Papers

We consider the Lorentz positive cone of $n\times m$ matrices that map the Lorentz cone in $\Rbb^m$ into the Lorentz cone in $\Rbb^n$. The Lorentz positive cone and its dual, the cone of Lorentz separable matrices, are shown to provide polynomial-time algorithms for the problem of minimizing a bilinear objective over variables contained in ellipsoids in $\Rbb^n$ and $\Rbb^m$. We also demonstrate how these cones can be used to strengthen SDP relaxations of other nonconvex quadratic optimization problems, including the two-trust-region subproblem.

Expressive speech synthesis has outpaced expressive speech perception: systems now render fine-grained vocal performances that no public benchmark can score. Most benchmarks for this inverse problem stop at six to nine basic emotion categories, largely on acted speech. This paper introduces VoiceNet, a human-annotated representation-level benchmark for voice performance understanding on permissively-licensed in-the-wild speech. VoiceNet has two subsets: VoiceNet-Emo applies a 40-emotion taxonomy with three expert ratings per item, and VoiceNet-Ext, a preliminary subset, scores 57 talking-style attributes including speaking rate, vocal tension, breathiness, and register. The paper also releases Emolia, an emotion-annotated version of the Emilia corpus, with a curated rebalanced subset enriched by dense MOSS-Audio Thinking annotations. Two voice-text contrastive baselines train on this data: a 110M-parameter VoiceCLAP-Small for fast large-scale data filtering and a 7B VoiceCLAP-Large for state-of-the-art performance. Both outperform existing CLAP baselines, which sit near chance on VoiceNet-Emo. On VoiceNet-Emo, VoiceCLAP-Large aligns more closely with the aggregate expert consensus than individual experts agree with one another: a comparison against the majority label rather than evidence of surpassing human emotion perception. All systems evaluated here are voice-text embedding models: VoiceNet scores representation-level attribute recognition and retrieval, not end-to-end spoken-dialogue behaviour. Clustering and filtering uncurated speech corpora into subsets that span diverse talking styles and emotions remains an open challenge; VoiceCLAP embeddings offer a promising tool for this task. VoiceNet, Emolia, and VoiceCLAP are publicly available for research use.

We propose Compact Shielded CSV, a private client-side validation blockchain for peer-to-peer payments designed for the postquantum era. Upgrading existing blockchains to quantum-resistant cryptography substantially increases on-chain overhead. By keeping all large cryptographic artifacts off-chain, Compact Shielded CSV keeps a minimal on-chain footprint independent of the size of the underlying cryptographic proofs and signatures. For a single input transaction, the onchain footprint is just 3 hashes (3x32 bytes): a nullifier, a degriefer, and a commitment to the transaction. We introduce the degriefer - a novel mechanism that enforces ownership and prevents double-spending using only hash commitments, eliminating the need for on-chain signatures entirely. These properties make Compact Shielded CSV a promising foundation for private, scalable, post-quantum digital payments.

Context. Agentic software engineering requires mechanisms to coordinate and govern agents' work. The framework motivating this study proposes a methodological harness with eight mechanisms: context engineering, persistent shared knowledge, executable specifications, N-version mindset and parallel agents, normative specifications, structured consultation, evidence-based acceptance, and graduated autonomy. These are realized through artifacts such as rule or context files, specifications, and architectural decision records, but the framework had not been empirically validated at repository scale. Objective. We analyze to what extent and how this harness is observable in repositories with agentic activity. RQ1 characterizes adoption through artifact prevalence, breadth, co-occurrence, and temporal evolution; RQ2 examines rule files - a key observable artifact and persistent source of agent instructions - to assess how their content reflects the proposed mechanisms. Method. From the AIDev dataset of 116,211 GitHub repositories, we analyzed 5,435 using a design stratified by visibility, measured through stars as an indicator of popularity and/or reputation. For RQ1, we detected and quantified artifacts and introduction dates for the seven observable mechanisms. For RQ2, we qualitatively coded rule files from 150 repositories. Results. Population prevalence of at least one mechanism is 21.7%, versus 65.3% among the most visible repositories. Multi-mechanism configurations are rare. Where rule files exist, they almost always guide the agent and state norms, while persistent shared knowledge, executable specifications, structured consultation, and graduated autonomy appear only in a minority. Conclusions. The harness is empirically observable, but mainly through isolated mechanisms rather than the integrated system proposed by the framework.

