Computer Vision
Papers
Fresh from arXiv, typeset properly. Highlighted words trade on the market.
All of arXiv
Monday 14 September
Deep networks trained on structural MRI for Alzheimer's disease (AD) staging often reach reasonable accuracy while attending to anatomically irrelevant regions, and multimodal models that add clinical tables frequently rely on variables that were used to assign the diagnostic label in the first place. We study both issues with a deliberately lightweight slice-based encoder (ResNet18 with a one-layer Transformer over slices) on 1,075 baseline T1-weighted scans from ADNI-1. First, we use FastSurfer segmentations as an anatomical reference: YOLOv8 models trained on segmentation-derived labels localize Alzheimer-relevant structures with mAP_50 above 0.96, and a Grad-CAM comparison shows that the image-only classifier frequently attends to the skull, orbits and background. Second, we adapt a CLIP-style image - tabular contrastive framework and organize ADNIMERGE variables along a label-leakage spectrum. Fusion with cognitive scores yields 87.3% three-way accuracy, which we treat as a leakage-driven upper bound rather than an imaging result; fusion with regional volumes yields 73.0%. We observe that the choice of contrastive target changes what the image encoder learns: on MCI vs. CN, the image-only head reaches 52.4% when the encoder is aligned to cognitive scores and 73.8% when aligned to volumes, although no tabular input is used at inference. Third, restricting the input to a per-subject crop of the medial temporal lobe raises image-only three-way accuracy from 58.7% to 65.1%. All results come from single runs on a small balanced test set, and we report confidence intervals and the protocol differences that prevent direct comparison with published numbers.
Cryptography and Security
The Model Proposes, the Code Disposes: A Pre-Registered Ablation of a Verifier-and-Acceptance Stage in an LLM-Orchestrated Offensive-Security Agent
We evaluate whether a verifier-and-acceptance stage - a model verifier whose verdicts are enforced by deterministic code - changes what an LLM-driven offensive-security agent reports. We report a 15-run exploratory pilot, a pre-registered 20-run confirmatory ablation, and a pre-registered 2 x 2 factorial study with 40 runs across two deliberately vulnerable lab targets. In the confirmatory study, removing the stage eliminated pre-report suppression (median 2 versus 0 findings per run; exact one-sided p = 0.00003) and reduced model-blinded shipped precision (median 0.471 versus 0.353; p = 0.0087). Recall against a frozen but incomplete ground-truth list did not differ significantly (two-sided p = 0.158; equivalence was not established). The factorial study attributed suppression to the model verifier (Holm-adjusted p = 0.004); deterministic acceptance rules alone suppressed no false positives, and no interaction was detected (p = 0.72). The full design retained 93.8% of model-adjudicated true candidates but did not meet its pre-registered non-inferiority criterion because the lower one-sided 95% bound was 0.875, below the 0.90 floor. Across the confirmatory and factorial studies, an instrumented canary recorded zero contacts in 60 of 60 runs, with incidental external contacts disclosed separately. Independent human adjudication of the retained blind packets is pending, so precision and sensitivity endpoints are supporting rather than final evidence. Six audit-trail failures, including one in the evaluation tooling, are also disclosed. The results support a narrow conclusion: the verifier changes what the system ships, while deterministic code supplies enforcement and auditability; they do not establish superiority to other agents or generalization beyond lab targets.
Read the full abstract
Computation and Language
Inoculation Midtraining with Learned Neologisms
Large language models (LLMs) often learn both desirable and undesirable properties during post-training. We study whether midtraining, an earlier training stage, can shape which of these properties later generalise. We introduce Inoculation Midtraining, a technique that teaches a base model that unsafe behaviour belongs to a designated <quarantine_token> context, as indicated by the <quarantine_token> neologism (a new token) introduced during midtraining, and then post-trains the model on unsafe data within that context. We then evaluate the model outside the context, with the <quarantine_token> neologism excluded from the system prompt. Across supervised fine-tuning and reinforcement learning post-training regimes, we find that Inoculation Midtraining can reduce misalignment while preserving the transfer of benign data properties (e.g., speaking in German or Shakespearean prose). However, our approach does not outperform standard Inoculation Prompting, is sensitive to training configuration, and produces a leaky boundary that nearby contextual cues can reactivate. These results show that inoculation with a learned association introduced via midtraining can shape selective generalisation. Still, more work is needed before this approach can become a load-bearing component in a developer's safety framework.
