Papers

A set of experiments on merging dynamics of localized magnetized plasma pressure filaments in close proximity has been carried out in a large linear magnetized plasma device. A two filament configuration is used to determine the separation scale of cross-field interaction. Fluctuation analysis of the ion saturation current confirms the presence of unstable drift-Alfvén modes on each filament which are predominantly driven by the steep electron temperature gradient. When the filaments are separated a distance of approximately five times the size of a single filament or greater, no interaction is observed and the drift-Alfvén modes remain uncoupled. When brought closer together, significantly stronger interaction occurs with transfer of charge and energy, leading to the generation of inter-filament electric fields. This has the effect of rotating the filaments and influencing the merging dynamics. When the filaments are asymmetrically heated, the cooler filament becomes wrapped around the other, and the drift-Alfvén modes become fully coupled in a global mode structure. Transport between the filaments is explained in terms of E x B drift and the local potential wells of the filaments, and drift-Alfvén mode coupling occurs when the filaments are brought within the distance of a collisionless skin depth.

We study the sharp $L^2$-Poincaré inequalities on Euclidean balls and with respect to the Gaussian measure, and we focus on their stability. In both settings, the sharp constant is the first nonzero eigenvalue of a self-adjoint operator and the set of optimizers is a finite dimensional linear space. Using the spectral decomposition, we establish sharp $L^2$-stability estimates and sharp gradient stability estimates, together with the stability of the stability inequalities in both norms, with explicit optimal constants and with the complete characterization of the equality cases. We then show that this process can be continued. The Poincaré deficit is equal to an infinite sum of $L^2$-distances to the increasing spaces of optimizers of the successive stability inequalities. It is also equal to an infinite sum of $L^2$-gradient distances to the same spaces. The coefficients of the first sum are the spectral gaps $\mu_j-\mu_{j-1}$, and the coefficients of the second sum are $\mu_1$ times the gaps of the reciprocal spectrum, $\mu_1\left(\mu_{j-1}^{-1}-\mu_j^{-1}\right)$. On the Euclidean ball, all the constants are expressed through the roots of Bessel functions, and we prove that the first two spectral levels are given by the degree one and the degree two Neumann modes in all dimensions, while the third level is given by the first radial mode when $N=2,3$ and by the degree three modes when $N\geq4$. In the Gaussian case, all the constants are rational numbers and the $L^2$-sum has all coefficients equal to $1$; we prove that this property characterizes the arithmetic spectrum.

Pull requests (PRs) underpin pull-based software development by enabling distributed code review and collaborative contribution in open-source projects. Yet many become inactive before integration and are eventually abandoned or closed, wasting contributor and maintainer effort. Although prior work has examined PR abandonment and review delays, less is known about the contribution types, discussion-level barriers, and post-stalling collaboration patterns associated with inactivity. We investigate which PR types most often stall, why inactivity occurs from authors' and reviewers' perspectives, and how stalling relates to later contributor and reviewer engagement. We analyzed 14,234 stalled PRs and 164,562 review comments from 19 popular GitHub repositories using stale-bot workflows. An LLM-based voting classifier categorized PRs by contribution type, while quantitative analysis was combined with qualitative coding of general and inline review discussions. Feature-enhancement and issue-fixing PRs formed the largest share, together exceeding 77% of classified stalled PRs. General comments linked stalling mainly to communication and coordination breakdowns, including missing interaction, delayed feedback, and unclear follow-up. Inline comments showed that inactivity does not always reflect disengagement: many PRs were blocked by technical or dependency issues, including failing checks, configuration problems, compatibility concerns, and environment mismatches. Only 39.56% of contributors later submitted another PR, and reviewer re-engagement with the same contributors was approximately 21%. PR inactivity is a socio-technical coordination problem involving communication, technical readiness, review ownership, and automation practices. We recommend type-aware triage, clearer review feedback, explicit ownership of next actions, CI blocker management, and cause-aware stale-bot interventions.

We provide explicit bounds on the number of zeros for $\zeta$ in short intervals near the edge of the critical strip. Ultimately we show that even in short intervals a positive proportion of zeros will not be near the edge of the critical strip. For example with $h>100$ and $t>10^{100}$ the proportion of zeros of $\zeta$ with imaginary part in $[t-h,t+h]$ which have real part in $[\frac{1}{16},\frac{15}{16}]$ is at least $13\%$ and a positive proportion is already obtained at $t>10^{43}$.

