Papers

Counter-UAV systems based on thermal infrared detection must stay accurate as operational datasets evolve, yet sequential fine-tuning causes catastrophic forgetting of prior tasks, a problem that remains insufficiently characterized in this domain. This continual-learning study measures the stability-plasticity trade-off in YOLOMG, a YOLOv5-based detector run as a single thermal-infrared stream with the motion channel disabled, trained sequentially across three anti-UAV benchmarks of rising scale difficulty: Anti-UAV-RGBT, Anti-UAV410, and CST Anti-UAV. Naive fine-tuning on CST yields a Forgetting Measure of -0.605 against the Stage 1 ceiling, corresponding to a 90% capability loss, with -0.572 occurring in Stage 3 alone. In contrast, knowledge distillation from a frozen teacher is associated with FM = -0.033 +/- 0.004 across three seeds, corresponding to 95% retention. Because no Stage 2 no-KD control is included, this result establishes retention under KD training rather than a causal KD effect. Per-stratum analysis shows large-target detection collapsing to near zero within the first epoch, despite an inter-stage cosine similarity of 0.987 over the gradient-updated weights, pointing to scale-conditioned gradient imbalance, rather than weight drift, as a candidate mechanism. Scale-Stratified Herding (SSH), a 300-exemplar buffer balanced across four UAV size strata, roughly halves the forgetting (FM = -0.605 to -0.311) and keeps large-target detection non-zero. An ablation attributes the gain primarily to scale stratification rather than herding: random-stratified replay performs at least as well (FM = -0.221 versus -0.311 for SSH). These replay results are single-seed and should therefore be treated as preliminary.

Linear regression and kernel smoothing offer tractable explanations of in-context learning: in both, the features determine the weight assigned to each context label. However, whether this fixed-weight account describes pretrained tabular foundation models (TFMs) remains unclear. Testing this account using derivatives runs into a standardization trap: public TFM packages standardize the labels before the model sees them, yet ordinary derivatives also reflect behavior outside the set of standardized labels, making a model appear nonlinear even when every prediction it makes agrees with a fixed-weight map. We propose two certificates that depend only on predictions at standardized labels and can reject two distinct explanations: fixed-weight prediction and sums of independent nonlinear label transformations. Across the five public TFMs that we evaluate, our certificates show that changing one context label alters how other labels influence the prediction, a behavior we call joint processing. We further find that joint processing emerges with training and that attention scores carry most of the measured interaction. Together, these findings motivate TFM explanations that account for how context labels change the influence of individual examples.

Continual learning (CL) is essential for Large Language Models (LLMs) to sequentially adapt to evolving tasks. To mitigate catastrophic forgetting, recent advances implement low-rank adaptation with orthogonal projections (e.g., O-LoRA) to isolate task parameters. However, we reveal that such strict geometric constraints trigger an "Orthogonality Dilemma": rigid parameter isolation impedes the transfer and accumulation of shared representations across semantically related tasks. In this work, we propose a new replay-free method, called Consolidation and Decoupling LoRA (CoDe-LoRA), for CL of LLMs. CoDe-LoRA disentangles the learning process into Consolidating Universal Knowledge and Decoupling Task-Specific Knowledge. To achieve this, CoDe-LoRA leverages an adaptive null space projection mechanism and semantic routing to balance knowledge accumulation with task-specific adaptation. Experimental results across four backbones and three CL benchmarks show that CoDe-LoRA achieves the best average accuracy. Our code is available at https://github.com/Estrellajer/CoDe-LoRA.

This paper studies distributed information aggregation and Nash equilibrium seeking over open multi-coalition networks, where both agents and coalitions may dynamically enter and leave the network. The varying network population poses a fundamental challenge to conventional gradient-tracking methods, as agent departures may cause the loss of gradient information required for distributed aggregation. To overcome this issue, a mass-preserving gradient-tracking method is designed, in which the gradient information carried by departing agents is redistributed to the remaining agents within the same coalition. The proposed method preserves the aggregate gradient information of the coalition objective over the intra-coalition communication network and achieves gradient tracking under dynamic population changes without imposing additional initialization conditions beyond those of conventional gradient-tracking methods. Based on the proposed gradient-tracking mechanism, a distributed Nash equilibrium seeking framework is developed, and its convergence is established under both uncoordinated fixed and diminishing stepsizes. Furthermore, to alleviate the slowdown of convergence caused by classical diminishing stepsize schemes ($\frac{\eta_0}{t}$, where $\eta_0>0$ is a constant), an exponential-interval diminishing stepsize scheme is designed based on the agents' exploration intervals. Finally, a numerical simulation based on an open connecting control game demonstrate the effectiveness of the proposed information-tracking and Nash equilibrium seeking framework.