Read the full abstract
Machine Learning
Privacy-enhanced federated learning via asynchronous aggregation and local differential perturbation
This study proposes a privacy-enhanced federated learning framework to address secure collaborative training in distributed data environments. The framework integrates Dynamic Differential Privacy (DDP), lightweight Homomorphic Encryption (HE), and Local Differential Privacy (LDP) mechanisms to ensure data privacy protection during model training. Additionally, the framework employs an asynchronous aggregation strategy with version control to support distributed training in asynchronous environments. Experimental validation on the CIFAR-10 and Purchase-100 benchmark datasets demonstrates that the method maintains high classification accuracy (up to 82.6%) even under stringent privacy constraints ({\epsilon} = 0.1), while reducing communication overhead by 21.3% compared to FedAvg. Experimental results demonstrate that this framework effectively balances privacy protection and model performance in distributed machine learning scenarios, providing a scalable technical foundation for large-scale distributed collaborative computing.
Read the full abstract
Data Structures and Algorithms
Thin-shell stability of Gaussian cooling: logconcave sampling with sesteric complexity from a cold start
We show that logconcave probability measures along the Gaussian cooling path have thin-shell stability, generalizing the thin-shell theorem. This result leads to improved complexity for the fundamental problem of sampling an arbitrary logconcave distribution from a cold start. For (near-)isotropic logconcave distributions, the complexity is nearly $n^{2.5}$, improving the previous bound of $n^{2.75}$, and matching the complexity of the abstract Speedy walk.
Read the full abstract
Machine Learning
Learning Multimodal One-step Flow Policy via Value-weighted Optimal Transport
Offline reinforcement learning aims to learn a policy solely from fixed datasets, which often contain multimodal action distributions. Flow policies can naturally represent such multimodal behaviors, but learning an efficient one-step flow policy remains challenging: standard value guidance often leads to mode collapse or exploits overestimation bias in out-of-distribution regions. To address this, we introduce One-step Flow policy via Optimal Transport (OptiFlow), a framework for one-step flow policy learning as a structured sample-allocation problem. OptiFlow jointly trains a value-aware reference flow policy and an efficient one-step policy, coupling their action samples through state-wise entropic optimal transport. For each state, critic-estimated values define the priority of distillation target actions, while the action-distance cost ensures geometrically compatible pairings. By avoiding direct critic maximization, our transport-guided approach enables in-distribution exploitation by anchoring the one-step policy to high-value, dataset-supported modes without the risk of out-of-distribution divergence. Experimental results demonstrate that OptiFlow effectively captures optimal multimodal behaviors and achieves strong performance across diverse offline RL benchmarks. Our code is available at https://github.com/Yonsei-DILLab/OptiFlow.
Read the full abstract
Analysis of PDEs
Uniqueness of positive solutions of double-power nonlinear stationary Schrödinger equations: the fixed frequency case and the fixed mass case
The existence of a positive radial solution to the problem \begin{align*} -\Delta u + \lambda u = u^p + u^q, \quad u \in H^1(\mathbb{R}^N), ~~ N \ge 2, \end{align*} where $\lambda > 0$ and $1 < q < p < 2^* - 1$, has been known for a long time. For the pure-power nonlinearity, it is well known that this solution is unique. In contrast, the uniqueness problem for the double-power case is substantially more delicate and depends on the dimension and the exponents. It has been known that the uniqueness result is in general not true for the equation with double-power nonlinearities in dimension three and it is a long-standing open question to prove the uniqueness of positive radial solutions throughout the full subcritical range for $N \ge 4$ since the work of H. Berestycki, P.-L. Lions [2] (Arch. Ration. Mech. Anal. 1983). In this paper we provide the uniqueness results for all $\lambda > 0$ and all $1 < q < p < 2^* - 1$ when $N \ge 5$. In dimensions $N = 2,3,4$, with additional conditions on $p$ and $q$, we obtain the uniqueness results for all $\lambda > 0$. Our results are in sharp contrast to those of J. Dávila, M. del Pino and I. Guerra [6] (Proc. London Math. Soc. 2013), where some non-uniqueness results were obtained for $N=3$. At the same time, we give a positive answer and rigorous proof to the implication of their numerical simulation, clarify the open issues left by [6]. Finally, we completely resolve the uniqueness problem of the above equation with $L^2$-mass constraints in the mass-subcritical and mass-critical regimes for all $N \ge 2$.
