Papers

We study how a minimal length can affect the chiral vortical effect, in which rotation produces a current in matter made of chiral fermions. We focus on an isotropic modification of the momentum-space density-of-state (DoS) and determine the vortical conductivity from the static, long-wavelength correlation between the current and momentum density. The analysis is carried out first for a single right-handed Weyl fermion and is then extended to general massless chiral fermion systems with mutually commuting conserved charges. We find that, for a broad class of analytic isotropic DoS modifications, the familiar leading contributions from chemical potential and temperature remain unchanged at first order within the DoS sector. Minimal-length effects instead appear through higher-order thermal and chemical-potential contributions. Some of these corrections contain the same chiral charge combinations that appear in gauge and mixed gauge-gravitational anomalies, while another contains an independent higher-order charge combination. Their coefficients depend on the form of the DoS modification and are therefore not fixed by the usual four-dimensional anomaly structure. For the Kempf-Mangano-Mann deformation, the exact result shows that a positive minimal-length parameter always reduces the DoS contribution to the vortical conductivity. We also find that a Planck-suppressed correction is extremely small under standard heavy-ion chemical freeze-out conditions. These results show that isotropic minimal-length modifications can change higher-order contributions to chiral vortical transport while leaving the familiar leading anomaly-related response unchanged within the DoS sector considered here.

This paper is concerned with a natural variant of the contact process that places a social cluster of individuals, rather than a single individual, at each lattice point, and distinguishes between symptomatic and asymptomatic individuals. In particular, the epidemic model depends on four infection parameters: the external (inter-cluster) and internal (intra-cluster) infection rates for both symptomatic and asymptomatic individuals. Newly infected individuals begin as asymptomatic before transitioning to the symptomatic state at a spontaneous mutation rate, and infected individuals recover at a baseline rate of one. We study the phase structure of the deterministic nonspatial mean-field model when each cluster reduces to a pair of individuals, and prove that the stochastic spatial model exhibits the same phase transitions with respect to the four infection rates for all cluster sizes. We also prove the existence of a phase transition driven by the cluster size itself for the spatial model.

This work investigates a class of exactly solvable quantum systems described by a generalized Dunkl–Schrödinger equation with energy-dependent interactions. A two-parameter extension of the Dunkl differential operator is introduced, incorporating both the conventional reflection contribution and an additional derivative-reflection coupling. The resulting formalism leads to a modified Heisenberg algebra, from which the continuity equation, conserved probability current, and generalized normalization condition are derived for energy-dependent potentials. As an application, an energy-dependent harmonic oscillator is studied in detail. Exact analytical expressions for the eigenfunctions and energy spectrum are obtained in terms of associated Laguerre polynomials. The combined effects of the Wigner parameter, the deformation parameter, and the energy-dependence parameter are shown to produce parity splitting, nonlinear energy spectra, and nonuniform level spacing while preserving exact solvability. The thermodynamic properties of the model are then examined through the canonical partition function using the Euler–Maclaurin summation formula. Analytical expressions for the partition function, internal energy, entropy, and specific heat are derived and analyzed numerically. The results demonstrate that the generalized Dunkl deformation and energy-dependent interaction provide effective control over both the spectral and thermal characteristics of the system, extending the class of exactly solvable quantum models with reflection symmetry.

We study the head-on collision of two plane neutrino waves in the coupled Einstein-Weyl system, treating the massless Weyl spinor as an odd Grassmann valued field. We show that the mixed components of the Einstein equations force the two counter propagating waves to be built on a single Grassmann mode, whose bilinear $\varepsilon=\eta\eta^{*}$ is nilpotent. With this structure the full coupled system is solved exactly in the interaction region for arbitrary amplitude and phase profiles, and the metric and spinor fields are matched to the single wave regions by the non-impulsive O'Brien-Synge junction conditions. The metric functions contain no terms coupling the two waves; the Weyl scalar $\Psi_2$ vanishes, and $\Psi_0$, $\Psi_4$ are proportional to the neutrino energy fluxes. The waves therefore traverse each other without mutual scattering; we trace this to the nilpotency of the single mode bilinear. We discuss the limits of this result and indicate where cross couplings can arise.

