Papers

We provide a presentation of the braid group in which the generating set is the set of all positive half-twists, the relations are balanced positive relations, and every relation is conjugate to one in a finite list. We then prove that for any such presentation, its positive monoid is not isomorphic to the quasipositive braid monoid. Finally, we show that the quasipositive braid monoid has no proper Garside family. These results settle a problem of Calegari from the K3 problems list.

We introduce the Phantom Zernike Wavefront Sensor (phantom ZWFS), an engineered reference extension of phase-contrast wavefront sensing in which controlled pupil plane modulation generates off-axis coherent speckle structures that undergo localized focal plane phase shifts. These references may replace the central ZWFS reference when the on-axis PSF core is required for another optical function, such as photonic science coupling, or may augment a conventional central ZWFS response to provide additional wavefront information. Using a monochromatic scalar model, we derive the resulting carrier structure and show how the engineered references can be recovered through spatial demodulation and incorporated into calibrated reconstruction. A simple four-reference realization demonstrates that the reference geometry redistributes photon sensitivity across spatial frequency magnitude and orientation, producing mode dependent changes in linearity and cross-talk rather than a uniform improvement over a classical ZWFS. We then consider differential piston on a seven segment aperture as an application specific example. A hybrid sensor retaining the central ZWFS phase region and adding six engineered references that match the segment geometry gives similar precision near the phased state but substantially improved performance at larger piston excursions. Under the adopted photon-noise-limited benchmark, the hybrid reduces the nonlinear reconstruction error and provides better conditioned optical encoding of the piston state than an idealized vector ZWFS, allowing more precise piston estimation from the same detected photon flux at larger piston excursions.

We study byte-level autoregressive language models whose hidden state is a small quantum register, whose transitions are learned token-conditioned quantum channels, and whose next-byte probabilities come directly from Born-rule measurements without a learned classical vocabulary projection or softmax. In the fixed-width recurrent setting, unitary dynamics cannot contract or reset stored distinctions; experimentally, unitary recurrence loses to its retrained one-token control in every tested variant. Introducing learned dissipation, then token-conditioned dissipative gating, yields useful memory beyond one token. The gated model improves test negative log-likelihood by 0.20 +/- 0.03 nats/byte over retrained parameter-matched one-token controls at context length 12, and by 0.29-0.33 at longer contexts, but never beats retrained two-token controls. A key-value recall benchmark shows a width-dependent threshold: one stored pair meets the pre-registered accuracy bar at 4-5 memory qubits, two pairs at 6-7, and three pairs remain unsolved. This is consistent with, but does not establish, a capacity explanation for the short effective memory horizon. As a secondary result, one-qubit models reach lower stationary memory entropy than a heuristically optimized classical predictive frontier at matched error in 9/9 seeds. On IBM Heron hardware, computational-basis memory-register distributions show signal against a uniform null in 12/12 circuits across prefixes of up to 12 tokens, next-bit probabilities pass pre-registered accuracy bars at four-token prefixes, and bytes are generated autoregressively by sampling physically measured bit probabilities. The tested models are classically simulable and substantially worse than classical n-gram baselines. We claim no computational quantum advantage.

We study strictly positive multivariate binary distributions by separating one-body marginals from dependence. Local positive reweighting defines a dependence nucleus, completely characterized by the higher-order Walsh log-linear coefficients $J_S$, $|S|\ge2$. Any such nucleus and any strictly positive one-body marginals determine a unique joint distribution. In finite-spin Gibbs systems, these coefficients are the many-body interaction couplings.

We study the difficulty of finding a new solution when a list of distinct solutions is already supplied. For any fixed finite collection of list lengths, we can prescribe the additional-solution degrees while preserving a given $1$-paddable domain and first-solution degree. We also prove an exact compatibility condition for prescribing an arbitrary positive-list degree at the same time: it must dominate every prescribed fixed-level degree, and this condition suffices. All degrees use polynomial-time reductions with at most one oracle query and correct behavior on empty fibers; the realization needs no assumption about P versus NP. The combined construction gives each underlying witness exponentially many short counter copies. A list that exhausts these copies is long enough to pay for exhaustive search through the underlying witness bound. Separately, we construct a total problem whose first-solution task and every fixed-length additional-solution task are polynomial-time solvable, while its arbitrary-list task is FNP-complete. Assuming P differs from NP, the domain, first degree, and entire fixed-level degree sequence therefore do not determine the arbitrary-list degree. A constant fiber cap restores determination by finitely many fixed levels. The constructions change witness sets and do not prescribe an infinite sequence of degrees or enumerate all remaining solutions.