Read the full abstract
Materials Science
High-Temperature Hydrogen Sensors Based on Gallium Oxide Heterojunction Diodes
Long-term, high temperature operation of Ga2O3 devices is a crucial hurdle that must be overcome before widespread adoption of the technology can be achieved, but is largely absent from the overall body of work. Demonstrations up to this point show devices are either limited by material or dopant instability that leads to performance degradation with time. Herein, Ga2O3-based hydrogen sensors employing Pt Schottky and Cr2O3/Ga2O3 p-n diodes (Mg- and N-doped) were fabricated and evaluated for long-term stability at 600C for 800-1,800 hours, with cyclic exposure to N2 and low-concentration H2 (500-1,500 ppm). Transient current density (measured at -0.1 V) and periodic J-V characterization were used to track performance. Despite gradual declines in sensor signal and sensitivity, devices distinguished hydrogen concentrations throughout weeks of operation. Degradation was architecture-dependent: Cr2O3:Mg degraded gradually, consistent with known Mg migration; the Pt Schottky diode showed dramatic changes after 1,000 hours; and Cr2O3:N showed the lowest but most stable performance before failing at 800 hours. Thermionic emission and Lambert W-based modeling confirmed hydrogen exposure reduces interfacial barrier height via a proton-induced dipole mechanism common to both diode types. TEM of aged Pt Schottky diodes revealed Pt grain growth and microvoid formation as key degradation mechanisms. TOF-SIMS confirmed nitrogen dopants remain confined to the Cr2O3:N layer, supporting N-doping as a stable, lower-performance alternative to Mg-doping
Read the full abstract
Differential Geometry
Ricci flow with metric torsion on surfaces of positive Euler characteristic
We study an adapted Ricci flow of connections with metric torsion on surfaces with positive Euler characteristic. We first prove that there do not exist any nontrivial solitons of the flow on the $2$-sphere thus confirming a conjecture of Branding–Kröncke (J. Geom. Anal. 27.3 (2017), arXiv:1606.09121). We give an explicit family of torsion data for which the corresponding global solutions fail to converge on $\mathbb{S}^2$. Nevertheless, we provide several sufficient conditions for the convergence of the flow to a stationary point. We first prove that the normalized adapted Ricci flow always converges on $\mathbb{RP}^2$, which completely answers a question in the paper of Branding and Kröncke. Using this, we deduce that the flow converges on $\mathbb{S}^2$ whenever the initial metric and the torsion one-form are antipodally symmetric. We also prove a Łojasiewicz–Simon gradient inequality for the flow and use it to prove convergence to a stationary point provided the solution is close to an arbitrary stationary point.
Read the full abstract
Logic in Computer Science
Finite Dependence and Invariance Hierarchies for Finitely Supported Structures
Finite support indicates that the dependence is limited, but not how each finite context controls what can be observed. For a data symmetry G \leq Sym(A) and a finitely supported G-set X, we study the finite-dependence profile X^S = {x : Fix_G(S) fixes x}, indexed by finite contexts S \subseteq A. For relations on A^n this yields Boolean algebras B^{(n)}_{G,S} = Inv_n(Fix_G(S)), forming the invariance hierarchy. The hierarchy separates several nominal principles. Its meet law is the relation-algebra form of the Bojanczyk-Klin-Lasota least-support criterion, while level injectivity is fungibility. The independent join law governs composition across unions of contexts; it holds for locally oligomorphic automorphism groups of ultrahomogeneous structures in purely relational languages of arity at most two, but can fail for unary observations over ternary data. Freshness remains valid sort by sort, and its unrestricted some/any form characterizes full equality symmetry among closed groups. With least supports, orbit-finiteness implies finite context levels and a uniform support bound; under local oligomorphicity the converse also holds. Pointwise closure does not change the profile, and standard atomic-site presentations show that suitable ultrahomogeneous structures with the same age have equivalent action topoi, even though their material FSS universes may retain different external cardinal data. Contextual orientability gives a complementary pointed invariant whose threshold is the least support size of choice from unordered pairs. For countable omega-categorical structures with degenerate algebraic closure, the meet law is equivalent to weak elimination of imaginaries.
Read the full abstract
Algebraic Topology
Logarithm of the Universal Two-Valued Formal Group
We solve the long-standing problem of determining the exact denominators of the coefficients of the logarithm $B(x)=x+\sum_{n\geq1}b_nx^{n+1}$, obtained from the universal formal group of complex cobordism by the modulus square construction: $$ b_n=\frac{C_n}{d_n},\qquad d_n=\frac{n+1}{2}\operatorname{lcm}(1,\ldots,2n+2), $$ where $C_n\in\Omega_{\mathrm U}^{-4n}$ is primitive and undecomposable. We prove that $C_n$ belongs to the coefficient ring $\Lambda$ of the universal two-valued law and to the subring $\Lambda_{\mathrm{St}}$ generated by quotient Stong manifolds, and give an explicit integral Stong manifold formula for it. The rational lift of $b_n$ to quaternionic cobordism modulo torsion has exact denominator $2d_n$. Consequently, $C_n$ has no integral quaternionic lift, whereas $2C_n$ does. Every Chern number of $C_n$ is divisible by $d_n$, and $c_{2n}(C_n)=d_n$. We introduce odd genera $\mathrm{Bc}_N$. For $N\geq n$, $\mathrm{Bc}_N$ detects the odd part of $d_n$. Its restriction to the Stong ring is integral exactly for $N\leq3$, while the universal odd genus $\mathrm{Bc}_{\infty}$ takes values there with only odd denominators. For every $n\geq5$, we construct undecomposable classes in $\Lambda^{-4n}$ with equal Ochanine genera and top Chern numbers but distinct $\mathrm{Bc}_3$ values. Finally, the oriented extension of $\mathrm{Bc}_3$ is integral on closed spin manifolds of real dimension below $24$ and on closed string manifolds of real dimension at most $24$. The spin bound is sharp: an Anderson-Brown-Peterson spin $24$-manifold has nonintegral $\mathrm{Bc}_3$ genus.