Efficiently transferring kinetic energy from a slow, high-force driver to a light payload, with a near-uniform output force, requires a transmission whose mechanical advantage rises along a steeply convex profile. Continuous-contour mechanisms (variable-radius drums, cams) become prohibitive at scale through diameter span, inertia and rim burst. The RopeComb bypasses these limits by synthesising profiles ($G = 0 \to 100+$) from discrete constant-radius sheaves deflecting tension members into staggered spans. Each fold's ratio vanishes at first contact, enabling snatch-free engagement at 10 m/s without clutches or dampers. Because the array ratio is closed-form, bounded least squares matches uniform-force targets to within 0.6-1.5% across mass ratios of 100:1 to 10,000:1 using 8-13 members. To carry large carriage reactions without the overturning moment a single array imposes, two independent arrays flank a central guide and drive a partitioned fixed-ratio stage ($n_1 + n_2 = k$). Sizing array ratios and member counts in the fall-count ratio $n_2 : n_1$ cancels 97.7% of the single-array reaction and 95% of its overturning moment in simulated 1,000:1 launches at $k=7$, reducing compliant peak-to-mean payload force from 1.40 to 1.13. All results are numerical.

Transformer models such as BERT and Vision Transformer~(ViT) achieve strong performance via densely parameterized attention backbones. However, the least significant bits~(LSBs) of their 32-bit floating-point weights can be abused as covert channels to conceal malicious payloads, posing a serious threat to the AI model supply chain. We propose \GS (\GSabbr), a lightweight, post-training, zero-data sanitization method that completely replaces the declared mantissa-LSB channel with a Gray-code-guided low-transition sequence. Complete payload-independent overwrite, whether keyed or public, makes the sanitized target bits independent of the embedded payload and gives that declared channel zero capacity. Gray coding supplies overwrite structure, while a keyed per-tensor phase supplies pattern diversity. Benchmarked against seven post-training defenses on four Transformer model presets and two real-world malware payloads, \GSabbr maintains sub-$1\%$ accuracy impact and achieves $49.96\pm0.66$ percentage-point Recovery Reduction (RR) under five implemented attacker variants. Because pre-defense recovery is effectively $100\%$, RR near 50 percentage points corresponds to post-sanitization bit accuracy at binary chance. Its main empirical advantage is stable near-chance sanitization with substantially smaller weight-distribution shift than the evaluated near-chance baselines PatternMask (PM) and Post-Training Quantization (PTQ).

Efficient long-video understanding with vision-language models (VLMs) is often framed as selecting informative frames or visual tokens at a fixed native resolution. We show that per-frame resolution can instead be traded for denser temporal coverage, while front-end decoding latency depends on the size of the candidate pool rather than the final token budget. An empirical study across multiple VLMs and long-video benchmarks yields three findings: dense low-resolution sampling outperforms sparse native-resolution sampling at matched token budgets; resolution-sensitive tasks benefit from selected high-resolution frames; and front-end decoding dominates wall time for hour-long videos. Motivated by these findings, we introduce LoHi, a training-free, single-pass framework that combines a dense low-resolution video stream with sparse high-resolution image frames through the VLM's native video and image pathways. LoHi-Anchor selects high-resolution frames using codec I-frame metadata, while LoHi-SemDiv uses query relevance and visual diversity over CLIP features. Across three long-video benchmarks, LoHi improves average accuracy by 10.6 percentage points over the native-resolution baseline at a matched token budget and by 5.2 percentage points over the strongest prior efficiency method. It also reduces front-end decoding latency by up to 7x on hour-long videos. Project page: https://sixundong.com/projects/lohi

In 1978, Amitsur and Small asked whether, for every division ring $D$, maximal left ideals of $D[x_1,\ldots,x_n]$ contract to maximal left ideals in smaller polynomial subrings. Chapman and the author showed that the answer is negative in general and called $D$ an Amitsur-Small ring when this contraction property always holds. Earlier work showed that Hamilton's real quaternion algebra is Amitsur-Small, that division algebras of degree three are not Amitsur-Small, and that degree-two examples are restricted to a specific quaternionic form; subsequent work excluded cyclic division algebras of odd prime degree. We give a complete resolution in the centrally finite case. If $D$ has finite dimension over its center $F$, then $D$ is an Amitsur-Small ring if and only if either $D=F$, or $F$ is real closed and $D$ is the Hamilton quaternion algebra $(-1,-1)_F$. In every other noncommutative centrally finite case, failure already occurs in two variables: there is a maximal left ideal $M\subseteq D[x,y]$ such that $M\cap D[x]$ is not maximal in $D[x]$.

