted convex baffle configuration emerges as the most efficient configuration for sloshing suppression, achieving a well-balanced reduction across all modes.
Read the full abstract
Cryptography and Security
Counter-Swarm Doctrine: Containing Coordinated Agent Intrusions
Agents can turn shared infrastructure into a channel for coordinated intrusion. The Hugging Face incident and a separate public-wiki investigation show why a security assessment may need evidence from several executions and the artifacts they leave behind. We argue that the operational unit of defence should be a revisable coordination episode linking observed transfers, task authority, and response history. The central research problem is prospective episode discovery: finding which actions belong together before an evaluator supplies their membership. We define unsanctioned coordination relative to collaboration and delegated-authority policy, connect storage-mediated coordination to stigmergy, and specify the evidence needed to distinguish influence from common causes. First-contact signals are one possible input to discovery; the design also follows inherited state and later use. A proposed evaluation compares isolated actions, rolling windows, known groups, and prospectively discovered episodes at matched review cost and false-alert workload. It measures harmful outcomes across all assigned population runs and tests recurrence after channel closure and state quarantine. A checksum-verified reconstruction of the public wiki export separates the decline in retained writes from later administrative cleanup. The contribution is an incident-grounded position, descriptive analysis, and evaluation design. It makes the recommendation to monitor across executions testable without claiming a new detector or a measured containment benefit.
Read the full abstract
High Energy Physics (Phenomenology)
Improved analyses of the electroweak phase transition and its phenomenology in the Georgi-Machacek model
The Georgi-Machacek model is a promising new physics scenario with isospin-triplet scalar fields, whose effects can qualitatively change the nature of the electroweak phase transition compared with the Standard Model. We examine the electroweak phase transition in this model by using the thermally resummed one-loop effective potential. We also incorporate effects of custodial symmetry breaking due to radiative corrections by evolving the couplings and classical fields according to the renormalization group equations. We identify viable parameter regions by Bayesian analysis using theoretical requirements and the latest experimental bounds. We then investigate the electroweak phase transition in the resulting parameter regions and discuss its phenomenological implications, such as Higgs boson decays into neutral gauge bosons, di-Higgs production at high-energy colliders, and stochastic gravitational waves associated with a first-order electroweak phase transition. We find that parameter points for the strong first-order electroweak phase transition can be verified by these observables at next-generation experiments.
Read the full abstract
Classical Physics
Repeated Binary Direct Collinear Impacts Under Incremental Contact Laws With Residual Indentation: A Hybrid Systems Formulation
Incremental contact laws specify the normal contact force through a differential relation, possibly carrying an internal state, driven by the indentation and its rate. The force may vanish at a nonzero indentation, through plasticity, an elastic aftereffect, or, under some laws with a viscous component, a load drawing the bodies apart rapidly, leaving a residual deformation at separation. Such a deformation sits uneasily within rigid body dynamics, which admits none. The tension is tolerable when the indentation is small relative to the bodies, so that it may be carried constitutively rather than geometrically. Even then, the law alone does not determine the interaction: because force and indentation no longer vanish together, conditions for the commencement and termination of contact must be supplied separately, as must the fate of the deformation and internal state at separation, both absent from the equations of motion. This article formulates the repeated direct collinear impact of two convex bodies under external forces as a hybrid dynamical system. The contact interface is modeled as a massless element carrying the contact law and its state, coupled to the bodies through relative velocity and an interaction force dictated by the contact state. Consequently, all switching and resets are confined to the interface, leaving the geometry and equations of motion of the bodies unaltered. The principal analytical properties of the formulations are established, among them passivity and completeness; the branching of solutions at the onset and termination of contact is examined. The framework is demonstrated through numerical simulations.
Read the full abstract
Computational Finance
Unbiased Monte Carlo Greeks for Discontinuous Payoffs
Pathwise differentiation of Monte Carlo estimators fails at payoff discontinuities, producing zero or biased sensitivities for barriers, autocallables, and digital options. The industry workaround — smoothing the indicator functions — introduces bias and requires per-product calibration. We derive a correction formula that restores unbiased Greeks without smoothing. For a payoff $F(Z,\theta)$ that is piecewise smooth with discontinuities on surfaces $\{g_i = 0\}$, we show that the sensitivity decomposes into a pathwise term (computed by standard AAD) plus a sum of boundary corrections, each involving the payoff jump, the Gaussian density at the boundary, and the sensitivity of the boundary to the parameter. The correction is computed by Newton root-finding in the normal-random space, with the jump evaluated by two forward replays of the pricing kernel. The implementation uses AADC (pip install aadc), whose tape replay and automatic discontinuity tracking make the method fully automatic — the quant writes standard pricing code, and the correction driver identifies and handles all discontinuities. We prove the formula for arbitrary compositions of smooth functions and indicator functions (not just outer products), covering real autocallable payoff structures with recursive alive/dead logic. Benchmarks on QuantLib models (GBM, Heston, Hull-White) show all Greeks within 0.1–4% of analytic or bump-and-revalue references.