Balmer absorption is common among little red dots (LRDs), but absorbers at or redward of systemic are rare, occurring in only $\sim10-15$% of H$\alpha$ absorbers. In this paper, we study two such exceptional cases with deep JWST/NIRSpec spectroscopy: 15 hr of high-resolution (G395H) observations of RUBIES-EGS-49140 (z=6.68), resolving the absorption in all four transitions from H$\alpha$ through H$\delta$, and medium-resolution spectroscopy (10 hr of G235M, 2 hr of G395M) of UNCOVER-A2744-45924 (z=4.46; 1.7x magnification). Both sources are among the optically reddest and most luminous LRDs known, and both show deep, near-systemic Balmer absorption troughs. We find two systematic trends along the Balmer series: the absorption centroids become more redshifted toward higher-order transitions, while the absorbed equivalent widths decline only weakly with increasing order, far less than expected from the atomic optical-depth ratios for a single attenuating screen. Ca\,{\sc{ii}}\,K is detected in absorption in both sources, whose offset follows the H$\alpha$ trough rather than the more redshifted higher-order Balmer lines. We further report the detection of a broad base in [Ne\,{\sc{iii}}]\,$\lambda$3870, along with broad [O\,{\sc{iii}}]\,$\lambda$4364, [O\,{\sc{iii}}]\,$\lambda5008$, and He\,{\sc{i}}\,$\lambda5877,\lambda7067$, while He\,{\sc{ii}}\,$\lambda$4687 remains undetected or weak. Standard AGN photoionization models cannot reproduce the observed line ratios, whereas AGNs with high gas densities provide a consistent explanation, as also indicated by the anomalously high He\,{\sc{i}}\,$\lambda7067/\lambda5877$ ratio. A possible explanation for the relative strengths of the Balmer absorption lines could be a dense, optically thick medium whose re-emission modifies their apparent absorption strengths, while the velocity progression may arise from stratification in the absorbing gas.

Electrodermal activity (EDA) is widely used for pervasive stress monitoring because of its sensitivity to sympathetic arousal and its computational efficiency. However, research remains fragmented: studies concentrate on a small subset of available datasets, preprocessing and evaluation protocols are inconsistent, and widely adopted methodological conventions remain unvalidated. Existing reviews summarize these practices without evaluating them under a common protocol. To bridge these gaps, this paper presents a comprehensive survey and benchmark for EDA-based stress detection. We review the technical pipeline from preprocessing to modeling and compile 26 public datasets and 22 representative unimodal studies to inform future dataset selection and methodological design. Under a subject-independent protocol spanning five diverse datasets, we benchmark decomposition, filtering, and channel-input choices alongside a broad range of machine learning (ML) and deep learning (DL) models in both in-domain and cross-domain transfer settings. Our results provide the first multi-dataset validation of cvxEDA, the default decomposition method, show that feature-based ML outperforms end-to-end (E2E) DL, and identify ecological validity as a key factor in cross-domain transfer. We conclude with directions toward robust, data-efficient, and trustworthy deployment.

We perform direct numerical simulations (DNS) to investigate how cubic-lattice porous substrates influence momentum and heat transfer in turbulent channel flows. The simulations span friction Reynolds numbers from 260 to 1500, Prandtl numbers of 0.5, 1, and 2, and substrate porosities of 50%, 71%, and 87%. We show that theories developed for rough- wall turbulence can be extended to porous surfaces by replacing the roughness height with the inverse of the streamwise Forchheimer coefficient. The shifts in the mean velocity and temperature profiles follow existing fully rough momentum and thermal theories, enabling their prediction with rough-wall models. Combining these models with synthetic temperature and velocity profiles, we derive analytical formulas for the friction coefficient and Stanton number that agree with our DNS data to within 5%. The performance enhancement factor, which measures heat-transfer augmentation relative to the pressure-drop penalty at constant pumping power, is comparable to that obtained for rough surfaces. This suggests that porous substrates provide an alternative method for enhancing heat transfer in turbulent flows.