Output-only safety monitoring sees only the end of a model's computation, yet the model computes its answer before emitting it: what it is internally poised to say is safety-critical. Jacobian-space (J-space) readouts, linear maps from hidden states to the output vocabulary via the model's input-output Jacobian, have been proposed as a window on behaviorally accessible internal content, and a first safety protocol (JADR) showed that danger recognition is readable there. What remains unknown is how this accessible content relates to the latent-space safety geometry studied by representation engineering, and how safety training shapes it. We introduce two quantitative bridges: (i) a transport-amplification profile $A_\ell$, measuring how strongly a latent safety direction is carried toward output space by each layer's Jacobian, and (ii) a paired base-vs-tuned protocol that attributes accessibility to training provenance. Across four small model pairs (135M to 1.5B, three lens-fit seeds), tuned checkpoints concentrate recognition in upper-middle layers, and at 0.5B DPO installs refusal while J-space recognition drops to chance: training can widen the accessibility gap exactly where behavioral safety looks best. A deployment audit closes the loop: monitor-aware GCG-style suffixes suppress the prompt-side monitor at zero behavioral cost at every scale; only a learned monitor rung resists its own adaptive re-attack; the training-time defense fails its re-attack at every penalty weight; steering shows the amplification profile is descriptive, not causal; and continuous-prefix optimization fails where discrete search succeeds. Safety monitoring, training, and evaluation must operate on accessibility itself, not on outputs alone.

Modelling molecular absorption spectra from scalable ground-state representations is challenging because conventional approaches often require explicit excited states or time propagation. Here, we introduce a method to obtain Absorption Using Recursive Operator Response with vibronic Augmentation (Aurora), which combines neural quantum states (NQSs) with continued-fraction memory-kernel coupling theory (CF-MKCT). Aurora maps the zero-temperature Liouvillian response hierarchy onto repeated wavefunction-space applications of a shifted Hamiltonian to a dipole-prepared state, enabling direct spectral reconstruction from variational ground states. The implementation combines variational Monte Carlo, implicit-restarted-Lanczos-assisted NQS optimization, matrix-free Hamiltonian application, and adaptive recursion control. The resulting electronic spectra closely agree with full-configuration-interaction and conventional MKCT references for tractable molecules, with the restarted-Lanczos strategy improving difficult NQS cases. For electron-phonon Hamiltonians, Aurora combines Huang-Rhys mode selection with local singular-value-decomposition (SVD) phonon encoders. The H2 benchmark indicates that compression retains the dominant vibronic envelope and improves upon bare local-basis truncation, while paired calculations for N2, H2O, H2S, and NH3 yield similar principal features with and without encoding. We further obtain vibronic spectra for C6H6 and a 10-electron, 10-orbital C10H8 active space. Overall, Aurora provides a unified route from variational molecular ground states to electronic and vibronic spectra without constructing excited-state manifolds.

We study how the topology and geometry of a compact Kähler manifold relate to the zeros of its one-forms. We determine all implications among several closely related conditions. In particular, we construct a smooth projective variety for which the Aomoto complex is exact for every nonzero holomorphic one-form and every semisimple local system, although every closed real one-form, and hence every holomorphic one-form, has a zero.