Read the full abstract
Information Theory
Fundamental Limits of Adaptive Beamforming Under Finite Training
Finite training reduces the output signal-to-interference-plus-noise ratio (SINR) of an adaptive beamformer, and a natural question is how much of this loss is unavoidable. This paper determines this question, providing a beamforming counterpart of the Cramér–Rao bound in spectral estimation. An exact identity expresses the SINR loss as a bounded function of the error in the clairvoyant minimum-variance distortionless-response (MVDR) weight. It yields a local asymptotic minimax lower bound over all measurable data-dependent beamforming rules, including biased and irregular rules. The first-order coefficient is $\tr(\Mb\Jb_{\rm eff}^{-1})$, where $\Jb_{\rm eff}$ describes the information in the training data and $\Mb$ measures the sensitivity of the output SINR. Matching constructions determine this coefficient in two complex-Gaussian models. For an $N$-sensor uniform linear array with $K$ distinct point interferers and $2K+1\le N$, a data-driven split one-step beamformer attains the coefficient $C_\theta\le K$ at every interior scene of a fixed compact regular parameter set. For unrestricted covariance matrices, sample matrix inversion (SMI) attains the coefficient $N-1$ through the classical Reed–Mallett–Brennan law. The difference quantifies the first-order value of finite-source structure. Geometric formulas and numerical results describe the dependence on interference power and array geometry, and the finite-sample departure near a weak-source boundary.
Read the full abstract
Computational Physics
Physical-Field Reconstruction from Sparse Observations: When Are Diffusion Models Preferable to Deterministic Regression?
Reconstructing physical fields from sparse observations is central to system identification, forecasting, and control, yet sparse measurements generally underdetermine the full field. This makes reconstruction an ill-posed inverse problem rather than simple interpolation. Although many deterministic and generative methods have been developed, there is still no clear consensus on when a single point estimate is sufficient and when a distribution of plausible reconstructions is more useful. We conduct a fair comparison of a deterministic U-Net, conditional diffusion, and prior-guided diffusion under matched experimental settings, including 2D Poisson equation, 2D Navier-Stokes flow, and 1D Kuramoto-Sivashinsky dynamics. Through this comparison, we make three observations. First, accuracy is field- and regime-dependent, with no systematic advantage for diffusion under higher complexity or sparser observations. Second, ensemble means improve phase-aligned accuracy, whereas individual samples better preserve variability and can retain high-wavenumber power in selected regimes. Third, conditional diffusion provides more reliable uncertainty estimates at lower cost, while prior-guided diffusion is more robust to mask-distribution shifts but requires substantially higher inference cost and guidance tuning. These results clarify when generative reconstruction is useful and provide guidance for improving uncertainty estimation, fine-scale sample fidelity, robustness, and computational efficiency in sparse field reconstruction.
Read the full abstract
Dynamical Systems
Observable functions of rational ODE models and how to find them
Consider a parametric ODE control model. A function of the states and parameters is called observable if its value can in principle be reconstructed from input-output data. The observable functions form a field, called the observation field, represented naturally by a set of generators. Even when the model is not fully observable, this field captures the information still accessible from input-output data. We present an algorithm for computing a concise generating set for the observation field of a model with rational dynamics. The algorithm relies on two new results: one allows observable functions to be extracted from the coefficients of repeated Lie derivatives of the outputs, while the other reduces the required orders of differentiation by exploiting identifiable parameter combinations. We implement the resulting algorithm in StructuralIdentifiability$.$jl (https://github.com/SciML/StructuralIdentifiability.jl). For computational efficiency, we employ recent techniques for differential elimination and rational function field simplification. Using models from epidemiology, chemical kinetics, and cancer modeling, we show that the algorithm produces generators with domain-specific interpretations that can inform model analysis and development.
