Papers

Sodium chloride is a major component of extraterrestrial oceans, but it is generally assumed to be excluded from the high-pressure ice mantles of water-rich planets. Here we show that a concentrated NaCl solution, when quenched in an amorphous state before compression, crystallises in a metastable NaCl-bearing ice VII phase. Synchrotron x-ray diffraction and neutron diffraction reveals an expanded and strongly disordered ice VII lattice with no detectable crystalline NaCl. Ab initio calculations reproduce lattice expansion by substituting water molecules for Na$^+$ and Cl$^-$ ions and show that the structural response increases strongly with salt concentration. Upon decompression at 300 K, the phase expels crystalline NaCl at $5.1 \pm 0.3$ GPa and does not reform upon recompression, demonstrating that its formation is controlled by the thermodynamic pathway. Despite an 8.5% larger unit-cell volume, NaCl-bearing ice VII has a density close to that of pure ice VII and is substantially less dense than an equivalent mechanical mixture of ice and halite. These properties suggest that salty ice may be transported through deep convecting mantles but release salt near the 5 GPa boundary, where the denser products could inhibit further upward transport. Pathway-dependent salt trapping may therefore provide a previously unrecognised control on chemical exchange between rocky interiors and subsurface oceans.

We study leading parity-even four-derivative corrections to the linear Christodoulou regularity criterion for a neutral, massless, minimally coupled spectator scalar in four-dimensional charged de Sitter black holes. We derive the corrected static solution, cold extremality relation, near-horizon geometry, Cauchy-horizon surface gravity, and neutral scalar quasinormal spectrum within Einstein-Maxwell effective theory. Near cold extremality, the scalar equation reduces analytically to the finite-temperature anti-de Sitter two-dimensional throat problem. We find that the dominant nonstationary spherically symmetric mode retains its cold-limit regularity ratio through first order, while higher angular sectors respond to the deformation of the relative throat radii. The finite-distance neutral-scalar Christodoulou boundary nevertheless shifts because the relevant exterior scalar spectrum and the unequal horizon surface gravities enter the regularity ratio. At smaller and intermediate cosmological scales, positive effective coefficients move the Christodoulou threshold to larger normalized separation of the Cauchy and event horizons. Toward the larger cosmological-scale part of the cold branch, the same coefficient directions move that threshold back toward the cold endpoint. The reversal accompanies a transition in the quasinormal family that supplies the spectral gap. We further show that the coefficient combination governing the fixed-charge extremal mass parameter differs generically from those governing the cold throat and the finite-distance neutral-scalar boundary. Extremality therefore does not determine a universal direction for the neutral-scalar Christodoulou-threshold correction. Our results identify cosmological scale and spectral-family competition as essential elements in relating effective gravity, black-hole extremality, and Cauchy-horizon regularity in this spectator sector.

We propose a two-level filtered subspace iterative (TLFSI) method for computing all eigenpairs whose eigenvalues in the interval of interest $[a,b]$ after finite element discretization of self-adjoint PDE eigenvalue problems. The method combines a two-level approach with a modern filtered subspace iteration framework. In this paper, we use contour integral filtering as an example. The two-level method constructs a coarse space according to the spectral location of the eigenvalues of interest and obtains from it an approximation to the target invariant subspace. We prove that the distance between the approximate and target subspaces depends only on the separation between the spectral region of interest and the remainder of the spectrum, together with the coarse mesh size \(H\). Numerical experiments on benchmark elasticity problems confirm the theoretical predictions.

Online students often face emotional challenges such as frustration, isolation, and fluctuating motivation, which can hinder sustained engagement in online learning. While affective chatbots have shown promise in healthcare and wellness, most learning support chatbots focus primarily on cognitive assistance with limited attention to learners' emotional and motivational experiences. In this work, we investigate how adaptive affective chatbot support can be designed for online learning through an iterative design process. Through a mixed-methods needfinding study (n=38), we identified substantial variation in how learners respond to affective support, with no single strategy universally preferred. This led us to develop three adaptive design alternatives: user-controlled, AI-driven, and hybrid co-controlled adaptation, and examine key trade-offs between automation, transparency, and user control. A final evaluation (n=25) showed that 88% of participants favored the adaptive hybrid design over a non-adaptive alternative, highlighting the value of balancing transparency, user control, and interaction simplicity. We discuss design implications for adaptive affective chatbot systems in emotionally challenging online learning contexts.

