Papers

We use replica-exchange molecular dynamics (REMD) to map the conformational free-energy landscape of the macrocyclic drug lorlatinib in explicit water and chloroform. In water a single dominant basin is recovered; in chloroform two conformational states are resolved. We use these conformers as the starting point to calculate proton chemical shifts from first principles and a continuum solvent model. The resulting population-weighted spectrum is compared directly to the experimental CDCl$_3$ spectrum, benchmarking the computational approach against measured NMR data.

Electromagnetic counterparts to gravitational-wave (GW) sources probe the properties, environments, and evolution of compact-object mergers. S251112cm belongs to an emerging class of GW candidates with possible subsolar-mass components. We present an optical counterpart search for S251112cm with the Wide Field Survey Telescope (WFST). Observations began 20.2 hr after the GW trigger and lasted three nights, covering approximately 780 deg$^2$ and 51% of the localization probability in the updated skymap. Our search for newly emerging, rapidly evolving, off-nuclear optical transients yielded four candidates with host-galaxy distances broadly consistent with the GW distance estimate. All four evolve substantially more slowly than AT 2017gfo, ruling out an AT 2017gfo-like origin and disfavoring their association with S251112cm as rapidly evolving kilonova counterparts. Combining WFST and DECam increases the localization probability coverage to approximately 67%. Assuming the counterpart lies within this footprint, we constrain a grid of binary-neutron-star kilonova models. For viewing angles $\theta_{\rm obs}<60^\circ$, with other parameters fixed to the AT 2017gfo best-fit values, models with dynamical ejecta mass $M_{\rm dyn}\gtrsim0.01 M_\odot$ or wind ejecta mass $M_{\rm wind}\gtrsim0.05 M_\odot$ are disfavored over most of the covered GW probability. Our phenomenological analysis further constrains rapidly evolving optical transients with $T_{\rm rise,1/2}\leq2$ days to peak absolute magnitudes fainter than approximately -13 mag. To our knowledge, WFST provides the strongest optical constraints to date in the northern sky on possible rapidly evolving counterparts to this new class of GW candidates with subsolar-mass components. The comparable depths and complementary sky coverage of WFST and DECam extend these constraints across more of the localization region at similar sensitivity.

Complementing an earlier paper, which focused on $3$-sumfree sequences, we here consider only $2$-sumfree sequences: starting with positive integers $f$ and $g>f$, the infinite, increasing 2-sumfree sequence $S_{f,g}$ is constructed as follows. After any initial segment, the next entry is the smallest positive integer exceeding all previous ones that differs from all sums of distinct pairs in the sequence. It follows from a theorem in the previous paper that for every $f\geq 1$ and all $f+1\leq g<2f$ the sequence $S_{f,g}$ exhibits ultimately periodic behaviour. In this paper we state precise conjectures that, if true, would imply that every $2$-sumfree sequence is ultimately periodic. Here periodicity of an increasing sequence is understood to mean periodicity of the sequence of first differences, or, equivalently, of its characteristic sequence. We supply much computational evidence to support the conjectures.

Speech retrieval has become increasingly important as spoken content continues to grow across meetings, lectures, podcasts, and videos. Existing benchmarks and models have advanced semantic search over spoken content, but largely focus on what is said while overlooking who says it. In many real-world scenarios, however, users need to retrieve speech based jointly on semantic content and a target speaker, where the speaker may be specified naturally through a reference speech utterance rather than a predefined identity. To address this gap, we introduce VoiceTrace-Bench, a benchmark for hybrid speech retrieval in which each query combines text specifying what to retrieve with reference speech specifying who to retrieve. This setting requires models to integrate complementary semantic and speaker information directly from heterogeneous query inputs. Motivated by the joint audio-text modeling capabilities of audio-language models (ALMs), we develop VoiceTrace, a two-stage retrieval framework consisting of VoiceTrace-Emb, an embedding model that learns unified representations for efficient large-scale retrieval, and VoiceTrace-Reranker, a reranking model that jointly examines each query–candidate pair for fine-grained relevance estimation. Experiments show that VoiceTrace achieves state-of-the-art performance on established semantic speech retrieval benchmarks, while substantially outperforming cascade-based approaches on VoiceTrace-Bench, demonstrating its effectiveness for both conventional semantic retrieval and the new hybrid retrieval setting.