We present TriO, a multi-modal unsupervised world model that predicts 4D occupancy, obstacle segmentation, flow and LiDAR. In contrast to prior work, TriO utilizes three distinct sensor modalities (camera, LiDAR, and RADAR) as both inputs and sources of self-supervision, eliminating the need for additional human annotations. Thanks to its novel supervision, the model is able to segment any occupancy from the drivable surface, overcoming the limitations of existing open-set methods in handling long-tail objects. TriO achieves state-of-the-art results in multiple 3D and 4D tasks, including occupancy, flow, and LiDAR prediction, as well as zero-shot road obstacle segmentation across multiple datasets such as Argoverse 2, and Spotting the Unexpected.

Strongly correlated chiral materials have been proposed to host a chiral Weyl-Kondo semimetal (cWKSM) state, in which Kondo hybridization pins chirality-induced Kramers-Weyl point crossings, together with additional symmetry-enforced band crossings, near the Fermi level. Realizing this state requires a material that combines crystal chirality, non-symmorphic symmetry, and Kondo correlations, and currently there are no known materials that combine all these requirements. Here we report evidence for a cWKSM state in CeGaGe, a member of the RXY (R = rare earth; X = Al, Si; Y = Ga, Ge) family. Magnetization measurements confirm long-range antiferromagnetic order below T$_N$ = 4.7 K with a complex, canted magnetic structure. Unique to CeGaGe among the known RXY systems is a structural transition from an achiral tetragonal I4$_{1}$md structure (at high temperatures) to the chiral tetragonal P4$_{3}$ structure (at low temperatures). The structural transition in CeGaGe allows us to highlight the role of chirality in stabilizing the Kramers cWKSM state. Hall resistivity measurements with $H || c$ reveal an anomalous Hall conductivity (AHC) that is constant below the Kondo temperature $T_K \approx 7$ K and drops sharply above it. This crossover tracks the onset of Kondo coherence rather than the magnetic ordering, which occurs at T$_N$ = 4.7 K. Together with first-principles calculations for the $P4_3$ structure, this identifies the AHC as an intrinsic, Berry-curvature-driven contribution generated by Kondo hybridization, rather than a consequence of extrinsic scattering or dynamic scalar spin chirality. These results establish CeGaGe as a rare experimental platform in which crystal chirality, strong electronic correlations, and non-trivial band topology coexist, providing the first material realization of a Kondo-driven cWKSM state.

The Erdős–Sós theorem states that every graph of average degree greater than $t-2$ contains every tree on $t$ vertices. A short counting proof was discovered by GPT-6 Astra in 2026. We give a visual, reader-centered exposition of that argument whose presentation differs substantially from the original. We recast the counting objects as early neighbors in a reveal-and-stop procedure, organize the induction through explicit partitions and reversible swaps, and develop the proof through worked examples with consistent drawings. We also give direct and probabilistic conclusions, compare this formulation with other recent expositions, and record the classical consequence $R(T;q)\le q(t-2)+2$ for multicolor Ramsey numbers of trees.

Emergency departments (EDs) run some of the least standardized processes in healthcare, and long stays are often attributed to demand rather than measured on data. We analyze an anonymised event log of 1{,}820 ED stays (25{,}115 events) through a full process mining pipeline: preprocessing resolves burst logging and missing values, performance analysis quantifies flow at the case and transition level, and Inductive Miner with token-based conformance checking assesses process structure. The filtered log (1{,}754 cases, 16{,}376 events) exhibits a mean throughput of 6.58 hours with a heavy tail, 884 distinct control-flow variants whose most frequent one covers only 4.3% of cases, and an inversion of clinical priority in which urgent patients (acuity 2, 8.03 h) stay 38% longer than critical ones (acuity 1, 5.83 h). Perfect fitness (1.0) combined with low precision (0.71) reveals absence of normative pathways rather than deviance. Two redesign scenarios, acuity-based care pathways and anticipatory discharge, are derived with quantified targets.

We study transfer systems on partially ordered sets whose Hasse diagrams are rooted trees. We prove there is a bijection between transfer systems on a tree and order-preserving functions to the natural numbers that are bounded by the rank function. We then develop a recurrence for the set of functions and use this to enumerate transfer systems on some families of trees. In addition to our study of trees, we prove for any finite poset $\mathcal{P}$, the lattice of transfer systems on the opposite poset $\mathcal{P}^{op}$ is isomorphic to the opposite lattice of transfer systems on $\mathcal{P}$. We show this by establishing a bijection between weak factorization systems and transfer systems, which builds on the results for lattices shown in previous work of Franchere, Ormsby, Osorno, Qin, and Waugh.