Context: Fine-grained emotion classification of mobile app reviews enables requirements engineering activities that go beyond polarity-based opinion mining, including emotionally informed issue prioritisation and feature-oriented feedback analysis. However, automatic fine-grained emotion extraction from app reviews remains understudied. Objectives: Building on a previously published annotation framework and human-labelled ground truth adapted from Plutchik's taxonomy, this paper investigates how large language models can be leveraged for automatic multi-label emotion classification under severe class imbalance. Methods: We compare encoder-only fine-tuning under multi-label and binary-ensemble formulations, decoder-only zero- and few-shot prompting across open-source and proprietary models, and a catalogue of imbalance mitigation strategies (loss reweighting, resampling, generative data augmentation), with the synthetic-review generator and prompting strategy selected via an intrinsic augmentation-utility ranking. Results: Fine-tuned encoders trail the best decoder-only few-shot prompting (macro-F1 0.642) by a wide margin at baseline (multi-label: 0.387; binary ensemble: 0.450); pairing the best multi-label encoder with generative data augmentation and positive-weighted loss closes most of this gap (+0.204) at up to three orders of magnitude lower inference latency than the decoders, with the largest gains on the rarest emotions, from undetected to gains of up to +0.501 F1. Conclusion: Large language models make fine-grained, multi-label emotion classification of app reviews feasible for requirements engineering pipelines, with modest macro-F1, and the best formulation and mitigation strategy are backbone- and formulation-dependent. We release the experimental pipeline, synthetic corpora, and fine-tuned checkpoints for replication and reuse.

We reconsider the Bardeen black hole surrounded by a Kiselev fluid, obtained by adding a Kiselev term to the Ayon-Beato-Garcia-Bardeen metric. Contrary to claims in the literature, the solution is regular at the origin only for $c=0$: for $c\neq0$ curvature invariants diverge as $r\to0$, and the spacetime is asymptotically de Sitter-like rather than flat, with a cosmological horizon $r_c$ for $0<c<c_N(q)$. We map the full $(q,c)$ domain of horizon structures in closed form, locating the extremal and Nariai-type boundaries and their cusp. Using the Ma-Zhao prescription for the mass-dependent nonlinear-electrodynamics source, we derive the corrected entropy and verify the extended first law and Smarr relation to machine precision by computer-algebra differentiation, not through the entropy's own definition. The heat capacity shows a single Davies-type divergence, located as a closed-form quartic and expanded analytically in $c$. In thermodynamic geometry, the Quevedo-metric curvature has a spurious second singularity generated by the entropy's integration constant and the non-uniqueness of its conformal factor; we isolate this artefact and show only the Davies-point divergence is physical. The Ruppeiner curvature, identically zero for pure Bardeen, becomes non-zero once $c\neq0$. For the shadow, we compute the critical impact parameter $b_c(q,c,w)$ and show the nominal bound $c\lesssim10^{-2}/M$ from the EHT image of Sagittarius A* is not self-consistent: the same $c$ places the cosmological horizon nine orders of magnitude closer than Earth's distance to Sgr A*, and the bound is fourteen orders of magnitude larger than $H_0/c$, the true cosmological dark-energy scale. We reinterpret the EHT constraint as bounding $c$ only near the horizon, not as an astrophysical or cosmological bound on the quintessence parameter.

Coherent alias sampling implement an important subroutine for block encoding and qubitization based quantum simulation using quantum lookup table. Related constructions exploit data lookup and matrix structure to implement block encodings for use in quantum singular value transformation. For a state on $n$ qubits, constructing and storing the alias table requires both classical and quantum resources exponential in $n$. We show that when the amplitudes are samples of a polynomial of degree at most $d$, the table can be represented and evaluated coherently with cost polynomial in $n$, $d$, exponentially reduced from $\Theta(2^n)$ classical and quantum cost. We construct a quantum circuit for coherently computing the alias table for polynomial states using reversible integer arithmetic. By employing fast arithmetic, we construct a deterministic circuit for preparing polynomial states over Clifford+$T$ with a complete gate count of $\tilde{O}(ndB+B^2)$, where arithmetic width $B=\tilde{O}(dn+d\log(d+1))$. In the high precision limit for state fidelity $\epsilon$ with fixed $n$ and $d$, it is sufficient have $B=\Theta(\log(1/\epsilon))$ and the complete gate count becomes $O(\log^3(1/\epsilon)\log\log(1/\epsilon))$. Additionally, the effective representation of the alias table can also be used for classical sampling algorithms. Furthermore, we develop a novel technique for uncomputing the auxilary register generated by coherent alias sampling. This technique opens a research direction in sampling based initial state preparation.

