Solar and Stellar Astrophysics
Papers
Fresh from arXiv, typeset properly. Highlighted words trade on the market.
All of arXiv
Wednesday 9 September
The characterization of Sun-like stars, including determining whether those stars are single or have stellar companions, has become increasingly important for missions seeking planets beyond the Solar System. With upcoming missions like the Habitable Worlds Observatory (HWO) on the horizon, vetting stars for binarity and confirming potential targets as viable is essential precursor science. In this paper, we present an automated algorithm that searches for double-lined spectroscopic binaries (SB2s) in high-resolution spectra. Using MOLUSC, we simulate realistic stellar populations to quantify the capability of algorithm for detecting stellar companions around nearby stars, with potential applications to HWO target stars and Gaia DR4 planet-host candidates. In addition, we explore how the detection rate is influenced by complementary observations like high-resolution adaptive optics imaging and time-series radial velocities. We find that our SB2 detection algorithm alone can rule out 46.1% of stellar-mass companions. High-resolution adaptive optics alone rules out 79.6% of stellar companions. Radial velocity surveys alone ruled out 98.4% and 98.7% for 5-year and 10-year surveys, respectively. When all three methods are combined, the probability of a typical nearby star having a stellar companion is <1%. The combination of the methods allows us to better understand and resolve target lists for future surveys.
Quantum Gases
Fractional Chern Insulators in Twisted Bilayer Optical Lattices
Twisted bilayer materials provide a versatile platform for realizing correlated topological states. Motivated by the recent experimental realization of atomic Bose-Einstein condensates in twisted bilayer optical lattices, we investigate the emergence of fractional topological phases in such systems. For experimentally realistic parameters, the single-particle spectrum hosts a pair of quasi-degenerate nearly flat moiré bands. We show using self-consistent Hartree-Fock calculations that atomic interactions spontaneously break time-reversal symmetry, lift the quasi-degeneracy, and generate an isolated nearly flat band with a nonzero Chern number. At fractional filling, exact diagonalization of the Hamiltonian projected to this band reveals a threefold degenerate ground-state manifold and a characteristic particle entanglement spectrum, providing strong evidence for a fractional Chern insulator. The resulting many-body gap is of order \(10^{-3}E_R\), corresponding to a nanokelvin energy scale accessible to state-of-the-art cold-atom experiments. Our scheme requires neither spin-orbit coupling, as in transition metal dichalcogenides, nor a finite magnetic field, as in existing twisted-bilayer-graphene experiments, providing a highly tunable route to strongly correlated topological phases in twisted bilayer optical lattices.
Read the full abstract
Representation Theory
A counterexample to the periodicity conjecture for finite-dimensional algebras
The periodicity conjecture asks whether a finite-dimensional algebra is periodic whenever all its simple modules are periodic. We construct a $36$-dimensional counterexample whose simple modules have period four. The algebra has a $2$-dimensional nonperiodic module, which shows that the algebra itself is not periodic.
Read the full abstract
Information Theory
Conditions for Global Optimality in Quantum Arimoto-Blahut Algorithms
Generalized Arimoto–Blahut (AB) algorithms are widely used in information theory and quantum optimization, but monotonic objective decrease and numerical stabilization do not guarantee global optimality. We establish necessary and sufficient conditions for the global optimality of full-rank AB fixed points for convex differentiable objectives under linear constraints. An AB fixed point is globally optimal if and only if the AB update direction and the objective gradient differ by an element of the constraint normal space at that point. The same compatibility condition characterizes agreement between individual AB and mirror-descent (MD) updates, whereas pathwise equivalence requires it along the entire common trajectory. Thus, an AB algorithm may follow a trajectory different from MD and still reach the global optimum. We also develop a posteriori optimality certificates based on feasible directional derivatives, including finite-difference upper bounds on the objective gap that require only objective evaluations. For channel relative entropy between dephasing and depolarizing channels, we analytically identify the global minimizer as the unique full-rank AB fixed point, although pathwise equivalence fails. Numerical experiments illustrate convergence of AB and MD along different trajectories and validate the certificates. By contrast, an amplitude-damping example shows that a monotone AB iteration can stabilize at a suboptimal fixed point, whose nonoptimality is detected by the finite-difference certificate. These results provide structural and computable criteria for assessing global optimality in quantum AB algorithms.