Existing visual navigation policies are inherently bound to fixed camera configurations, creating a fundamental barrier to zero-shot deployment across heterogeneous robot sensor layouts. To overcome this limitation, we present an embodiment-informed navigation policy capable of generalizing across diverse depth sensor configurations on a specific aerial platform. Instead of implicitly learning spatial alignments, our approach explicitly unprojects depth measurements from arbitrary depth sensor payloads, varying in sensor count, mounting extrinsics, and intrinsics, into a shared robot-centric frame, stitching them into a unified spherical range image and a binary validity mask. This mask allows the downstream policy to explicitly distinguish covered space from unobserved blind spots. Trained via reinforcement learning with aggressive camera randomization, our policy generalizes zero-shot to unseen layouts featuring up to seven cameras, scaling success rates from 78% to 95% as total spatial sensing coverage increases. Finally, real-world flight trials on a physical quadrotor, conducted in an obstacle-filled corridor and an outdoor forest, validate the policy's zero-shot transfer across camera configurations and its resilience to sudden online sensor dropouts.

The COVID-19 pandemic emphasised the importance of remote emotional communication and highlighted a gap in existing technologies that lack haptic channels. While video calls maintain visual and auditory presence, they cannot convey the emotional depth of physical contact, especially hugging, a universally recognised form of intimacy and support. To explore this challenge, we developed the Hugging Suit, a pneumatically-actuated wearable system that enables users to simulate and receive remote hugs in real time. Unlike prior haptic systems that focus solely on tactile sensation, our approach integrates both technical and experiential considerations. The system integrates a programmable Air Actuator Matrix (AAM), a portable high-pressure pneumatic unit, and a fabric-based pressure sensor layer, enabling precise, wearable haptic feedback while remaining lightweight and mobile. Guided by a Research through Design (RtD) methodology, we iteratively refined the prototype to improve tactile resolution, overall usability, and user comfort, while introducing modular components to support flexible and scalable haptic configurations. In parallel, we explored how specific contextual factors, such as lighting, privacy, and the visibility of the remote partner, might shape the emotional and perceptual experience of receiving a remote hug. Preliminary feedback suggests that low-lit, private spaces, especially when users could also see their remote partner via video, improved emotional engagement and comfort during mediated haptic experiences. Our low-cost, maker-space-friendly approach contributes to affective haptics applications in long-distance relationships, emotional regulation, and remote therapeutic support.

Current evaluation of multilingual Large Language Models (LLMs) rests on an implicit Translation-Isomorphism Assumption (TIA): that semantic structures across languages are congruent and mutually mappable without loss of information. We argue that this assumption is not merely violated in practice, but ill-posed in principle for a typologically identifiable class of concepts, including pragmatic markers, honorifics, and diachronically stratified terms. We formalize this failure using a usage-cloud framework, representing concepts as point sets of contextualized embeddings. We define $\alpha$-unalignability as the impossibility of any mapping that simultaneously preserves lexical faithfulness (centroid correspondence) and structural faithfulness (local neighborhood topology). We provide three layers of evidence. Behaviorally, we show that FLORES-200 translation failures are predicted by language family and resource class but not by script, and that LOBSTER reasoning scores vary by family. Mechanistically, we report a Representation-Intervention Gap (RIG) in a nine-model case study on Yami: the models' activations encode a regularity along which Yami groups with other low-resource and Austronesian languages, yet interventions on language-specific neurons show no demonstrated advantage over random masks: the regularity is visible but not usable by this intervention. Finally, we operationalize these findings into a multidimensional diagnostic profile: Cycle-Consistency, Pragmatic-Load Disagreement, Manifold-Curvature Mismatch, and RIG. We argue that collapsing cultural competence into a single scalar incentivizes "probabilistic flattening," and that recognizing the unalignable class is a precondition for AI that respects, rather than erases, cultural divergence. This suggests that multilingual alignment is not a single well-defined objective, but a set of mutually incompatible projections.