Scanned mail, uploaded PDFs, and consolidated attachments often arrive as page streams that must be split into individual documents before downstream classification, extraction, or routing. Zero-shot large language models can detect document boundaries without task-specific training, but standard Page Classification (PC) and Boundary Decision (BD) formulations resolve only one boundary per model call. We introduce Multi-Split Boundary Decision (MSBD), which predicts multiple boundaries within a page window in a single call, reducing the number of inference requests. We evaluate MSBD across multiple language models, document collections, input modalities, and window sizes. The results reveal a model- and corpus-dependent operating range in which MSBD preserves strong segmentation accuracy while substantially improving inference efficiency, followed by a sharp decline at larger windows. MSBD provided the strongest overall accuracy–efficiency trade-off, while large windows expose distinct over- and under-segmentation behavior across models. These findings show that multi-boundary prediction can make zero-shot page stream segmentation more efficient when the window size is selected for the target corpus.

Tracking recovery of walking function requires detecting meaningful gait change across rehabilitation sessions, yet objective 3D measurement remains confined to specialized motion-capture laboratories. Small camera sets and body-worn inertial sensors broaden access, but reliability varies across joints and time, allowing sensing failures to masquerade as patient change. We present GaitVista, a reliability-aware measurement layer whose lightweight gate assigns joint- and frame-specific visual contributions using camera coverage, local visual quality, cross-modal disagreement, and root-motion continuity, and exposes them for inspection. Across seven clean and degraded sensing conditions on TotalCapture, GaitVista reduces average full-body and lower-body error by 27.7% and 27.8%, attains the lowest worst-condition error among fusion methods, and reduces the gap to a joint-frame oracle from $2.76$–$5.33$~cm for condition-blind baselines to $1.11$~cm. On MoVi with image-derived keypoints, it is the only deployable fusion method to improve over both unimodal streams, reducing marker-supported error by 6.4% relative to the strongest learned fusion baseline. On TotalCapture, it improves bilateral knee-flexion waveform accuracy by 18.9%. Raw inertial measurements from five TotalCapture participants show location- and time-varying magnetic disturbance, supporting the design's reliability premise. Both benchmarks contain neurologically healthy participants in controlled settings and retain participant-specific IMU calibration; we therefore report progress toward accessible gait assessment, not validated clinical deployment.

Building on the recently developed tensor-backflow method, which accurately describes energies and stripe states in the Fermi-Hubbard model, we investigate $d$-wave pairing in the two-dimensional $t$-$t'$-$U$ Hubbard model. For the $8\times8$ lattice, we improve the wave function using symmetry projections followed by one Lanczos step, whereas for the $12\times 12$ and $16\times 16$ lattices, we apply one Lanczos step without symmetry projection. We focus primarily on the parameter regime with filling $n=0.875$, on-site repulsion $U=8$ and next-nearest-neighbor hopping $t'=-0.2t$, a regime in which enhanced $d$-wave pairing has been reported. The resulting energies are competitive with those obtained using state-of-the-art neural quantum states. For example, on a $16\times 16$ lattice with periodic boundary conditions, the tensor-backflow wave function after one Lanczos step without explicit symmetry enforcement achieves a relative energy difference of $4.4\times 10^{-3}$ from the symmetry-preserving neural quantum state result. From the real space pair correlations, for $t'=-0.2t$, we observe stronger $d$-wave pair correlations together with the characteristic relative sign between horizontal and vertical bond-pair components. Compared with $t'=0$, finite negative $t'$ increases the spectral weight of the corresponding pair-density matrix, which indicates the enhanced overall $d$-wave pair fluctuations. The leading eigenvalue is enhanced on the $8\times8$ lattice for both with and without symmetry projections. For larger lattices such as $12\times 12$ and $16\times 16$, the leading pair-density matrix spectral weight is distributed among several nearly degenerate eigenmodes rather than being concentrated in a single dominant mode. This indicates enhanced and more broadly distributed $d$-wave pair fluctuations, but no single eigenvalue exhibits the extensive scaling required for ODLRO.

