AI and science frontiers · 2026-09-24

The work introduces SRMA-Mamba, a Mamba-based network that uses the Spatial Anatomy-Based Mamba (SABMamba) module to perform selective scans across sagittal, coronal, and axial anatomical planes for a global spatial context, and the Spatial Reverse Mamba Attention (SRMA) module to progressively refine boundaries from a coarse segmentation map and hierarchical encoder features, reporting better segmentation metrics than SegResNet, UX-Net, MedNeXt, SwinUNETR, SwinUNETRv2, and SegMamba on CirrMRI600+ T1W and T2W, with fewer parameters and GFLOPs than SegMamba.

The study proposes a "gravity-awareness" framework in which two models trained on parabolic-flight literature — CorticalG, a Fourier-feature-augmented multilayer perceptron predicting g-dependent EEG band-power change, and PhysioG, eleven independent Gaussian processes predicting heart-rate variability, electrodermal activity and motor measures — are complemented by Claude 3.5 Sonnet generating first-person narratives of awareness at 0 g, lunar 0.16 g, Martian 0.38 g and hypergravity; results show alpha (DMN) and mu (SMN) suppression in microgravity, elevated beta (PFC) and gamma in hypergravity, a V-shaped electrodermal response (right-hand conductance rising over 200% at 1.8 g) and an inverted-U trunk-activity pattern (dropping over 50% at 0 g).

The work proposes AGGRNet, which embeds a Feature Extraction and Aggregation (FEA) module and a C2PCA block into a YOLOv11 classification backbone; spatial and channel attention plus a learnable threshold τ split feature maps into informative and non-informative parts that are then aggregated by cross-attention, yielding results above the compared SOTA models on five public datasets, including a 5.48% accuracy gain on Kvasir and a 2.2% gain on LIMUC.

The work proposes a zero-shot 3D CT super-resolution framework: it first trains a diffusion model on abundant 2D X-ray data and uses DDNM/DDNM+ to upsample low-resolution CT projections into high-resolution projection priors, then applies a new Negative Alpha Blending Gaussian Splatting (NAB-GS) that models positive and negative Gaussian densities to learn the signed residual between diffusion-generated HR projections and upsampled LR projections for HR volume reconstruction; on the two public datasets UHRCT and MELA it achieves higher PSNR and SSIM than trilinear, cubic, NeRF, and CuNeRF zero-shot methods, is competitive with the supervised ArSSR, runs in about 15 minutes per volume, and two domain experts judged the 4× results to have clinical potential while 8× still needs improvement.

The work presents CleanSurvival, a reinforcement-learning (Q-learning) based automated data-preprocessing framework tailored to time-to-event (survival analysis) models for censored data: it selects among combinations of data imputation, outlier detection, and feature extraction techniques to optimize performance for a Cox, random forest, neural network, or user-supplied time-to-event model; on real-world dataset benchmarks, this Q-learning-based preprocessing improved predictive performance relative to simple baselines, runtime behavior was condition-dependent and most clearly interpretable in the best-covered benchmark cells, and a simulation study showed effectiveness across different types and levels of missingness and noise.

This survey defines parallel reasoning as a three-stage inference paradigm of decomposition, parallel processing, and aggregation, gives a formal formulation and distinguishes it from Chain-of-Thought and long-thinking, then organizes representative methods along three lines—non-interactive (self-consistency, Best-of-N ranking, structured reasoning), interactive (intra-model and multi-agent interaction), and efficiency (parallel decoding, parallel function calling, speculative decoding)—and summarizes applications, core challenges, and future directions.

Starting from the BIOMEDICA corpus, the authors used DAB-DETR-based subfigure detection, Qwen2.5-VL-32B-Instruct subcaption extraction, and Qwen2.5-14B-Instruct summarization of inline context to curate and release OPEN-PMC-18M, a dataset of 18 million subfigure-text pairs, then trained vision-language encoders evaluated on 6 retrieval and 19 zero-shot classification tasks across radiology, microscopy, and visible light photography, reporting an average recall of 21.64, a 27% relative improvement over the previous best model OPEN-PMC, and an overall zero-shot average F1 of 39.77.