Read the full abstract
Software Engineering
Using Agentic AI for contextualized and multifaceted code review at Ericsson
Context: Conducting effective code reviews is increasingly challenging due to the growing complexity of software systems and the accelerated code generation by AI coding agents. LLM-based approaches for code reviews have shown promising results in identifying defects and improving code quality. However, existing approaches rarely consider project-specific contextualized knowledge, and few have been evaluated in industrial settings. Objective: In this study, we propose a multi-agent-based solution that provides multifaceted assessments of code changes. Method: Following the Design Science Research Process, we developed and evaluated our solution in an industrial setting. Our solution combines specialized agent skills with context-specific knowledge to identify antipatterns in code changes across four dimensions: readability, maintainability, reliability, and performance. Using our solution, we generated reviews for several code commits and identified more than 200 issues. These issues were then manually validated by the developers of the case company for their correctness and importance. Results: The evaluation results show that our solution achieves 96% accuracy in correctly identifying issues in the investigated code commits. Furthermore, around 69% of the correctly identified issues were rated as important, with approximately 33% rated as severe issues that must be fixed and 36% as important issues that should be fixed. Qualitative feedback from developers corroborates these findings and highlights the usefulness of the generated reviews. Conclusion: Our findings provide empirical evidence from an industrial evaluation that combining specialized agent skills with context-specific knowledge yields accurate, practically useful code reviews.
Read the full abstract
Metric Geometry
On the covering of $n + \epsilon$ square with $n^2 + 1$ unit squares for $n \geq 4$
In 2006, Alexander Soifer conjectured that one cannot fully cover a square of side length $> n$ with $n^2 + O(1)$ unit squares. For small $n \in \{2, 3\}$, in 2009, Janusz Januszewski proved that it is impossible to fully cover a square of side length $> n$ with exactly $n^2 + 1$ unit squares. Recently in 2023, Baek and Lee proved that it is impossible to fully cover an equilateral triangle of side length $> n$ with exactly $n^2 + 1$ unit equilateral triangles whose sides are parallel to it. There were also some progress on a related problem: what is the largest square with side length $S(k)$ that can be fully covered by $k$ unit squares. However, there have been no improvement on the original conjecture. In this work, new tools have been developed that led to the proof of this conjecture for $n = 4$, with potential applications to related problems.
Read the full abstract
Information Theory
Private Information Retrieval With Arbitrary Privacy Requirements: Introduction and Capacity Results
In this paper, we introduce the problem of private information retrieval (PIR) under arbitrary privacy requirements, in a graph-based storage system. This formulation is motivated by the server storage limitations, abundance of data (messages) and heterogeneous data privacy requirements. Under the arbitrary privacy requirement, each message has to be retrieved privately from a pre-specified subset of servers, where the subset always includes the servers storing it. Thus, each server is associated with a privacy set, which pre-specifies the message indices that should be privately retrieved from it. This setting is a generalization of the classical PIR setting, where the required message index needs to be kept private from all servers, i.e., there, the privacy set of each server comprises all message indices. Our setting is also a bridge between the newly formulated local PIR (LPIR) setting and the classical PIR setting, where in the former, the privacy set is exactly the set of stored message indices. In this paper, we derive general lower and upper bounds on the PIR capacity for general graphs, under certain privacy requirements, that capture the essence of both LPIR and classical PIR. Then, we focus on path and cyclic storage graphs under these and more fine-grained settings, for which we derive capacity results for certain cases, and establish lower and upper bounds for others. Their low degree allows for a more in-depth understanding of the new privacy formulation and admits more privacy requirement settings compared to other simple graphs. Finally, we introduce a new graph structure, the pyramid storage graph, to model server storage. Although this graph has never been investigated in the literature in any PIR context, it enjoys a nice symmetric structure for message storage and replication patterns.
Read the full abstract
Representation Theory
Integral quiver presentations for $\mathscr R$-trivial semigroup algebras
Semigroups in which right divisibility is a partial order, that is, $\mathscr R$-trivial semigroups, have been studied by people in algebraic combinatorics in connection with Markov chains (especially left regular bands). Margolis and the author showed that the quiver of the algebra of an $\mathscr R$-trivial monoid is field independent. We prove here that the integral semigroup ring of an $\mathscr R$-trivial semigroup has a bound quiver presentation provided that it is unital. This then gives a uniform quiver presentation over any field via extension of scalars.