Read the full abstract
Machine Learning
Compressed Recurrent Feedback in Tsetlin Machines: A Reproducible Boolean-FSM Study
Sequential inference on small devices requires a model to retain useful history without repeatedly processing a long input record. A Recurrent Tsetlin Machine (RTM) provides this memory by returning Boolean clause outputs from one time step as inputs to the next. Direct feedback, however, grows with the clause bank and can make the recurrent input unnecessarily wide. This paper investigates a fixed-width alternative. We combine clause activations by exclusive-OR (XOR) folding, retain the folded bits at two time scales, and threshold them back to a binary state. The resulting design reduces 480 clause activations to 96 recurrent bits. We evaluate the method on a reproducible Boolean finite-state-machine benchmark with explicit transition rules, data splits, and random seeds. Across 144 runs, the compressed model obtains $61.47 \pm 6.74\%$ and $62.94 \pm 9.92\%$ accuracy on the two task families. Raw clause feedback changes these means by less than one percentage point, while increasing the recurrent width tenfold and measured host execution time by $4.38\times$ and $3.71\times$. Gated neural models remain more accurate, and a no-feedback control retaining only short input history achieves comparable or slightly higher accuracy. On this benchmark, folding matches raw feedback within small empirical margins at a much narrower interface; these findings also underscore the critical necessity of no-feedback recurrence controls when benchmarking sequence models.
Read the full abstract
Probability
Generic Strong Well-Posedness for McKean–Vlasov SDEs with Jumps in a Krylov-Stable Coefficient Space
We prove that strong well-posedness is generic for McKean–Vlasov stochastic differential equations with jumps in a complete coefficient space that allows genuinely discontinuous state dependence. For every $T>0$, there exists a dense residual subset $\mathfrak R_T$, independent of the initial state, such that every coefficient $a\in\mathfrak R_T$ generates a unique strong solution for every deterministic initial condition $x\in\mathbb R^d$. Moreover, the solution map $x\mapsto X^{a,x}$ is continuous in $\mathcal S_T^2$, and every approximation by spatially Lipschitz coefficients converges locally uniformly in $x$ to the same canonical solution family. The coefficients are uniformly bounded, uniformly non-degenerate in the diffusion component, and Lipschitz continuous in the law with respect to the $2$-Wasserstein distance, while their state dependence may be merely measurable. The ambient space is obtained by completing a spatially Lipschitz core under a Krylov-stable metric based on local $L^{2(d+1)}$ coefficient errors. A stopped conditional Krylov–Melnikov occupation estimate converts convergence in this metric into occupation convergence along non-degenerate Itô–Lévy trajectories. The key probabilistic ingredient is a set-valued stability theorem: near a core coefficient, every strong solution of every nearby solvable equation remains close to the unique core solution, without any uniqueness assumption on the nearby equation. Together with closedness of the strong-solution relation, this permits a Baire-category argument on the completed coefficient space. We also give an intrinsic spatial-translation criterion showing that the completion contains natural classes with essential spatial discontinuities. A projective-limit argument yields the corresponding global-time generic well-posedness result.
Read the full abstract
Artificial Intelligence
IIns-VAE+: A Robust Transfer Learning Framework for Environmental Identification in Wireless Sensing
Environmental identification in wireless sensing is essential for 6G integrated sensing and communication (ISAC) systems to achieve reliable situational awareness. However, deep learning (DL) models for this task often fail to generalize under domain shift across diverse environments. While the Inter-Instance Variational Auto-encoder (IIns-VAE) learns features of rich representation, its neural classifier remains vulnerable to these distribution changes. In this paper, we propose IIns-VAE+, a hybrid model that combines the IIns-VAE framework with Minimax Risk Classifiers (MRC) to improve adaptability in transfer learning scenarios. We use real-world datasets to evaluate our framework across three transfer learning scenarios, including general to specific room environments, high to low label resolutions, and mixed to specific environments. The experimental results indicate that IIns-VAE+ significantly outperforms baselines, demonstrating its critical value in building adaptable and robust perceptive networks in future 6G systems.
Read the full abstract
Quantum Physics
The Future-Record Model of Everettian Probability
I recently showed how objective Everettian probability can be defined over the records borne by an observer's later continuations, without unique actualization or additional premeasurement subjects \citep{Baxter2026}. Here I extend that model to all decoherent physical record states, including records that support no such continuation. The observer-indexed model is then a restriction of an exhaustive physical-record model.