Read the full abstract
Superconductivity
Record-Breaking Elemental Superconductivity in Tetralayer Kagome Borophene
Superconductivity above the liquid-nitrogen temperature remains rare in two-dimensional elemental crystals, where strong covalent bonding often yields high phonon frequencies but insufficient electron-phonon coupling. Here, using first-principles calculations and fully anisotropic Migdal-Eliashberg theory, we predict tetralayer kagome borophene (TKB) stabilized by ABAB covalent stacking, as a liquid-nitrogen-temperature elemental superconductor. With a predicted critical temperature of 102 K, TKB sets a record-high value among previously reported elemental superconductors. Unlike known high-Tc boron-based superconductors dominated by in-plane sigma-bonding states and high-frequency in-plane B-B stretching modes, TKB realizes an out-of-plane s-pz-bonding-mediated pairing mechanism, in which interlayer s-pz bonding states at the Fermi level are strongly coupled to low-frequency out-of-plane vibrations of boron atoms. These results reveal a distinct out-of-plane pairing channel in multilayer borophene and establish covalent stacking engineering as a potential route for high-Tc superconductivity in two-dimensional materials.
Read the full abstract
Astrophysics of Galaxies
CROCODILE-SIDM: Tidal Formation of Dark Matter-Deficient Galaxies as a Test Case
We introduce CROCODILE-SIDM, a framework for treating self-interacting dark matter (SIDM) with the $N$-body part of GADGET4-Osaka code, as part of CROCODILE simulation family. As a test case, we investigate the impact of SIDM on the tidal formation of dark matter-deficient galaxies (DMDGs) with velocity-dependent cross-section models. We demonstrate that our implementation reproduces the analytic scattering rate in isolated halos. Including dynamical friction self-consistently, we evolve a dwarf satellite with $M_*=2\times10^8\,\mathrm{M}_\odot$ in a $\sim10^{11}\,\mathrm{M}_\odot$ halo on a decaying orbit around a massive host, comparing CDM with four SIDM cross sections for two initial satellite density profiles: a cuspy Navarro–Frenk–White (NFW) profile and a cored Burkert profile. We find that self-interactions primarily regulate the amount of DM retained between pericentric passages, and that the sign of this effect depends on the initial profile: a larger cross section retains more DM for the Burkert initial condition but less DM for the NFW initial condition. We show that this opposing behavior reflects the direction of SIDM heat conduction, which is set by the halo's evolutionary state at infall. Core formation in the cuspy profile assists DM stripping, whereas tidally accelerated gravothermal contraction in the cored profile suppresses tidal mass loss. Consequently, SIDM can either assist or hinder DMDG formation, depending on the satellite's inner structure, making DMDGs a potential probe of SIDM cross section.
Read the full abstract
Machine Learning
EEGBind: Detecting Source-Level Interictal Epileptiform Discharges via EEG-Centric Multimodal Binding
Source-level analysis of interictal epileptiform discharges (IEDs) is relevant to presurgical evaluation and treatment planning because it helps characterize where epileptiform activity is likely to arise. Beyond detecting whether an IED is present, this setting requires assigning IED-positive activity to clinically meaningful brain-region categories. This setting is challenging because source-region evidence in short electroencephalography (EEG) windows can be subtle, partial, and affected by subject variability, class imbalance, and imperfect multimodal context. We present EEGBind, an EEG-centric multimodal binding framework for five-class source-level IED classification. EEGBind treats EEG as the primary modality and binds synchronized video-context features around an EEG-centric representation. Instead of relying on early or overly strong multimodal fusion, which may perturb the source-sensitive EEG representation, EEGBind uses video context as auxiliary evidence for robust classification. A view-consistent repair stage is further used to improve hidden-set robustness while preserving the learned source-class boundary. On the NeuroMM 2026 Grand Challenge Track 3 NMM-Source-IED benchmark, EEGBind achieves 0.8395 on weighted-F1 and outperforms strong competitors. These results support EEG-centric multimodal binding as a practical strategy for source-level IED classification. The open-source code is available at https://github.com/HKUSTGZ-ML4Health-Lab/NeuroMM2026_IED_Detection.
Read the full abstract
Signal Processing
A Fully Wave-Domain Wideband MU MIMO OFDM Transmitter via Stacked Intelligent Metasurfaces
This paper proposes an advanced realization principle for wideband multiuser multiple-input multiple-output orthogonal frequency-division multiplexing (MU-MIMO OFDM) transmitters, where the conventional transmitter-side baseband chain is physically synthesized in the wave domain. For design and optimization purposes, this fully wave-domain wideband MU-MIMO OFDM transmitter implemented by a cascaded SIM structure is functionally partitioned into two cascaded SIM blocks. The first block, denoted as SIM_1, integrates symbol loading and channel-adaptive MU-MIMO precoding updated at the channel-coherence timescale, mapping the user streams to a virtual port-subcarrier representation. The second block, denoted as SIM_2, acts as an offline-configured sampling-rate modulator that materializes the inverse discrete Fourier transform (IDFT) and cyclic prefix (CP) insertion directly in the wave domain. This baseband-free architecture establishes a virtual-to-physical transition from information bits to radiated CP-extended OFDM waveforms. To account for practical nonidealities, SIM_2 is optimized to fit the ideal multi-port CP-OFDM operator, and its residual response is mapped into an effective coupling matrix. Then, SIM_1 is optimized in a communication-oriented manner by jointly adapting discrete phase shifts and stream-subcarrier power loading to maximize the sum spectral efficiency. Results demonstrate the convergence, architecture trade-off, wave-domain OFDM materialization accuracy, and competitive performance of the proposed baseband-free transmitter.