Sub-fractional Brownian motion is self-similar and long-range dependent but has no stationary increments, so the increment covariance is not Toeplitz and no spectral density is available. We show that a complete finite-sample likelihood theory survives nonetheless. The model is a linear trend observed at $N$ equidistant times through a superposition of $m$ independent sub-fractional Brownian motions with known Hurst indices and a common scale. Nondegeneracy follows from realising the process as the even part of a two-sided fractional Brownian motion, and the maximum likelihood estimators of the trend and of the scale are explicit. The statistics on which inference rests are pivotal, their laws depending on the sample size alone, so intervals and tests of exact level are available at every $N\ge2$, together with complete sufficiency, minimum variance unbiasedness and attainment of the Cramér–Rao bound. An explicit variance bound gives strong consistency and asymptotic normality, and simulations confirm the exact coverage and the predicted effect of a misspecified Hurst vector.

A joint guidance and robust observer-based output-feedback synthesis framework is developed for discrete-time nonlinear systems subject to bounded process disturbances and measurement noise. The method co-designs a reference trajectory, time-varying Luenberger observer and observer-based linear output-feedback gains, and coupled ellipsoidal funnels that certify conditioned-invariant bounds on tracking and estimation errors. Nonlinear residuals are bounded by incremental quadratic constraints (QCs) parameterized by local Lipschitz constants, estimated via sampling approaches. The coupled stability conditions are formulated as linear matrix inequality (LMI) surrogates and solved using sequential convex programming with proximal regularization (prox-SCP) via semidefinite programming (SDP). Simulations of a unicycle model under persistent sensory noise and system disturbances demonstrate robust tracking and funnel-certified constraint satisfaction with respect to the jointly designed reference trajectory.

The Chinese remainder theorem (CRT) is highly sensitive to remainder errors. Robust CRT addresses this problem by recovering the folding integers correctly from erroneous remainders, so that the final reconstruction error is bounded by the remainder error level. The existing dynamic range–robustness tradeoff results mainly assume a uniform remainder error bound, where all remainders share one scalar error bound. This paper studies the exact tradeoff when the remainder error bound is a vector in the sense that different remainders may have different error bounds. For any fixed candidate range length, which specifies the integer interval over which an unknown integer is to be determined from its erroneous remainders, and any fixed remainder error bound vector, we derive a scalar necessary and sufficient condition for robust determination. Based on this condition, we characterize the largest admissible range length, called the dynamic range, for a given remainder error bound vector. Conversely, for a given range length, we characterize all remainder error bound vectors under which remainder errors can be tolerated. We also give an exact folding integer vector decoding algorithm. We show that the proposed results reduce to the classical CRT in the error-free case and coincide with the known result when all remainder error bounds are equal. Numerical results show that the dynamic range depends on the full remainder error bound vector. They also show that the remainders are coupled in the robustness, which cannot be captured by a single uniform error bound.

Process model forecasting (PMF) aims to predict the process model that will characterize a future period, thereby providing a process-level view of how behavior is expected to evolve. Existing PMF methods, however, forecast directly-follows graphs, which cannot explicitly represent concurrency. We extend PMF to causal nets by forecasting time series of relation and binding counts and using these forecasts to reconstruct future process models with AND/XOR semantics. To evaluate the resulting models, we introduce a protocol that accounts for partial traces and constructs the workflow nets required for conformance checking. Experiments on four event logs show that the forecasted models achieve conformance levels close to those of models re-mined from observations in the corresponding future windows. They also outperform static discovery baselines, which retain high precision on the structurally stable log but exhibit substantial precision losses on the other three logs. Filtering infrequent bindings improves most conformance metrics, although it also removes much of the concurrent behavior captured by the models.

We calculate the $\theta$ dependence of the vacuum energy in an infinite family of cousins of QCD, where the vacuum structure can be analyzed exactly. These theories are $\mathcal{N}=1$ EFTs that descend from $\mathcal{N}=2$ $SU(N_c)$ gauge theory with $N_f$ flavors of fundamental hypermultiplets, that are then deformed by anomaly mediated supersymmetry breaking (AMSB). We reproduce results found in previous work that used the full $\mathcal{N}=2$ machinery but via a much simpler $\mathcal{N}=1$ method. Our method easily extends to infinitely many more theories for which the full $\mathcal{N}=2$ approach becomes intractable. We identify values of $\theta$ for which the theory undergoes a phase transition, and find that they can occur at both fractional and irrational values of $\theta/\pi$.