Read the full abstract
Computation and Language
Protocol Compression Changes Which Party Pays: Bilateral Cost in Cross-Organization LLM Agent Communication
Agents that talk across organizations exchange long messages billed by the token. A shorter notation therefore looks like a saving that costs nothing but an agreement to use it. Recent work reports the saving is conditional. Compressed notation can instead raise total tokens by 8% to 11% over a JSON baseline, when parsing failures force extra model calls. That is measured for one payer. Between two organizations neither side can install a decoder at the other end, and each pays under its own tokenizer, price, and cache state. We measure both sides. A preregistered token-level study covered 198 content-matched item pairs across six vendors, for 2,376 native-usage cells. We then overlay an English baseline, runtime schema negotiation followed by compression, and injected-schema compression on a two-party procurement bargain with an exactly enumerated feasible set. The overlay covers 1,053 completed dialogues of a 1,215-cell grid across 3 model pairs, plus a 405-dialogue rerun of the negotiated condition. Compression amplifies cross-vendor cost dispersion by a factor of 1.078, with a 95% CI of [1.066, 1.091], and two vendor pairs reverse which endpoint is cheaper. Runtime negotiation succeeds as a protocol and fails as a bargain. The parties agree a schema in 121 of 135 headline dialogues, none of them the schema we would have supplied. They settle the task in only 9 of those dialogues, and they reach impasse in 106 of them. The negotiated sessions average 10.8 turns against 17.6, and cost 52% of the English total because sessions end sooner, not because the handshake is repaid. Break-even horizons run from 20 to 70 turns, the low end only under the conditional accounting, and all of them lie above every observed English session. On one cross-vendor pair both parties keep about half their cost. On the other the receiving party pays more at a high cache-hit rate.
Read the full abstract
Artificial Intelligence
Substrate-Portable Execution for Production LLM Workflows
Production LLM agents execute tool-calling loops, retrieval chains, and compositional workflows in multiple modes, yet execution semantics are often coupled to one runtime. We encountered this portability problem in Rufus, a conversational AI assistant with a large tool catalog that serves millions of Amazon customers. Rufus supports real-time serving, asynchronous background tasks, and high-volume batch workloads such as evaluation and content pregeneration. Each mode has distinct service-level objectives and typically uses a separate runtime. Reusing streaming orchestration makes asynchronous and batch workloads blocking and prevents use of batch inference APIs, which offer a 50 percent discount at published prices. We present a binding-adaptive agent execution platform that separates workflow definition from execution substrate. Developers define a workflow once as a typed dataflow graph. The platform compiles the graph to in-process streaming for real-time serving, durable AWS SWF orchestration for asynchronous execution, or distributed Apache Flink stream processing for batch inference. No workflow code changes are required. LLM inference is represented as a suspendable graph node whose behavior depends on the substrate: streaming delivery online, durable retry asynchronously, and batched submission offline. We validated dozens of production agent configurations across five orchestration patterns: single-inference RAG, iterative ReAct, compositional PreAct, conditional routing, and multi-agent deep research. Across all three bindings, we found no detectable difference in output quality. Batch execution reduced per-query inference cost in line with published batch API pricing while operating alongside the streaming path at production scale.
Read the full abstract
Representation Theory
PAGR: Proof-Carrying Algebraic-Geometric Retrieval: A Quiver-, Provenance-, and Sheaf-Theoretic Framework for Grounded LLM Retrieval
Retrieval-augmented generation is usually formulated as a statistical information-retrieval problem. Graph-based variants add relational structure, but the mathematical status of that structure is often left underspecified. Three distinct questions tend to be conflated: which statements are certified as knowledge, which latent representations are useful for retrieval, and which multi-hop compositions are semantically admissible. We propose Proof-Carrying Algebraic-Geometric Retrieval (PAGR), a framework that separates these questions mathematically. Its symbolic layer is a many-sorted relational theory generated by a typed quiver, path equations, and positive Horn inclusions. A quiver representation assigns inner-product spaces to entity types and linear operators to relations. A cellular sheaf measures local-to-global consistency. Semiring provenance records derivations and supports machine-checkable certificates. The central principle is epistemic separation: learned geometry may rank and organize evidence, but cannot promote a hypothesis to certified ground truth. We show the certification criterion is invariant under arbitrary replacement of learned components. Further results include a conditional completeness bound, identification of the isometry group as the relevant symmetry for residual-based retrieval, a cohomological consistency diagnostic, and a bounded-bisimulation index for admissible-path expansion. PAGR is a mathematical architecture for separating where a system should look from what it is allowed to treat as knowledge.