Collective intelligence is a collaborative autonomy paradigm in which multiple agents pursue shared objectives through local perception, information exchange, and coordinated action. UAV swarms embody this paradigm by coordinating multiple vehicles in tasks such as search, inspection, and tracking. Recent advances in large language model (LLM) agents have strengthened natural-language task understanding and high-level planning, providing a flexible semantic interface between mission descriptions and collective behavior. While these advances expand semantic reasoning, applying LLM agents to UAV swarms raises challenges in grounding model decisions in executable capabilities, reconciling global task reasoning with distributed execution, and using mission-specific experience for continual adaptation. To address these challenges, we introduce AeroWeaver, an embodied-agent harness that weaves individual UAV skills into coordinated mission-level behavior. AeroWeaver connects semantic decisions to governed skills, organizes role-conditioned local agents for distributed coordination, and uses role-indexed state-action-reward experience to refine skill selection online. Experiments and runtime validation show that AeroWeaver maintains valid skill execution under tested conditions and supports body-local multi-UAV operation without a central agent generating joint actions from global context, while reward-guided online updates provide a training-free path for adaptive learning swarm agents from accumulated execution experience. Code: https://github.com/Admire-ljb/AeroWeaver.

With the rapid advancement of deep learning technology, shared GPU clusters receive an increasing number of deep learning training (DLT) jobs. Yet resource fragmentation make such clusters underutilized and forces the DLT jobs running on them to endure long turnaround times. Extensive research has been devoted to quantifying fragmentation and developing scheduling algorithms that alleviate its impact. However, existing fragmentation measures break down in the absence of workload distribution information, while current schedulers cannot continuously maintain resource fragmentation at a low level. To tackle these problems, we first introduce Scheduler-Induced Fragmentation (SIF), a metric built on the notion of partial-nodes that is independent of historical workload knowledge. We then propose COMPASS-ABS, which employs the COMPact-ASSured (COMPASS) algorithm to confine the cluster state within a tight Anchor-Based Space (ABS), whose construction fully leverages the topological alignment between dominant workload size and node capacity. Moreover. We also prove that it ensures SIF is bounded by $\frac{2}{N}$ under a workload composition condition that matches both theory and production. Evaluations implemented on a physical cluster and a simulated cluster demonstrate COMPASS-ABS effectiveness at improving resource utilization, reducing DLT job completion time by reducing fragmentation.

Google reCAPTCHA is the most widely deployed visual CAPTCHA service, protecting hundreds of thousands of websites from automated bots. It serves as a critical line of defense against automated attacks, including credential stuffing, bulk account creation, and automated form abuse. It has proven largely effective since its introduction in 2007. However, the rise of accessible AI now threatens its efficacy. Prior work has demonstrated that commercial cloud-based vision-language models (VLMs) can solve visual CAPTCHA challenges, but at non-trivial monetary cost per attempt. In this paper, we show that free and locally-run models can break Google reCAPTCHA. We conduct a comprehensive study of reCAPTCHA and present a taxonomy of its challenge types: Type A (independent image tiles, with static and dynamic sub-variants) and Type B (a single image partitioned into a 4x4 grid), each demanding a distinct solving strategy. We design zero-shot, no-cost solvers built entirely on open-source local models, specifically CLIP (58% per-challenge accuracy on Type A) and OWLv2 (43.5% on Type B), requiring no model training and no API access. Our end-to-end automated solver achieves a 92.6% per-session success rate across 500 real-world reCAPTCHA sessions. We further demonstrate that reCAPTCHA can be defeated by a non-technical adversary, using only natural-language instructions to a commodity AI assistant. This collapses the practical attacker skill floor to near zero and fundamentally changes the threat model for challenge-based CAPTCHAs. Although reCAPTCHA increasingly favors reputation-based verification, visual challenge-based fallback persists as a safety net that, paradoxically, has become the weakest link in the defense chain, suggesting that challenge-based visual CAPTCHAs may have reached the end of their useful life.