Read the full abstract
Nuclear Experiment
Reinvestigation of the Yield of the Trinity Device
Previous published analyses of trinitite to determine the yield of the Trinity nuclear explosion used various estimates of the fraction of fission products that became incorporated in trinitite. Furthermore, these studies utilized nuclear data that has subsequently been revised. Utilizing a planar germanium detector, we observed gamma rays produced by the decays of both 137Cs and 239Pu from four samples of trinitite. From the observed 137Cs/239Pu ratios of and a simple analysis method based on current nuclear data and weapons debris studies, we find an average yield of 9.1+-4.5 kilotons from our four samples. While this result is substantially lower than the established total yield of 21 kilotons, it is in reasonable agreement that attributable to fissions in the 239Pu core alone.
Read the full abstract
High Energy Astrophysical Phenomena
Aarmed with Data: Bumps, Outflows, and Disk-like Emission in TDE 2025aarm
The origin of the optical emission in tidal disruption events (TDEs) remains one of the major outstanding questions in the field, in part due to the limited number of nearby events with high-cadence monitoring to track their evolving photometric and spectroscopic properties. We present multi-wavelength observations of the nearby ($z=0.01368$) TDE\,2025aarm, including near-daily spectroscopic coverage prior to the optical peak. Its proximity makes it one of the brightest TDEs discovered, reaching a peak magnitude of $m_r\sim15.5$ ($M_r\sim-18$). The light curve deviates from a smooth evolution, exhibiting multiple rebrightening episodes visible in both the individual filter light curves and the bolometric luminosity. Blackbody modelling reveals that these rebrightenings are associated with an increase in temperature of $> 5,000-10,000$\,K, while the inferred photospheric radius remains approximately constant. Simultaneously, the H$\alpha$ line not only increases in blueshift but also broadens, suggesting a link between the continuum rebrightenings to changes in the kinematics of the line-forming gas. We identify a persistent absorption component at $\sim-3900$\,km\,s$^{-1}$ in multiple Balmer lines, providing further evidence for outflowing material. The H$\alpha$ profile also exhibits excess flux compared to a Gaussian on both sides of the line, inconsistent with simple scattering-dominated outflow models. Disk-profile modelling provides evidence for the emergence of a disk-like component least $\sim20$ days after peak, with substantial changes in the disk properties between $\sim50$ and 60 days. These observations highlight the complexity of TDE emission processes and demonstrate how dense multi-wavelength monitoring can disentangle the roles of accretion, reprocessing, and outflows in shaping TDE emission.
Read the full abstract
Analysis of PDEs
Radial solutions of quasilinear Hardy–Emden–Fowler equations: exact decay rates, Delaunay solutions, and nonexistence
Let $N>p>1$, $s>-p$, $0<\mu<\bar\mu:=\big(\frac{N-p}{p}\big)^{p}$ and $q>p-1$. We determine the behaviour at infinity of every positive solution of the quasilinear Hardy–Emden–Fowler equation \[ -\big(r^{N-1}|v'|^{p-2}v'\big)'=\mu\,r^{N-1-p}v^{p-1}+r^{N-1+s}v^{q}, \qquad r>R_{0}. \] Three rates compete: the self-similar rate $r^{-\gamma_{*}}$ with $\gamma_{*}=\frac{p+s}{q-p+1}$, and the two Hardy rates $r^{-\nu_{\pm}}$, where $\nu_{\pm}$ are the roots of $h(\nu):=\nu^{p-1}(N-p-(p-1)\nu)=\mu$. Writing $p^{*}_{s}=\frac{p(N+s)}{N-p}$ and $q_{S}=p-1+\frac{p+s}{\nu_{+}}$ for the Hardy–Sobolev and Serrin exponents, we obtain a complete picture in four regimes. If $q>p^{*}_{s}-1$, every solution realises exactly one of the three rates, with a logarithmic correction in the borderline case $h(\gamma_{*})=\mu$, and we compute the leading constants. If $q=p^{*}_{s}-1$ the system becomes integrable: besides the singular self-similar solution and a heteroclinic “bubble” there is a one-parameter family of Fowler–Delaunay solutions, for which $r^{\gamma_{*}}v(r)$ is a non-constant periodic function of $\ln r$. If $q_{S}<q<p^{*}_{s}-1$ only two orbits survive, and if $q\le q_{S}$ there is no positive solution at all in an exterior domain. This is the quasilinear, weighted counterpart of the picture obtained by Franca and Sfecci for $p=2$. We work with the slope variables $x=-rv'/v$, $Y=r^{p+s}v^{q-p+1}$, in which the equation becomes a planar system that is real-analytic on the half-plane $\{x>0\}$ for every $1<p<N$, so that no restriction of the type $p\le2$ is needed; the analysis then rests on a universal barrier $x<\nu_{+}$, on an explicit Dulac function which degenerates into a first integral exactly at $q=p^{*}_{s}-1$, and on an explicit Lyapunov function.