Recovering executable CAD programs from 3D meshes is challenging due to the compositional nature of CAD construction and the interaction between discrete modeling choices and continuous parameters. Many learning-based methods predict complete programs in a single pass and rely predominantly on sketch-extrude representations, limiting operation diversity and opportunities to correct geometric errors during reconstruction. We introduce StepCAD, a generative optimization approach that combines a state-conditioned CAD policy with geometry-guided search. Given an input mesh, the policy predicts construction actions conditioned on both target and intermediate geometry, and an IoU-guided tree search refines the resulting program through local edits. We also introduce ARCADE-1.5M, a large-scale dataset of 1.5M executable CAD programs spanning diverse operations, sequences with a maximum length of 150+ counted operations, and 12.5M intermediate state-action transitions. Experiments across multiple CAD reconstruction benchmarks show that StepCAD achieves state-of-the-art geometric reconstruction accuracy with consistently high validity, yielding up to 87.2% relative IoU improvement over the strongest evaluated baseline, with particularly large gains on complex shapes. Project page: https://ghadinehme.com/stepcad.github.io/

We study how a radially oscillating thin shell outside a black hole (BH) affects gravitational wave ringdown and derive matching conditions for polar and axial perturbations from the linearized Israel junction conditions on the moving worldtube. Periodic shell motion produces Floquet sidebands in the scattered wave. For a distant shell, the motion has little effect on the prompt ringdown and only weakly modifies the late time echoes. When the shell is closer to the BH, a replica of a long lived shell mode enters the incident frequency band, increasing the late time echo amplitude by roughly two orders of magnitude. We also identify an exceptional point where two polar modes coalesce and find the contribution of its second order pole in the late time waveform.

World Action Models (WAMs) leverage pretrained video foundation models for robot manipulation, but their large backbones and video-action co-prediction are expensive. Although existing acceleration techniques offer many ways to reduce this cost, selecting and composing them requires substantial engineering for each model and hardware platform. To tackle this bottleneck, we present WAMJET, an agentic harness that accelerates WAM inference by equipping coding agents with reusable optimization guidance and measurement and validation tools. WAMJET follows a bottleneck-driven workflow where the agent profiles inference, modifies targeted code, validates effects, and iteratively refines the acceleration stack as bottlenecks shift, while preserving action quality. Experiments span six WAMs, three coding agents, and two GPU architectures. WAMJET achieves up to 9.95x lossless speedup over upstream implementations. Approximation and hardware-aware optimization yield additional latency reductions, with comparable success rates. The results show that WAMJET can produce effective acceleration stacks for WAM deployment.

Rigorous schedulability analysis is essential for the design of hard real-time systems, yet errors in pen-and-paper proofs threaten the safety of critical applications. The Prosa initiative addresses this by offering a foundation for building machine-checkable schedulability analysis proofs in the Rocq proof assistant. However, the substantial time and expertise required to construct such proofs remain a major barrier for wider adoption of Prosa. This work presents ProsaBuddy, an LLM?based agent system designed to lower the effort needed to develop mechanized real-time schedulability proofs. ProsaBuddy employs a ReAct loop with retrieval over the Prosa codebase, access to Rocq tools and optional human-written hints. It uses a subgoal?delegation architecture, decomposing a lemma into subgoals and dispatches them to subagents for proof. We evaluate ProsaBuddy on a mini benchmark drawn from real-time scheduling literature. Experiment results show that ProsaBuddy significantly outper?forms state-of-the-art LLM-based Rocq automated proving agent systems and a general coding agent OpenCode

Let $G$ be a compact connected semisimple Lie group, let $Q$ be a bi-invariant metric and let $\rho$ be its Cartan three-form. We prove that every left-invariant generalized Ricci flow with $[H_0]=[\rho]$ exists for all positive time and converges smoothly to a Bismut-Ricci-flat pair without rescaling or pullback. The standard pair uniquely maximizes generalized scalar curvature in its torsion class, and every left-invariant Bismut-Ricci-flat pair on a compact simple group is standard, answering questions of Fusi, Lafuente and Stanfield and of Lauret and Will. On the product of two copies of a compact simple group, nonstandard stationary pairs in the Cartan class are known; we construct nonstationary flows with the exact initial form $[H_0]=[\rho]$ and nonstandard limits, disproving the general semisimple case of the convergence conjecture of Garcia-Fernandez and Streets. The proofs use a proper nonincreasing energy and a decomposition of Killing fields that parametrizes the stationary pairs by nilpotent Lie algebra endomorphisms.

