arXiv:2606.30219v2 Announce Type: replace
Abstract: LLM evaluation and AI safety face a shared measurement problem: benchmark scores, reward-model signals, and reported safety metrics can improve while the latent properties they are meant to represent remain difficult to verify. This paper combines a hybrid survey - a systematic search paired with narrative synthesis and separately tracked grey evidence - with a conceptual framework and a structured ten-model audit. The synthesis spans eight evidence streams: benchmark validity, dynamic evaluation, LLM-as-judge reliability, safety evaluation, jailbreak/refusal robustness, reward hacking, mechanistic interpretability, and governance/auditability, covering 2018-2026 evaluation-safety measurement work. We introduce EvalSafetyGap as an organizing hypothesis for comparing evaluation-side and alignment-side proxy failures under optimization pressure, using Goodhart's Law together with two constructs we develop here - an Instability Decomposition and an Alignment Trilemma - as tools for generating testable comparisons. The audit shows how conclusions shift when capability, behavioral safety, and governance are measured separately. In this sample ($n = 10$), the association between capability and sustained adversarial robustness is statistically indeterminate using the displayed Table 3 inputs (Pearson $r = +0.232$, $p = 0.520$), and the apparent open-closed safety gap is modest, driven mainly by governance and disclosure rather than behavioral robustness, and sensitive to how a single borderline model is classified; attempt-budget results are protocol dependent. Because the public evidence uses heterogeneous protocols, the audit is diagnostic rather than rank-generating. The contribution is a shared vocabulary and evidence map to support dynamic evaluation, transparent source reporting, multi-attempt safety measurement, and auditable alignment practice.
Science Journals
arXiv:2405.03529v5 Announce Type: replace
Abstract: This paper contributes to the study of optimal experimental design for Bayesian inverse problems governed by partial differential equations (PDEs). We derive estimates for the parametric regularity of multivariate double integration problems over high-dimensional parameter and data domains arising in Bayesian optimal design problems. We provide a detailed analysis for these double integration problems using two approaches: a full tensor product and a sparse tensor product combination of quasi-Monte Carlo (QMC) cubature rules over the parameter and data domains. Specifically, we show that the latter approach significantly improves the convergence rate, exhibiting performance comparable to that of QMC integration of a single high-dimensional integral. Furthermore, we numerically verify the predicted convergence rates for an elliptic PDE problem with an unknown diffusion coefficient in two spatial dimensions, offering empirical evidence supporting the theoretical results and highlighting practical applicability.
arXiv:2607.14185v1 Announce Type: new
Abstract: Feedback-driven loops support iterative improvement in large language models, reinforcement learning, and autonomous discovery, yet their gains often diminish under repeated internal feedback. We study why closed-loop knowledge systems saturate and what external information can move them beyond their current attractors. We introduce a three-level operational framework in which knowledge states $x_t$ evolve through transition kernels $K_{\theta}$ indexed by a structural parameter $\theta$. The governing structure is defined as the observational equivalence class of $\theta$ induced by these kernels, while attractors and basins are properties of the fixed-$\theta$ dynamics. A structural intervention changes $\theta$ and produces a detectable kernel discrepancy on pre-specified probe states, making structural change falsifiable. Using a Lyapunov drift condition, we show that stable internal dynamics approach bounded stability regions with exponentially attenuated transients and a noise-controlled residual floor. We characterize escape through a metric condition on intervention-induced attractor displacement and a baseline-relative KL lower bound for increasing escape probability. This analysis also explains why conditional mutual information alone cannot certify escape: it measures variation among intervention-conditioned updates rather than departure from the no-intervention law. Case studies in LLM code repair, sparse-reward reinforcement learning, and Bayesian optimization use matched continuation controls to illustrate how feedback strength and alignment affect quality-improving escape. Our contribution is an operational connection among stability tools, measurable intervention effects, and cross-domain diagnostics.