Traditional plasma density modulation schemes can extend the dephasing length but suffer from wakefield degradation as plasma density increases. To address this limitation, a drift-like phase-shift acceleration scheme is proposed, introducing drift sections between plasma cavities to enable controlled phase shifting and avoid dephasing. Idealized simulation results show that electrons are accelerated from 16 MeV to 700 MeV within 1.65 m with an energy spread of 5%. Furthermore, segmented simulations introducing realistic active plasma lenses verify the practical feasibility of this scheme, electrons are accelerated up to 720.5 MeV with a reduced energy spread of 3.3% within 1.67 m. The plasma wakefield maintains a robust amplitude throughout the propagation, suggesting the potential for continued acceleration beyond the simulated region. This scheme demonstrates a promising way to enable the witness beam to gain energy continuously from the tail to the head of the driver beam and provides a new approach for generating high-energy, high-quality beams in heavy ion driven plasma wakefield acceleration.

While Large Language Models excel in natural language processing, efficiently extending their capabilities to spoken input remains a significant challenge. Existing methods for building SpeechLLMs often rely on computationally expensive full-model fine-tuning, or employ parameter-efficient projectors that suffer from inefficient token sequence lengths and costly full-model supervision. In this paper, we introduce Aligned Continuous Integrate-and-Fire, a highly efficient framework for zero-shot speech processing. Our method dynamically compresses continuous acoustic frames into the exact discrete token length of the target text utilizing explicit Dynamic Time Warping alignments. This allows our initial training stage to establish a robust acoustic-to-semantic bridge using lightweight distance metrics, entirely bypassing the computationally expensive LLM forward pass. For subsequent fine-tuning, we propose a memory-efficient knowledge distillation objective that targets a single LLM layer, performing competitively with full-model cross-entropy training at a fraction of the computational cost. Through extensive evaluations on Automatic Speech Recognition and Speech Translation, we demonstrate that our method achieves superior performance compared to prior parameter-efficient baselines.

Safety alignment teaches large language models (LLMs) to recognize harmful requests and reject risky instructions. Yet aligned models can fail when harmful intent is concealed within seemingly benign contexts. Robust safety therefore requires both knowledge of safety boundaries and vigilance: the ability to detect unusual premises, misleading reasoning, and latent risks beneath surface-level semantics. Vigilance requires models to scrutinize a request's underlying intent and assumptions before acting. To cultivate this capability, we introduce cunning questions, which are not necessarily safety-related but contain misleading premises, atypical reasoning, or subtle inconsistencies. We hypothesize that learning to look beyond such reasoning traps can transfer to safety-critical scenarios. Experiments show that Cunning training improves robustness to out-of-distribution jailbreak attacks and strengthens subsequent safety fine-tuning. Furthermore, augmenting an existing state-of-the-art safety alignment pipeline with Cunning establishes a new state of the art across our evaluated settings, reducing mean ASR across nine backbone–benchmark combinations from 17.40% to 15.05%. Trace analysis after matched safety fine-tuning suggests that safety judgments are more likely to govern responses before harmful planning begins. A conditional theoretical analysis further characterizes when invariance learned from cunning data can transfer to safety-related inputs. These findings suggest that cunning data can strengthen model vigilance and complement conventional safety alignment.

Active perception is essential for robotic manipulation when fixed viewpoints leave task-relevant information occluded or unobserved. However, enabling vision-language-action (VLA) models to reason across changing viewpoints and actively acquire informative observations remains challenging. We present ActiveScale, a framework that advances active perception through coordinated model, data, and hardware designs. Our model augments a VLA with historical video observations and explicit camera-pose supervision, using per-frame pose tokens and a lightweight prediction head to associate observations across viewpoints and support a coherent understanding of the scene. To learn from the camera motion naturally present in human activity, we introduce a scalable human–robot mid-training recipe using 1000 hours of egocentric and robotic data, adapting the model to temporal inputs and pose supervision. We further introduce Active-perception Mobile-manipulation Platform (AMP), a robotic platform that supports active perception and mobile manipulation through single-operator teleoperation, enabling scalable collection of demonstrations that coordinate viewpoint changes and manipulation. Experiments demonstrate improved success rates on active-perception tasks, while ablation studies validate the contributions of camera-pose-aware modeling and egocentric mid-training. Together, these components provide an integrated foundation for studying and developing active perception in robotic manipulation.