Recent advances in sensor technology have enabled more effective human activity recognition (HAR), particularly in real-time systems with limited computational resources. However, Ultra-Wideband (UWB) radar data remain challenging due to high dimensionality, noise, complexity, and nonlinear characteristics. This research proposes a framework to efficiently reduce data size, uncover significant patterns, and classify six activity types (Standff, Liedown, Noactivity, Sit, Stand, and Walk) from UWB signals with high accuracy. Two novel dimensionality reduction techniques are introduced in this paper. The first, Clustered Polynomial Expansion with Incremental PCA (CPE-IPCA), combines clustering and polynomial feature expansion with Incremental PCA, preserving 100% of the variance in only 50 components. The second, Post-PCA Standardization Approach (PPSA), standardizes data after PCA and retains 99.1% of the variance in 80 components, achieving superior compression and computational efficiency compared to conventional nonlinear methods. Frequent patterns are identified using Apriori and FP-Growth, which are then classified with Random Forest and a Vector Space Model (VSM). The framework achieves 100% accuracy with Random Forest on CPE-IPCA and 99% on PPSA, while VSM attains 100% precision, recall, and F1 on PPSA and near-perfect performance on CPE-IPCA (precision 1.00, recall 0.98-1.00, F1 0.99-1.00), demonstrating a fast, interpretable, and robust HAR system suitable for healthcare, assisted living, and smart environments.

We study interactive proofs of proximity (IPPs) for model evaluation, where a resource-limited verifier interacts with an untrusted prover, typically the model owner, to certify statistical properties of a model under an unknown input distribution. Our formulation separates sampling the input distribution from querying the model and evaluating its output; distinguishes real audit data (black-box sampling) from generated data (chosen-randomness, or gray-box, access to the sampler); and allows the prover and verifier to use different evaluators. We focus on doubly-sublinear IPPs, where both the verifier and honest prover use sublinear resources, and on (weighted) Hamming weight properties. For ordinary Hamming weight, we give a tolerant doubly-sublinear IPP. For completeness and soundness radii $\varepsilon_c<\varepsilon_f$ and gap $g=\varepsilon_f-\varepsilon_c$, a logarithmic-round instantiation uses $\widetilde{O}(1/g)$ verifier queries and $O(1/g^2)$ honest-prover queries, improving the cubic dependence of Amir, Goldreich, and Rothblum (ITCS 2025). We prove matching query lower bounds up to polylogarithmic factors. For distribution-weighted Hamming weight, black-box sampling requires $\Theta(1/g^2)$ verifier samples but only $\widetilde{O}(1/g)$ evaluations; the quadratic sample complexity is necessary in the interior regime. With chosen-randomness access, the problem reduces to ordinary Hamming weight, yielding $\widetilde{O}(1/g)$ calls and evaluations. If the parties' evaluators disagree arbitrarily on a $\rho$-fraction of the distribution and by at most $\gamma$ elsewhere, our protocols remain doubly sublinear whenever $g>2\kappa$, where $\kappa=\rho+(1-\rho)\gamma$. Applications include auditing accuracy, group fairness, calibration, harmlessness, usefulness, and average-case robustness.

Prediction-powered inference (PPI) typically relies on point predictions on unlabeled data. When predictive distributions are available as in a wide range of applications, including predictions from LLMs, we introduce distribution-informed prediction-powered inference (DiPPI), a general framework for further improving statistical efficiency by using predictive distributions as auxiliary information in the spirit of PPI. We characterize the optimal use of this information through score calibration, derive the oracle efficiency for a finite-dimensional representation of the predictive distribution, and provide theoretical guarantees for positive learning with the cross-fitted DiPPI estimator. Through simulations and three real-data applications, we show that DiPPI achieves better efficiency than PPI methods based on point predictions. These gains arise when the predictive distribution contains score-relevant information that is partially lost in point predictions.