arXiv:2607.15257v1 Announce Type: new
Abstract: Recent advances in Tool-Integrated Large Language Models have made web search a core capability of information-seeking agents. However, as interaction histories grow, agents increasingly struggle to track task progress. When search attempts fail to yield useful evidence, current single- and multi-agent systems can become trapped in repetitive loops, wasting search budgets and ultimately compromising the quality and completeness of the final output. We introduce SearchOS, a system-level multi-agent framework that turns fragile, implicit search progress into explicit, persistent, and shared state. First, we formulate open-domain information seeking as relational schema completion with grounded citations, where agents discover entities, populate attributes across linked tables, and anchor each value to source evidence. Then we design Search-Oriented Context Management (SOCM), which externalizes the evolving state into Frontier Task, an Evidence Graph, a Coverage Map, and Failure Memory. Built on SOCM, SearchOS applies a pipeline-parallel scheduling mechanism that overlaps the execution of sub-agents and continuously refills freed slots with tasks targeting unresolved coverage gaps to improve utilization and throughput. To schedule and control the execution of search agents, SearchOS introduces a Search Tool Middleware Harness that intercepts model and tool interactions to record grounded evidence and react to stalls or budget exhaustion, and provides a reusable hierarchical skill system comprising strategy and access skills to augment the agents' search process and avoid repeating failed search patterns across runs. On WideSearch and GISA, SearchOS leads all metrics among the evaluated single- and multi-agent baselines, paving the way toward robust information-seeking collaboration.
arXiv:2607.14143v1 Announce Type: cross
Abstract: The notion of circuit refers to a cyclic oriented influence between the elements of a dynamical system. There are two classes of circuit: positive and negative. R. Thomas conjectured that a necessary condition of multi stationarity is the existence of positive circuits. In this paper we use dynamical system tools and planar analysis to find conditions for which the conjecture holds for planar systems.
arXiv:2607.00724v3 Announce Type: replace
Abstract: Multilingual fluency often invites a stronger assumption: a model that can speak a user's language must also understand the culture encoded by that language. We call this the Illusion of Cultural Alignment. To test this assumption directly, we introduce MSQA, a benchmark of 1,064 natively sourced questions across 11 language groups, five cultural dimensions, and three difficulty tiers. Unlike translated benchmarks, MSQA targets locally grounded knowledge and reduces shortcuts from English-centric cross-lingual transfer. Evaluating 18 LLMs, we find substantial cultural degradation and a pronounced Locality Effect: cultural competence tracks pre-training exposure more closely than general reasoning ability. We further show that common inference-time remedies do not dissolve the illusion. Models remain overconfident on unfamiliar cultural questions, repeated sampling yields unstable rather than reliable correctness, and retrieval augmentation helps unevenly on long-tail facts. These findings indicate that cultural alignment cannot be inferred from multilingual ability alone and requires deeper intervention than calibration, sampling, or retrieval at inference time
arXiv:2606.01772v2 Announce Type: replace
Abstract: The impact of applying generative AI tools to requirements engineering (RE) in industrial practice remains poorly understood. This paper examines how AI-assisted RE tools are used in industrial practice at XITASO, a medium-sized enterprise for high-tech software engineering, and how they reshape workflows, tool integration, and PO--developer relationships. We combine a 2024 company-wide use-case survey with two rounds of semi-structured interviews with eight product owners (POs) in late 2025 and spring 2026, covering an in-house chatbot and seven commercial AI tools. We identify 15 distinct use cases across four categories: product backlog management, tender management, requirements and domain understanding, and document and artifact creation. Three findings emerge. First, the effect of AI on PO--developer interaction is mixed: the prevailing single-user interaction model can substitute for collaborative dialogue, and developers do not always welcome AI-generated artefacts. Second, tool integration -- not tool capability -- is the binding constraint: where integration is in place, time savings are dramatic; where it is missing, POs fall back on manual workarounds. Third, AI advances faster than the surrounding organisational systems, so its benefits accrue to individual POs while team processes and customer readiness remain the bottleneck. The empirical GenAI-RE literature remains dominated by early-stage, lab-oriented evaluations of isolated tasks while practice has moved into territory it has not yet studied: practitioners are already assembling cross-tool integrations, navigating customer governance, and renegotiating role boundaries. From these patterns we derive a set of questions practitioners considering AI-assisted RE may ask of their own situation.