Read the full abstract
Combinatorics
Local measures of interval edge-uncolorability
An interval edge coloring of a graph is a proper edge coloring by integers such that the colors on the edges incident with any vertex form an interval of integers. Not all graphs are interval colorable; a simple counterexample is $K_3$. The (interval coloring) deficiency of a graph $G$ is the minimum number of pendant edges whose addition to $G$ yields a graph with an interval edge coloring. In this paper, we introduce and study further measures of how far from being interval colorable a graph is. The local deficiency of a graph $G$ is the smallest number of pendant edges that needs to be added at every vertex of $G$ in order to obtain a graph with an interval edge coloring; we can think of the colors of these added edges as ”locally missing” at a vertex. We also study a weaker version of this notion, the weak local deficiency, which informally is the size of a largest set of consecutive integers ”locally missing” at a vertex in a proper edge coloring of $G$ minimizing this size. We compare weak local deficiency, local deficiency, and deficiency, and show that the difference can be arbitrarily large in both cases. Moreover, we give concrete examples of graphs whose weak local deficiency (and thus local deficiency) grows with the number of vertices as well as with the maximum degree. We also prove some constructive results on graphs with small weak local deficiency. In particular, all complete multipartite graphs have weak local deficiency at most $2$, and many complete multipartite graphs have weak local deficiency at most $1$. Moreover, bipartite graphs with maximum degree at most $6$, and Eulerian bipartite graphs with maximum degree at most $8$ both have weak local deficiency at most $1$. We conclude the paper by pointing to several open questions for further research.
Read the full abstract
Astrophysics of Galaxies
Rubin Observatory Reveals a Dust-Shrouded Halo Globular Cluster in Ophiuchus
We present the discovery of the Milky Way satellite Rubin-GC1 (Rubin J1628-1735) using data from the NSF-DOE Vera C. Rubin Observatory's Early Data Preview 2. The system was discovered in the direction of the Galactic bulge with $(l, b) \sim (359^{\circ}, 21^{\circ})$ in a region of high stellar density and a foreground reddening of $E(B-V) \approx 0.7$. We simultaneously fit the morphology and isochrone parameters of Rubin-GC1, determining that it is a faint ($M_V = -2.4^{+0.5}_{-0.7}$) and compact ($r_{1/2} = 5^{+1}_{-1}$ pc) Milky Way satellite at a distance of $D_\odot = 31^{+3}_{-3}$ kpc; its stellar population is intermediate-age ($\tau=9.2^{+1.2}_{-0.7}$ Gyr) and moderately metal-rich ($\mathrm{[Fe/H]_{iso}}=-0.73^{+0.04}_{-0.04}$). Together, these properties suggest Rubin-GC1 is most likely a star cluster in the faint tail of the luminosity function of globular clusters surrounding the Milky Way. We identify a distinct proper motion signature from Rubin-GC1 in Gaia DR3, and from a comparison of the system's 5D phase space information to $N$-body models of the Sagittarius dwarf spheroidal galaxy's tidal stream, we determine that Rubin-GC1 is likely an accreted cluster associated with the Sagittarius leading arm. The discovery of Rubin-GC1 highlights the potential of Rubin to uncover faint, resolved satellites in dust-shrouded regions of the Milky Way, including remnants of the accretion events that have shaped the Galaxy's hierarchical assembly.
Read the full abstract
Machine Learning
LLM-Based Schema-Aware Split Learning for Privacy-Preserving Mental Distress Prediction Across Heterogeneous Surveys
Rising societal and lifestyle complexity has been linked to a growing prevalence of mental distress worldwide. Educational institutions, workplaces, clinics, etc. collect large volumes of mental health survey data to understand and reduce this burden. Collaborative analysis of such data could yield effective generalizable predictive models. Privacy constraints and varied survey designs (i.e., different questions, scales, and formats) hinder direct integration. We propose a schema-aware split learning (SL) framework that preserves privacy, using a large language model (LLM) as a shared semantic encoder to harmonize heterogeneous survey schemas across institutions. We serialize each survey record into a natural-language description, unifying disparate survey schemas into a common format. The LLM is fine-tuned for mental distress assessment via Low-Rank Adaptation (LoRA) and partitioned across client and server. Clients retain the raw survey responses locally and run only a lightweight front-end, so original records never leave the institution that collected them. The resource-intensive backbone runs on the server, minimizing client-side computation. Using LLaMA-3.2-3B-Instruct, the framework attains an average ANLS of 0.708 with only 2,000 training samples, surpasses federated learning (FL) in eight of nine settings, and cuts per-client computation by three orders of magnitude, while generalizing to unseen datasets. Overall, it enables accurate, privacy-preserving, and resource-efficient collaborative learning from heterogeneous mental health survey data.