Read the full abstract
Mesoscale and Nanoscale Physics
A discrete Stiefel-Whitney invariant with twofold rotation symmetry
We review Stiefel–Whitney invariants and give a discrete formulation of an additional $\mathbb{Z}_2$ invariant for two-dimensional spinful insulators with twofold rotation and time-reversal symmetries (layer group $p1121'$). A singular gauge transformation specifies a real bundle over the quotient of the Brillouin torus, whose second Stiefel–Whitney number defines the invariant. We derive its dependence on the choice of auxiliary gauge function and prove invariance under changes of Bloch frame and additivity under direct sums. The discrete formula uses overlaps of independently chosen Bloch frames and symmetry-compatible $\mathrm{Pin}_+$ lifts on the boundary of a half Brillouin zone. Atomic and topological-insulator models illustrate the formula and distinguish the additional invariant from the Kane–Mele index. Together with three quantized $\mathbb{Z}_2$ partial polarizations and the filling number, it labels the known stable classification $\mathbb{Z}\oplus\mathbb{Z}_2^{\oplus 4}$.
Read the full abstract
Solar and Stellar Astrophysics
The FU Ori outburst of PR Ori B
We report the discovery of a nearby FU Ori-type outburst (FUor), PR Ori B, in the L1641 cluster of the Orion star-forming region. The high-amplitude variability was first identified in the NEOWISE (3-5 $\mu$m) photometry of the unresolved PR Ori binary system. Long-term, resolved optical photometric monitoring demonstrates that PR Ori B is the driver of a $\Delta G=5~$mag outburst, while PR Ori A has remained constant over the last 20 years. The near-IR spectrum of PR Ori B changes from a late K-type spectral type during quiescence to a viscously heated disk during outburst, including deep absorption in $^{12}$CO and H$_2$O bands. The optical spectrum also exhibits features that are commonly associated with FUors, including P Cygni profiles in Na I D lines and absorption in the Ca II infrared triplet. The luminosity of the outburst (L$_{acc}\sim$30–40 L$_\odot$) is similar to that commonly observed in FUors. The comparison of Spitzer/IRS and VLT/VISIR spectroscopy shows some evidence of silicate crystallisation during the outburst. PR Ori B is one of the closest and brightest FUors discovered over the last few years, only one magnitude fainter than the archetype of the class FU Ori. The proximity and brightness will allow for future high angular resolution observations to probe the physics of the inner disk and to evaluate changes in the disk due to the increased luminosity.
Read the full abstract
Machine Learning
EFQ-Softmax: Exp-Free Quantization for Softmax
Low-bit attention accelerates Transformer inference by moving the $QK^\top$ and $PV$ matrix multiplications to FP8 or FP4 matrix engines. However, the softmax path often evaluates shifted-score exponentials in higher precision, forms a temporary probability block, and quantizes it before low-bit $PV$ multiplication. This exp-then-quantize path creates a mismatch between a high-precision probability producer and a low-bit matrix consumer. We propose EFQ-Softmax (Exp-Free Quantization for Softmax), a low-bit probability-generation method that directly maps shifted attention scores to block-scaled E2M1 operands. For each microscaling block, EFQ-Softmax selects an exponent-only scale from the local maximum, maps the shifted scores to a normalized residual domain, and generates nonnegative E2M1 probability codes using a single affine rule. The resulting operand is used consistently in both the $\widetilde{P}V$ numerator update and the $\widetilde{P}\mathbf{1}$ denominator update. The FlashAttention-style row-maximum update, historical rescaling, high-precision accumulation, and final normalization remain unchanged. We evaluate end-to-end quality on Qwen3-8B, Qwen3-VL-8B-Instruct, and WAN2.2-TI2V-5B, and separately measure kernel-level performance on the A5 vector unit. EFQ-Softmax improves the Qwen3-8B seven-task mean from 0.6749 with MXFP4 to 0.6773 and the Qwen3-VL nine-task mean from 0.7826 to 0.8000. On WAN2.2, it maintains temporal consistency and visual quality comparable to the FP16 and MXFP4 baselines under VBench. On the A5 vector unit, EFQ-Softmax reduces the vector-stage latency of the fused probability-generation kernel by 40.33% on average across sequence lengths from 16K to 128K. These results show that direct low-bit probability generation can replace the conventional exp-then-quantize path while preserving end-to-end model quality.