Read the full abstract
Artificial Intelligence
CWF: A Collaborative Writing Framework for Personalized and Reliable Popular Science Writing
We introduce Personalized and Reliable Popular Science Writing, a novel task that requires adapting scientific explanations to audiences with different cognitive levels while preserving factual accuracy. However, improving personalization often introduces simplifications that increase the risk of hallucination and factual distortion. To address these challenges, we first construct a dataset of 39,134 entries and a reader-centric Personalized Science Communication Benchmark (PSCB) that jointly evaluates audience adaptation and factual accuracy. To reduce data and computational requirements while improving generalization across domains and audiences, we introduce DA-MoE, which explicitly decouples audience adaptation from domain knowledge through separate modeling. To enable robust verification and revision in evidence-scarce scenarios, a multi-agent fact-checking mechanism that augments limited evidence with role-specific agent debate and propagates confidence over a graph is proposed. Experiments on PSCB show that our approach achieves state-of-the-art performance. Our code is open-sourced at https://github.com/DPInnovationWorks/CWF.
Read the full abstract
Hardware Architecture
WaferTrans: Enabling IOMMU-free Distributed Virtual Address Translation for Wafer-scale GPUs
Wafer-scale GPUs (WSGs) provide sufficient on-wafer bandwidth to make near-lossless Unified Memory feasible. However, existing designs still rely on a CPU-IOMMU to translate remote virtual-address accesses. This centralized mechanism scales poorly to tens of GPU dies: translation requests must traverse costly off-wafer hierarchies and contend for limited CPU-side resources, making address translation a critical bottleneck. We propose WaferTrans, an IOMMU-free distributed virtual-address translation design for WSGs. WaferTrans introduces PTE Presence Consistency (PTE-PC), a lightweight consistency model that tracks PTE insertions and removals, and equips each GPU with a PTE Presence Directory (PPD) that locates the GPU holding a requested PTE. It further employs a distributed PTE-PC mapping and a cooperative query mechanism to localize PTE-PC maintenance while preserving complete lookup coverage. Together, these mechanisms enable the GPU array to resolve remote translations within the wafer, eliminating its dependence on the CPU-IOMMU. Compared with the SOTA Trans-FW design, WaferTrans improves performance by 2.5x on average.
Read the full abstract
Artificial Intelligence
SAP: State-Guided Data Synthesis with Argument Provenance for Multi-Turn Tool Use
High-quality multi-turn tool-use data is essential for training agentic models, yet existing data synthesis methods often underrepresent the argument-level dependencies that are critical to long-horizon tool use. As a result, even when a model selects the correct tool, task execution may still fail because the model fills tool arguments with fabricated, stale, or weakly grounded values. To address this problem, we propose State-Guided Data Synthesis with Argument Provenance (SAP). SAP combines state guidance, tool-argument provenance constraints, and turn-level validation to efficiently construct tool-use trajectories with long-range dependencies and high accuracy. Using data generated by SAP, we build SAP-4B, which is highly competitive even when compared with much larger models across multiple benchmarks. Source code, synthesized data, and trained weights are available at https://github.com/Zichen1024/SAP.
Read the full abstract
General Economics
Optimal harvesting under annuity and compound interest laws: economic-ecological trade-offs in a logistic growth model
The relationship between investment policy associated with species growth profile is essential in seeking the most appropriate strategy for a policymaker. Balancing maximum profit with the sustainability of species remains a central issue in both ecological and economic contexts. This study presents a comparative analysis of two interest principles, annuity and compounding, within the framework of optimal control. Various investment policies are examined from the perspective of capital theory, incorporating concepts such as future value, accumulation function, and force of interest, each contributing to the formulation of an optimal control strategy. Our analysis is based on a one dimensional logistic model incorporating linear harvesting. Key parameters include the species growth rate and interest rate, and optimality is evaluated with respect to these variables. Using Pontryagins Maximum Principle, we derive the optimal harvesting policies under both discounting laws and characterize the resulting steady state equilibria. The principal finding indicates that for species with low intrinsic growth rates and low annual interest rates, the annuity law of interest yields optimal outcomes. Conversely, for any annual interest rate, species exhibiting moderate or high growth rates maximize profit under the compound law of interest. The study also addresses maximum net revenue and optimal strategies for varying growth rates under each interest law.