Transaction fee mechanisms (TFMs) allocate scarce block space in blockchains. Prior work focuses on incentive compatibility (IC): honest behavior maximizes the payoff of each strategic player. We introduce rational-harm proofness (RHP), which rules out deviations that harm honest participants without reducing the deviator's utility relative to honest behavior. IC alone does not exclude such costless harm: in a second-price auction, a losing bidder can raise the winner's payment without changing its own payoff. For finite block capacity k, we characterize TFMs satisfying RHP alongside user incentive compatibility (UIC) and miner incentive compatibility (MIC). In the plain model, where a single miner implements the auction, we prove a tight tetrilemma: no TFM simultaneously achieves positive miner revenue, UIC, MIC, and RHP against miner-user coalitions, while any three are jointly achievable. In the MPC-assisted model, where a committee of miners implements the auction using multi-party computation, we construct a randomized TFM with positive miner revenue satisfying UIC, MIC, and RHP against users, miners, and miner-user coalitions containing at most k users. Randomness is necessary to confirm transactions under congestion: any deterministic TFM satisfying UIC and user RHP in this model confirms no transactions when the number of submitted bids exceeds k. Finally, IC and RHP are incomparable for every strategic role in both models: mechanisms can satisfy either property without satisfying the other. These results extend transaction fee mechanism design beyond incentives to protection against costless harm.

Encoded demographic information in internal model representations is a commonly assumed risk factor for algorithmic bias, with demographic representation invariance often being touted as the ideal state. However, while demographic shortcut learning is a genuine threat, some degree of encoding is necessary when demographics correlate with target labels. Here, we show, mathematically and empirically, that enforcing demographic invariance can actually hamper bias mitigation and even create new biases. We distinguish marginal from class-conditional representation invariance, and show that they imply the standard group fairness notions of demographic parity and equalized odds, respectively. We evaluate the effects on predictive performance and fairness of enforcing both invariance types, both theoretically and empirically across five tabular and two chest X-ray imaging datasets. Our findings support our mathematical argument that demographic representation invariance is neither desirable nor sufficient for fairness.

The $b$-invariant of a normal graded ring was defined recently by Okuma, Watanabe, and Yoshida. We study this for invariant rings of finite groups, proving that it is nonpositive in the case of abelian groups, but not necessarily otherwise; more generally, we give a description of the anticanonical module for invariant rings of finite groups. Turning to infinite groups, we record the anticanonical module and the $b$-invariant for families of determinantal rings. We also construct standard-graded Cohen-Macaulay rings of characteristic zero, with isolated singularities and negative $b$-invariant, that are not of strongly $F$-regular type, thereby disproving a conjecture of Okuma-Watanabe-Yoshida.

The rank $r$ in LoRA is widely treated as a capacity control: a smaller rank is assumed to yield a simpler model that generalizes better. We show that, under hard per-factor norm budgets—the idealization of the weight decay and norm control used in practice—this intuition breaks down. The reason is structural: under such budgets, the updates LoRA can reach are exactly the matrices of rank at most $r$ inside a nuclear-norm ball, and every complexity and displacement functional we analyze is maximized over this set by a rank-one update—so the rank cap never binds. The consequences follow directly. The linear-readout model class we study is identical for every $r \ge 1$, its Rademacher complexity carries no dependence on $r$, and the distance the adaptation can move the source distribution obeys a rank-independent upper bound that we show is sharp. If rank does not control capacity, where does it act? We identify two places. Statistically, replacing the per-factor budgets with a joint budget on the product restores a data-dependent, rank-sensitive complexity bound—though the gain appears only for well-spread feature distributions, and the worst case remains rank-free. Spectrally, rank sets the price of adaptation: canceling the leading singular directions of the pretrained weight requires both sufficient rank and sufficient budget. We bound the smallest rank achieving a desired source–target alignment, with upper and lower bounds that match under two-sided spectral decay. Together, these results recast rank as governing which updates are reachable and what cancellation costs—not how much capacity the model has.

We prove that, in every closed Riemannian three-manifold $(M^3,\bar g)$, there exists a constant $C=C(M,\bar g)>0$ such that every closed smoothly embedded minimal surface $\Sigma\subset M$ satisfies $b_1(\Sigma;\mathbb{Z}_2)\le C\bigl(\operatorname{Ind}(\Sigma)+\operatorname{Area}(\Sigma)\bigr)$, establishing Song's additive genus-index-area conjecture. For closed connected orientable two-sided minimal immersions, we prove the universal inequality $\gamma(\Sigma)\le8\operatorname{Ind}(\Sigma)$ in positive Ricci curvature and the estimate $\gamma(\Sigma)\le C(M,\bar g)\operatorname{Ind}(\Sigma)$ in positive scalar curvature. In the latter case, we also bound both area and total curvature by $C(M,\bar g)(1+\operatorname{Ind}(\Sigma))$.