Long-context decoding is limited by memory bandwidth, because every output token reads the KV cache of every layer. Sparse decoding reduces this cost by reading only part of the KV cache. We observe that the number of pages a query needs varies widely across KV heads, layers and steps. Fixed budgets are simple, but they are sized for demanding cases and tuned per workload; adaptive budgets follow this variation more flexibly, but existing designs pay for it with extra selection cost or training. At the kernel level, FlashAttention-3 (FA3) and FlashInfer are designed for rows of similar length: with page lists whose length differs per KV head, they either pad the lists (forfeiting much of the sparse saving), leave thread blocks unbalanced, or rely on a host-side plan that runs outside the CUDA graph. We propose MOIRA, a training-free sparse decode path in vLLM whose budget adapts per KV head and per layer. For every request, layer, KV head and step, a coverage rule keeps the smallest set of pages whose estimated attention mass reaches a fraction $\gamma$. A new kernel, self-planning attention, lets each thread block derive its own share of the work from the list lengths, so the whole decode step stays inside the CUDA graph. On an H200, at RULER's 128k context, MOIRA with $\gamma=0.99$ matches dense accuracy while reading about 30% of the pages and reduces the time per output token (TPOT) by 2.2-2.5$\times$ relative to dense FA3; with $\gamma=0.98$ it reduces TPOT by 2.7$\times$ and stays within the noise of dense. Under high serving load it raises throughput by up to 51%. These results suggest that a budget adapted per head and layer, paired with a kernel that keeps such budgets inside the CUDA graph, makes sparse decoding both flexible and fast.

Let $C \subset \Pbb^2_k$ be a smooth projective plane curve of degree $d \ge 5$ over a number field $k$. We prove that any cubic arithmetic or geometric progression sequence on $C$ is finite. The proof rests on a prime-$p$ tower of cyclic Kummer covers controlled by the divisor of a linear-form quotient on $C$, combined with the Kadets–Vogt classification of curves with infinitely many cubic points, the Abramovich–Harris gonality descent, a Castelnuovo–Severi factorization for the elliptic maps supplied by Kadets–Vogt independently at consecutive levels of the tower, and a deck-transformation obstruction. We also obtain a conditional quartic result under a cyclotomic non-splitting hypothesis. The case $d = 4$ with $C(k) \neq \emptyset$ admits infinite cubic progressions, so no analogue of the main theorem can hold there.

We prove that prediction for the synchronous two-dimensional majority rule is $\mathrm{P}$-complete under logspace many-one reductions, resolving a problem open for almost three decades. In 1997, Moore established $\mathrm{P}$-completeness in dimension three and higher and conjectured that the two-dimensional case admits an efficient parallel algorithm. We consider an $n\times n$ torus on which each cell follows the majority of its four nearest neighbors and retains its current state in a tie. Given an explicitly specified initial configuration and a time $T$, prediction asks whether a designated cell is in state $+1$ at time $T$. The central challenge is to make independent information streams cross in the plane under a homogeneous, monotone, diffusive local rule. We overcome this obstacle through a temporal encoding of Boolean values: both values generate activity, but are distinguished by signal arrival times. This encoding yields a crossover that preserves both values and composes with wires, duplication, and AND and OR gates to simulate arbitrary monotone Boolean circuits. Thus a local rule that favors agreement can nevertheless transport, combine, and cross independent information in two dimensions. We also prove $\mathrm{P}$-completeness for deciding whether a designated cell ever reaches $+1$, without a prescribed time horizon. The prediction result extends to every uniform symmetric signed majority rule on the same neighborhood, including the minority rule.

Machine unlearning is seen as a promising approach to enable users to exercise the "right to erasure" in the context of AI models. We ask how users might influence the behavior of models when exercising this right. We define two types of behaviors that users might adopt when requesting the deletion of their data: adaptivity and collectivity. Drawing connections between the goals of users in this context and results in stochastic optimization, we demonstrate theoretical gaps between the potential effects of groups of users who do and do not display these behaviors. We then show how techniques from data valuation might be used to design deletion requesters that can significantly alter model behavior in realistic settings. In experiments on computer vision tasks, we demonstrate the differential effects of different models of user behavior and attempt to isolate the impacts of adaptivity and collectivity.