Read the full abstract
Systems and Control
Multimodal Large Language Model-guided Constrained Optimization for RAN Intelligent Control
Artificial intelligence (AI)-based radio access network (RAN) controllers are commonly designed for predefined operating scenarios and optimization tasks, limiting their adaptability when network conditions and operator requirements change after deployment. This paper proposes multimodal large language model (MLLM)-guided constrained optimization for RAN intelligent control (MLLM-coRIC), a requirement-adaptive hierarchical Open RAN (O-RAN) framework for joint resource and power allocation. MLLM-coRIC jointly exploits the operator's natural-language specification and radio-frequency (RF)-derived network context as multimodal inputs, enabling a single deployed framework to address different optimization problems without redesigning task-specific algorithms or retraining control policies. At the Non-Real-Time RAN Intelligent Controller (Non-RT RIC), the MLLM performs holistic, longer-timescale reasoning over the operator requirement, predicted network evolution, and measured optimization outcomes to design the numerical control loss. The relative constraint penalties are iteratively refined through a closed-loop process, allowing the optimization criterion to reflect both the intended network behavior and the expected operating context. At the Near-Real-Time RIC (Near-RT RIC), a loss-conditioned hybrid executor translates the synthesized criterion into fast radio actions by combining learned ramp-constrained resource allocation with model-based interference-aware power control. A multi-cell evaluation environment integrating the CARLA urban mobility simulator and the Sionna RT ray-tracing-based wireless propagation simulator is implemented for validation.
Read the full abstract
Statistical Mechanics
Nanothermodynamics: stable thermal equilibrium and nanoscale fluctuations
Nanothermodynamics describes the process where large systems subdivide into equilibrium distributions of small subsystems. A key ingredient is Hill's subdivision potential (E) that ensures adherence to the 1st and 2nd laws of thermodynamics in systems of any size. In this review and reassessment, it is emphasized that nanothermodynamics gives new insight into many measurements, theories, and simulations. Measurements establishing the need for nanothermodynamics show thermodynamic heterogeneity from multiple effective temperatures (T_i) inside most types of materials. One theoretical result that requires E=0 is the stable solution of Ising's original model for finite chains of interacting spins, a solution Ising could not have found 40 years before Hill's work. Another result is a novel solution to Gibbs' paradox that makes the entropy of the semiclassical ideal gas exactly extensive. Molecular dynamics simulations reveal how a standard fluctuation relation is modified when local degrees of freedom fluctuate faster than their coupling to the heat bath, consistent with the measured thermodynamic heterogeneity. Simulations of a Creutz-like model, comprised of Ising spins coupled to an explicit heat bath of Einstein oscillators, are used to study the 2nd law. It is found that maximizing the total entropy (S_t) requires an intrinsically irreversible step, providing a counterexample to the usual claim that statistical mechanics emerges from reversible dynamics. Furthermore, fluctuations of this model are best described by Einstein's reversal of Boltzmann's relation and the 2nd-law, not by recent fluctuation theorems.
Read the full abstract
Soft Condensed Matter
Coupled self-charging and self-propulsion of evaporating sessile droplets
Evaporating sessile water droplets universally deposit surface charges while simultaneously becoming self-charged—a phenomenon that remains under-appreciated despite its potentially important consequences. Here, we investigate one such consequence: droplet self-propulsion driven by self-charging. This self-propulsion arises spontaneously in the absence of external stimuli and produces self-avoiding trajectories that can displace droplets by many times their own diameter. Our results have important practical implications for processes requiring spatial precision, such as inkjet printing. The coupling between self-charging and self-propulsion produces mosaic surface charge patterns characterized by alternating polarities, which we explain with a physical model based on properties of the electric double layer.
Read the full abstract
Optics
Helicity-engineered nonlinear optical responses in photo-excited topological semimetals
Topological materials provide a transformative arena for light-driven control of electronic motion; yet, direct manipulation of electron dynamics on sub-cycle timescale remains a significant challenge. We demonstrate that the strong-field-driven high-harmonic generation of a Weyl semimetal can be controlled and manipulated through bicircular pump-probe driving fields. While a lone probe pulse generates exclusively odd-order harmonics, the addition of a pump pulse triggers a series of sidebands arising from nonlinear frequency mixing of pump and probe photons. Our results reveal that the sideband intensities are highly sensitive to the relative helicity of the pulses and the orientation of the polarization plane. This sensitivity stems from the chiral nature of the Weyl nodes, which couples efficiently to the light's helicity only when the node-separation axis is perpendicular to the polarization plane. Furthermore, the significant suppression of sideband intensity with increasing pump-probe delay identifies these features as a potential clock for electron-hole decoherence. These findings establish frequency-mixed high-harmonic generation as a sensitive probe of chiral quantum dynamics and suggest a robust framework for manipulating topological currents via structured light, with implications for lightwave electronics and ultrafast quantum information processing.