The growth of the Internet of Things (IoT) has raised concerns over the privacy of data collected by resource-constrained sensing devices. Homomorphic encryption (HE) addresses this by letting a device encrypt its data once and an untrusted cloud server compute on the ciphertext without seeing the values. In practice, HE's memory and computational cost have kept it out of reach of microcontroller-class devices. Prior work, SEAL-Embedded, made this feasible using CKKS, but left three gaps: its arithmetic is entirely scalar, even on hardware with a vector instruction set; it never decrypts on the device, so the client cannot consume a result; and it does not explore BFV, whose encoding uses only integer arithmetic and decrypts exactly. We present Zeppelin, the first HE library to use an embedded vector instruction set, the first to support both encryption and decryption on an embedded device, and the first BFV implementation on MCU-class hardware. Zeppelin vectorizes the number-theoretic transform, HE's main bottleneck, for ARM's Helium extension, and includes a decryption procedure that avoids large-integer arithmetic and timing leakage of the secret key. A server-side adapter converts Zeppelin's ciphertexts into a format compatible with Microsoft SEAL, so the device handles encryption and decryption while the server performs all homomorphic computation. On an STM32N6 MCU with an ARM Cortex-M55, Zeppelin encodes and encrypts 4096 packed values in 4.76 ms (seeded symmetric, 128-bit security per the HE standard) and decrypts and decodes the result in 5.38 ms, using under 500 KB of RAM, with the vectorized NTT engine ${\sim}2.7\times$ faster than an equivalent scalar implementation on the same core.

Long-term conversational memory is becoming an integral component of modern LLM systems. Proposed architectures group records by topics and events, construct hierarchies and graphs, and connect facts through causal and temporal relations. We experimentally study the interaction between two memory parameters: structural depth and the width of context supplied to the answer model. Using EverMemBench, we evaluate depths D1-D4, core budgets of 1,024/2,048/4,096 tokens, and additional Production and Oracle conditions up to the full archive. Increasing width from 1K to 4K improves Accuracy by 10.11-17.98 percentage points, whereas increasing depth provides no monotonic gain. Beyond 8-16K, Production performance reaches a plateau while tokens per correct answer continue to increase; Oracle preserves quality on full archives of 68-71K tokens. These results motivate further investigation of large, coherent context blocks instead of progressively deeper memory structures.

Pauli decomposition of an $n$-qubit operator is expensive if each of the $4^n$ Pauli strings is treated as an unstructured object. Labeling those strings by the configuration space $(\mathbb{Z}_2)^n$ makes them displacement operators $X^{\otimes x}Z^{\otimes z}$ on the computational space: shifts by~$x$ and multiplications by characters $Z^{\otimes z}$ in the computational basis. That labeling organizes the operator into $2^n$ fibers of length $2^n$, one per shift~$x$, each collecting the computational-basis amplitudes that hop by~$x$. Expanding each fiber function in the character basis of $(\mathbb{Z}_2)^n$ — a length-$2^n$ Walsh–Hadamard transform — recovers the Pauli coefficients at total cost $O(n\cdot4^n)$. The same fiber / shift–character picture extends to $(\mathbb{Z}_p)^n$ (Heisenberg–Weyl qudits). paulikit is a memory-bounded realization for qubits: for matrix-sparse inputs, peak memory need not densify to a full $2^n\times2^n$ operator. Performance is checked on a stated low-end GNU/Linux platform for a coupled-oscillator sweep beyond a billion Pauli terms.

Teleoperation serves as the fallback solution to autonomous driving but reliable functions of the teleoperation require a certain amount of mobile network resources, which cannot be guaranteed at all times. Therefore, predictive quality of service (pQoS) is introduced as a concept to increase the resilience of the teleoperation. In this paper, based on a data measurement campaign, we propose a prediction framework to prediction two important network KPIs of teleoperation: uplink data-rate and round-trip latency. Furthermore, we introduce a method to alleviate the performance degradation of machine-learning-based prediction models on previously unseen data due to concept drift by incorporating historic data into the prediction pipeline. Additionally, we introduce the metric of critical scenario detection to evaluate the prediction performance specifically for teleoperation.