We examine log-optimal portfolio allocation when the long-run price of an asset follows a power-law trajectory, $P(t)=At^\alpha$, and its instantaneous return variance decays as $\sigma^2(t)=\sigma_0^2 t^{-2\gamma}$. Under a zero risk-free-rate benchmark, the continuous-time Kelly fraction scales as $K^\ast(t)=(\alpha/\sigma_0^2)t^{2\gamma-1}$. Exact temporal invariance therefore occurs when $\gamma=1/2$, whereas deviations from this value produce systematic age dependence in the allocation. We propose a scaling hypothesis connecting growth in network participation, effective market liquidity, and declining volatility. Under a specified set of scaling assumptions, this model predicts the benchmark exponent $\gamma=1/2$. Using historical daily Bitcoin prices, we estimate the power-law price exponent and examine the sensitivity of the volatility exponent to the length of the rolling window. For windows of four to nine years, the estimated volatility exponents have an arithmetic mean of 0.53 and a cross-window standard deviation of approximately 0.03. Because these estimates are obtained from overlapping observations and the same underlying price history, this spread is interpreted as a measure of model sensitivity rather than a formal confidence interval. Finally, we show that a time-dependent multiplicative contribution to return variance generally breaks exact Kelly invariance. We illustrate this result using a scenario in which Bitcoin transaction-fee variability affects the effective variance process. The results identify the conditions under which log-optimal allocation can remain stable under non-stationary power-law asset dynamics and clarify the assumptions required when applying this result to Bitcoin.

Humanoid robots can acquire complex skills by imitating kinematic humanoid motion references, yet reliable references for contact-rich interactions remain difficult to obtain: motion capture deteriorates under occlusion and close physical contact, while retargeting introduces additional contact and geometric inconsistencies. We present HIGenNTO, a framework that synthesizes humanoid-scene interaction motion references by optimizing the initial noise of a pretrained text-conditioned motion model under sparse spatiotemporal and scene constraints. The same formulation satisfies desired contacts, avoids collisions, and maintains stable support while retaining the prior's realism and temporal coherence, generating interaction motions from scratch and composing long-horizon behaviors stage-wise. Across robot-environment and robot-object tasks, HIGenNTO produces motions that can be executed by tracking policies in simulation and used to train depth-conditioned visuomotor policies operating solely from onboard sensing. We deploy these policies on a Unitree G1 across four contact-rich tasks. Finally, the task specifications themselves can be written by a coding agent, which proposes interaction tasks and compiles them into prompt, constraint, and scene programs, authoring three of our eight evaluated tasks and four further behaviors. Together, these results establish a scalable path from high-level task descriptions to physically executable humanoid interactions.

Modern code review is central to software quality, but its effectiveness depends on reviewer expertise, change characteristics, and how tools present changes. Most platforms display modified files alphabetically by default, although our prior work shows that developers find this ordering cognitively misaligned with how they understand multi-file pull requests. Whether these ordering-related attention patterns affect outcomes at scale remains unclear. We present a large-scale study of file ordering and review effectiveness, mining 330,343 multi-file pull requests and 756,814 file instances from 182 GitHub projects in five programming languages. We examine whether file position is associated with later bug-fixing changes, whether pull request size moderates this relationship, and whether reviewer attention, proxied by comments, aligns with latent bug outcomes. Results show statistically significant but modest associations among file position, pull request size, review activity, and latent bug likelihood. Latent bug rates rise from 56.7% at position 1 to 61.5% at position 30. Pull request size has a non-linear relationship with latent bug likelihood: pull requests of about ten files have the lowest risk, while very small and very large ones have elevated rates. Hurdle models show that attention is diluted as pull request size grows: each additional modified file reduces the odds of receiving any review comment by about 8.7%. These findings reveal an attention-effectiveness gap: visible review activity does not necessarily prevent defects. Alphabetical ordering is therefore not a neutral interface default, but a structural feature shaping attention allocation, review coverage, and confidence in review outcomes. We propose context-aware file ordering, dependency-aware grouping, risk-aware prioritization, and per-file coverage indicators to make review attention more visible and actionable.