Assuming that $\zeta(s)$ has no nontrivial zero with $\mathrm{Re}\, s > 7/8$, as asserted in a recent unrefereed preprint of OpenAI, we obtain $\Lambda \le 13963/100000 = 0.13963$ for the de Bruijn-Newman constant. A published Polymath boundary computation supplies a finite-region strip of half-width $1/5$. We derive an approximation on $3 \le \mathrm{Im}\, z \le 6$ from its general Riemann-Siegel coefficient bounds. At height $61/20$, a fixed cutoff permits two differentiations, and comparison with $\zeta$ bounds both the zero-density kernel and its vertical derivative. A new conjugate-pair inequality and a twenty-four-step strip comparison give the stated bound. The finite scalar and polynomial certificates are included in the paper; earlier comparisons giving $0.17973$, $0.17172$, and $0.16408$ are retained to show the source of the gain. The argument uses neither Gordon's analytic inputs nor Gomila's new computations. The seven-eighths hypothesis remains an assumption, and the published numerical boundary input is retained.

A conversational speech model has to decide, at every pause, whether to speak. Existing evaluations test boundary detection, policy adherence, or aggregate statistics, not this decision. We place real-time speech systems at 427 pauses and 482 turn ends from 176 unscripted conversations in Seamless Interaction, give each the conversation so far, and score what it generates against what the person did: whether and how soon it speaks at a pause; its latency, length, and wording at a turn end; and whether it can answer a question about the preceding 30 to 120 seconds. Across six systems, the hosted APIs (GPT Realtime, Gemini 3.1 Flash Live, Grok Voice) respond at almost every pause whichever endpointer they use; only their latency, about 1 to 4 s with silence-based endpointers and about 8 s with semantic turn detection, decides whether the words land inside the pause. The full-duplex models (PersonaPlex, Moshi) are the only systems that regularly stay silent and the only ones that interrupt within a human reaction time. Hosted systems reply longer than the person. Given the history as text, hosted systems recall 41 to 60% of conversation facts under a Claude Opus 5.5 judge and 19 to 26% under a strict word matcher; hearing it as audio, full-duplex models recall 17 to 22% and 12 to 15%. We release the labels, points, harness, and every output.

Approximating the effective anisotropic hyperelastic response of a representative volume element (RVE) requires constitutive expressiveness, mechanical consistency, and computational efficiency. This work develops a polyconvex physics-augmented neural network (PANN) construction based on paired directional features whose orientations, stretch powers, and determinant exponents adapt to the RVE response. The features are composed with convex, componentwise nondecreasing cores, realized as an input-convex neural network (ICNN) or an input-convex Kolmogorov-Arnold network (ICKAN). The complete energies satisfy objectivity, an exactly stress-free reference, two-dimensional polyconvexity, and prescribed growth conditions by construction, and nonnegative energy is certified separately for the selected models. On a multicavity and a single-cavity periodic cell, these energies are compared with an Unconstrained energy, trained identically without the curvature and growth constraints, and with three hyperreduced projection-based reduced-order models (HPROMs), which serve as physics-based references. On the first cell, learning the features lowers the stress error sevenfold, and six features suffice on both. The constrained energies remain below one percent in stress error and in the displacement error of a tensile-coupon FE$^2$ application with a full-order reference, whereas the Unconstrained energy and the HPROMs are more accurate inside the data. Beyond the data, however, the Unconstrained energy loses rank-one convexity, and under confined compression it stops below half of the 500 MPa that the constrained energies reach. These results quantify what the guarantees cost in accuracy and what they secure beyond the data.