Formal verification of quantum circuits becomes difficult as circuit size grows because exhaustive reasoning over the Hilbert space is computationally expensive. In previous work [1], we introduced a bit-vector abstraction that separates superposition tracking from basis-state behavior, enabling SMT-based verification of fixed-size error-detection and entanglement circuits built from H, X, and C-NOT gates. This paper extends the framework in five directions. First, we generalise correctness properties to parametrised repetition and concatenated phase-flip codes with up to 101 logical elements. Second, we verify syndrome extraction for CSS codes specified by arbitrary parity-check matrices, including the Steane, [[15,7,3]] Hamming, [[15,1,3]] Reed-Muller, and rotated surface codes with distances up to 101. We consider arbitrary error patterns, detection up to a specified weight, and recovery using a classical decoder, with counterexamples beyond each code's correction or detection capability. Third, we characterize the {H,X,C-NOT} fragment for which the abstraction faithfully represents Hilbert-space semantics. We prove soundness, demonstrate necessity with counterexamples, validate the characterisation against state-vector simulation, and prove the six lemmas from the original work. Fourth, a mutation study using 13 operators identifies equivalent mutants and detects all 622 non-equivalent mutants within the evaluated fault model. Fifth, a direct Z3 implementation verifies GHZ circuits with up to 262,144 qubits across three topologies on a 16GB laptop. Random H+C-NOT networks show verification cost rises sharply with depth when outputs depend on many inputs. The approach does not target general stabilizer circuits, Y-type stabilizers, fault-tolerant protocols, or physical noise; we identify extensions needed to address these cases.

We present the largest study to date of star cluster mass and age distributions in nearby galaxies. The analysis is based on a uniformly selected catalog of over 15,000 (36,000) human (machine) classified compact clusters across 38 spiral galaxies in the PHANGS-HST survey, with improved photometric SED age dating that mitigates key degeneracies with reddening and metallicity. The galaxies span a factor of ~100 in star formation rate (SFR) and in their surface density of star formation and molecular gas (Sigma_SFR, Sigma_H2), and cover a broad range of morphologies. We find that cluster mass functions are well described by a power law, dN/dM ~ M^beta, with beta=-1.9 +/- 0.2, and that this shape is remarkably similar across all galaxies in our sample and eight composite groups of galaxies with similar SFRs, with no significant dependence on SFR, Sigma_SFR}, or Sigma_H2. Cluster age distributions decline steeply and can be described by dN/dt ~ t^gamma with gamma=-0.83 +/- 0.24, again with little dependence on host-galaxy properties or cluster mass. The independence of the mass and age distributions implies a separable joint distribution, g(M,t) ~ M^beta t^gamma, and suggests that cluster populations in spiral disks are dominated by strong, mass-independent dissolution over their first ~Gyr. These results establish a benchmark for interpreting cluster demographics across diverse galactic environments and provide context for studies of young cluster populations at high redshift.

Vehicle-to-everything (V2X) sidelink and uplink broadcast must deliver safety-critical and application-layer data reliably under fast time-varying fading, intermittent packet erasures, and exposure to passive eavesdroppers. Fixed-rate physical-layer codes are poorly matched to these conditions because link quality and the number of successfully received symbols vary across vehicles and over time. This paper proposes a unified vehicular communication framework that combines rateless Construction-D' Raptor lattices, nested-lattice physical-layer security (PLS) with direct-sequence spread spectrum (DSSS), and blockchain-assisted trust management for decode-and-forward (DF) cooperative relaying. At the physical layer, Raptor codes provide incremental redundancy over unreliable V2X links; multilevel lattice decoding enables power-efficient joint coding and modulation; nested shaping and spread spectrum limit fine-lattice recovery by eavesdroppers; and a distributed ledger records decoding success at relays to support reputation-aware relay selection. We model the vehicular air interface as a block-Rayleigh fading channel with additive Gaussian noise and random symbol erasures, representative of urban V2X sidelink behavior. The proposed code concatenates Luby-transform (LT) codes with quasi-cyclic low-density parity-check (QC-LDPC) precodes within Construction-D', with girth-4 cycle removal on the Tanner graph. AWGN benchmarks show competitive symbol error rate (SER) and improved frame error rate (FER) versus matched-length LDPC and LDLC lattices. Under block-Rayleigh fading with erasures, for n=995 at average VNR of 1.5 dB and erasure probability 0.15, rateless Raptor reception achieves 0.3% outage with about 2.3 packets on average, outperforming LDLC and LDPC under identical channels, while an eavesdropper with 8 dB combined SNR and spreading loss fails in 62% of sessions.