We study the rare decay $\Lambda_b \rightarrow \Lambda(\rightarrow p\pi^-)\nu\bar{\nu}$ of polarized $\Lambda_b$ baryons produced at the $Z$ pole. We find that the differential decay distribution is characterized by five independent angular observables, four of which require initial $\Lambda_b$ polarization. While standard observables in mesonic $b \rightarrow s \nu \bar\nu$ decays ($B \rightarrow K^{(*)} \nu \bar{\nu}$, $B_s \rightarrow \phi \nu \bar{\nu}$) probe only two combinations of short-distance Wilson coefficients, $|C_L + C_R|$ and $|C_L - C_R|$, the angular observables in $\Lambda_b \rightarrow \Lambda(\rightarrow p\pi^-)\nu\bar{\nu}$ provide access to the orthogonal combinations $|C_R|^2 - |C_L|^2$ and $\text{Im}(C_L C_R^*)$. In particular, we identify a $T$-odd CP asymmetry, $A_5$, that directly probes CP-violating phases in the short-distance effective Hamiltonian. We present Standard Model predictions for the branching ratio and all angular observables using lattice QCD form factors, and evaluate the projected new physics sensitivity at the FCC-ee. Our results demonstrate that baryonic angular observables are essential for resolving parameter degeneracies and for completely characterizing the chiral and CP structure of possible new physics in $b \rightarrow s \nu \bar{\nu}$ transitions.

This study demonstrates an overall view of recent quantum mechanical based modeling effort using empirical tight-binding method to expand the ability of calculation of electronic band structure for strained-layer superlattices composed of ternary InAs/InAs1-xSbx and InAs/InAs1-xSbx/AlAs1-xSbx materials. Using a modified sp3s* empirical tight-binding method and building the superlattice Hamiltonian, we demonstrate that combining the virtual crystal approximation with a bowing parameter for the s-on-site energy accurately predicts ternary superlattice behavior. We provide a comprehensive overview of our numerical simulations, which model diverse superlattice configurations ranging from long wavelength infrared range to wide-bandgap barrier architectures. Notably, the simulated energy levels strongly agree with experimental bandgap measurements of the ternary superlattices and their structural variants.

Machine Learning (ML) methods promise a fast solution process for Optimal Power Flow (OPF). While inconsistent test cases, implementations, and evaluation metrics across existing studies make it challenging to determine which algorithmic advances are most critical for real-world deployment. To this end, we propose ML-OPF-Bench, a unified benchmark for AC- and DC-OPF that evaluates representative ML algorithms under a consistent pipeline, stress-tests them across system sizes, distribution shifts, and resource budgets, and ranks them with a multi-objective framework. We find that prediction accuracy alone is not a reliable indicator of operational feasibility. Under heavily loaded, congested conditions, even the strongest in-distribution performers lose their advantage, while feasibility is maintained largely by post-processing that enforces the target constraints rather than by the underlying pure ML predictor. Data scaling shows that prediction accuracy and constraint violations follow different trajectories, whereas compute scaling shows that returns diminish and that larger models do not consistently perform better. These results expose critical trade-offs among ML methods' speed, accuracy, and feasibility, and offer practical guidance for future ML-OPF design. We open-source the benchmark as an extensible Python package for integrating new learning-based OPF algorithms and evaluating them under the same standard as the existing baselines.

The minimum fault weight of a logical readout depends jointly on data-error persistence, check redundancy and time-boundary constraints. For any binary check matrix $H$ and nonzero parity support $w$, we prove that a repeated-readout model with a zero-syndrome endpoint has exact distance $\min_{w\cdot y=1}[\operatorname{wt}(y)+q\operatorname{wt}(Hy)]$ for at least $q$ wrong readouts. This reduces the search over fault histories to one weighted spatial minimum and gives the minimum repetition count for each adjacent-column check in square Bacon–Shor and generalized Shor readouts. Combining those checks with redundant Gauss-check pair products yields an exact joint distance and an outcome-count optimum within the specified gauged-readout family. At target distance $m$, one round uses $m^2$ outcomes, compared with $m(2m-1)$ for the repeated baseline in the same phenomenological model. For a one-epoch bare-ancilla CNOT implementation, we prove a schedule-dependent distance upper bound by constructing an accepted wrong readout from one gate fault and outcome flips. The bound identifies necessary timing conditions for retaining the target distance and gives matching upper bounds of three for the best tested schedules at both $m=3$ and $m=5$.