Frustrated magnets host rich emergent behavior because competing interactions generate dense manifolds of low-energy states. The Ising Shastry-Sutherland model is a paradigmatic setting, motivated experimentally by Ising-like rare-earth tetraborides such as TmB$_4$ and ErB$_4$, whose field-induced magnetization plateaus have motivated extensive microscopic modeling. Yet how tunable quantum fluctuations reorganize this material-motivated phase diagram remains largely unexplored. Here, we realize a transverse- and longitudinal-field Ising model on the Shastry-Sutherland lattice using a programmable Rydberg-atom array. We map its quantum phase diagram using density matrix renormalization group calculations and probe corresponding signatures experimentally in arrays containing more than 110 atoms. Beyond the established classical phases, the calculations identify two regimes absent in the classical limit and stabilized at finite transverse field: a $2/5$ filling plateau whose finite stability region emerges from a classically degenerate phase boundary and a symmetry-preserving quantum-dimer regime replacing classically selected collinear order. Experiments on finite open arrays reveal signatures consistent with these regimes across three boundary-pinning conditions, with the resulting response maps agreeing qualitatively with the calculated phase diagram. Our results establish programmable Rydberg-atom arrays as tools for exploring frustrated, material-motivated Hamiltonians and motivate searches for new, quantum-driven magnetization plateaus and quantum-dimer behavior in ErB$_4$ under a transverse field or in strain-tuned TmB$_4$.

Molecular dynamics simulations of solid polymer electrolytes require force fields that are fast enough to converge ion transport over hundreds of nanoseconds, yet accurately describe the ion-polymer interactions that govern it. Fixed-charge force fields meet the first requirement but neglect electronic polarization, an effect often approximated by scaling ionic charges using empirically chosen factors. Here we present AutoDrude, a differentiable framework that parametrizes polarizable Drude force fields directly from DFT energies, forces, and dipoles, with no empirical tuning and at only 1.20 times the cost of a fixed-charge simulation. For PEO-LiTFSI, the dipole-constrained (hard-constraint) AutoDrude model reproduces the experimental ionic conductivity as accurately as the best charge-scaled model (17% versus 16% mean absolute percentage error). We compare the polarizable forcefield to fixed charge ones models by fitting a fixed charge model. Unlike the charge-scaled models tested here, however, AutoDrude also reproduces the dilute-limit cation transference number, its increase with salt concentration, and the absence of Li-TFSI contact ion pairs at low concentration. A penalty-based (soft-constraint) variant yields a closer fit to the DFT reference yet underestimates the conductivity by up to an order of magnitude, demonstrating that closer agreement with DFT training data alone does not guarantee accurate ion transport. These results indicate that although charge scaling can reproduce ionic conductivity the explicit inclusion of polarization is necessary to capture the concentration dependence of cation transport in PEO-LiTFSI.

The staggered order parameter of the gapped XYZ spin-$1/2$ chain follows from the eight-vertex solution, but in elliptic variables that fix the Hamiltonian couplings only implicitly. We derive a closed form for the coefficient governing the leading quadratic response of the spontaneous staggered magnetization to the XY anisotropy at fixed longitudinal anisotropy in the gapped antiferromagnetic regime. Starting from the Baxter–Kelland–Jimbo–Miwa–Nakayashiki polarization formula, we show that at fixed longitudinal anisotropy both the modular nome and the crossing-parameter variable of the formula move with the XY anisotropy, and that the motion of the crossing-parameter variable reverses the sign of the expansion at fixed elliptic variables. The coefficient is a rational term plus a Lambert series. It is strictly negative and bounded above by a rational function of the longitudinal anisotropy, to which it is asymptotic in the Ising limit. At the isotropic point it diverges as the inverse five-halves power of the distance to that point, and since the Lambert sum has an asymptotic expansion with exponentially small remainder, the subleading terms of the divergence also follow in closed form. When the XY anisotropy vanishes as the five-fourths power of the distance to the isotropic point, the ratio of the perturbed to the unperturbed order parameter tends to a Gaussian in the scaling variable, and the quadratic truncation is controlled when that variable is small. A fourth-order perturbative calculation in the spin chain, with no input from the eight-vertex solution, reproduces the leading strong-coupling behaviour, and infinite-DMRG calculations at four anisotropies in the gapped regime are consistent with the closed form.