Let $k$ be a field of characteristic zero and $D_N(a_n)=a_{n+1}a_{n+2}$ the adjacent-quadratic locally nilpotent derivation on $A_N=k[a_0,\dots,a_N]$. We determine the invariant ring at $N=5$, where deleting $a_0$ gives the polynomial base kernel $B^{D_5}=K=k[p,q,J_1,J_2]$, so localization gives the polynomial-line invariant algebra $(A_5^{D_5})_{pq}=K_{pq}[W]$ with $W=F_0/(p^2q^4)$, but this does not by itself determine the global kernel. An invariant $H$ of weight $24$ and $a_0$-degree $2$ lies outside $K[F_0]$, and adjoining it gives the six-generator hypersurface $C_5$; the obstruction is explained by algebraic residue behavior defeating naive denominator descent. An invariant $F_{46}^{int}$ of weight $46$ and $a_0$-degree $4$ lies outside $C_5$ and satisfies $p^2F_{46}^{int}\in C_5$. To pass from this candidate to the full invariant ring, a second slice chart arising from $f_{13}$, together with exact $q$- and $p$-saturation, yields $A_5^{D_5}=C_5[F_{46}^{int}]$, hence finite generation, with a codimension-two complete-intersection presentation and explicit Hilbert series. At $N=6$ the transported $B$-level ring is exact, while the $A_6$ analysis is established through total degree $13$ and the full $A_6$ invariant ring remains open.

A cardinality constraint written on an edge of an entity-relationship diagram admits two opposite readings. Under the reading used by UML and by Chen's original model, the label is read with the entity set on the same side of the relationship; under the reading used by standard database textbooks, it is read with the entity set on the opposite side. The two readings are exact opposites, so a reader who assumes the wrong one takes away the opposite of what the designer meant. For binary relationships the difficulty is confined to interpretation, because the two constraints a binary relationship set admits can both be drawn. For relationships among three or more entity sets the picture is worse: a ternary relationship set admits twelve cardinality constraints, and a diagram with three edges can carry at most three of them. This paper presents a notation, Card(R; p; q) = (lower, upper), that removes the ambiguity without taking a side in it, and that is not limited to one constraint per edge. We give the number of constraints an n-ary relationship set admits, show which of them a design determines without ever writing them down, and state two inference rules, decomposition and augmentation, that derive one constraint from another, together with the side conditions under which each is sound. We close with a worked case study that turns three business requirements into three explicit constraints and nine more that the design decides on its own.

We prove that the wedge local model for a ramified unitary group of signature $(r,s)$ ($r>s$) at self-dual level is flat when $p\geq 2s-1$, confirming Pappas's conjecture in this case. The proof first reduces the problem to determining the defining equations of a nilpotent orbit closure of self-adjoint matrices, and then to computing the global sections of the corresponding Kempf resolution. The computation is carried out by reducing to the type $A$ case using the Koszul complex.

Event cameras have microsecond-level temporal resolution and high dynamic range which make them more resilient to motion blur and poor illumination than standard frame-based cameras. Event-camera visual odometry (VO) pipelines maximize these benefits when they process the asynchronous event stream at the native temporal resolution. Continuous-time Gaussian process (GP) regression and a white-noise-on-acceleration (WNOA) prior can handle asynchronous measurements but result in a prohibitively large estimation state when applied naively. This paper presents a continuous-time event-camera stereo VO pipeline that maintains the native measurement times of asynchronous events while also running in real time. It reduces the estimation states to keytimes while maintaining full temporal resolution by interpolating measurements to their exact timestamps with a physically founded WNOA prior. This decouples the state size from the dense number of measurements without discarding their asynchronous nature. The real-time continuous-time VO pipeline is evaluated on the MVSEC and DSEC datasets. It provides estimates in real time that are more accurate than ES-PTAM, a state-of-the-art discrete estimator, in all but one of the tested sequences. The pipeline respectively provides estimates at 22 Hz and 6 Hz on MVSEC and DSEC and RMS relative errors of 0.46 cm and 0.038 degrees across all valid sequences, which were 11 and 15 times better than ES-PTAM, respectively.