The work proposes PrefixAgent, a two-phase large-language-model-driven framework for prefix adder optimization: in Phase I a fine-tuned large reasoning model iteratively optimizes the backbone via regroup tool calls, and in Phase II it performs local timing repair with level-opt, fanout-opt, and node clone tools; the authors use e-graph equality saturation and explanation to generate interpretable rewrite trajectories as supervision data, and under the NanGate45 and OpenROAD flow PrefixAgent produces smaller-area adders than DP, MCTS, PrefixRL, CircuitVAE, and PrefixGPT in nearly all configurations, achieving up to 11.3% area reduction over the best baseline at 64 bits and also improving area in a commercial flow and a 256-PE systolic array.

Using Fisher information and random matrix theory, the authors derive a closed-form expression for the noise-induced error of chaotic/diffusive reconstructive spectrometers, linking the variance bound σ²_ε Tr[(AᵀA)⁺] to the spectral correlation length Γ_corr, mean transmittance T₀, and the numbers of frequency channels N and measurement channels M, establish conditions for super-resolution, and validate the theory with a random matrix model and full-wave FDTD simulations, where the predicted optimal on-chip cavity size is about 17.9 µm.

Using 20 eighteen-month-old pregnant crossbred replacement heifers averaging 383 kg, this study fed 0.0%, 0.5%, 1.0%, and 2.0% North Atlantic brown seaweed (Atlantic GRO®, made of Laminaria longicruris and Fucus vesiculosus) on a dry matter basis over a 50-day performance trial, then measured gas emissions in 16 of them with headbox metabolic chambers for two 24-h periods; no supplementation level adversely affected growth performance (p > 0.51), while the 1.0% and 2.0% groups emitted about 8.2% and 8.7% less methane than controls (p < 0.05), with significantly lower carbon dioxide emissions and oxygen consumption as well (p < 0.05).

The work proposes TG-OT, a fully automatic CCTA-IVUS registration framework: lightweight CNNs first predict calcifications, bifurcations, and lumen radii on the topological (θ, z) cylinder (with the IVUS network additionally detecting guidewire artifacts), and the frozen detectors are then integrated directly into a differentiable registration pipeline that optimizes centerline warping parameters driven by an unbalanced Sinkhorn optimal transport loss on the cylindrical geometry plus a Dice term, complemented by a lumen radius matching term; on N=47 paired CCTA-IVUS cases from the IMPACT study at Erasmus University Medical Center in a 5-fold cross-validation setup, it reaches longitudinal Dicectl=0.99, rotational Sc=0.96, and lumen DiceL=0.

Across four abdominal CT datasets (WORD, AMOS, CT-1K, AbdomenAtlas), the authors generated pseudo-label variants with seven models (nnU-Net, MedSAM, TotalSegmentator, and four STU-Net sizes), then trained DynUNet under identical deterministic settings for in-domain training and for pretraining followed by fine-tuning; in-domain performance rose strongly and non-linearly with label quality and dataset volume partly compensated for quality, whereas fine-tuned models significantly outperformed a no-pretrain baseline in the vast majority of settings and pretraining label quality no longer clearly affected downstream results.

Using soil moisture from 17 CMIP6 models bias-corrected against GLDAS and combined into a multi-model ensemble, the study characterizes 2015–2050 global soil moisture droughts with a non-parametric SSMI under SSP2-4.5 and SSP5-8.5 and assesses agricultural exposure with a Drought Exposure Index overlaying cropland, finding intensified drought characteristics toward mid-century, longer durations and wider spatial extent under SSP5-8.5, pronounced drying in South America, southern Europe, South Asia and parts of North America, and a global mean DEI rising from 0.71 under SSP2-4.5 to 0.78 under SSP5-8.5.

The authors propose the Cryptographically Verifiable Agent Authorization (CVA) hypothesis, formalizing authorization as a relation RCVA that jointly binds an agent principal, a concrete authorization request, an execution context, and policy satisfaction, and provide candidate security properties plus an executable zero-knowledge proof of concept over Groth16 zk-SNARK; the prototype implements principal binding and plan-level request binding while context binding and runtime execution binding remain unimplemented.