We establish a Carathéodory-type extension theorem for premeasures with values in general separated abelian net convergence groups. Starting from hereditary exhaustivity on the source ring and intrinsic conditions for combining null errors, we construct a countably additive extension on a regulated Boolean sigma-algebra, with full upper and lower regularity in the original convergence. Completeness is required only on the limit-stable hull of the source range. Under closed summation for arbitrary directed rows and finite-response reflection, we identify the regulated domain with the splitting class of Sion's inner and outer processes. For a Dedekind-complete real vector lattice, we show that sequential closed summation on a bounded interval is equivalent to weak sigma-distributivity of the generated band, giving a converge-theoretic interpretation of the concept. Source representations allow us to recover the partially ordered-group extension of Volauf and the finite-variation extension of Kusraev and Malyugin from the general theorem. We then treat finite semivariation by adjoining dominated coordinate variations to an additive lift of the premeasure. The resulting extension is defined on the full semivariation regulator domain and preserves semivariation exactly, without requiring decomposability or the countable sup property. It includes measures of infinite vector variation.

Audio-visual speech enhancement (AVSE) improves speech intelligibility in noisy environments but often relies on computationally expensive models unsuitable for resource-constrained edge devices. This paper introduces a reinforcement learning-driven neural architecture search (RL-NAS) framework to discover compact, efficient AVSE architectures. The search space preserves temporal alignment between audio and visual streams, while a novel reward function jointly optimises speech enhancement performance, measured by SI-SNR improvement, and model compactness. Using proximal policy optimisation with constraint-aware action masking, the framework evaluates only 35 candidate architectures and identifies a model with 2.9 million parameters, compared with 22 million in the baseline. The proposed architecture achieves a 14.5 dB SI-SNR improvement, only 0.7 dB below the baseline, while reducing model size by approximately nine-fold. Evaluation on commercial mobile CPUs demonstrates a 4.3-fold reduction in real-time factor, supporting practical on-device deployment. Compared with random search, RL-NAS achieves 92% higher search efficiency. Architectural analysis reveals that shallower but wider temporal convolutional blocks and aggressive visual pathway pruning can maintain enhancement performance while substantially reducing computational requirements. These findings demonstrate the potential of reinforcement learning-guided architecture search for developing lightweight, efficient and deployable multimodal speech enhancement systems.

Hardware malware detectors (HMDs) classify programs as benign or malicious based on hardware performance counter (HPC) data. Previous studies have shown that ensemble classifiers outperform other classifiers on detection metrics. However, the area, power, and timing costs of HMDs remain open questions. We address these questions across a comprehensive design space comprising multiple classifier families, variable HPC sets, and ensemble learning methods. We designed a pipeline to reconstruct selected trained models, generate synthesizable register-transfer-level (RTL) implementations, validate each design against a bit-accurate golden model, and evaluate the implementations. The results show that tree- and rule-based detectors achieve the highest detection performance with nearly an order of magnitude less area and several times lower energy than function-based classifiers. Reducing the number of HPC inputs to fit a processor's concurrent-counter limit preserves detection performance while substantially reducing area, power, and delay.

Large Language Model (LLM)-guided evolutionary optimization has successfully improved prompts, programs, and even entire AI systems. Most optimization methods select a candidate from the search frontier, then mutate it in one LLM pass based only on its failures. This focus on the frontier can stall on production AI systems, where isolated failures reveal little about complex, hidden interactions. We introduce CoCoEvolve (Cortex Code Evolve), an evolutionary optimizer built around what we call meta-reflection, wherein an agentic mutator (Cortex Code) selectively retrieves evidence from the entire evolutionary history and probes the live system before verifying and proposing mutations. We apply CoCoEvolve to optimize three classes of production AI systems: data agents, LLM-powered data pipelines, and data engineering workflows. On DABStep, CoCoEvolve tops the public leaderboard, reaching 92.8% held-out hard accuracy compared to 78.4% for Meta-Harness and 30.4% for GEPA under matched conditions. Across six LLM data pipeline workloads, it outperforms every evaluated baseline. On data-eng-bench, CoCoEvolve lifts test-set solve rate from 59.1% to 77.4%, past the best baseline of 73.6%. It also has 11 percentage points higher accuracy than GEPA on HotpotQA while remaining competitive with specialized optimizers on established mathematical and systems benchmarks. Matched ablations confirm that each component of CoCoEvolve contributes to these gains, demonstrating the value of meta-reflection for optimizing production AI systems.