arXiv:2607.06503v2 Announce Type: replace
Abstract: Large language model (LLM) agents often waste inference compute by continuing multi-step trajectories that are already doomed to fail. We study early failure prediction and inference-time early stopping for LLM agents using hidden-state probes. Lightweight linear probes on internal activations predict eventual task failure from the first interaction round, substantially earlier than agent-monitoring methods based only on observable behavior. We turn this signal into a recall-controlled abort cascade for reducing LLM agent inference costs. The cascade applies a distribution-free calibrated failure detector at each early interaction round and jointly optimizes per-round recall budgets. This design ensures that eventually successful episodes survive all early-stopping gates at a user-specified global recall rate. After selection, the cascade is frozen and certified on independent data, providing an exact post-selection recall guarantee. We evaluate the method on TextCraft and WebShop with Qwen-2.5-7B, Llama-3.2-3B, and Qwen3-1.7B. The proposed LLM agent early-stopping cascade outperforms the best single-gate baseline in every model-environment pair, saving 1.5-8.8 times more compute at a 90% recall target. Achieved recall remains within one standard deviation of its target in all 24 configurations. The strongest settings reduce generated tokens by 60.2% on TextCraft and 54.9% on WebShop at 90% recall, while retaining savings of 45.0% and 41.5% at 95% recall. Behavior-only monitoring is consistently weaker, and adding behavioral features to hidden-state probes provides no further gain. We also characterize the sample complexity required to certify high-recall early-stopping policies. The code will be released soon.
arXiv:2607.14491v1 Announce Type: new
Abstract: State-backed influence operations are routinely measured as high-prevalence sources of ``hate'' and ``toxicity.'' We argue those rates rest on a measurement error: the detectors behind them are validated to catch a broader definition inclusive of hostility or divisiveness aimed at an out-group, and so over-attribute hate to content better described as partisan or geopolitical invective. Across 25.08M tweets from seven government-attributed campaigns in the Twitter Information Operations archive (8,275 accounts), we separate hate from the other forms of divisiveness. We first validate a two-prompt LLM-based detector, matching human labels at Cohen's $\kappa=0.82$, to identify the broader hostility; we then develop an auditable rule, agreeing with an expert at $\kappa=0.52$, to further classify this content (5,457 posts) into three sub-categories. About 50.1% are identity-based attacks on people, whereas 30.4% are partisan attacks and 19.5% invective against states and their foreign policy. Reporting all of it as hate therefore overstates hate roughly twofold; only 18.7% is both identity-based and dehumanizing or inciting. Six of seven campaigns sort into three regimes that a single ``hate'' rate flattens, namely identity hate (RU-op and IRA, both Russia-attributed), geopolitical invective (both Iran operations), and partisan divisiveness (both Venezuela operations). We call the shared product $manufactured divisiveness$. The line to separate these constructs itself remains unsettled: on the hardest cases three independent human experts agree only moderately (pairwise $\kappa=0.37$--$0.50$), and the best of nineteen LLM models tops out at $\kappa=0.601$ against the experts' majority. Our findings can help redefine the study of hate in the context of influence campaigns and broader online discourse.
arXiv:2607.07688v2 Announce Type: replace-cross
Abstract: How close to singularity can an $n \times n$ unimodular matrix be? For ternary cases as $n$ increases, exact expressions are unlikely, but upon fixing $n=4$ and assessing $(2k+1)$-ary cases as $k$ increases, we make significant progress; similarly for $(k+1)$-ary cases of $4\times 4$ nonnegative unimodular matrices.
arXiv:2505.09809v2 Announce Type: replace-cross
Abstract: In this short note, we use flag algebras to prove that the number of colour alternating 6-cycles in a red/blue colouring of a large clique is asymptotically maximized by a uniformly random colouring. This settles the first open case of a problem of Basit, Granet, Horsley, K\"undgen and Staden.