A subset $X$ of a finite group $G$ is $k$-power-avoiding if for all $g\in G$ we have that $\{g,g^k\}\not\subseteq X$. The paper shows that if $G$ contains a $k$-power-avoiding subset $X$ with $|X|\geq |G|-c$, then the group $G^k$ generated by the $k$th powers of elements of $G$ has bounded order (in $k$ and $c$). We provide more detailed structural results when $c\leq 2$, and in particular we classify the groups which arise when $c\leq 2$ and $k$ is prime.

Fault-tolerant agreement protocols fail if replicas share a common flaw that simultaneously affects more replicas than the tolerable threshold. Therefore replicas should ideally fail independently, which can be achieved through diversification. However, in practice, often the same protocol implementation is shared by all replicas which is not surprising given that the provision of multiple diverse implementations is difficult and highly laborious. This poses a major risk, as a shared protocol implementation is a prime candidate for common bugs due to its complexity. With PatchyBFT, we demonstrate how, given a reference implementation, Large Language Models (LLMs) can be utilised for the automated and scalable generation of code that compiles, passes tests, and crucially differs semantically/binary-wise, that can re- place code in the reference implementation, thereby significantly reducing diversification costs. We demonstrate the feasibility of diversification of replication protocol implementations using LLMs by diversifying three implementations: PBFT, HotStuff, and Raft, showing how up to 65% of the codebase can be diversified.

In autonomous driving, perception models often struggle to generalize to new environments due to domain shifts. While unsupervised model adaptation offers a feasible solution without labor-intensive manual labeling, existing methods that rely solely on the ego-vehicle's data often lead to inferior pseudo-labeling performance. To address this critical issue, we propose LDE, Learning from Distributed “Eyes", a novel framework that transforms collaborative perception (CP) into a source of high-quality supervision for model adaptation. This pseudo-labeling approach is hyperparameter-insensitive and relatively reliable, assuming CP often outperforms single-agent's perception. However, naively implementing this approach encounters (1) the communication bottleneck of sharing rich features under time and bandwidth constraints, (2) the view discrepancy between the CP view and the learner's Field of View (FoV), and (3) the unreliability even in CP-generated labels. To address these issues, we design an adaptation-oriented feature sharing mechanism that selectively transmits the most critical information for adaptation, an FoV filtering method that meticulously eliminates mismatched labels, and a curriculum learning strategy to progressively exploit pseudo labels. Extensive experiments on 3D object detection tasks demonstrate that LDE consistently outperforms both the pre-trained models and state-of-the-art unsupervised adaptation methods.

We present DiT-Garment to model dynamic 3D clothing over human body models in arbitrary motion. Unlike existing methods, DiT-Garment can animate garments with unseen designs and physical materials, while allowing for direct inference of deformations for any target pose. To achieve this, we leverage a 2D diffusion transformer architecture to learn 3D deformations in a 2D UV-space. As the result is non-deterministic, our generative model learns the distribution of possible outcomes. The template garment is represented as a 3D triangle mesh spatially aligned with a 3D human body model in a standardized pose. To work with different garment designs without the need of a common template or complex graph convolution operations, the diffusion transformer is conditioned on a 3D position map of the template, represented in UV-space, which allows to implicitly learn a deformation of the 3D space around the body in standard pose. Further conditioning on body motion and physical parameters allows to physically ground the model. We quantitatively and qualitatively evaluate DiT-Garment on both synthetic and real data. While only trained on synthetic simulations of automatically generated cloth designs, our method generalizes to captured and artist-made garment designs. Code and data are available for research purposes at https://dumoulina.github.io/dit-garment/.

This paper analyzes the stochastic dynamics of a two-link pendulum that models lower limb segments (the thigh and shank) during the swing phase of the gait cycle. We describe the model using a system of non-linear Ito stochastic differential equations in a four-dimensional phase space under both external and internal random perturbations.We perform a spectral analysis of the discrete Kolmogorov operator, compute the stationary invariant measure, and estimate diffusion parameters using Neural Stochastic Differential Equations (Neural SDEs).