Read the full abstract
Materials Science
First-principles investigation of structural, mechanical, vibrational, thermal, electronic and optical properties of Na3Bi: a topological Dirac semimetal
Exploring the properties of topological materials has remained an active area of research in condensed matter physics for the past few decades. Here, we present a first-principles investigation of the structural, elasto-mechanical, thermal, electronic, and optical properties of Na3Bi, one of the earliest discovered topological Dirac semimetals, using density functional theory. The hexagonal phase of Na3Bi is found to be both mechanically and dynamically stable, exhibiting moderate elastic anisotropy with a universal anisotropy index of 0.832. The compound displays a moderate machinability index of 2.386 alongside a low hardness value of 2.005, consistent with its predominantly brittle, covalently bonded character accompanied by a slight metallic contribution. Its comparatively low Debye temperature of 181.39K points to a mechanically soft lattice with loosely bound atoms. Electronic structure calculations confirm semimetallic behavior, with well-defined Dirac points pinned at the Fermi level and a Fermi surface composed of point-like pockets. Optical analysis across a wide spectral range reveals non-selective, moderate reflectivity of 40 to 45% and, in the visible region, an unusually high refractive index reaching between 2 and 4.5. Na3Bi is additionally found to be an efficient absorber of ultraviolet radiation. Together, these findings establish a comprehensive baseline for the mechanical, thermal, and optical behavior of hexagonal Na3Bi and point to its potential relevance for acoustic damping, UV-sensing, and optoelectronic applications
Read the full abstract
Computation and Language
StreamAlign: Streaming Text-Aligned Speech Tokenization
Text-aligned speech tokenization methods have emerged to better align speech tokens with LLM token spaces, enabling more effective utilization of pretrained LLMs. However, they rely on offline automatic speech recognition (ASR), leading to two key limitations: (i) the need for complete utterances before tokenization, precluding real-time streaming, and (ii) vocabulary mismatch between ASR and LLMs, which reduces acoustic granularity from the subword to the word level. We introduce StreamAlign, a text-aligned speech tokenization framework that enables streaming tokenization for real-time speech-text joint modeling. StreamAlign performs online speech-text alignment by combining character-level RNN-Transducer alignment with word-level ASR guidance, mitigating ASR-LLM vocabulary mismatch while preserving recognition accuracy. A proactive word boundary classifier anticipates word completion at chunk boundaries, reducing tokenization latency from 560 ms to 270 ms. On LibriSpeech, StreamAlign achieves the lowest WER and highest UTMOS among evaluated tokenizers. Furthermore, StreamAlign-SLM, a spoken language model trained on StreamAlign units, outperforms other end-to-end spoken language models in speech continuation while achieving the strongest overall consistency on SALMon and spoken StoryCloze.
Read the full abstract
Quantum Physics
Enhancing Pauli Correlation Encoding for quantum optimization via systematic expressivity analysis
Quantum approaches for combinatorial optimization problems have attracted considerable attention in recent years. Among these approaches, Pauli Correlation Encoding (PCE) has emerged as a promising framework for quantum devices with limited qubit resources because it embeds optimization variables in expectation values of Pauli strings. However, the mechanisms underlying its performance and the reasons for its saturation remain unclear. In this work, we investigate these questions through a systematic analysis of expressivity and trainability. First, we compare PCE with classical surrogate models based on tensor networks whose structures progressively approach the topology of the PCE circuit. The results show that PCE attains comparable solution quality with substantially fewer trainable parameters, indicating strong parameter efficiency. Second, to determine whether the performance saturation of conventional PCE is caused by insufficient expressivity or by optimization difficulty, we perform a diagnostic expressivity test in which the circuit is trained toward reference configurations for Max-Cut. The results show that even shallow PCE circuits can represent strong solutions, indicating that the main bottleneck is not the representational power of the ansatz, but the trainability under the relaxed objective function. Motivated by this finding, we propose a multistage continuation framework that gradually transforms a smooth relaxed objective into a sharper objective that more closely approximates the target discrete problem. Numerical experiments on G-set instances with 800 vertices show that the proposed method consistently outperforms conventional PCE and is competitive with representative graph neural network (GNN) methods. These results clarify the main factors behind PCE performance and provide a practical strategy for improving PCE on quantum devices with limited qubit resources.