Insertion is a fundamental operation in robotic construction assembly, where variations in material properties and assembly conditions make it difficult to select contact forces that complete the task without exceeding the assembly's capacity. Although construction documents encode engineering knowledge about materials and their conditions, translating this knowledge into load limits for a specific assembly remains difficult. This paper presents SAGE (Source-grounded Assembly Gating and Execution), a system that converts documented material evidence into capacity estimates for robotic insertion. SAGE restricts a large language model (LLM) to extracting tensile and compressive strengths from retrieved passages and tables and records their sources. A response model then interpolates offline finite element (FE) solutions to convert these strengths and the assembly conditions into axial load capacity. For fits with positive clearance, the estimated capacity sets the policy's axial force limit; for interference fits, it is compared with measured support demand to determine admission. On the primary benchmark, SAGE reduces mean capacity error from 80.65% for direct LLM estimates based on the same evidence to 10.74%. Without refitting, the mean error remains 8.00% on 16 additional geometries. Under the assigned support release model, SAGE correctly classifies 59 of 62 scored simulation runs, with only conservative errors. In recorded xArm6 demonstrations, SAGE takes material documents as input and completes physical insertion in 9 of 13 trials. These results show that assigning document interpretation to the LLM and force calculation to an explicit mechanical model produces accurate capacity estimates and traceable insertion decisions.

Projectile interception is a challenging dynamic manipulation problem. Intercepting a thrown object with a robot arm requires reaching a point on the object's path as the object passes through it. Slicing also fixes the blade's velocity and orientation at contact. The goal is therefore a subset of the states of the robot and arrival times that moves as the object falls, and the arm must reach it within its actuator limits in milliseconds. We present FRUITNINJA, an anytime sampling-based planner that grows a tree on the GPU in batches toward the interception manifold. Each edge is an exact cubic whose travel time is found by a parallel search against the arm's dynamics, so every edge satisfies the actuator limits. Plans are ranked by a risk-aware objective over the uncertainty in the object's position and the arm's arrival time. We evaluate on a Franka Research 3 against six baselines in a calibrated real-time simulator, where FRUITNINJA cuts 96.7% of tosses in the open and 68.3% among five obstacles, versus the best baseline's 68.3% and 35.0% respectively.

We argue here that the current dominant practice in LLM human simulation: prompting instruction-tuned assistant language models to role-play personas, is inaccurate and produces stereotyped predictions (lacking natural diversity). It has previously been shown that LLMs can be bound to personas using naturalistic, freetext dialog avoiding stereotyping. Here we show that binding can also be achieved using short, individual samples of dialog from specific people. Demographics can be added later without negative effects by simply querying the model. We use the term Persona Mixture Models (PMMs) for well-calibrated human models, currently realized as pretrained base models. We show that PMMs produce more accurate predictions than instruction-tuned models and retain more of the lexical, semantic, and pragmatic diversity found in human dialog. We measure realism and diversity of LLMs simulating human interlocutors across a diverse set of corpora spanning open-domain text, human-AI chat, and task-oriented dialogue between human speakers. However, base pretrained models can produce out-of-domain dialog and may lose some of the human's internal state over long contexts. We propose and explore tandem models which combine a pre-trained model with an instruction-tuned supervisor. Tandem models achieve the best overall accuracy and diversity in our experiments.

Vision-Tactile-Language-Action (VTLA) models have demonstrated clear advantages over Vision-Language-Action (VLA) models in contact-rich manipulation. However, developing VTLA models is severely constrained by the massive amounts of vision-tactile data and computational resources required. To address this bottleneck, we propose VT-Bridge, a lightweight residual adaptation strategy that bridges pretrained foundation VLAs to VTLAs. Rather than training a VTLA model from scratch or modifying the original architecture of a pretrained VLA, VT-Bridge employs an identical lightweight residual-adapter architecture across VLA backbones and uses backbone-specific weights to refine actions at the robot execution frequency. This design substantially lowers the data and training barriers. Specifically, it requires up to 50 vision-tactile demonstrations per task to fine-tune a VLA backbone and train a 0.98M-parameter residual adapter. Experiments with three representative VLA backbones ($\pi_0$, $\pi_{0.5}$, and SmolVLA) across four contact-rich manipulation tasks further demonstrate its consistent effectiveness across VLA architectures. On average, VT-Bridge raises the task completion rate from 11.7% with task-level VLA fine-tuning alone to 62.9%. Together, these findings demonstrate the broad applicability, effectiveness, and accessibility of VT-Bridge for contact-rich manipulation. The project page is available at https://hoxnocha.github.io/vt-bridge-web/.