Read the full abstract
Robotics
WLA$^3$: World Latent Action Modeling for Semantics, Dynamics, and Kinematics
Scaling generalist policy models with heterogeneous data is limited by the lack of unified, low-noise action supervision. Human egocentric videos are abundant, but only a small fraction comes with high-quality hand-action labels. Observed world transitions offer a common source of action-related supervision across data sources. We introduce WLA$^3$ (World Latent Action Modeling for Semantics, Dynamics, and Kinematics), a unified generalist policy model framework built around representations learned by a World Latent Action Model (WLAM). WLAM first learns how multimodal world states change over a local interval, encoding synchronized camera views and available embodiment-state changes into a compact local latent action and a richer transition feature. Reconstruction from partial modalities and consistency across overlapping windows encourage robust transition representations. WLA$^3$ reuses them across semantics, dynamics, and kinematics: local latent actions support action-sensitive physical-dynamics modeling, segment-level features directly supervise the VLM through a Semantic Latent Aggregate (SLA), and an action expert jointly predicts latent actions together with embodiment-specific robot controls. Human videos provide scalable transition supervision, while robot trajectories ground the shared representation in executable native controls. On LARYBench, the final 32D latent action reaches 67.89% average classification accuracy. WLA$^3$ achieves 81.9% average success across six real-robot tasks versus 66.2% for $\pi_{0.5}$. Performance improves as generalist policy model mid-training data scales, and human videos support human-to-robot transfer. Project page can be found at https://wla-3.github.io/.
Read the full abstract
Logic in Computer Science
Bridging the Gap Between Plain VASS and Branching VASS
Vectors addition systems with states (VASS), a model equivalent to Petri nets, are finite-state machines with finitely many counters ranging over the natural numbers. The decidable reachability problem for VASS has many applications in logic, automata, and verification. In this paper we study the reachability problem for BVASS, a branching generalization of VASS. We show that BVASS reachability sets are very similar to VASS reachability sets, namely that they are sections of VASS. Our proof relies on a new well-quasi-order (wqo) on BVASS runs that generalizes the well-known wqo on VASS runs. By leveraging an amalgamation property, we prove that every BVASS run can be transformed into an equivalent one of bounded branching complexity. This allows us to derive several results on the geometry of BVASS reachability sets. As an application we obtain that reachability sets of 5-dimensional BVAS are effectively semilinear, as is the case for 5-dimensional VAS.
Read the full abstract
Quantum Physics
Generation of entanglement and magic via continuous homodyne monitoring of a qubit pair
Continuous weak measurements of quantum systems are of great relevance in quantum foundations and applications. They can be achieved by probing the quantum system of interest by repeatedly measuring an auxiliary system weakly coupled to it. Here we study two qubits coupled in different points to a common one-dimensional electromagnetic field and simultaneously monitored in the right- and left-propagating output channels by homodyne detection. Using a collision-model description, we derive an analytical Stochastic Master Equation (SME) governing the resulting diffusive quantum trajectories, including the interference between the two measurement channels. For nonlinear functions of the quantum state, such as entropies, averages over quantum trajectories generally differ from the corresponding quantities evaluated on the unconditional state. Through this mechanism, we show that continuous monitoring generates entanglement, absent in the unconditional dynamics, and enhances quantum magic in the qubit pair during the decay. Both resources can be tuned through the optical phase accumulated between the qubits and the phases of homodyne local oscillators. Our results establish continuous homodyne monitoring of multiple emitters as a tunable mechanism for generating quantum resources.
Read the full abstract
Human-Computer Interaction
On Edge in the Dental Chair: Designing VR Support for Moments of Dental Anxiety
Dental anxiety can change as a procedure unfolds, yet dental virtual reality (VR) commonly provides continuous distraction or relaxation. We investigate how support can be coordinated with specific simulated dental events. Stakeholder interviews (N=36), participatory design with three returning dentists, and patient walkthroughs of a no-intervention prototype (N=12) informed five Anxiety Events and an intervention-module framework. Drawing on cognitive vulnerability and emotion regulation, we implemented a standardized event-contingent VR system with predefined event-module assignments and shared agency and safety controls. A randomized study (N=24) compared the intervention package with no-intervention VR. The adjusted intervention-minus-control difference averaged -12.83 VAS-A points across events (95% CI [-24.42, -1.70]). Physiological, behavioural, and qualitative measures contextualized participants' experiences. The findings inform timely, comprehensible support and reassuring social presence in simulated dental VR; they concern the complete package rather than individual modules or clinical effectiveness.
Read the full abstract
Mesoscale and Nanoscale Physics
Up/down-conversion of infrared light by few-layer graphene polytypes
Optical nonlinearity of materials with broken inversion symmetry enables two-photon processes where an incoming pair of photons can generate an up-converted photon with the combined frequency, or a high-energy photon can be split into a correlated pair of down-converted photons. Here, we identify few-layer graphene films that offer up/down-conversion capability in the infrared spectral range. For mixed stacking tetralayers (ABCB), which are non-centrosymmetric 2D crystals, we find highly efficient nearly resonant up/down-conversion of [$\omega_1,\omega_2$] photon pairs into/from a photon with $0.7$eV$<\Omega=\omega_1+\omega_2<1.1$eV. We also note that a pronounced second-order nonlinearity in the spectral range of $0.7$eV$<\Omega<0.9$eV can be promoted in rhombohedral tri- and tetralayers by asymmetrical encapsulation. Potentially, for fibers coated with few-layer graphene, this opens the door for “in fiber" production of photon pairs with correlated polarizations.