Read the full abstract
High Energy Physics (Experiment)
Extended High-Mass Axion Search with an Auxetically Tuned Higher-Order-Mode Cavity
Conventional high-mass axion haloscopes based on the TM$_{010}$ mode lose detection volume as the resonant frequency increases. We report an extended axion search using a dielectric-restored TM$_{020}$ cavity haloscope with symmetry-preserving auxetic tuning based on a single-degree-of-freedom mechanical architecture. Using a near-quantum-limited microwave receiver, we searched a frequency range of 4.98-5.07 GHz and exclude axion-photon couplings with sensitivity approaching the KSVZ benchmark. Together with two earlier searches, the present scans extend a multi-scan program based on this architecture, yielding nearly 300 MHz of contiguous high-mass axion coverage over 4.98-5.27 GHz, the first broad search reported with a single higher-order-mode haloscope. This work establishes higher-order-mode cavities as a practical, scalable route beyond the TM$_{010}$ volume penalty.
Read the full abstract
Computation and Language
Scaling E-Commerce Attribute Extraction with Parallel Decoding
Customers rely on specific product attributes to compare products and make purchasing decisions, but e-commerce catalogs are messy and unstructured, making it difficult to identify which attributes matter most and extract them at scale. Standard Attribute Value Extraction (AVE) systems treat all attributes equally, producing large, inconsistent attribute sets that do not reflect the factors consumers use to differentiate products. We introduce a two-stage LLM pipeline that first discovers a compact, ranked schema of purchase-discriminative attributes for each product category, then extracts their values from catalog text using a fine-tuned compact LLM (Qwen3-4B) with Hyper-Parallel Decoding (HPD). This pipeline achieves 85% extraction accuracy, on par with the foundational LLM it was distilled from, while reducing inference costs by 92% over foundational LLMs, enabling production-scale use for product discovery and catalog enrichment. The resulting category-level structured representations effectively constitute automatically constructed product knowledge bases, providing consistent, comparable attributes across varied product categories that can ground downstream knowledge-intensive applications.
Read the full abstract
Combinatorics
NP-Hardness of the $H$-Free Edge-Deletion Problem
For a graph $H$, the $H$-freeness edge-deletion problem is the algorithmic problem of finding, for an input graph $G$, the minimum number of edges of $G$ whose deletion turns $G$ into an $H$-free graph. We show that for every graph $H$ containing a cycle, this problem is NP-hard. This proves a conjecture of Gishboliner, Levanzov and Shapira, and completes the characterization of the complexity of the $H$-freeness edge-deletion problem, answering a question of Alon, Shapira and Sudakov.
Read the full abstract
Number Theory
From Pythagorean Runs to Pell-Generated Cubic Runs
A classical Pythagorean run is an identity in which a block of consecutive squares is equal to the immediately following block of consecutive squares. Boardman's construction gives such a run for every prescribed length. We consider a cubic analogue in which the second block is allowed to have common difference 2. We prove that there are infinitely many positive integer triples \((m,A,B)\), with \(B>A+2m-1\), such that $$ \sum_{j=0}^{2m-1}(A+j)^3 = \sum_{j=0}^{m-1}(B+2j)^3. $$ An explicit infinite family is obtained from the Pell-type equation $$ 48329z^2-156m^2=161, $$ which reduces to a generalized Pell equation in normalized form. The ordering condition ensures that the step-2 progression begins strictly after the consecutive block ends. The resulting sequence of half-lengths satisfies an explicit second-order linear recurrence.
Read the full abstract
Human-Computer Interaction
How Far Do Capability Cues Travel? Anthropomorphism and Differentiated Trust in a Platform-Embedded AI Assistant
Visible AI capabilities need not translate into broader judgments of trustworthiness. In a randomized 2 x 2 experiment with 270 U.S.-based Reddit users, an embedded assistant displayed one or three functions, with or without a brief rationale. Displaying three functions increased perceived multifunctionality; no other randomized main effect survived correction across the six outcomes. Rationale availability did not reliably increase perceived intelligence. Exploratory analysis indicated stronger uptake of the functional display at higher objective AI literacy. Among concurrently measured judgments, perceived multifunctionality was associated with perceived intelligence, which was associated with anthropomorphism and all three trust dimensions. After accounting for perceived intelligence, anthropomorphism was positively associated with benevolence, but not reliably with integrity or ability. These findings separate interface effects from relationships among users' perceptions and show why ability, integrity, and benevolence should be evaluated separately.