We study Hamiltonian stationary Lagrangian surfaces in complex space forms with nowhere-zero mean curvature vector $H$. We prove that if the Gaussian curvature is constant and $\grad|H|$ has positive constant length, then both the surface and the ambient space are flat. Moreover, the immersion is locally congruent to the known surface obtained as the product of a Cornu spiral and its reflection.

Let \pi: \mathcal X \rightarrow \Delta be a smooth proper family of compact complex manifolds such that the central fiber \mathcal X_0 is Kähler. Then all the fibers close to 0 are Kähler and {\pi} is a weakly Kähler morphism, even if \mathcal X is not a Kähler space. In this paper we show that if dim \mathcal X_0 = 4 and K_{\mathcal X_0} is nef, then K_{\mathcal X_t} is nef for all t in a neighborhood of the origin.

We prove that a complete noncompact Kähler surface with nonnegative Ricci and nonnegative quadratic orthogonal bisectional curvature is contractible, and hence homeomorphic to $\mathbb{R}^4$, if it is simply connected at infinity. Under positive bisectional curvature, this removes the contractibility assumption from the conditional uniformization theorem of Datar–Pingali–Seshadri: strong Steinness and simple connectivity at infinity suffice to identify the surface biholomorphically with $\mathbb{C}^2$. We also derive bounded-gradient strictly plurisubharmonic exhaustions and uniform holomorphic kernel estimates for complete $U(n)$-invariant Kähler metrics on $\mathbb{C}^n$ with nonnegative bisectional curvature.

In practice-based design courses such as knit yarn design, students must turn visual ideas into feasible material outcomes. This is difficult because creative decisions are tied to yarn properties, stitch structures, machine operation, and limited opportunities for physical sampling. This study presents an integrated pedagogical framework that combines flipped learning, exemplar-based reference, GenAI-assisted visual prototyping, and studio feedback in an undergraduate knit yarn design course. The framework was implemented through a cross-device platform with pre-class micro-videos, formative checks, a curated gallery, and a GenAI-supported ideation module. An exploratory course-based evaluation compared a historical control cohort (N = 12) and an intervention cohort (N = 16), supplemented by questionnaire responses and brief interviews. The findings are interpreted as context-specific indicators rather than confirmatory causal evidence. Exploratory comparisons showed higher scores in creativity thinking, design skills, problem solving, and total course score in the intervention cohort. Student and instructor responses suggested that flipped learning supported studio readiness, while GenAI mainly supported early-stage visual exploration rather than precise technical guidance. Overall, the study offers a practice-based instructional framework for integrating flipped preparation, GenAI-assisted visual prototyping, and studio feedback in design education.

Grasp synthesis aims to generate stable and physically plausible hand–object interactions, and has become a fundamental problem in both human hand modeling and robotic manipulation. However, a unified representation across human and robotic hands is still lacking, mainly due to differences in hand morphology and surface modeling. Prior methods typically rely on either contact maps or dense implicit descriptors to represent interaction, but these representations are often incomplete or computationally expensive and redundant. We propose InterMASH, a unified geometric representation that establishes cross-embodiment correspondence using sphere-fixed anchors. At each anchor, low-degree spherical harmonics compactly encode local hand geometry, object geometry, and contact, forming an explicit and interpretable token sequence. Building on this natively tokenized structure, we introduce a conditional Diffusion Transformer that operates directly in the proposed InterMASH representation space and jointly generates hand geometry and contact, improving consistency and physical plausibility. Our method achieves competitive performance with state-of-the-art methods on key physical feasibility metrics in a large-scale ShadowHand benchmark, supports joint training across multiple hands, and shows that cross-embodiment fine-tuning with human grasp data can improve robotic grasp success and diversity. Project page is available at https://inter-mash.github.io/.

We study the asymptotic behavior of a sequence of Galton–Watson processes with inhomogeneous immigration when the limit of the means of the offspring distributions is less than $1$ or equal to $1$. Under growth conditions on the expected values of the immigration distributions and the variances of the offspring distributions, and assuming the weak convergence of properly scaled immigration processes towards a non-negative stochastic process $\mathcal Y$ with càdlàg or continuous sample paths, we establish functional limit theorems for the sequence of Galton–Watson processes with inhomogeneous immigration in question. The limit stochastic processes can be represented as a constant multiple or an integral functional of $\mathcal Y$ .