Generative models have the potential to accelerate inorganic materials discovery by enabling inverse design, but generating experimentally realisable crystals remains challenging. Oxidation states are widely used to assess the compositional validity of crystals and guide inorganic materials discovery. While existing generative models for crystals can generate materials with charge-neutral oxidation-state assignments, they poorly reproduce the distributions of oxidation states observed in synthesised materials. To address this limitation, we propose OxiGen, an oxidation-state-aware crystal diffusion model that explicitly represents oxidation states during generation. OxiGen enforces global charge neutrality by construction using a structured output layer with exact inference over a finite-state automaton. Empirically, OxiGen substantially improves oxidation-state fidelity, generates the highest rate of stable, unique, and novel crystals among evaluated methods, and maintains high compositional validity even under property conditioning.

Morphing has recently gained renewed interest with the emergence of generative models, particularly in audio and image generation. In musical sound synthesis, morphing can generate intermediate sounds between two targets, helping musicians and sound engineers explore new sounds with interesting perceptual properties. As morphing is inherently defined in perceptual terms, evaluating this task is challenging. In this work, we introduce Sobolev Distances to Ideal Morphing (SDIM), a novel objective metric to quantify the regularity of audio morphing trajectories in perceptually relevant audio embedding spaces. Leveraging a physics-based sound synthesizer, we evaluate the discriminative power of SDIM on controlled morphing trajectories with varying degrees of regularity and compare it with that of existing audio morphing metrics. Results show that, contrary to state-of-the-art metrics, the proposed metric reliably discriminates desirable trajectories from adversarial ones.

In this paper, we investigate connections between dependence logics, viewed as a subclass of relating logics, and three-valued logics. More specifically, we identify common features of Epstein's subject-matter semantics and the variable-inclusion conditions characteristic of some infectious many-valued logics. Inspired by Del Cerro and Lugardon's sequent calculi for dependence logics, in which classical connectives are combined with connectives satisfying Epstein-style conditions, we introduce two three-valued dependence-like logics that combine classical conjunction and disjunction with infectious negation and implication. We also provide sound, complete, and cut-free bisequent calculi for these logics.

Many two-sample tests for populations of persistence diagrams assess global differences without identifying the regions of the birth-death plane that contribute to them. We study simultaneous inference for local mean contrasts when the number of available diagrams is fixed. They are differences in expected weighted feature mass within $\ell_\infty$ neighborhoods at several centers and radii. We estimate these contrasts using additive landmark responses. A Gaussian multiplier bootstrap calibrates simultaneous confidence intervals while allowing unequal group covariances. The neighborhoods whose intervals exclude zero form a map with approximate family-wise error control, and selecting a subset of original intervals for display preserves their joint coverage guarantee. On the simultaneous coverage event, every reported neighborhood lies within twice its radius of the support of the mean-measure difference. A geometric result gives sufficient radius conditions for a displaced feature to produce a nonzero contrast. A comparison of sufficient detection thresholds quantifies the tradeoff between reducing the number of tested coordinates and reserving observations for an independent pilot. In simulations with 40 to 120 diagrams per class, the bands achieved 94%-98% simultaneous coverage under both the strict null and equal means with unequal covariances. In the latter setting, a permutation maximum and the pooled-t implementation of the two-stage persistence-image test of Moon and Lazar rejected in up to 32% and 26% of runs, respectively. In the fixed-budget simulations, spending a third of the observations on a pilot to choose landmarks or radii located changes less often than a prespecified grid at a single radius. On the MUTAG benchmark, the localized region concentrates on rings of fused-ring systems, an exploratory reading.

Understanding the structural relationships among attributes in tabular data is fundamental to machine learning and pattern recognition. While functional dependency (FD) discovery has been extensively studied, scalable discovery of logical dependencies (LDs), particularly as the number of attributes and dependency order increase, remains underexplored. These dependencies capture non-deterministic, condition-specific relationships among pairwise or multiple attributes. Furthermore, existing approaches do not provide a unified framework for extracting multi-attribute LDs and FDs. To address these limitations, we propose LDTool and HLDTool for extracting and visualizing multi-attribute LDs and FDs from tabular data. LDTool extends dependency discovery beyond pairwise relationships, while HLDTool enables scalable extraction through hypergraph-guided search-space reduction. Experiments on three simulated and eleven real-world datasets demonstrate that the proposed framework extracts meaningful LDs and FDs while improving scalability. LDTool recovers the same FDs as existing FD discovery methods with lower runtime in high-dimensional feature spaces, whereas HLDTool enables dependency discovery in datasets with hundreds of features. The proposed framework provides interpretable visualizations of dependency structures and supports applications in exploratory data analysis and the quantitative evaluation of synthetic tabular data.