Read the full abstract
Differential Geometry
Obata-type equations under the Bakry-Émery Ricci curvature conditions
In this paper, based on the warped product structures determined by the Obata-type equations with Robin boundary condition, we establish some rigidity results for compact manifolds with smooth boundary under appropriate Bakry-Émery Ricci curvature conditions and some other assumptions.
Read the full abstract
Computer Vision
VFNet: Multi-View Spatio-Temporal Model for Void Fraction Estimation in Gas-Liquid Two-Phase Flow
Void fraction, which quantifies the proportion of the fluid flow volume occupied by the gas phase, is a key parameter in the characterization of gas-liquid two-phase flow. Existing estimation methods either rely on flow assumptions that do not generalize across different fluids or on intrusive sensing that disturbs the flow behavior. We propose VFNet, a dual-branch spatio-temporal neural network for void-fraction prediction from synchronized multi-view videos of two-phase flow. A local branch extracts features from confined spatial regions and fuses the synchronized dual views, while a spatio-temporal branch captures the global evolution of the flow across space and time to refine a coarse geometric estimate. Trained on simulated computational fluid dynamics (CFD) data with known ground-truth void fractions and evaluated against both learning-based and traditional baselines, VFNet achieves the best performance across a broad range of metrics and also improves downstream flow-pattern classification on real two-phase flow data.
Read the full abstract
High Energy Astrophysical Phenomena
The Evolution of Thermal and Non-thermal Emission Components in GRB 250920C
We present a temporal and time-resolved spectral analysis of GRB 250920C using Fermi/GBM and Swift/BAT data. The prompt emission consists of two distinct episodes, EI and EII, separated by a significant quiescent interval. We perform Bayesian spectral fitting with empirical, thermal, composite, and physical synchrotron models. In EI, the spectra show clear evidence for an additional thermal component. The BB+PL model is preferred in most bright time bins, and joint Fermi/GBM+Swift/BAT fits further support the presence of this component. The blackbody temperature generally decreases with time, the blackbody flux follows the pulse profile, and the thermal flux fraction remains high. Fireball-parameter estimates give a photospheric Lorentz factor of a few hundred and an initial radius of $r_{0}\sim10^{9}$–$10^{10}$ cm, with $1+\sigma_{0}$ of order unity and $\eta\gg1$, supporting a thermally dominated baryonic outflow. In contrast, EII is dominated by non-thermal emission. Its low-energy photon indices do not significantly exceed the synchrotron line of death, and the spectra are well described by fast-cooling synchrotron radiation in a decaying magnetic field. The magnetization constraint gives $\sigma_{\min}\sim1.1$–$4.3$. Since these values exceed unity, they suggest that EII may be Poynting-flux dominated. These results therefore suggest a possible transition in GRB 250920C from a fireball-dominated EI phase to a Poynting-flux-dominated EII phase.
Read the full abstract
Atmospheric and Oceanic Physics
Boundary-layer scheme selection outweighs microphysics in WRF simulations of the August 2018 Kerala extreme rainfall event
Because precipitation is regulated by intricate interactions between topography, cloud microphysics,cumulus convection, radiation, and boundary-layer processes, it is still difficult to accurately simulate orographically enhanced monsoon rainfall over the Western Ghats.
Signal Processing
Efficient Graph Neural Networks for Multicarrier Wideband Hybrid Beamforming Optimization
6G wireless technology is poised to adopt higher and wider frequency bands, leveraging highly directional beamforming. However, the vast bandwidths amplify the impact of beam squinting. Traditional solutions, such as adding a true-time-delay filter to each antenna, are cost-prohibitive due to the required hardware scale. This paper proposes a signal processing alternative using Graph Neural Networks (GNNs) to optimize hybrid beamforming in multicarrier wideband systems. Using a bipartite graph to represent a shared analog beamformer among multiple subcarriers, we develop three GNN structures with distinct digital beamformer representations (i) at the subcarrier nodes, (ii) at the edges, or (iii) integrating traditional singular-value decomposition solutions. By designing an efficient message-passing mechanism, these structures offer insights into the impact of different GNN designs on communication system performance and computational complexity. Extensive analysis and ablation studies show that our proposed GNN structures outperform traditional optimization methods and existing ML-based solutions. Furthermore, the proposed GNNs exhibit strong resiliency to beam squinting and better robustness against imperfect CSI than even fully digital beamforming and all existing hybrid designs. These GNNs can also be extended to multi-user scenarios and demonstrate excellent generalization capabilities, allowing trained models to adapt to diverse multicarrier and multi-user settings without retraining.