arXiv:2502.17602v2 Announce Type: replace-cross
Abstract: We study a class of stochastic nonsmooth optimization problems in which an outer variable minimizes the expectation of a pointwise maximum. This minimization--expectation--maximization (minEmax) problem arises in Wasserstein distributionally robust optimization and adversarially robust training, and it cannot in general be reformulated as a finite-dimensional minimax problem when the underlying distribution is not empirical. We propose a stochastic smoothing proximal gradient method based on log-mean-exp smoothing of the value function. Under compactness and Lipschitz-type assumptions, we present nonasymptotic analysis in terms of Goldstein stationarity and show that every almost-sure cluster point generated by our method is a Clarke stationary point; by Clarke regularity, such a point is also directional stationary for the original problem. Numerical experiments on newsvendor, robust regression, and adversarially robust learning problems show that the proposed method is competitive with existing baselines.
A system-level approach to generalized feedback Nash equilibrium seeking in partially observed games
arXiv:2503.24159v2 Announce Type: replace-cross
Abstract: This work proposes an algorithm for seeking generalized feedback Nash equilibria (GFNE) in noncooperative dynamic games. The focus is on cyber-physical systems with dynamics which are linear, stochastic, potentially unstable, and partially observed. We employ System Level Synthesis (SLS) to reformulate the problem as the search for an equilibrium profile of closed-loop responses to noise, which can then be used to reconstruct a stabilizing output-feedback policy. Under this setup, we leverage monotone operator theory to design a GFNE-seeking algorithm capable to enforce closed-loop stability, operational constraints, and communication constraints onto the control policies. This algorithm is amenable to numerical implementation and we provide conditions for its convergence. We demonstrate our approach in a simulated experiment on the noncooperative stabilization of a decentralized power grid.
arXiv:2504.00978v2 Announce Type: replace-cross
Abstract: Over the course of the past decade, advances in the radial velocity and transit techniques have enabled the detection of rocky exoplanets in the habitable zones of nearby stars. Future observations with novel methods are required to characterize this sample of planets, especially those that are non-transiting. One proposed method is the Planetary Infrared Excess (PIE) technique, which would enable the characterization of non-transiting planets by measuring the excess infrared flux from the planet relative to the star's spectral energy distribution. In this work, we predict the efficacy of future observations using the PIE technique by potential future observatories such as the MIRECLE mission concept. To do so, we conduct a broad suite of 21 General Circulation Model (GCM) simulations with ExoCAM of seven nearby habitable zone targets for three choices of atmospheric composition with varying partial pressure of CO$_2$. We then construct thermal phase curves and emission spectra by post-processing our ExoCAM GCM simulations with the Planetary Spectrum Generator (PSG). We find that all cases have distinguishable carbon dioxide and water features assuming a 90$^\circ$ orbital inclination. Notably, we predict that CO$_2$ is potentially detectable at 15 $\mu\mathrm{m}$ with MIRECLE for at least four nearby known non-transiting rocky planet candidate targets in the habitable zone: Proxima Cenaturi b, GJ 1061 d, GJ 1002 b, and Teegarden's Star c. Our ExoCAM GCMs and PSG post-processing demonstrate the potential to observationally characterize nearby non-transiting rocky planets and better constrain the potential for habitability in our Solar neighborhood.
arXiv:2504.02500v2 Announce Type: replace-cross
Abstract: Quantum processing units (QPUs) built on superconducting Josephson junctions remain the most industrially mature route to fault-tolerant quantum computing, but the field has moved substantially since early 2025. This paper provides an updated overview of Josephson-junction QPUs, grounded in the quantum-mechanical principles, superposition, entanglement, and decoherence, that any qubit implementation must contend with, and in the physics of Cooper pairing and quantum tunneling that make a Josephson junction behave as a controllable qubit. We examine the engineering challenges of scaling these devices, including crosstalk, classical control-interface bottlenecks, and quantum error correction, and discuss the first below-threshold surface-code demonstrations on superconducting hardware. We survey recent materials advances, near-term computational results, and applications beyond computing, and compare Josephson-junction QPUs against trapped-ion, photonic, and neutral-atom alternatives, the latter having emerged as a serious fourth architecture since our original analysis. We close with the current public roadmaps toward fault-tolerant superconducting quantum computers and a calibrated assessment of how much progress those roadmaps still require.