Pretrained vision-language-action (VLA) policies fail on memory tasks, behaviour that depends on events observed long before the current frame, because their training data is overwhelmingly Markov and mapping episodic memory directly to action chunks is hard. Yet the same policy, finetuned once to follow a privileged oracle's grounded-subgoal text, nearly solves these tasks: the difficulty lies entirely in mapping memory to a grounded symbolic subgoal. Existing work entrusts this stage to a single VLM that selects keyframes and reasons over raw frames, and falls far short of the oracle: key identities and dynamics are buried in noisy pixels, the reasoning spans hundreds of frames in an unbounded context, and hallucination corrupts the memory itself. We instead present GAMMA (Grounded multi-Agent Memory with Mechanical Audit), which builds the stage as a collaborating two-agent memory system: a detector tool (SAM-3) for spatial grounding, a VLM writer that distils each replan window into one grounded text line, an append-only text memory bank, and a VLM reasoner that reads the bank and emits the next grounded subgoal at the oracle's cadence, wrapped in a propose-verify harness that admits, defers, corrects, or rejects every agent claim against mechanical visual evidence, at training-corpus construction and at deployment alike. On a sixteen-task memory-manipulation suite GAMMA recovers a large fraction of the oracle ceiling with the policy held fixed, reaching the ceiling on several tasks, and transfers with unchanged contracts to a second benchmark.

Technologies like large language models make it easier to convert cultural material into analyzable variables, while LLM-driven agentic systems extend machine assistance to other parts of the research workflow. Rather than a full how-to guide or an exhaustive review of recent LLM use cases, this chapter looks at some crucial aspects of this new paradigm as well as the bigger picture. I argue that these technologies expand the scale, methodological range, and scope for delegation, while the basic logic by which those measurements support cultural claims remains mostly unchanged.

Altermagnetism, as a novel magnetic paradigm, not only possesses a zero net magnetic moment, but also exhibits a substantial nonrelativistic spin-splitting effect. Currently, developing a simple and experimentally feasible method for designing altermagnets is eagerly pursued. Moreover, achieving 100% spin polarization along certain directions in valence or conduction bands for altermagnets remains challenging. To address these issues, based on a tight-binding model, we propose a viable chemical strategy for constructing organometallic altermagnets by using ambidentate ligands. Furthermore, in these organometallic altermagnets, direction-dependent 100% spin polarization along the [110] direction in band edges can be induced via the coexistence of altermagnetically spin-split dispersive and flat bands, where the flat band suppresses electron transport in one spin channel and the dispersive band enables conduction in the other spin channel. Then, via first-principles calculations, a screening of 90 two-dimensional ambidentate-ligand based square organometallic lattices leads to the identification of three altermagnetic materials, namely Fe(taz)2Cl2 (taz = 1,2,4- triazine), Co(app)2Cl2, and Fe(app)2Cl2 (app = 1,4-azaphosphinine), all of which possess 100% spin polarization along the [110] direction in valence bands, as well as notable piezomagnetic behavior. This study paves the way for achieving direction-dependent fully spin-polarization in altermagnetic materials.

Reliable robotic manipulation requires maintaining a firm grasp while strictly avoiding both contact slip and structural damage to the object. Satisfying these inherent secure-and-gentle grasping constraints is challenging for unknown objects. We present a novel online data-driven predictive control framework for safe dynamic robotic grasping of unknown objects. The proposed system features a pair of compact, single-channel tactile-sensing modules that characterize the gripper-object dynamic interaction in real time, entirely bypassing the need for prior knowledge or offline training. A finite-state-machine-based grasping force probing mechanism is developed to establish the initial grasping force and explicitly identify grasping force constraints. Through online dynamics learning, a Gaussian Process (GP) model captures gripper-object interaction dynamics. The model uncertainties are further used to form robust grasping constraints, which are embedded into a model predictive control scheme. We demonstrate error convergence in a probabilistic sense. The proposed control framework runs at 100~Hz in real time to dynamically manipulate and handle objects rather than static holding. Extensive experimental evaluations across diverse contact surface topologies and various types of objects validate superior performance and efficiency compared to baseline strategies. The proposed method is able to manipulate objects ranging from delicate items, such as raw tofu and marshmallows, to rough objects, such as rocks.