We determine the torsion subgroup of $H_{m-1}(\mathbb{B}_mK_{m+1,m+r-1};\mathbb{Z})$ for $m\ge2$ and $r\ge0$: top homology with arbitrary coefficients is the kernel of an unsigned subset-inclusion matrix, and its integral diagonal form determines all primary summands. Every finite order occurs, with explicit representatives. Generalized theta classes span an embedded copy of the cokernel of the inclusion matrix, containing all torsion; for $r\ge m$ they generate the torsion, each of order $\operatorname{lcm}(1,\ldots,m)$. For every prime power $q$ and $m\ge q$, the graph $K_{m+1,m+q-1}$ is minimal in the minor order for order-$q$ torsion in $H_{m-1}(\mathbb{B}_m)$. In particular, odd torsion first appears in $H_2(\mathbb{B}_3K_{4,5})\cong\mathbb{Z}^{155}\oplus(\mathbb{Z}/2)^4\oplus\mathbb{Z}/3$, and no proper minor of $K_{4,5}$ has odd torsion in $H_2(\mathbb{B}_3)$. For arbitrary part sizes, we give a multiplicity-free decomposition of $H_m(\mathbb{B}_mK_{a,b};\mathbb{Q})$ under vertex permutations and prove that $H_{m-1}(\mathbb{B}_mK_{a,b};\mathbb{Z})$ has no $p$-primary torsion when $a,b\ge2m-1$ and $p\ge m$ is an odd prime. The explicit order-$q$ class retains its order under every enlargement of the second part of the graph, while for $m=q=p$ an odd prime it is killed by a specified enlargement of the first part.

This paper investigates the task of predicting job experience levels in recruitment texts, aiming to automatically identify the qualifications required for positions. Unlike traditional text classification, recruitment texts typically possess explicit internal structures, with different paragraphs playing disproportionate roles in conveying experience clues. To address this, we propose a structure-aware Section-Aware BERT approach that segments and encodes key paragraphs (titles, responsibilities, requirements) for integrated modeling, building upon rule-based systems and classical baselines TF-IDF. Simultaneously, we evaluate large language models under both few-shot and fine-tuning settings on the same dataset to compare the capability boundaries of different modeling paradigms. Experimental results demonstrate that explicitly leveraging text structure significantly improves experience level prediction performance, particularly in scenarios with ambiguous job titles. Further error analysis reveals systemic challenges in this task, including confusion between Entry and Senior levels and the blurred boundaries of Mid-level positions. This research provides an effective modeling approach and analytical framework for understanding structured recruitment texts.

Recursive computation repeatedly compresses or reuses intermediate states, creating a simple tension: information that must remain useful across longer recursive paths is also exposed to more opportunities for loss before reaching the final prediction. Existing reconstruction or local-prediction objectives provide tractable supervision, but do not ensure that the retained information remains sufficient for subsequent recursive computation. We identify local predictive sufficiency with recursive predictive closure: controlling local predictive deficiencies at individual interfaces controls the resulting discrepancy at the root. We then turn this principle into a tractable training procedure. Starting from a variational characterization, we derive finite predictive tests and an empirical predictive deficiency that measures predictive value retained across compression. Its predictive sensitivities define margin-relaxed half-space constraints on parameter updates, and we project the host optimizer's proposed update onto their intersection only when predictive preservation would otherwise be violated. Across temporal graphs, language memory, vision-language-action control, and recursive self-improvement, the method matches or improves the corresponding host models under matched compression budgets, with larger gains under heavier recursive or memory demands, while better preserving predictive information across successive transformations. Crucially, the same task-agnostic predictive-preservation principle is instantiated across all four settings through host-compatible interventions while keeping the endpoint task, backbone, and evaluation protocol fixed. These results establish predictive preservation at recursive interfaces as a general training principle for recursive compression.

Sequential recommenders can generate synthetic interaction sequences and retrain on the augmented corpus in a recursive self-improvement loop. To limit error accumulation, current methods verify each generated sequence remains predictive of the user's real interactions and discard those that drift away from it. Verification does not, however, determine which verified sequences should train the next model. With every verified sequence used for training, source sequences yielding more verified sequences or longer continuations have more influence, although neither quantity indicates how much those sequences will help the next model. We formulate the decision of which verified sequences are used to train the next model as post-verification acquisition and introduce Disagreement-Aware Recursive Self-Improving Recommendation (DA-RSIR). DA-RSIR caps each source sequence's contribution and ranks its verified sequences by how much the model's predictions disagree over their augmented interactions. It uses a score derived from Bayesian Active Learning by Disagreement (BALD) and estimated with Monte Carlo (MC) dropout. DA-RSIR requires no extra labels, teacher model, or quality scorer. Across four datasets, three recommender models, and two metrics, it improves on the retain-all approach in all $24$ comparisons and attains the highest mean in $23$ of $24$ overall; the aggregate improvement is statistically significant on both metrics. A single DA-RSIR round exceeds the retain-all approach's best gain over five recursive rounds. These findings establish post-verification acquisition as a separate control point in recursive self-improvement, separating which sequences pass verification from which verified sequences are used to train the next model.