Read the full abstract
Signal Processing
CoFi-CLM: A Coarse-to-Fine Channel Language Model for Finite-Bit CSI Feedback
In frequency-division duplex massive multiple-input multiple-output (MIMO) systems, the user equipment (UE) must convey high-dimensional downlink channel state information (CSI) under a stringent finite-bit feedback budget. Deep learning (DL) has emerged as a powerful tool for CSI compression due to its ability to capture complex channel correlations and learn compact CSI representations. However, recovering fine-scale channel structure from limited feedback remains challenging. To address this challenge, we propose the Coarse-to-Fine Channel Language Model (CoFi-CLM), a large AI model deployed at the base station (BS) that explicitly learns dependencies between coarse and fine CSI representations. Specifically, the UE reports coarse-token indices from a learned codebook, while CoFi-CLM predicts distributions over unreported fine-token indices conditioned on this feedback in a single Transformer forward pass. A dual-path decoder fuses the direct coarse reconstruction with the generated fine-scale reconstruction. Concentrating fine-token prediction and fusion at the BS allows the CLM capacity to scale without increasing the computational or storage cost at the UE. Across 64 to 160 feedback bits, CoFi-CLM consistently outperforms the compared methods on both seen and unseen scenarios. At 128 bits, it improves the normalized mean square error (NMSE) over the recent large AI model baseline by approximately 1 dB on both sets. The performance gap between seen and unseen scenarios further supports its generalization under the considered channel model. The source code is publicly available at https://github.com/doovvv/CoFi-CLM.
Read the full abstract
Human-Computer Interaction
Towards Scalable Measurement of Durable Skills
Durable skills, such as collaboration, creativity and critical thinking, are instrumental to success in the modern workforce. Yet, measuring these skills remains a persistent challenge. Moreover, because what is not measured is often not taught, these skills are often overlooked in mainstream educational curricula. Designing effective assessments for these skills necessitates balancing two often-conflicting requirements: ecological validity and psychometric rigor. On the one hand, the assessment environment should emulate natural real-world human interaction between humans. On the other hand, it should be scalable, controllable and reproducible. Here we argue that LLMs can be used to better capture both of these aims. Concretely, we develop a framework where the subject converses with AI teammates in a way that resembles human-human interaction for authenticity, while also offering the psychometric control required for informative and robust assessment. Importantly, the AI participants not only act as teammates but also, in an "Executive LLM" setup, steer the conversation towards eliciting a high density of observable evidence for skill proficiency. We complement this with an AI evaluator that can be used to measure skill proficiency in such interactions. We evaluate our assessment protocol based on transcripts of interactions of human participants with our AI framework, for multiple durable skills. For the skill of creativity, we further demonstrate the efficacy of an autorater for evaluating complex tasks performed by real students. Our analysis shows that the use of the Executive LLM significantly increases elicited evidence and that LLM-automated scoring of conversations largely agrees with that of expert annotators. This research demonstrates the utility of orchestrated LLMs approaches for measuring complex social and cognitive constructs in a scalable and controllable manner.
Read the full abstract
Computer Vision
LynnReal-Omni: Native multi-modal Video Generation for Agentic Visual Workflows
Video diffusion models are stochastic and hard to control: precise content often requires repeated sampling without guaranteed success, and long-horizon scenes drift in appearance, interactions, and temporal coherence. Agentic visual creation provides explicit references, editable 3D scenes, or executable game states for stable control, but does not by itself guarantee high object or character fidelity. Combining the two can enable stable, high-quality generation. To realize this combination, we present LynnReal-Omni, a native multimodal video generation framework built on a 32B shared multimodal diffusion transformer that unifies text-to-video, image-conditioned generation, reference-guided generation, structural control, editing, degraded video restoration, and long-video generation. It accepts heterogeneous visual inputs, including appearance references, editable 3D renders, and game recordings, allowing agents to compose visual conditions within a unified model. We also train a dedicated 27B Flash shared multimodal diffusion transformer for real-time rendering. We build a systematic data pipeline for video cleaning, subject association, multimodal annotation, and aligned control construction, yielding a curated corpus of multi-shot audiovisual segments, and introduce MSAVP, a 100-prompt, 20-metric evaluation design that separates instruction following, generating plausibility, visual quality, temporal behavior, and audio coordination. LynnReal-Omni-Flash further reduces inference cost through model and decoding acceleration, including a lightweight VAE decoder; on one H100, warm generation and decoding of a 22-frame 540p video take 843 ms with LynnReal-Omni and 377 ms with Flash. These results provide a foundation for real-time streaming video generation, making LynnReal-Omni a unified, controllable, and efficient basis for agentic visual creation.