This paper addresses the challenge of automatic, frame-accurate event detection in table tennis videos. Current methods for estimating 3d ball trajectories and ball spin typically require that key events, such as ball-racket contacts, have already been identified in advance. This requirement makes it difficult to apply these methods to longer, unedited video recordings. To overcome this limitation, we propose EventNet, a two-stage pipeline to detect key events: (1) 2d keypoints are extracted of the upper-body poses for both players, table corners and ball center. A small keypoint transformer combines them into a compact representation that is robust to changes in viewpoint, lighting, and background clutter. (2) The temporal sequences of these frame-based representations are processed by a transformer encoder that predicts two time-to-event values for each frame, indicating how close the current frame is to the next and previous ball-racket contact. One novelty is a new, temporal cosine-like target signal. Furthermore, we introduce viewpoint augmentation via 3D reprojection and frame-rate augmentation to improve robustness and generalization. Our extensive ablation study gives deeper insights into the importance of various architectural and training aspects. Experimental results show that the proposed approach achieves an F1 score of 91.16% and a mean frame deviation between ground truth and predicted frame of 0.42 on the Latte-MV dataset and 73.08% / 1.16 on the challenging TTHQ dataset. Overall, our work demonstrates that 2d keypoint-based temporal modeling with our EventNet architecture is a promising and practical approach for automatic event detection in table tennis videos.

Music Source Restoration (MSR) seeks to recover original, unprocessed instrument stems from mixed, mastered, and possibly degraded recordings. Unlike conventional source separation, which treats the mixture as a linear sum of clean sources, MSR must additionally invert nonlinear production effects, such as equalization and compression, and transmission-related degradations, such as codec artifacts. We propose a three-stage framework built around this distinction: (1) a mixture restoration model that addresses degradation before separation, (2) a single model that separates the restored mixture into eight target stems (vocals, guitars, keyboards, synthesizers, bass, drums, percussion, and orchestra), and (3) stem-specific restoration experts fine-tuned on the separator's own residual artifacts. Each stage improves restoration quality over the previous one on the MSR Challenge test set. We release code and models to support future research in MSR at https://github.com/theMoro/music_source_restoration.

We study isomorphism decision and basis construction for finite Abelian groups of known order $n$. We measure complexity by the number of additions performed in the groups and by the total running time. We give a randomized isomorphism decision algorithm that performs $\tilde O(n^{1/4})$ group additions and runs in $\tilde O(n^{1/4})$ time. This improves the $\tilde O(\sqrt n)$ upper bound of Chen and Fu and matches Bshouty $\Omega(n^{1/4})$ lower bound up to polylogarithmic factors. We also prove that finding a basis with success probability at least $2/3$ requires $\Omega(\sqrt n)$ group additions in the worst case. This matches the $\tilde O(\sqrt n)$ upper bound of Chen and Fu up to polylogarithmic factors. These results separate isomorphism decision from basis construction in finite Abelian groups: their optimal worst-case complexities are $\tilde\Theta(n^{1/4})$ and $\tilde\Theta(\sqrt n)$, respectivel

Facial affect models are trained almost entirely on adults, yet are increasingly applied to children in education, health, and developmental research. We present a controlled, multi-model audit of five AffectNet-pretrained expression models (EmoNet, EmotiEffLib, DDAMFN++, OpenFace 3.0, LibreFace) on children, across four child image datasets, the AffectNet-8 validation set, and two spontaneous child video datasets, through one shared harness. Three findings emerge. First, the child gap is model-agnostic: every architecture degrades from posed to naturalistic faces and shares the fear$\rightarrow$surprise confusion. Second, it is concentrated and corroborated across all five models: open-mouth faces (read as surprise, correlating with the AU26 jaw drop) and South-Asian children degrade systematically, with a smaller averted-gaze penalty, while closed-mouth faces, White and Black children, and direct gaze do not; the bias tracks expression morphology and specific populations, not skin tone. Third, the gap is diagnosable: a linear probe on frozen features reaches 0.75-0.91 on unseen children versus 0.48-0.66 zero-shot, so it lies largely in the classifier head, not the representation, whereas dimensional valence/arousal regression degrades sharply under domain shift. Building on this, recalibrating only the head on a little target data recovers $+0.13$ to $+0.28$ on the two largest child sets across all five models at negligible adult cost, though the gain is in-distribution and does not transfer across child collections. We will release the harness, per-sample predictions, and analysis code; the child face data stays license-locked and is never redistributed.