Deep learning has improved inertial measurement unit (IMU) sensing for mobile and wearable applications. However, an IMU model trained in one domain often becomes unreliable when it is used with a new user, device, or body position. Existing methods usually treat this problem as a static model-design task: they pretrain a stronger representation, add data augmentation, or select one adaptation method before deployment. In practice, the target domain is only gradually observed, labels are scarce, and different domain shifts require different sensing actions. This paper presents SenseAgent, an LLM-guided sensing agent for cross-domain IMU activity recognition. Instead of asking an LLM to classify raw IMU signals, SenseAgent uses the LLM as a runtime planner over sensing tools, source-domain experience memory, online target memory, and verifiers. The agent builds a label-free diagnosis report from the target stream and uses it to decide whether to keep raw inference or invoke specialized tools, including gravity-aware sensing, prototype transfer, and style normalization. Verifiers check source calibration, target-memory reliability, and no-harm criteria before accepting high-risk tool decisions. SenseAgent also supports scarce feedback without retraining the backbone or replacing the label-free route. This design converts cross-domain IMU sensing from a fixed inference pipeline into a closed-loop sensing process that diagnoses target shifts, selects suitable sensing actions, and rejects unsafe adaptations. We evaluate SenseAgent across multiple IMU datasets and deployment shifts. Results show that its verified route selection improves cross-domain sensing, especially under harder placement and compound shifts, and further benefits from limited user feedback.

Whereas previous results on abilities and limitations of transformers have restricted the definition of transformers in various ways, here we study softmax-attention, multi-layer transformers operating on real values, with very few additional assumptions. Applying results from real geometry, we obtain upper bounds on the VC dimension and split VC dimension of such transformers ($O(n^4)$ and $O(n^6)$, respectively, where $n$ is the input length). Conversely, we also construct specific transformers witnessing lower bounds on these quantities ($\Omega(n)$ in each case). These results have some notable consequences. For example, within the class of symmetric (permutation-invariant) functions, we show that transformers can uniformly express all functions over an alphabet of one symbol and non-uniformly express all functions over an alphabet of two symbols, but cannot (even non-uniformly) express some functions over an alphabet of six symbols. We also prove limitations on how many bits of a real number a transformer can access.

Analog dial meters are widely deployed in industrial applications and utility sites, where environments and meter types vary and inspection data may be sensitive. Currently, automatic meter readers must be individually developed and deployed for each environment and meter type in practice. Deep learning could handle this variability but requires diverse labeled data that are costly to collect and update. In practice, meter images are distributed across independent sites, each with limited labels, while raw images often cannot be pooled because of ownership, governance, or privacy constraints. To address these challenges, we present a federated framework for visual analog meter reading that enables multiple sites to collaboratively train a reading model without sharing their raw images. Our framework consists of a four-stage pipeline: (1) dial localization, (2) thin-structure segmentation trained federatively across clients, (3) polar unwrapping, and (4) tick-counting decoding for final reading. To enable systematic evaluation of this setting, we release MeterFL, a 1,382-image mask-annotated dataset organized into deployment-motivated pseudo-clients derived from visual attributes via deterministic rules, with dHash near-duplicate control between the segmentation train and test splits. We evaluate both segmentation quality and end-to-end reading accuracy. MeterFL is publicly available at https://github.com/weidazhaoooo/Meter-FL.

Decoding imagined speech is an important goal for brain-computer interfaces but remains challenging due to weak neural responses, low signal-to-noise ratio, and limited imagined-speech datasets. Language models provide strong contextual cues for text prediction, but how much they can help neural decoding and whether their contribution differs for decoding perceived and imagined speech remains unclear. To investigate this, we use a paired listened-imagined MEG dataset and incorporate language-model information at two stages. First, we train a contrastive neural decoder that aligns MEG representations with acoustic and contextual language representations, improving cross-subject word decoding for both listened and imagined speech. Second, at inference, we introduce a neural-constrained beam-search framework that combines neural evidence with language-model next-word probabilities. We find that imagined-speech decoding benefits more from the language model than listened-speech decoding. For Imagined speech, the best-performing balance between neural and language-model evidence shifts toward the language model, and the gain over neural-only decoding is larger. Together, these results suggest that language priors are most useful when neural evidence is weaker, making them particularly valuable for imagined-speech BCIs.