We explain how to derive the Robinson-Schensted-Knuth (RSK) correspondence, a fundamental bijection in algebraic combinatorics, from the Yang-Baxter equation. The Yang-Baxter equation arose in the study of quantum many-body systems and later became a cornerstone of the theory of solvable lattice models, particularly vertex models. In a vertex model, arrows occupy the edges of a grid, and each vertex carries a Boltzmann weight determined by the arrows on the four edges meeting at it. The weight of a configuration is the product of these local weights, and a partition function is the sum of the weights of all configurations with prescribed boundary conditions. For a vertex model whose partition functions are the Schur polynomials, the two sides of each instance of the Yang-Baxter equation admit exactly one weight-preserving matching of their summands. Carried across a grid, this forced matching is the classical RSK correspondence in the form of Fomin's growth diagrams. In natural coordinates the local matching becomes the combinatorial three-dimensional R, a set-theoretic solution of the Zamolodchikov tetrahedron equation. Its periodic closure returns the combinatorial R-matrices of one-row crystals. The forced matching is special to the Schur weights. For the Hall-Littlewood and q-Whittaker deformations and their spin versions, at generic parameter values no deterministic matching works for all boundary data. Reading each instance of the Yang-Baxter equation probabilistically, we replace the matching by a coupling of the two sides - a bijectivization, or probabilistic bijection - and obtain Markov operators that transport probability measures attached to vertex models. Iterated over the grid, these operators produce randomized RSK-type dynamics and interacting particle systems, including q-PushTASEP and the stochastic six-vertex model.

Distributed database management systems (DDBMSs) introduce new challenges for assessing their reliability due to distribution-specific characteristics that affect query execution and optimization. Existing testing approaches, largely designed for centralized DBMSs, often fail to explore diverse distributed execution behaviors and suffer from low executability of generated test queries, thereby limiting their effectiveness in bug detection. We propose DAT (Distribution-Aware Testing), a novel automated approach for detecting query-processing bugs related to distribution strategies and distributed optimizations in DDBMSs, by systematically leveraging distribution-aware information throughout the testing pipeline. DAT builds on a set of techniques that capture diverse combinations of logical schemas and data distribution strategies, and performs guided query mutation to trigger a wide range of distributed query execution behaviors and optimizations, while improving query executability via historical feedback. We implement our approach in a tool, DistRanger, and evaluate it on four widely used production DDBMSs. It uncovers 31 previously unknown bugs, including 28 related to distributed query processing and optimization, and outperforms state-of-the-art testers.

For every sufficiently large N, we prove that the regular N-gon uniquely minimizes the first Dirichlet eigenvalue among convex polygons of prescribed area with at most N sides, up to rigid motions. Quantitative Faber–Krahn stability localizes minimizers near the disk but leaves their discrete geometry undetermined. We encode that geometry by a centered measure of exterior angles and prove an inverse-cubic Fourier inequality whose equality case is the uniform root configuration. Its defect gives a lower bound for the quadratic spectral excess over the regular polygon. Cyclic cancellation and a common material-coordinate comparison make the nonlinear difference small relative to the same defect. Exact constraint coordinates extend this comparison to arbitrarily small positive defects. Minimality then forces zero defect, which characterizes the regular polygon.

RF jamming poses a severe threat to 5G/6G networks, increasing packet latency being especially disruption to mission-critical URLLC services. This paper introduces a proactive Jamming Detection xApp (JD-xApp) for the Open RAN architecture that detects attacks by monitoring the moving-average Block Error Rate (BLER) via the E2 interface. Upon detection, the algorithm overrides standard link adaptation, enforcing a robust upper ceiling on the Modulation and Coding Scheme (MCS) to stabilize latency. To counter O-RAN platform adoption challenges, we transition the framework to an open-source Centralized Unit/Distributed Unit implementation (OCUDU). Evaluated with high-end Keysight lab equipment including the UXM 5G Wireless Test Platform and the PROPSIM F64 channel emulator, the JD-xApp reduces the expected number of packet retransmission attempts by about 67.3%. Finally, the system's feasibility is validated via over-the-air deployment using the POWDER lab infrastructure.