Deciding whether two samples come from the same distribution is a classical problem in statistics, and generative transport offers a new way to approach it. We build a stochastic interpolant directly between the two samples and observe that, for a symmetric schedule, its law is invariant under the time reflection $t \mapsto 1-t$ whenever the two distributions coincide. We therefore test whether the marginals at times t and 1-t agree by computing their Jensen–Shannon divergence. Both marginals are explicit mixtures over all cross-pairs of observations, so nothing is learned, and permutation calibration gives an exact finite-sample level. For Gaussian noise, this divergence equals a time integral that pairs the reflection defects of the velocity field and of the score, so the test compares transport dynamics rather than endpoints alone. With a narrow-plus-broad noise design, the test attains the minimax separation rate n^{-2s/(4s+d)} over bounded, compactly supported densities whose difference has Sobolev smoothness s > 3d/4, with no lower bound on the densities. Fusing a dyadic grid of noise scales through their permutation ranks, without sample splitting, preserves exact level and adapts to unknown s at an iterated-logarithmic cost. Empirically, the test matches or outperforms state-of-the-art kernel two-sample tests.

Algorithmic recommendations can help participants navigate large matching markets. For example, recommendations for school and college choices may reduce information frictions and disparities in access to high-performing programs. At scale, however, recommenders in capacity-constrained settings can be self-defeating: if they steer too many users toward the same items, then even users who were originally predicted to have a high chance of matching to an item may not, due to increased competition. In this paper, we formalize this phenomenon of recommendation-induced congestion; motivated by the NYC high school match, we show that naive recommendations can cause sharp decreases in program acceptance rates, most affecting applicants with the fewest nearby options. Next, we propose and theoretically analyze a congestion-aware, bilevel optimize-and-simulate approach to allocate recommendations and improve match outcomes safely, in equilibrium. Finally, we deploy this approach in the 2025-26 admissions cycle of the NYC high school match, aiming to reduce disparities by highlighting personalized lists of nearby, high-performing programs where an applicant has a high predicted offer likelihood. In a randomized controlled trial, we find that 16.4% of treatment applicants ranked a recommended program, versus 10.5% of control applicants who ranked a program they would have been recommended (57% relative increase; $p$=0.011); 5.6% of treatment applicants matched to such a program, versus 3.3% of control applicants (71% relative increase; $p$=0.071); further, no treatment applicant was rejected from a recommended program. Our findings suggest that recommenders should be analyzed and designed as market-shaping interventions.

Comprehensive visibility into Internet Background Radiation (IBR) requires diverse measurement infrastructure. Recent work showed that cloud IBR depends on the provider making single-provider telescopes systematically biased. These studies relied entirely on Western providers and left the Asia-Pacific region as a blind spot. In this report we extend a multi-cloud network telescope onto Tencent Cloud inside mainland China. We detail how Internet fragmentation, geo-politics, and provider rules complicate global measurement through strict onboarding barriers and an unreliable management plane. Once deployed, our nodes captured high-volume regional scanning invisible to our Western vantage points. We conclude by connecting our findings to the Responsible Internet framework. This framework relies on independent global observation to build a transparent Network Inspection Plane. We show that building this plane is bottlenecked by sovereign ecosystems and requires a collaborative community effort.