In the Directed Steiner Network (DSN) problem we are given a directed graph and a set of demands $(s_i,t_i)$, and asked to find a cheap subgraph connecting each terminal pair. In its online version, the demands arrive online and must be served by buying edges irrevocably. DSN is a fundamental hard problem in network design, heavily studied in both the offline and the online setting. Offline, it has a superpolylogarithmic hardness of approximation. However, offline hardness says nothing about online algorithms, which are computationally unrestricted. It has been an open question whether uncertainty itself (needing to commit to a solution without knowing future demands) rules out polylogarithmic-competitive online algorithms. In this work, we show the first such unconditional, information-theoretic hardness. Namely, we give an $\exp\!\bigl(\Omega(\sqrt{\log n})\bigr)$ bound on the competitive ratio, which holds even for randomized algorithms against an oblivious adversary, and on unit-cost DAGs. Our proof uses a novel connection between online network design and algebraic coding theory. We encode requests using a hidden low-degree polynomial, whose past evaluations reveal nothing about future ones. We then use list-recovery bounds to show that an algorithm cannot make cheaply reusable decisions without knowing those future evaluations.

In the realm of material analysis, identifying different material phases is of key importance. Artificial intelligence in the form of neural networks provides a very fast and, once trained, computationally inexpensive method for analysing large amounts of image data, like the sets of diffraction patterns generated during four-dimensional scanning transmission electron microscopy (4DSTEM) dataset acquisition. In this work, we train multiple network architectures on images of this type to distinguish between the phases of LiNiO2 and NiO, an important and challenging distinction in the lithium-ion battery community, since the formation of NiO limits battery capacity. We test both classical convolutional neural networks (CNNs) and different forms of vision transformers (ViTs). Our networks are trained on synthetic images and tested on experimentally recorded diffraction patterns. Additionally, we also investigate the decision-making of our networks using the GradCAM method. We test our networks on both synthetic as well as experimental diffraction patterns. Our networks exhibit very robust results, especially when dealing with highly varying data, an area where traditional template-matching methods typically struggle.

Summarization ships in countless production systems, making model selection a routine decision that depends on measuring summary quality. Existing metrics struggle to support this: ROUGE captures only surface overlap, while LLM-as-judge scores saturate to near-identical values that fail to rank models effectively. We observe this saturation across three public datasets, two proprietary datasets, and multilingual settings. Motivated by this, we introduce Semantic Scaffold, an evaluation framework that extracts a hierarchical representation of facts, questions, and entity attributes from a source text, labeling each as a main point or supporting detail, and reusing this structure as a fixed reference for scoring summaries. From this representation, we derive three diagnostic metrics: Fact Preservation Score (FPS), Question Preservation Score (QPS), and Entity Preservation Score (EPS), designed to reward the preservation of essential information while penalizing detail overload, and position them as interpretable diagnostics that remain informative where holistic axes collapse. Finally, we analyze four recurring failure modes of ROUGE and LLM-as-judge scores, demonstrating that scaffold-based evaluation remains informative where conventional metrics collapse.

In this paper we study an equivalence between the existence of a Sobolev inequality without gain, \[\|\varphi\|_{L^p(v,\Omega)} \leq S(p,1) \| \sqrt{Q}\nabla \varphi\|_{L^p(\Omega)},\] that holds for smooth functions of compact support and the existence of a degenerate weak solution $(u,\nabla u)\in QH^{1,p}_0(v,\Omega)$ to a Dirichlet problem for the $p$-Laplacian with a zero order term: \begin{equation*} -v^{-1}\text{Div}(|\sqrt{Q}\nabla u|^{p-2}Q\nabla u)+F|u|^{p-2}u = |f|^{p-2}f - v^{-1}\text{Div}(v|g|^{p-2}g\mathbf{t}),\; x \in \Omega, \quad \text{and} \quad u =0, \; x \in \partial \Omega, \end{equation*} More precisely, we use the Sobolev inequality to prove the existence of a degenerate weak solution to this equation and then use the existence of such a solution to produce a Sobolev inequality. Moreover, we show that solutions are unique.