On-water catalysis accelerates reactions between poorly soluble organic substrates in aqueous suspensions, but its molecular origin remains debated. This Account argues that hydrogen bonding and proton transfer can both enhance charge-transfer stabilization between the organic reactants. Hydrogen bonds from surface water polarize the reacting complex, whereas protonation can perturb the same donor-acceptor interaction more strongly without requiring identical reaction pathways. We connect simulations of the water/vapor interface, finite-temperature cycloaddition studies, and energy decomposition analysis based on absolutely localized molecular orbitals. Dangling OH groups establish reactive contacts, but an increased transition-state hydrogen-bond count alone does not explain activation. Electronic analysis instead identified hydration-enhanced charge-transfer stabilization between organic partners by approximately 30%. Comparisons of related dienophiles, a protonated limiting model, and oxygen/sulfur substitution link substrate-dependent electronic response to hydration and bond formation. A hydration-induced charge-transfer response metric provides a testable molecular hypothesis, not a standalone measure of catalytic acceleration. Activation depends on differential stabilization of reactant and transition-state ensembles. Increasing the accessible reactive interface can amplify this local mechanism without requiring exceptionally strong additional electric fields, connecting conventional on-water chemistry with selected microdroplet reactions while distinguishing adsorption, concentration, and intrinsic molecular activation.

Vision-Language-Action (VLA) models have demonstrated strong performance across diverse robotic manipulation tasks, yet their predominantly vision-centric perception and position-controlled execution remain insufficient for contact-rich manipulation. Visual observations alone often provide limited evidence of contact onset and interaction magnitude, while position-control policies cannot respond compliantly to rapidly changing contact dynamics. To bridge both the perception and control gaps, we propose TAO-Force, a force-conditioned VLA framework that combines force-aware policy learning with contact-regulated execution. For force-aware perception, TAO-Force introduces Force-conditioned Feature-wise Linear Modulation (F-FiLM) to inject encoded force feedback into the representations of a frozen pretrained visual-language backbone while preserving its semantic priors. For responsive control, it employs a contact-gated fast-slow architecture, with a slow position-control branch tracking nominal trajectories during non-contact phases and a fast admittance-control branch regulating physical interaction during contact phases. Detailed analyses on a force-perception task and real-world evaluations across four contact-rich manipulation tasks validate the effectiveness and robustness of TAO-Force.

Cybersecurity combines high-stakes analysis with complex technical language, making it an impactful and challenging domain for LLMs. We present MiST (Mid-trained Security Transformer), a suite of 8B and 32B models that achieve strong performance on public cybersecurity benchmarks. We use mid-training as an intermediate adaptation stage between general pre-training and cybersecurity training. Rather than performing continual pre-training over large volumes of raw domain text, we curate a compact, expert-vetted seed corpus, and transform it into high-quality domain-specific synthetic training data. The final MiST checkpoints improve mean cybersecurity accuracy by +13.1 and +8.6 absolute percentage points over the corresponding Qwen baselines for 8B and 32B, respectively, corresponding to relative gains of +27.0% and +15.8%. Ablation results further show that these cybersecurity gains arise in the mid-training and supervised fine-tuning stages through a combination of the synthetic data generation flows. Furthermore, we show that MiST provides a stronger initialization for downstream task-specific fine-tuning adaptation and reinforcement learning.

Acoustic lenses have been introduced recently to compensate for the phase distortions induced by the propagation across a human skull for ultrasonic deep-brain stimulation in humans. In this study, we present bifocal lenses that compensate for human skull aberrations and allow simultaneous targeting of multiple structures deep in the brain. We investigated the impact of phase unwrapping in the design of the lenses and how this process improves the distribution of pressure produced in N=5 human skulls for two different spatial arrangements of the targets. The results show that unwrapping the phase computed during the design increases the fidelity of the pressure field generated across the human skulls. The spatial precision is on average improved by 73%, and out of target energy deposition is on average reduced by 58%. The results presented in this study highlight the importance of phase unwrapping to optimize the safety and efficacy of future transcranial ultrasound stimulations targeting multiple regions.