In instruction-guided generative recommendation, LLM-based recommenders need to balance two goals: responding to the user's current request and aligning with the preferences in their interaction history. When the two conflict, history events can override the request. We show that turning the effect of individual history events into supervision faces two obstacles. First, the events that most influence a recommendation are not necessarily the ones that support the target item. Second, removing a misleading event can raise the target's score but a competing item's score even more, so a higher target score alone does not guarantee a better ranking. We propose Asymmetric Intervention-Guided Margin Supervision (AIMS), which converts the effect of removing individual history events into ranking supervision. For training requests already ranked correctly, a frozen reference model identifies request-specific deletions that improve both the target's score and its margin over a competitor near the recommendation cutoff. These margins serve as training targets, while the complete history is retained as input. Training combines cross-entropy with an asymmetric auxiliary loss that penalizes margin shortfalls and routes its gradient only through the competitor score. Inference is unchanged, requiring no history editing or deletion search. Across six LLM backbones on an industrial dataset and two public benchmarks, AIMS improves Recall and NDCG over strong baselines. Ablations support request-specific margins and asymmetric supervision, and the selected deletions preferentially remove constraint-violating history.

Sustainable protein discovery lacks the fast computational proxies, analogous to molecular docking or density functional theory, that accelerate drug and materials discovery. Evaluating whether a novel food tastes like its animal-based target requires expensive human sensory panels, bottlenecking the design-build-test loop. We introduce TasteBench, a multimodal benchmark and privacy-preserving competition for sensory prediction, spanning two tasks: a food-level ranking task built on 21K+ human evaluations across 215 plant-based foods in 24 product categories, yielding 935 within-category ranking pairs, and a supporting molecular-level taste classification task over 15K flavor molecules. To enable rigorous interpretation of model performance, we characterize the ground truth: inter-rater agreement among panelists is low (Krippendorff's $\alpha = .077$), and the split-half reliability ceiling of panel-aggregated rankings is .825, establishing the range within which ML systems on this benchmark should be assessed. We evaluate baselines across four input modalities; on the same pairs panelists rated, the best model achieves .661 pairwise accuracy, competitive with the median individual panelist (.650), and .683 across all within-category pairs. TasteBench provides the evaluation infrastructure and baselines for measuring progress on computational screening for sustainable protein discovery.

Deploying LLMs on consumer-grade GPUs with insufficient memory to hold their weights can result in prohibitively slow inference, because decoding repeatedly transfers offloaded weights from system RAM or flash storage into GPU at much lower bandwidth than local GPU-memory access. Activation sparsity reduces these transfers by skipping weights associated with zero or near-zero activations. However, as more activation contributions are omitted, model quality eventually degrades rapidly, indicating that weights associated with small-magnitude activations collectively influence model quality sharply. In this work, we improve the trade-off between model quality and decoding performance when exploiting activation sparsity. Our key idea is to replace the binary choice of whether or not to read a weight with three options: fully retain it, approximate it using a compressed weight representation, or omit it entirely. SpAx skips weights associated with activations closest to zero, reads approximate weights for smaller-magnitude activations, and reads original weights for the largest-magnitude activations. Smaller-magnitude activations attenuate the errors introduced by approximate weights, while compressed weight representations require fewer bytes to be transferred. With weights offloaded to CPU memory, SpAx speeds up decoding by 3.86X on average (up to 5.57X) with 16-bit weights and 2.06X (up to 2.74X) with 4-bit weights, at a WikiText-2 perplexity increase of at most 10%. With weights offloaded to flash storage, the speedups are 3.31X on average (up to 4.81X) and 1.54X (up to 2.03X).