This work studies the Lp error of approximating higher-order Korobov functions f∈K^{m+1}_p(Ω) by deep ReLU convolutional neural networks (CNNs), proving that for depth L≤Csd^4m^3N(log_2 N) there exists a network with inf‖f−f_L‖_{Lp(Ω)}≤C_{m,d}‖D^{m+1}f‖_{Lp(Ω)}N^{−m−1}(log_2 N)^{(m+2)(d−1)}, i.e. it improves the classical second-order mixed-derivative rate O(L^{−2+1/p}) to order (m+1) up to a logarithmic factor, and concludes that the higher-order expressivity of CNNs does not severely suffer from the curse of dimensionality.

This study assembled a dataset of 32,847 webcam videos from 1,888 participants (727 with Parkinson's disease) across 16 standardized clinical tasks and systematically benchmarked seven video foundation models (VideoPrism, V-JEPA2 and its SSv2 variant, ViViT, VideoMAE, VideoMAEv2, and TimeSformer) using frozen embeddings with a linear classification head, finding that task saliency is highly model-dependent: VideoPrism excels in visual speech kinematics and facial expressivity, V-JEPA2 variants are superior for upper-limb motor tasks, and TimeSformer remains competitive for rhythmic tasks like finger tapping, with overall AUCs of 76.4–85.3%, accuracies of 71.5–80.6%, specificity up to 90.3%, and sensitivity of only 43.2–57.3%.

This study empirically evaluates 7 representative diffusion LLMs (including the closed-source Mercury-Coder-Small) against 4 open-source autoregressive baselines on HumanEval/HumanEval+, MBPP/MBPP+, HumanEval-X, LiveCodeBench, RepoQA, RepoBench-C, and SWE-Bench Verified, finding promising but uneven code-generation ability: the closed-source diffusion model leads on most benchmarks (e.g., 79.1% on MBPP+ versus 73.

The authors introduce a general framework based on Lie-algebraic properties that maps a pool of Pauli operators to a binary matrix ΓA over F2, proves that pool completeness and minimality can be decided in polynomial O(N³) time via the rank and congruence relations of that matrix, uses it to construct minimal complete pools (MCPs), proposes MB-ADAPT-VQE (adding a batch of k operators per iteration) to cut measurement overhead and accelerate convergence in quantum chemistry, and extends the fixed-ansatz NI-DUCC-VQE method, previously limited to N≤14 qubits by the MCP construction bottleneck, to a 26-qubit H2O system.

The work proposes Dino U-Net: a frozen DINOv3 foundation backbone as encoder, combined with a dual-branch DINO Adapter and a Fidelity-Aware Projection Module (FAPM), which outperforms seven baseline methods on seven public medical image datasets spanning endoscopy, ultrasound, microscopy, MRI, fundus and CMR modalities, and shows performance improving as the backbone scales from S to 7B.

The work proposes a retrieval-augmented text-to-CT generation method: given a radiology report, a pretrained 3D vision-language encoder retrieves the semantically nearest case from a reference corpus, and that case's anatomical annotation is used as a structural proxy injected into a text-conditioned latent diffusion model through a ControlNet branch; on the CT-RATE dataset (27,514 training and 1,818 test volumes), retrieval-augmented generation improves image fidelity (FID 3D 0.004), clinical consistency (CT-Net AUC 0.787), and spatial controllability (Dice 0.772, HD95 3.072) over text-only baselines, while additionally enabling explicit spatial controllability that such approaches inherently lack.

The work introduces an exemplar-free continual learning framework for whole-slide-image-to-report generation: it builds a compact domain footprint per domain in a frozen patch-embedding space (a k-means codebook, a slide-level code histogram bank, patch-count statistics, and a report-style prototype), uses it to synthesize pseudo-WSIs whose pseudo-reports come from an immediate teacher snapshot for generative replay, and conditions the language model through a style prefix; across multiple public continual learning benchmarks the approach outperforms exemplar-free and limited-buffer rehearsal baselines and supports domain-agnostic inference without explicit domain identifiers.