arXiv:2504.04177v2 Announce Type: replace-cross
Abstract: We describe the physical processes that affect the formation, trapping, and outgassing of O2 at Europa and Ganymede. Following Voyager measurements of their ambient magnetospheric plasmas, laboratory data indicated that observed ions, mostly ejected from volcanic Io, would in turn impact and sputtering their surfaces, decomposing the ice producing thin oxygen atmospheres. Subsequently, Europa and Ganymede's O2 atmospheres were inferred from O aurora, condensed O2 bands identified at 5773 and 6225 Angstroms, and their atmospheres were shown to have a dusk/dawn enhancement, confirmed by recent Juno data. Although plasma produces these observables, processes that occur within the topmost surface are not well understood. Here, we note that the incident plasma particles produce nonequilibrium defect density locally in the surface ice grains. Defect diffusion within these grains leads to the formation of voids and molecular products, some of which are volatile. Although some volatiles are released into the satellite atmospheres, others are trapped at defect sites or trapped in voids, creating bubbles whose lifetimes are limited by the plasma-induced destruction rate. We discuss how trapping competes with annealing of the radiation damage, and how hemispheric differences at Europa and Ganymede, roughly determine the observed trend with latitude of O2 bands. We discuss the relative importance of condensed O2 and O2 adsorbed on regolith grains as atmospheric sources, accounting for dusk/dawn enhancements and temporal variability reported in condensed O2 band depths. Since plasma-induced damage and thermal annealing timescales drive oxidant variability on icy moons (likely also Callisto, Dione, and Rhea), they can help determine volatile downwelling, a potentially metabolic source for their oceans, and upwelling of other trapped oxidants (e.g. CO2) suggestive of ongoing geologic activity.
arXiv:2504.14659v3 Announce Type: replace-cross
Abstract: Minimum mean square error (MMSE) estimation is widely used in signal processing, information theory, and related fields. Despite its practical robustness, the MMSE can be discontinuous under standard notions of stochastic convergence. To bridge this gap, we review classical counterexamples to the continuity of the MMSE and observe that they share a common pathology: along the approximating sequence, the observation is strictly more informative about the limit estimand than the limit observation is. Motivated by practical acquisition mechanisms, we study MMSE continuity under two natural constraints: (1) continuity of the second moment, and (2) a degradedness (Markov) restriction ensuring that each approximating observation is no more informative than the limit observation is about the limit estimand. Under these conditions, we establish continuity of the MMSE and of the MMSE estimator. We provide complementary semicontinuity results and continuity guarantees in related settings and establish continuity under linear estimation. We further extend the analysis to the families of Bregman divergences and continuous metric cost functions, including the Kullback-Leibler and Jensen-Shannon divergences as special cases.
arXiv:2505.03982v4 Announce Type: replace-cross
Abstract: We reformulate the problem of alternating projections between two affine subspaces of a Hilbert space as the minimization of a least-squares functional associated with a bounded linear operator. This viewpoint reveals that classical alternating projections coincide with the unit-step Landweber iteration and enables the introduction of a general residual-state iteration framework that encompasses Landweber, its steepest-descent variant, and the conjugate-gradient method. Within this framework, we establish abstract convergence principles based on residual extinction and translation equivariance, allowing convergence analyses to be carried out once at the level of least-squares optimization and then transferred directly to alternating projection algorithms. As applications, we obtain new variants of alternating projections accelerated by steepest descent and conjugate gradients, together with convergence guarantees in both the consistent and inconsistent settings. We also establish linear convergence results under closed-range assumptions and express the convergence rates explicitly in terms of the Friedrichs angle and the largest principal angle between the underlying subspaces.