Vision foundation models such as DINOv2, SigLIP2, and MASt3R develop complementary capabilities from different pretraining objectives, yet their knowledge remains distributed across separate, specialized models. Multi-teacher knowledge distillation offers a path toward consolidating these capabilities into a single agglomerative backbone, but existing approaches assume a fixed set of teachers, and incorporating a new teacher requires repeating expensive joint distillation over the entire teacher set. We introduce GRAFT, a continual multi-teacher distillation framework that enables a unified backbone to progressively acquire capabilities from an open-ended sequence of foundation models. When a new teacher arrives, GRAFT treats the previously distilled model as a teacher for preserving learned capabilities, while the current student jointly learns from both the previous model and the incoming teacher. Furthermore, to reconcile the incompatible representation geometries of heterogeneous teachers, we introduce Teacher Specific Readout Tokens, which grant each teacher an independent read-out of the shared encoder, together with Geometry Agnostic Relational Loss that aligns a vision-language teacher by matching image-text similarity structures rather than raw feature values. We provide GRAFT model, which is a single, continually extensible backbone that unifies five domains, including image understanding, 2D dense prediction, 3D human pose estimation, 3D vision, and vision-language, delivering strong performance across all of them while acquiring each new capability at the cost of a single distillation rather than a full re-distillation.

We prove a variant of the $L^2$ Fourier restriction inequality for non-degenerate curves in the setting of orthonormal systems. To achieve this, we develop a Hilbert space approach to studying the $L^2$ theory for Fourier restriction for curves. Our arguments also rely on novel $L^p$-improving bounds for associated convolution operators, which take averages over neighbourhoods of cones generated by the curve. The $L^p$-improving bound are established using a variant of the high-low method, in particular using induction-on-scale.

Online education has enabled learners worldwide to take up courses from reputed institutions. However, a self-paced online course cannot guarantee the motivation and engagement that a learner experiences in a real-time classroom. Self-paced access also makes platform load unpredictable, which drives up the compute cost. We propose a Commitment Scheme for course access that aligns learner commitment with platform capacity. Learners book their study slots in advance; the instructor sets the budget of learning hours available for the course; and learners who use a full window earn additional watch hours. A booked window records an intention to study at a stated time, and a learner who appears in that window implements it. The byproduct is a platform load that can be forecast and bounded. The study involved two courses taken in sequence by the same learners, with the slot booking system activated only in the second. In-window study was observed on 86.8% of booked windows, and the median committer placed 95.5% of all study time inside self-booked windows. Among learners who studied across the launch, study regularity improved from 1.01 to 1.33 active days per week, with a supporting difference of +1.37 days per week against the same learners' preceding course. Commitments made on the same day as the study slot were honoured more often than advance bookings (88.8% against 62.5% two days ahead), which is consistent with the classic intention-behaviour gap.

Numerical measurements capture how a system behaves, but often leave the meanings of its variables unspecified. Some variables are measured but never labeled, and others are never measured at all. Existing methods assign semantics to such variables by consulting general human knowledge, but this inherits its biases where that knowledge exists and offers nothing where it does not. We bridge this gap between measurements and their meanings with causal structure instead, reading a variable's semantics from how it acts on other variables. We formalize this as structure-constrained semantic alignment, in which the embedding of each unnamed variable is solved under the dependence relations implied by the causal graph, with the embeddings of a few known names as anchors. Accordingly, we build CausalBridge, a framework that discovers the causal graph from the measurements, latent variables included, solves for the embeddings under those relations, and expresses them as names through a language model. The causal structure reflects the mechanism that generated the measurements and is recovered from the measurements alone, which may make it the one source of information free of bias from human knowledge. We evaluate CausalBridge on five questionnaires and three robotics scenarios, with 20 to 90% of the variable names masked. It recovers the semantics of observed and latent variables more accurately than existing methods that rely on association, and its lead widens as less of the system is documented. The graph it discovers names variables as accurately as the documented one, and a new system is named in minutes and at a fraction of the cost of sampling methods. Once the semantic gap is bridged faithfully, machines can understand the world and take actions causally.