Read the full abstract
Number Theory
Cauchy laws for zeta logarithmic derivatives and stationary Stieltjes transforms
We prove an unconditional Cauchy limit for a symmetrically truncated Stieltjes transform of the projected ordinates of the zeros of the Riemann zeta function. A suitable normalization of the logarithmic derivative of $\zeta$ on the critical line has the same limit provided that the sum of the distances to the critical line of the zeros lying to its right, counted with multiplicity up to height $2T$, is $o(T)$; we give an explicit bound on the comparison error. The proof rests on a convergence theorem for the Stieltjes transform of positive stationary point measures whose counting discrepancy satisfies an integrability criterion. This theorem is obtained by comparison with the transforms of periodic point measures, combined with truncation bounds and a transfer theorem. Over finite fields, a classical cotangent identity yields Cauchy limits for the logarithmic derivatives of zeta functions of varieties: the law is exactly Cauchy for nonconstant pure cohomological factors, and Poincaré duality gives explicit errors for full zeta functions, including Cauchy limits for smooth hypersurfaces of increasing degree, uniformly in the base field.
Read the full abstract
Robotics
DuctAM: A Duct-Assisted Quadrotor-Based Aerial Manipulator Enabling High-Force Push-and-Pull Interactions
Uncrewed Aerial Manipulators (UAMs) extend the capabilities of Uncrewed Aerial Vehicles (UAVs) from perception to physical interaction. Among various aerial interactions, push-and-pull operations are fundamental manipulation primitives that require sustained horizontal forces while maintaining stable flight. In this paper, we propose DuctAM, a compact aerial manipulation platform that enhances horizontal force capability for push-and-pull interactions using two ducted fans integrated along the quadrotor interaction axis. An attitude-force decoupled control scheme enables controllable horizontal forces without requiring large attitude changes. Extensive real-world experiments are conducted to validate the DuctAM. Figure-eight trajectory tracking experiments demonstrate stable flight and accurate motion control in both quad and duct modes. Force-measurement experiments quantify the decoupled longitudinal force capability of DuctAM. Finally, representative push-and-pull interaction tasks, including cart pushing, door closing, and drawer opening, verify the practical effectiveness of DuctAM. The results show that DuctAM achieves significantly improved horizontal interaction force capability while maintaining stable flight compared with conventional UAVs.
Read the full abstract
Quantum Physics
Probing Residual Noise at a Decoherence Sweet Spot in a $^{28}$Si/SiGe Spin Qubit
In $^{28}$Si/SiGe spin qubits with micromagnets, the longitudinal stray field gradient transduces charge noise into qubit frequency noise and limits coherence. We compare two neighboring qubits in the same device with 800 ppm residual $^{29}$Si, one near a decoherence sweet spot where the gradient is locally minimized and the other at a position with a larger gradient. At the sweet spot, $T_2^*$ reaches 67 $\mu\mathrm{s}$ and the Carr-Purcell-Meiboom-Gill coherence time reaches 4.6 ms, whereas $T_2^*$ is 5.2 $\mu\mathrm{s}$ at the neighboring qubit. Near 1 Hz, the frequency noise power spectral density at the sweet spot is nearly two orders of magnitude lower than at the neighboring qubit. The magnitude and low-frequency decay of the residual spectrum are compatible with the prediction for $^{29}$Si nuclear spin noise, and the weak interqubit correlation indicates that local noise becomes important for dephasing at the sweet spot. Despite a finite correlation with the charge sensor, sweet-spot operation strongly reduced the transduction of charge noise, bringing the residual frequency noise close to the level predicted for $^{29}$Si nuclear spin fluctuations.
Read the full abstract
Artificial Intelligence
LongAgent: History-Guided Agentic Search for Longitudinal Outcome Prediction
Extracting informative representations from longitudinal data that can predict future outcomes remains a critical challenge in medicine. Medical datasets are inherently heterogeneous, consisting of a large number of variables collected from different sources, sampled with different temporal spacings, and representing different aspects of human health status. This requires identifying those variables with predictive value, processing longitudinal information, and integrating multiple variables for outcome prediction. Here, we propose a novel agent-based approach, LongAgent, that can autonomously search over combinations of variable sets, temporal windows and longitudinal aggregation functions, and identify candidates with promising predictive performance. LongAgent utilises a history memory of previous searches and numerical evidence to guide subsequent exploration. On synthetic data, LongAgent achieves a mean prediction RMSE of 1.7376 and improves over the strongest non-agent baseline by 0.0151 (95% CI: [0.0045,0.0260]; p=0.0273). On a real clinical dataset, it performs comparably to the best baseline.