arXiv:2505.23869v4 Announce Type: replace-cross
Abstract: A proposition that connects randomness and compression is put forward via Gibbs entropy over set of measurement vectors associated with a compression process. In building this connection, we use a performance of a learning task as a probe of compression, over series of compression cycles within a cascade. The Gibbs entropy at each cycle measures the degree of randomness. Consequently a lossy compression process can be seen as an equivalent to {\it directed randomness} that preserves information content under certain bounds of Gibbs entropy and the performance of the learning task. The term directed means we guide the compression process with set of mathematical rules on how to reduce the model size. We formulate this connection with a theorem using a $\delta$ and $\epsilon$ bounds, and demonstrated a logical proof via comonotonic relationship within a very small decrease in compression ratio and the performance. We have showcase the validity of this proposition with a canonical vision task in deep learning with three different model compression processes as {\it a baseline model}. We use the following, simpler to more complex model compression approaches: (1) random pruning, (2) magnitude pruning, and (3) a more complex compression by using dual tomographic compression, which utilizes compressed sensing in dual fashion. We use remaining weights of deep learning network as a measurement vector where we measure the Gibbs entropy. The proposition is supported with the experimental evidence, resulting in very high correlation between learning performance and the Gibbs entropy over compression ratios for all different compression processes. We show case the idea that there is an inherent computable connection between compression probed by performance degradation and randomness from an entropy measure on the learned model.
arXiv:2506.03549v2 Announce Type: replace-cross
Abstract: Quantum position verification (QPV) enables multiple verifiers to certify a prover's location using quantum communication and physical assumptions. With experimental demonstrations of QPV becoming increasingly feasible, enhancing the practicality and security of QPV protocols is more important than ever. In this work, we make three key contributions toward this goal. First, we introduce a robust QPV protocol in which the verifier's state preparation is independent of channel loss, improving reliability in real-world conditions. Second, we refine existing security analysis techniques to bolster protocol resilience against experimental imperfections. Third, we identify and address some implicit assumptions present in existing security analyses, providing a framework to eliminate such assumptions. Additionally, as an example of QPV application beyond location verification, we illustrate how QPV can be leveraged for authenticating classical communication in quantum key distribution.
arXiv:2507.03209v2 Announce Type: replace-cross
Abstract: The discovery of new ionizable lipids for efficient lipid nanoparticle (LNP)-mediated RNA delivery remains a major bottleneck in RNA therapeutics development. Recent advances demonstrate the potential of machine learning (ML) models to predict transfection efficiency directly from lipid structure, enabling high-throughput virtual screening and accelerating lead identification. However, as new models for LNP transfection prediction continue to emerge, the lack of rigorous and standardized benchmarking poses a significant risk and may undermine confidence in their reliability for discovery. Here, we present a robust ML benchmarking framework for evaluating transfection prediction models based on ionizable lipid structures. This framework systematically benchmarks diverse molecular representations paired with a broad range of ML architectures spanning traditional models, feedforward neural networks, and state-of-the-art graph-based methods. In addition, the presented framework supports assessment of model generalization and evaluates prediction reliability beyond standard regression metrics. Using a curated dataset of 1,100 unique ionizable lipid structures derived from the HeLa transfection dataset originally reported by Xu et al., we show that within this framework, models leveraging explicit molecular substructure encoding consistently achieve the highest predictive accuracy and should serve as essential baselines for the development of new, more sophisticated models. In contrast, some current graph-based models, including AGILE, Chemprop, and KPGT, tend to show comparatively lower accuracy. The presented framework provides a standardized, transparent, and comprehensive benchmarking resource that enables meaningful comparison of emerging architectures and establishes strong baselines for future development of predictive models in lipid-based RNA delivery.
arXiv:2507.04268v2 Announce Type: replace-cross
Abstract: In this work, we present a compact analytical approximation for the quantum partition function of systems composed of quantum oscillators. The proposed formula is general and applicable to an arbitrary number of oscillators described by a rather general class of potential energy functions (not necessarily polynomials). Starting from the exact path integral expression of the partition function, we introduce a temperature-dependent Gaussian approximation for the high-temperature propagator and, then, invoke a principle of minimal sensitivity to minimize the error. This leads to a system of coupled nonlinear equations whose solution yields the optimal parameters of the Gaussian approximation. The resulting approximate partition function accurately reproduces thermodynamic quantities such as the free energy, average energy, and specific heat -- even at zero temperature -- with typical relative errors in the range of about 1\%--5\%. The accuracy deteriorates only moderately when the anharmonicity and coupling strengths are increased. We illustrate the performance of our analytical formula with numerical results for systems of up to ten coupled anharmonic oscillators. These results are compared to "exact" numerical results obtained via Hamiltonian diagonalization for small systems and Path Integral Monte Carlo simulations for larger ones.