The increasing reliance of autonomous AI agents on external and distributed knowledge sources introduces a fundamental challenge for decentralized information marketplaces: retrieval agents must evaluate the quality and relevance of data before purchase, while data providers must avoid revealing valuable information prior to guaranteed compensation. This paradox becomes particularly critical in trustless multi-agent environments, where no centralized intermediary can enforce fairness between parties. In this paper, we propose a fairly compensated protocol for distributed information retrieval and augmentation in autonomous agent networks. Our framework enables retrieval agents to securely evaluate and rank candidate documents without learning their plaintext contents, while ensuring that data providers are compensated only when valid information is successfully delivered. We further analyze the security properties of the protocol against malicious adversaries and evaluate its practical feasibility through implementations. Experimental results demonstrate that the proposed design is practical with current cryptographic infrastructures while preserving confidentiality, correctness, integrity, and fairness. We believe such mechanisms provide an important cryptographic foundation for trustworthy and economically sustainable decentralized knowledge marketplaces for future AI agent ecosystems.

LLM agents are beginning to write the rules they live under: policies, contracts and code that other agents review and approve. What happens when those rules must decide who is sacrificed from a group? We built GovSim-SelfGovern, a commons in which agents legislate in executable Python, test each law in a sandbox, vote on it, and live with the result. When the fishery is abundant, self-rule helps, raising the chance that the community survives without anyone starving from 52.5% to 75.0%. When it cannot feed everyone though, adding governance is not enough to save the commons. We trace much of this to the vote: agents readily draft laws that remove members, but then frequently vote them down, thereby dooming the commons. In a controlled replay, the same exile law passes 8% of the time when the sandbox preview names its victim at once and 55% when the removal is deferred until a future crisis. Showing voters the full source code does not close the gap, though most describe the removal in their own words. Nor is sacrifice a fixed trait. A single sentence of framing can swing self-removal from 2.5% to 91%, and due process deters agents more than considering the potential death of a peer. In contrast, give the community a treasury and the question of who must go almost disappears. Our work shows that what our collectives legislate is decided more by the form in which a choice is put to them as compared to their ethical beliefs.

Large language model (LLM) judges provide a flexible and scalable method for evaluating model and agent outputs, but their verdicts can be sensitive to incidental changes in the evaluated response, judge instructions, and scoring rubric. Existing systems examine important subsets of these failure modes, but auditing a configured judge requires testing both the judge instrument and the items it evaluates. We introduce MAWILE, a developer-facing workbench for auditing judge sensitivity across four surfaces: the judge prompt, judge rubric, target-system input, and target-system output. Given a user-supplied judge and representative evaluation items, MAWILE constructs and validates controlled perturbations, re-executes the judge, and localizes the resulting sensitivity. Each perturbation declares whether the verdict should remain invariant or change in a specified direction, allowing the same system to measure both robustness to irrelevant variations and sensitivity to meaningful changes. MAWILE audits binary, ordinal, and pairwise judges without requiring gold labels. The code for this tool is available at: github.com/megagonlabs/mawile-judge.

We compute the cohomology of the projective special linear Lie superalgebra $\mathfrak{psl}(n|n)$ over a field of characteristic zero in trivial and in adjoint coefficients in full detail through its associated Hochschild–Serre spectral sequences. Applications in physics exist for its real form, the projective special unitary Lie superalgebra $\mathfrak{psu}(n|n)$ — in particular, $\mathfrak{psu}(2,2|4)$ for AdS/CFT. We also find the cohomology of $\mathfrak{sl}(m|n)$ in trivial coefficients along the way and we point out a discrepancy with a previous statement in the literature. We present the explicit cohomology groups of low degree for $n\leqslant4$.