Large datasets and high capacity models have accelerated progress in vision and language. This work introduces a platform aimed at bringing comparable gains to embodied learning, world models, and robotics. We present EnvDreamer, a framework that uses large language and vision language models to generate Unreal Engine 5 environments for embodied AI and robot training. EnvDreamer enables sampling of large, diverse, interactive, customizable, and validator passed virtual environments for training and evaluation across navigation, interaction, and manipulation. We illustrate the platform with a large set of generated scenes and simple baselines. Policies trained on EnvDreamer generated environments, without explicit mapping or human task supervision, achieve competitive results on multiple embodied benchmarks spanning navigation, rearrangement, and manipulation. EnvDreamer also supports image-conditioned reconstruction for real-to-sim studies. Finally, we release EnvDreamer-20k, a dataset of 20,000 validator passed environments with task programs, scene graphs, trajectories, and metadata to support reproducible benchmarking.

The semantic information of objects in remote sensing images is typically invariant to geometric transformations from the dihedral group D4. However, SAM3-based open-vocabulary semantic segmentation (OVSS) methods often exhibit inconsistent responses to different geometric transformations. To exploit this property and improve the stability of OVSS for remote sensing images, we propose a feature-adaptive manifold repair method based on dihedral-group geometric transformations. First, we introduce multi-scale harmonic-guided D4 view selection (MH-D4VS) to select complementary candidate views from a set of geometrically transformed views. Next, we propose original-view-anchored adaptive manifold repair (OAMR), which uses the original view as an anchor and reliable cross-view information to selectively repair locally unreliable visual features. Finally, we develop pixel decoder test-time adaptation (PD-TTA) for SAM3, which fine-tunes only the parameters of the GroupNorm layers online during inference, thereby enhancing the model's ability to adapt to sample-level distribution shifts. Experimental results show that the proposed method achieves an average mIoU of 55.6% across eight remote sensing semantic segmentation benchmarks and delivers consistent performance improvements under different SAM3-based inference frameworks.

Self-evolving reasoning models learn from their own generated questions, yet repeated self-training can lead to performance collapse. In this paper, we investigate why performance deteriorates over successive rounds and how to sustain self-evolution. Our analysis identifies two recurring quality problems in self-generated questions: invalid questions and repeated variants of the same mathematical questions. First, invalid questions become more prevalent across rounds, and answer-consistency filtering further increases their proportion in training data. Second, existing question diversity controls based on lexical similarity can miss mathematically equivalent questions expressed in different ways, which leads to question diversity collapse in later training rounds. Building on these findings, we introduce R-Quest, which uses question validity and novelty feedback to guide self-evolution. We first train the solver to recognize and reject invalid questions, then use its judgments to guide questioner rewards and filter solver training data. To avoid question repetition, we use a frozen base model to compare sampled question pairs and provide novelty feedback. Empirically, our method consistently achieves the highest average performance on 12 benchmarks in mathematical reasoning, general-domain reasoning, and code generation across two model families. Additionally, R-Quest maintains stable performance gains over ten rounds of self-evolution, peaking in the final round and outperforming R-Zero by 17.32 points.