Data selection is already a central bottleneck in large-language-model training, where web-scale corpora are noisy and token budgets are finite. In continual pre-training (CPT), it becomes a forgetting-control problem: a poorly chosen target-domain corpus can overwrite capabilities encoded in the pretrained checkpoint. Existing CPT practice either scores candidates with parameter-agnostic scalars such as perplexity, or mitigates forgetting by spending many extra general-domain replay tokens. Neither strategy directly asks how training on a candidate will move the model parameters. We show that loss-based selection causes the post-CPT Fisher diagonal to drift downward on exactly the high-Fisher coordinates the pretrained model had committed to, while leaving low-Fisher coordinates largely untouched. This asymmetry exposes a parameter-space mechanism for catastrophic forgetting. Motivated by this observation, we propose a Fisher-aware CPT selector that decomposes each candidate's gradient into an anchor component, which measures perturbation along committed parameter directions, and a frontier component, which measures update capacity in unconstrained low-Fisher subspaces. We aggregate these signals with a log-determinant submodular objective and optimize it in a single pass using a scalable streaming data selection pipeline. On TinyLlama-1.1B and Llama-3.1-8B CPT over medical data, our selector improves target-domain quality while bounding forgetting on held-out pretraining benchmarks. Most importantly, it is substantially more token-efficient than forgetting-aware replay. 1B selected tokens already outperform the replay strategy trained with 10B tokens on both adaptation and forgetting, giving a 10x token-efficiency advantage.

Conventional catalytic descriptors evaluate symmetric, dissipative responses ($\mathcal{D}$), remaining blind to the antisymmetric, non-dissipative circulation ($\mathcal{L}$) associated with cyclic catalyst turnover. We address this by applying an augmented Onsager operator, $\mathcal{A} = \mathcal{D} + \mathcal{L}$, to catalytic interfaces. Through Onsager–Casimir parity rules, we show that electron spin governs this non-dissipative circulation within the reduced state space, as antisymmetric coupling necessitates a time-reversal-odd coordinate. We map these multi-physical couplings (Thermo-Hydro-Mechanical-Chemical-Electrical-Spin) using a Hasse diagram and introduce a basis-invariant Gateway number to quantify circulation relative to dissipation, thereby expanding standard one-dimensional volcano plots into a two-dimensional activity surface. Using metal–nitrogen single-atom sites on doped carbon as a model, published electronic-structure calculations supply the static response quantities from which both operator components can be constructed. Finally, we propose an explicitly falsifiable experimental test utilising magnetic field modulation near a spin-crossover to probe the effect of quantum spin on this non-dissipative regime.

Research on information operations has focused primarily on the coordinated accounts involved in such campaigns and in the content they promote. Bad actors manage the activity of inauthentic accounts through programmatic interfaces (APIs), but the role of this underlying infrastructure in enabling and uncovering such coordinated behaviors remains largely unexplored. This study investigates how the Twitter (now X) API was used to coordinate influence operations on the platform. Since API access is managed through developer apps, we extract app metadata from 43 operations taken down between 2018 and 2021 and apply three complementary computational approaches to study diverse forms of manipulation. Our analysis reveals several recurring patterns: API-as-a-service infrastructure reused across campaigns, gaming apps performing unauthorized engagement without user consent, spoofed app names to hide activity origins, and synchronized app networks indicating centralized control. Adding app usage as a coordination indicator provides a modest but consistent improvement in detection. These findings suggest that social media APIs play an important role in enabling online manipulation, and that the removal of app-level metadata by Twitter/X has eliminated an important tool for platform transparency and moderation.

Cosmic-string loops generate a broadband gravitational-wave background (GWB) that is typically modeled via two distinct formalisms: a discrete harmonic expansion, commonly used in pulsar-timing-array (PTA) studies, and a continuous burst superposition, employed in ground-based GWB searches. In this work, we formally unify the two formalisms, showing that they provide discrete and continuous representations of the same underlying radiation process and converge on the scale-invariant spectrum plateau. On the rising low-frequency branch, where the first few harmonics dominate, the discrete nature becomes important, causing the two formalisms to diverge. Within this unified framework, the spectral calculation is more consistent across frequency bands, enabling multi-band inference. Leveraging this physical insight, we perform joint Bayesian inference across four representative loop-distribution models using the NANOGrav, EPTA, and LIGO–Virgo–KAGRA (LVK) datasets. By jointly sampling the string tension $G\mu$ alongside the average cusp and kink numbers $\{N_c, N_k\}$, the loop-radiation uncertainties are accounted for. We find that employing the harmonic expansion strengthens the LVK bounds by $21\%-31\%$ for three of these models. Additionally, the stringent LVK bounds preclude three of these models from explaining the observed PTA common signal. For the sole surviving model, our multi-band analysis improves the inferred limit on $G\mu$.