Can out-of-distribution (OOD) generalization be predicted from a trained model's weights alone, without any target-domain data? Existing representational similarity metrics (CKA, SVCCA, RSA) compare activations rather than forecast generalization. We show they are provably insensitive to structural rerouting in the computational graph, the very change distribution shift induces. We close this gap with the Circuit Alignment Score (CAS), which compares class-specific circuits across domains via graph kernels, decomposed into same-class coherence and cross-class confusion. Casting CAS as a Lebesgue integral over the domain distribution, we prove its Monte Carlo estimate recovers the ground-truth ranking of learners by OOD accuracy, with pairwise inversion error vanishing at rate $O(1/M)$, where $M$ is the number of sampled domains. Across $48$ learners on PACS, CAS attains $0.88$ rank correlation with OOD accuracy, versus $0.58$ (CKA), $0.23$ (SVCCA), and $0.14$ (RSA), with similar trends on other benchmarks and even against data-dependent methods, making it the first provably consistent predictor of distributional robustness requiring neither target-domain data nor labels. The code is available at: https://github.com/ayanban011/ACE

Long-horizon coding agents receive verifiable rewards only after completing expensive sequences of tool calls. This increases inference cost, amplifies early wrong hypotheses, and can lead to sparse terminal reward and unstable training. We introduce Contextual Early Reward (CER), which predicts terminal reward through behavioral evidence in a trajectory prefix. CER synthesizes adaptive rubrics specific to the current task and stage through experiences summarized from related historical tasks. In test-time scaling on SWE-bench Verified, CER improves RM@8 over the strongest baseline by 4.2 percentage points (pp) on Nemotron 3 Ultra and 2.0 pp on Qwen 3.6 27B; on Nemotron, it takes only 15.3% tokens to match the best baseline performance. In RL training experiments, CER exceeds full-rollout TMax by 1.9 pp while using 52.7% fewer online policy-and-judge tokens. Together, CER provides an interpretable, efficient, and dense evaluation method for long-horizon coding agents.

Spectral risk measures, including conditional value-at-risk (CVaR), generate a family of convex criteria for portfolio estimation. We study how the criterion should be chosen when different criteria share the same population minimizer, and whether the efficient criterion can itself be learned from data. We develop a general asymptotic theory for empirical spectral-risk minimization under estimated linear constraints. Under normal scale-mixture elliptical returns, all spectral risk measures identify the same population efficient portfolio, but their empirical minimizers have different sampling distributions. Their asymptotic covariance decomposes into a common component and a positive-semidefinite component scaled by a functional of the spectral measure, reducing efficiency to an optimization over probability measures. We characterize the efficiency-optimal spectral measure and show that single-level CVaR is generally inefficient. We then construct a fully data-adaptive estimator that learns the radial distribution and optimal spectral measure from the same observations used for portfolio estimation, yet has the same first-order distribution as the infeasible oracle. Simulations and an empirical application illustrate the method.

We study how well gravitational-wave phase deviations can be identified using a neural network with classification and regression heads. The classification head distinguishes general relativity (GR) from six parametrized post-Einsteinian phase families with exponents $b\in\{-7,-5,-3,-1,+1,+2\}$; the regression head predicts the logarithm of the coupling magnitude, $\log_{10}|\beta|$. The network input is a response function quantifying the sensitivity of waveform mismatch to waveform deformations. We use stationary Gaussian noise with the Advanced LIGO design spectrum for 262 sources and a measured Hanford spectrum from the third observing run (O3) for 132 sources, and also test injections into recorded Hanford strain. On the two Gaussian datasets, the network identifies the exponent of deviations clearly above the noise with accuracies of $0.427$ and $0.329$. To interpret these results, we compare the network with an approximate classifier built from predicted waveform residuals for the six phase families and same sources. The two classifiers often confuse the same families. Examining residuals left by different phase corrections, we link these errors to similarities between the underlying deviations. On selected recorded Hanford strain, networks trained on simulated or recorded noise achieve overall classification accuracies close to those on simulated Hanford noise. Of thirteen neural-network variants tested on the Advanced LIGO design dataset, the recurrent network has the highest mean classification accuracy for loud deviations, although all variants remain below the approximate classifier. The network nearly matches that classifier's accuracy, indicating that identification is limited by similarities between neighboring phase families rather than by the network, for known intrinsic source parameters.