Self-evolving LLM agents typically improve how they act through accumulated experience, reflection, or skills. Yet even procedurally capable agents can fail when their beliefs about environment dynamics are incomplete, incorrect, or outdated. We study a complementary axis of self-evolution: evolving those beliefs during deployment. Such beliefs can support multiple tasks, but incorrect ones can also propagate across them. We introduce EvoBelief, a self-evolving world model that represents environment knowledge as falsifiable hypotheses. Unexpected outcomes and decision-relevant unknowns trigger hypothesis formation. Later observations and controlled experiments provide independent validation before a rule can guide predictions or plans within its supported scope. Counterexamples narrow or retire rules as the environment changes. The language model remains frozen, and each rule remains linked to its evidence and the decisions it changes. Across four benchmarks spanning static dynamics, hidden drift, and scientific discovery, EvoBelief improves task performance and reuses learned rules across tasks. On ALFWorld without actor-side transition memory, it improves success over the matched Frozen World Model by 23.9% and 31.3% with Claude Haiku 4.5 and GPT-5.4, respectively. On MiniGrid, it revises stale rules and recovers after unannounced dynamics changes. In DiscoveryWorld's Reactor Lab, it discovers and validates quantitative laws, completing 83.3% of Normal-tier and 58.3% of Challenge-tier instances, versus 0% for each evaluated baseline. These results demonstrate self-improvement through testable, revisable environment knowledge rather than model-parameter updates.

Which information actually shapes a language model's judgment when sources disagree? We study epistemic cue utilization through a controlled benchmark of 500 scenarios and five complementary tracks separating cue identification, cue influence, and competition between source, temporal, and presentation cues. Across 13 LLMs, nearly identical recognition of relative source reliability can accompany effects ranging from negligible to complete choice reversal when source histories are exchanged. The resulting profiles distinguish models that rely on source history from those more responsive to evidence lineage, event time, or presentation order. We then intervene on three Qwen checkpoints using matched sham material, fact restatement, structured evidence records, and an additional weighting principle. Restatement increases reliability sensitivity in the two 27B models, and adding the principle increases it by a further 39.2 and 47.2 percentage points relative to the ledger. The 9B model instead changes its sensitivity to evidence lineage without a comparable reliability response. Verified-state accuracy remains 100% throughout these interventions. Together, the benchmark and interventions show that identifying an epistemic cue and allowing it to influence judgment are distinct, experimentally measurable aspects of model behavior.

A normalized sequence $(y_n)$ in the dual $E^{*}$ of a Banach space $E$ is called normal if it is weak-star null, and strongly normal if [ G(y_n)=\left{x\in E:\sum_{n=1}^{\infty}|y_n(x)|<\infty\right} ] is dense in $E$. We prove that, within the class of Grothendieck Banach spaces, reflexivity is characterized by the subsequence behavior of normal sequences: a Grothendieck space $E$ is reflexive if and only if every normal sequence in $E^{*}$ contains a strongly normal subsequence. If $E$ is nonreflexive and Grothendieck, then $E^{*}$ contains a normalized weakly null sequence none of whose subsequences is strongly normal; more strongly, it contains a closed hereditarily $\ell_{1}$ subspace $Y$ with a normalized weakly null Schauder basis such that no sequence in $S(Y)$ is strongly normal. We also obtain an exact quotient criterion valid for arbitrary Banach spaces: a normal sequence has a strongly normal subsequence precisely when, for some subsequence $z=(z_n)$, the quotient $E/\bigcap_n\ker z_n$ is separable. The proofs combine a weak-star basic-sequence argument with results of Haydon, Johnson–Maurey–Schechtman and Popov. Concrete realizations are given in $(L_{\infty}(\mu))^{*}$ for nonatomic probability spaces (and, more generally, for $\sigma$-finite spaces with a nonatomic part) and in $\ell_{\infty}^{*}$.

3D reconstruction of building components using image-based scanning requires appropriate segmentation of the objects in the 2D building images. In order to benefit from consistent grouping of the 2D masks in 3D environments, vision foundation models such as the Segment Anything Model have been utilized by Group Anything with Radiance Fields (GARField) to perform 3D segmentation. This study fine-tuned SAM 2.1 on the HBD dataset, consisting of 2,679 annotated images from 42 buildings, to enhance multi-class segmentation of 2D building images, covering 11 classes (Wall, Window, Ceiling, Door, Column, Beam, Floor, Opening, Curtain Wall, Roof, and Lift). The model was evaluated using per-class IoU, Dice, and mIoU metrics, and the results show an improvement in mIoU from 49.9% to 69.6%, indicating a meaningful improvement in class segmentation and providing a more accurate model for distilling the masks in 3D. SegFormer was then used as a baseline, generating per-class masks for comparison with SAM 2.1, which indicates improvement in some object classes with larger surface area. The masks in the building images were grouped in the neural radiance field, and the pretrained SAM 2.1 was compared with the fine-tuned SAM 2.1 in the newly rendered images. This comparison shows that fine-tuning improves the grouping of building components at a user-selected grouping scale. This advances the application of Group Anything with Radiance Fields (GARField) in 3D, enabling better segmentation of building structures for 3D auto-reconstruction.