Finite Fibonacci words have recently been shown to contain exactly three net occurrences—occurrences of repeated factors whose one-letter left and right extensions are unique. This unexpectedly small constant raises a natural question: is it peculiar to the Fibonacci recurrence, or part of a broader morphic phenomenon? We study a family over the alphabet $\{0,1,\ldots,d-1\}$ determined by positive integers $e_0,\ldots,e_{d-1}$. For each letter $a$, let $L_a$ be the morphism that fixes $a$ and maps every other letter $b$ to $ab$; we consider the finite approximants $S_m=\mu^m(0)$ generated by $\mu=L_0^{e_0}\cdots L_{d-1}^{e_{d-1}}$. These words are the period-aligned finite approximants of the purely morphic regular epistandard family considered here. We prove a sharp dichotomy: for every $m\ge2$, $S_m$ has exactly three net occurrences when $e_{d-1}=1$, and exactly two when $e_{d-1}\ge2$; the initial approximant is also completely classified. The proof combines palindromic prefixes, return-word factorizations, and overlapping net-occurrence covers. As a consequence, when all exponents are equal to one—the standard $d$-bonacci case—every noninitial period-aligned finite approximant has exactly three net occurrences, placing the Fibonacci phenomenon in a wider epistandard framework.

Latent safety filters certify safety on a learned low-dimensional representation of the state, enabling constraints that resist analytic description. Because the encoder is lossy, a filter can report safe while the physical state is unsafe, with no detectable model error. Existing transfer conditions leave the effect of discarded safety information implicit. We ask when a lossy encoder admits a safety certificate that transfers to the physical system, and show the answer is governed by the detectability of the discarded safety-relevant dynamics. We construct a system whose latent model is exact and whose latent signals always report safe, while the physical state becomes arbitrarily unsafe. For this system no certificate exists and no monitor downstream of the encoder can detect the failure. When the discarded dynamics contract, a latent barrier certifies true safety up to two explicit margins, one for the latent-model error and one for the variation of safety across states the encoder cannot distinguish. In the linear case and under boundedness and non-degeneracy conditions, every calibrated barrier transfers with a finite margin when the safety-relevant subspace is detectable, and none does otherwise. On learned cartpole encoders, the model error does not indicate for which representations the estimated bound is non-vacuous, while the second margin does.

Learning-based models for radio resource management (RRM) are typically built for a single function and deployment, so each new setting repeats the development pipeline. RRM decisions, however, share a common structure: each is assembled from interdependent fields, defined by the standard, whose values are selected in view of the network state. We propose RRM-GPT, an autoregressive framework for RRM foundation models that generate these decisions as a language model generates text. An encoder maps heterogeneous network observations into a common token representation, and a decoder emits the decision one field at a time, each conditioned on the network state and the fields already committed. Pretraining on unannotated network logs teaches the model what makes a decision valid and how controllers choose among valid decisions; post-training then adapts it to deployment-specific operator objectives through imitation or reinforcement learning. The framework targets two forms of reuse: a function-specific model reused across deployments, and a model shared across RRM functions that generates their interdependent decisions as one sequence. In a 5G New Radio (NR) case study, we demonstrate that a single model generates complete scheduling grants spanning user selection, timing, link adaptation, resource allocation, and control signaling. The model captures dependencies among grant fields and transfers learned behavior to an unseen scenario without adaptation.

Large language models incur language-dependent representation and inference costs, but existing comparisons often conflate input language, assigned observable-trace language, and answer realization. We specify a prospective paired study that separates these interfaces while holding the semantic item, checkpoint, and answer oracle fixed. The initial design instantiates 240 exactly scored items rendered from templates in English and seven non-English languages, three distinct-lineage open-weight checkpoints, three trace-token budgets, 22 input- and trace-language conditions, a fixed answer reserve, and a separately counted delimiter: 47,520 initial core runs before prospective sample-size selection. RQ1-RQ3 estimate input and trace effects by intention-to-treat with failure-inclusive terminal accounting and test answer realization by cloning a sealed prefix and runtime-native KV state into eight crossed branches. Pre-freeze independent language review, fixed-form ASCII selectors, and code/surface/solver agreement constrain the realization test. H1-H5 share one Holm family and a global simultaneous component band. A secondary randomized experiment compares one-long-attempt and complete K-short-attempt policies at equal trace allowance under frozen seeds and oracle-blind aggregation; it is a full-policy contrast because answer capacity differs. Outcomes include exact token spans, correctness, latency, runtime-exposed memory, and qualified same-host operating-system-reported energy over prespecified hardware rails. The protocol separates tokenizer expansion, observable-trace cost, and answer-realization cost without treating visible traces as internal cognition or operating-system estimates as physical cross-device energy. No confirmatory model outcome is reported; result fields remain disabled until the frozen evidence ledger passes independent verification.