Decoding covert speech, such as mimed or imagined, from brain activity is harder than decoding vocalized speech. Cross-modal transfer, where information from one speech form helps decode another, is a promising remedy; yet how a decoder internally represents and processes brain activity from different speech forms remains unclear. In this work, we ask: which internal neurons of a decoder carry cross-modal information, and are these neurons shared across different speech forms? To answer these questions, we leverage mechanistic interpretability, using recordings of the same sentences in vocalized, mimed, and imagined input pairs for activation patching. We insert the decoder's internal activity for a sentence in one condition into its processing of the same sentence in another and measure the change in decoding accuracy. We find that no single neuron drives this benefit; instead, it arises from small groups of neurons, with vocalized speech as the most useful source. These groups are largely condition-specific in the early stage of the decoder but overlap in the later stage. These findings point toward more data-efficient covert speech decoders through training objectives that encourage shared later-stage representations learned mainly from vocalized data.

Let $F_2(C_n)$ be the $2$-token graph of the cycle $C_n$ and let $\rho$ denote the packing number. Gómez Soto and Ríos-Castro recently proved that $\rho(F_2(C_n))\ge a(n)$ for $n\ge 19$, where $a(n)$ is an explicit expression. In this note, we prove that \[ \rho(F_2(C_n))\ \ge\ \left\lfloor\frac{n(n-2)}{10}\right\rfloor+1\qquad\text{for every } n\ge 3, \] which improves $a(n)$ by one whenever $n\equiv 0,2\pmod{10}$.

Wi-Fi channel state information (CSI) has enabled device-free sensing applications such as human activity recognition. However, CSI sensing models remain brittle in cross-domain deployment, where changes in users or environments can produce incorrect predictions. Existing solutions usually treat this problem as an offline model-design problem, by pretraining a stronger representation or applying one fixed adaptation method to the entire target domain. In practice, labeled target data are scarce and different classes may fail in different ways under the same domain shift. To address this, we propose CSI-Agent, an evidence-seeking LLM agent that reformulates cross-domain CSI adaptation as a deployment-time decision-making problem. Rather than processing raw CSI or making sample-level predictions, CSI-Agent summarizes target-domain behavior into sensing-grounded class-level evidence. It establishes a strong target-adaptive default from complementary CSI views and uses an LLM planner to determine whether each class should retain the default or invoke a specialized action. Deterministic verification and bounded execution further reduce unreliable interventions. We evaluate CSI-Agent on four public datasets using five cross-domain splits covering device, user, environment, and compositional shifts. Under 1-shot adaptation, CSI-Agent achieves the best target-domain performance across all splits and improves the average Macro-F1 by about 16% compared to the strongest baseline method.

We study communication between two frozen large language models from different providers, with different tokenizers, accessed through their API endpoints. The two play a referential game: one sees an object and describes it in a short fixed-length message over a small alphabet; the other must pick that object out of a candidate set. Neither model's weights are updated. Each agent's prompt is rewritten by an isolated prompt optimizer whose reflection model reads that agent's scored interactions. In the positional setting, optimized prompts carry a shared code that generalizes to held-out objects above a measured no-codebook baseline, including when the memory window is removed. In a second setting, independent per-letter blocks no longer fit within the message, although a whole-object place value code does. The base system fails to establish reliable communication: the sender struggles to retain an injective rule, and the receiver has too few confirmed examples in view. A sender collision penalty, retention of successful interactions, and sequential optimization enable successful place value communication in some runs. Outcomes vary across runs and reflection models. In successful runs, the protocol is written into the optimized prompts, where it can be read and audited directly.

We investigate the applicability of Lagrangian and Hamiltonian Neural Network models to a dissipative system that has explicit time dependence in its Lagrangian, Hamiltonian, and total energy. To do so we consider these neural network models for simulated systems of a harmonic one-dimensional, one-component oscillator with damping, as well as without damping for comparison. We find that both the Lagrangian and Hamiltonian approaches are able to predict the empirical physical behavior of the damped oscillator systems and to effectively “learn” to varying degrees the underlying Lagrangians and Hamiltonians, as has previously been shown to be the case with undamped oscillator systems. These investigations elucidate important properties of Lagrangian and Hamiltonian mechanics, including properties that are not manifest when considering systems without explicit time dependence.