Read the full abstract
Artificial Intelligence
Which Tokens Should SFT Actually Learn? A Token-Trimming Perspective on Mathematical Reasoning
Supervised fine-tuning (SFT) applies a uniform cross-entropy loss to all target tokens, even though different tokens provide unequal learning signals for mathematical reasoning. This uniform treatment can over-sharpen already mastered tokens while amplifying learning pressure on uncertain, low-confidence tokens, leading to suboptimal training dynamics. We propose Trimmed Logit-Gap SFT (TrimSFT), a simple token-level reweighting method that scales the SFT loss according to the logit gap between the gold token and its strongest competitor. TrimSFT trims supervision away from both extremes: tokens already mastered (large logit gap) and tokens weakly supported by the current model (small or negative logit gap), concentrating learning within an intermediate logit-gap region between them. We instantiate this principle with a Gaussian weight centered at margin m with bandwidth {\tau}, requiring no reference model or additional forward pass. We evaluate TrimSFT on six base models from the Llama, Qwen, and DeepMath families across five mathematical reasoning benchmarks. TrimSFT consistently improves over standard SFT, achieving the best average performance on five out of six models, with gains of up to +26.9 points over SFT on MATH500. Further analyses show that the bandwidth {\tau} matters more than the exact margin location, and that half-trim variants that remove supervision pressure from only one side yield inferior trade-offs. A token-level logit-gap distribution analysis suggests that TrimSFT reshapes model confidence in a more balanced way than uniform SFT or monotonic reweighting methods. These results suggest that reasoning SFT can benefit from trimming both extremes rather than treating all tokens uniformly.
Read the full abstract
Applied Physics
A Predictive Design Framework for a Soft Robotic Ventricle using Contractile Actuators
The natural cardiac cycle is divided into the systole and diastole phases which encompass four distinct stages: isovolumetric contraction and ejection during systole, followed by isovolumetric relaxation and filling during diastole. Cardiovascular modeling of this cycle ranges from high-fidelity multiphysics simulations to reduced-order lumped-parameter (Windkessel) representation of the heart-artery coupling. However, current models do not relate the mechanics of the actuator driving the ventricle pump to the hemodynamics. In this study, we develop and experimentally validate a predictive design framework for ventricle-like pumps using various types of soft contractile actuators. We build a circulatory loop which reproduces the entire loop including the isovolumetric phases-where pressure changes occur without volume shifts. The framework is based on a lumped-parameter model, hereafter referred to as the phase-dependent Actuator-driven Windkessel 3-element (AWK3) model, to bridge soft actuator mechanics to the cardiac pressure-volume (P-V) loop. Unlike traditional models that require either pressure or volume as a fixed input to estimate the other, our proposed model predicts both variables when informed by the isometric characteristics of the actuators. We validate the model using a ventricle-inspired pump driven by a linear contractile series-elastic actuator or twisted and coiled polymer actuators (TCPA). We relate the actuator isometric testing protocol to the phase-dependent AWK3 model, which replicates the Frank-Starling law, accurately describing cardiac behavior under varying conditions of preload, afterload, and inotropy (contractility). This approach provides a robust platform for the design and high-fidelity control of bio-inspired soft robotic circulatory systems.
Read the full abstract
Computer Vision
Cross-Species Animal Re-Identification with Semantic Consistency Learning
Generalizable animal Re-Identification (ReID) aims to recognize individual animals across species with diverse morphologies and ecological contexts. Unlike person ReID, where different domains share similar body structures, animal species often exhibit drastically different anatomical structures and visual patterns, making it difficult to establish shared visual correspondences. As a result, representations learned across species tend to form fragmented embedding spaces, which severely limits cross-species generalization. To address this challenge, we propose Semantic Consistency Learning (SCL), a framework designed to learn representations that remain stable across appearance variations while preserving semantic structures shared across species. SCL consists of two complementary components. Foreground-Background Decoupled Spectral Normalization (FDSNorm) stabilizes feature statistics by suppressing environment-induced style variations in a region-aware manner, while Cross-species Neighborhood Modeling (CNM) captures transferable relational structures across species through dynamic feature neighborhoods. Extensive experiments on 11 public animal ReID datasets demonstrate that SCL consistently outperforms state-of-the-art methods under multiple cross-species evaluation protocols and generalizes effectively to previously unseen species and ecological domains. Code is available at https://github.com/Kemalau/ECCV-26-SCL.