We construct closed Levi non-degenerate CR manifolds with non-vanishing Chern monomials. For every $n \geq 2$ and $\lceil n / 4 \rceil \leq m \leq \lfloor n / 2 \rfloor$, there exists a $(2 n + 1)$-dimensional example of signature $(\lfloor n / 2 \rfloor, \lceil n / 2 \rceil)$ on which all Chern monomials with positive indices at most $m$ and total weight at most $n - m$ are non-zero in cohomology with real coefficients. Taking $m = \lceil n / 4 \rceil$ gives examples with $c_{1}^{\lfloor 3 n / 4 \rfloor} \neq 0$. For $n = 4$ or $n \geq 6$, the latter examples show that the Chern-class vanishing theorem for strictly pseudoconvex CR manifolds does not extend to the Levi non-degenerate case. In these dimensions, their underlying manifolds admit no strictly pseudoconvex CR structures. The construction uses circle bundles over products of integral pseudo-Kähler manifolds.

We propose a low-complexity SVD-based transceiver for zero-padded memory-$1$ ISI channels. Appending a single zero-padding symbol to an $n$-symbol block yields an $n\times(n-1)$ channel matrix with an exact symmetric tridiagonal Toeplitz Gram matrix. This structure provides closed-form singular values and vectors, enabling exact SVD precoding and equalization via FFTs without explicitly forming dense SVD matrices. The resulting implementation has $O(n\log n)$ complexity and incurs only a $1/n$ rate penalty. The resulting parallel scalar subchannels enable joint adaptive PAM and power allocation for SER minimization, followed by exact soft-information computation for LDPC decoding. At comparable code rates near $0.945$, simulations show that our proposed scheme outperforms a conventional TX-FFE/RX-DFE system with hard-decision Reed-Solomon (RS) decoding by approximately $5$ dB at FER $10^{-5}$. At high SNR, it achieves nearly the same FER performance as a SOVA-based receiver with LDPC decoding, while replacing trellis-based detection with fast SVD precoding and exact scalar soft-output computation.

Referring Audio–Visual Segmentation (Ref-AVS) is an emergent multimodal task requiring the precise identification and segmentation of a specific object based on a natural language expression grounded in both visual and auditory cues. Unlike traditional AVS, which segments generic sound sources, or referring video object segmentation (Ref-VOS), which relies solely on visual–textual grounding, Ref-AVS necessitates fine-grained tri-modal interaction across vision, audio, and language. However, existing methods often fail to maintain consistency when subjected to target shift, a phenomenon where the model erroneously transfers its focus between objects due to transient occlusions, acoustic fading, or identity ambiguities. To address this, we propose a novel Shift-Aware Identity-guided Latent (SAIL) refinement framework. It introduces a dual-token strategy that decouples frame-level dynamics from a global semantic anchor to explicitly model target-shift cues. We design a shift-aware refinement module that rectifies drift by aligning tokens with domain-specific evidence. Finally, we leverage a decoder that integrates the refined embeddings with a continuous memory update to maintain spatio-temporal coherence and identity consistency throughout the video sequence. Extensive experiments conducted on two Ref-AVS benchmarks demonstrate the effectiveness of our proposed method, significantly outperforming existing AVS, Ref-VOS, and Ref-AVS approaches. The code will be released after publication.