arXiv:2507.13782v2 Announce Type: replace-cross
Abstract: Ultra-high resolution 7 tesla (7T) magnetic resonance imaging (MRI) provides detailed anatomical views, offering better signal-to-noise ratio, resolution and tissue contrast than 3T MRI, though at the cost of accessibility. We present an advanced deep learning model for synthesizing 7T brain MRI from 3T brain MRI. Paired 7T and 3T T1-weighted images were acquired from 172 participants (124 cognitively unimpaired, 48 impaired) from the Swedish BioFINDER-2 study. To synthesize 7T MRI from 3T images, we trained two models: a specialized U-Net, and a U-Net integrated with a generative adversarial network (GAN U-Net). Our models outperformed two previous state-of-the-art 3T-to-7T models in image-based evaluation metrics. Four blinded MRI professionals judged our synthetic 7T images as comparable in detail to real 7T images, and superior in subjective visual quality to 7T images, due to the reduction of artifacts. Using both SynthSeg and NextBrain, automated segmentations of the synthetic 7T images were more similar to real 7T segmentations than automated segmentations from the 3T images that were used to synthesize the 7T images. Finally, synthetic 7T images showed similar performance to real 3T images in downstream prediction of cognitive status using MRI derivatives (n=3,168). In all, we show that synthetic T1-weighted brain images approaching 7T quality can be generated from 3T images, which may improve image quality and segmentation, without compromising performance in downstream tasks. Future directions, possible clinical use cases, and limitations are discussed.
arXiv:2510.16657v3 Announce Type: replace-cross
Abstract: Synthetic data has been increasingly used to train frontier generative models. However, recent studies raise key concerns that iteratively retraining a generative model on its self-generated synthetic data may keep deteriorating model performance, a phenomenon often coined model collapse. In this paper, we investigate ways to modify the synthetic retraining process to avoid model collapse, and even possibly help reverse the trend from collapse to improvement. Our key finding is that by injecting information through an external synthetic data verifier, whether a human or a better model, synthetic retraining will not cause model collapse. Specifically, we situate our theoretical analysis in the fundamental linear regression setting, showing that verifier-guided retraining can yield near-term improvements, but ultimately drives the parameter estimate to the verifier's "knowledge center" in the long run. Our theory further predicts that, unless the verifier is perfectly reliable, these early gains will plateau and may even reverse. Indeed, our experiments across linear regression, Variational Autoencoders (VAEs) trained on MNIST, and fining-tuning SmolLM2-135M on the XSUM task confirm these theoretical insights.
arXiv:2511.02177v3 Announce Type: replace-cross
Abstract: Metal halide perovskites are promising materials for optoelectronic applications owing to their outstanding optical and electronic properties. Among them, all-inorganic perovskites such as CsPbBr$_3$ offer superior thermal and chemical stability. However, obtaining high-quality CsPbBr$_3$ thin films via solution processing remains challenging due to the precursor's low solubility, and current additive or solvent engineering strategies are often complex and poorly reproducible. High-pressure recrystallization has recently emerged as a promising route to improve film quality, yet its impact on film properties remains insufficiently explored. Here, we systematically investigate the morphological, structural, and optical properties of CsPbBr$_3$ thin films prepared by high-pressure recrystallization, in comparison with standard non-recrystallized films. Optimized recrystallization at 300 bar produces smooth, pinhole-free, single-phase 3D perovskite layers with sub-nanometer roughness, while the film thickness is precisely tunable via precursor concentration. The process enhances both grain and crystallite sizes, leading to amplified spontaneous emission with a reduced excitation threshold and improved photostability. Temperature-dependent X-ray diffraction further reveals the orthorhombic--tetragonal--cubic phase transition, consistent with single-crystal behavior. This study provides fundamental insights into pressure-driven recrystallization and establishes a reproducible, scalable approach for fabricating high-quality CsPbBr$_3$ films for optoelectronic devices.