We introduce coset lifted product (CLP) codes, a family of quantum LDPC codes that combine the protograph structure of lifted product codes with the coset action of Aydin, Tamo, and Barg \cite{AydinTamoBarg2026}. Given a finite group $G$, a subgroup $H \le G$ that need not be normal, a protograph $A$ over $\mathbb F_2[G]$, and a protograph $B$ over $\mathbb F_2[N_G(H)]$, the two protographs act on the cosets $G/H$ from the left and from the right to produce quantum CSS codes. Further, for groups of odd order, we give an explicit formula for the dimension as a sum over the irreducible representations $\rho$ of $G$ of hypergraph product dimensions. A search over $14,843$ pairs $(G,H)$ with $H$ restricted to be non-normal and $120 \le n \le 400$ yields certified codes that exceed the best published value of $Q=kd^2/n$ at check weights $6,7$ and $9$, i.e., we obtain $[[360,20,20]]$ at $w=6$, a $[[360,24,20]]$ code at $w=7$, and codes $[[252,16,18]]$ and $[[288,16,20]]$ at $w=9$. Our search also produce high-distance codes with parameters $[[168,10,14]], [[240,8,20]]$ at $w=6$ and $[[216,8,18]]$ at $w=7$.

Aligning two independently pretrained unimodal encoders typically demands millions of paired examples. In this paper, we study the regime where only $N \sim 100$–$500$ pairs are available, while unpaired data on either side is abundant. Our method expresses every representation by its similarities to the $N$ paired anchors, which places both modalities in a common coordinate system; matches the two point clouds with a single entropic optimal transport problem solved over the entire pool, whose marginal constraints counteract the hubness that defeats nearest-neighbor matching; and distills the resulting plan into two lightweight encoders, so that the correspondence becomes inductive. We additionally put this marginal constraint in relation with the one-step hub corrections of the bilingual lexicon induction literature. Pairing is required only for the anchors: the unpaired collections may be drawn independently from the two marginals, so that no item outside the anchor set needs to have both of its modalities observed. We evaluate on image-text and molecule-text alignment, where our method outperforms existing approaches at the same anchor budget on Flickr30k, MS-COCO and ChEBI-20.

This paper presents the computational architecture and quantitative preclinical evaluation of MIRAGE, a multimodal image-guided robotic surgery framework for real-time visualization and navigation. It integrates stereovision scene reconstruction, augmented-reality (AR) video overlay, quantified heads-up display (HUD) navigation, and intraoperative ultrasound within the da Vinci console's TilePro display (Intuitive Surgical). Additional modules include optical/electromagnetic tracker registration, stereo calibration, fiducial rigid registration with RMSE feedback, and 3D point-cloud reconstruction. Video I/O and rendering run in a C++ backend, while a Python GUI coordinates calibration, registration, reconstruction, visualization, and user interaction through an in-process zero-copy binding layer. The framework formalizes spatial transforms, module states, and synchronization of heterogeneous video and tracking streams. Preoperative calibration uses synchronized checkerboard image pairs and optimization, with estimated intrinsics, extrinsics, and rectification reused intraoperatively. MIRAGE was validated in phantom experiments across da Vinci S, Si, Xi, and SP platforms. Across clinically relevant working distances, mean 3D reprojection error was 0.6mm, with best 0.1mm RMSE at 100mm on Xi. Results support MIRAGE as a preclinical technical-validation platform for accurate, responsive, and extensible multimodal surgical guidance.

Large-scale strategic systems, such as maintenance workflows, cyber campaigns, and multi-agent AI pipelines, consist of many interactions whose outcomes determine which later interactions become possible, who takes part, and with what resources. We introduce game Petri nets, formalized as outcome-driven compositional game Petri nets (OCGPNs), to model and analyze such systems. Local games are attached to places of a colored Petri net, and a game-to-token interface converts each resolved game into typed tokens that enable, synchronize, or block subsequent games. For general, concurrent nets, we give an operational semantics and a structural theory based on invariants and Farkas certificates, which shows that guards can certify unreachability whereas incentive instruments can only change values. For sequential nets, the net induces a stochastic game on its game places. We prove that global equilibria are exactly collections of equilibria of local games augmented by continuation values, characterize through successor-level pivotality when downstream games change upstream behavior, and compute values place by place via a decomposition into strongly connected components, at a cost additive in local game sizes. A composition calculus based on place fusion glues certificates across subnets and summarizes downstream subnets by their entry values. Two running examples, a maintenance workflow and an intrusion campaign derived from multi-resolution cyber-defense games, illustrate the theory, including the composition of a newly discovered attack scenario with an existing campaign.