Many social applications of machine learning exhibit performative effects: population behavior changes in response to deployed models. Performative prediction studies this interaction through a distribution map that relates each model to the population distribution it induces. One of the main results in this framework showed that repeated risk minimization (RRM), which updates models by retraining on the most recent data, can converge to a stable model that minimizes risk on its own induced distribution. However, existing analyses typically assume access to the complete distributions of features and labels after model deployment, ignoring the possibility of selective labels: observing labels only for the accepted subset of the population. In this work, we formalize performative prediction with selective labels and show that retraining only on observed data can misguide the retraining procedure and undermine the guarantees of convergence to a stable solution. We then propose a worst-case objective based on knowledge of a confidence interval on the probability of a positive label. Applying RRM to this objective permits us to remain within a bounded distance to the true stable point. Under a sensitivity assumption on the conditional label distribution, we further show how previously accepted data can tighten these confidence intervals over time. Experiments in a lending application with fairness regularization show that our robust optimization approach closely matches the performance of RRM with complete label access.

Classical reinforcement learning (RL) assumes that a reward is observed for every visited state-action pair. However, in real-world applications such as autonomous driving, such fine-grained feedback can be costly or difficult to collect, whereas trajectory-level feedback may be too sparse for efficient learning. To provide a general feedback model bridging these two extremes and handle large state spaces, we study RL with segment reward feedback under linear function approximation. Our work answers how the granularity of segment feedback and the choice of segmentation influence learning. For equal-length segments with known transitions, we design algorithms $\bitssegd$ and $\edlinucbsegd$ for binary and sum feedback types, respectively. They adopt posterior sampling with planning to achieve computational efficiency and the E-optimal experimental design to attain near-optimality. Nearly matching lower bounds are established. For equal-length segments with unknown transitions, we develop a unified $\seglsvits$ framework with two instantiations for binary and sum feedback, which carefully integrates the posterior estimated reward parameters into least-squares value iteration. These results reveal a fundamental insight: under binary feedback, increasing the number of segments significantly reduces the regret through an exponential factor, while surprisingly, under sum feedback, the granularity of segments does not affect learning much. Finally, to investigate whether segmenting according to state-action features can further expedite learning, we design an algorithm $\uneqsegbitsd$ that allows arbitrary segmentations. The resulting regret bound shows that under the usual elliptical potential analysis, the influence of state-action features on the regret appears only through logarithmic factors, and equal segmentation achieves the best performance.

Pervasive intelligent applications are increasingly deployed on mobile and Internet of Things (IoT) edge devices. Consequently, Large Language Models (LLMs) are increasingly used to support these applications. Yet, due to their high resource demands, LLMs are mostly deployed in the cloud. Layer-wise edge-cloud inference lets resource-constrained edge devices contribute computation to LLMs they cannot host in full. However, heterogeneous split points introduce two coupled inefficiencies. First, edge execution and communication create idle gaps between cloud invocations. Second, requests arriving at different model depths cannot be conventionally batched. We present DySCo, a collaborative runtime that keeps KV caches local and introduces dyForward, a model-aware layer-range executor that runs configurable contiguous layer ranges from resident model shards without reloading weights. For multi-edge serving settings, we introduce depth-synchronized batching (DSB), which advances heterogeneous requests to the deepest cut and batches their common suffix. Experiments across heterogeneous devices, two model families, and local and wide-area links show that idle gaps increase the latency of subsequent GPU forward calls even when waiting time is excluded, adding up to 25 ms of additional cloud-side suffix latency per decoding step in our measurements. At an average concurrency of eight, DSB improves throughput by 275% over FIFO, 48% over exact-match batching, and 79% over round-robin interleaving while reducing mean per-session latency. Together, these results show that requests with different edge-cloud splits can reuse resident cloud weights and share batched suffix computation. The artifact repository for this work is publicly available at: https://github.com/Large-scale-Sustainable-Computing-LSC/dysco-artifact

Artificial intelligence capabilities are often measured as resources accumulated within firms, yet they also circulate across organizational boundaries through worker mobility. Drawing on a database of approximately 535 million employment records across 58 countries, we reconstruct the inter-firm network of AI talent flows over 2010–2022. We find that AI talent inflows are becoming increasingly concentrated among a small set of leading firms, with stronger preferential attachment than in general labor mobility. Yet the network core remains contestable: central positions in AI networks change hands far more often than in non-AI networks. Network position also carries information that goes beyond workforce size. This network position is economically consequential: AI network centrality is associated with higher enterprise value, after accounting for workforce scale, assets, and fixed effects. Event-study estimates further indicate that sharp improvements in network position are followed by higher enterprise value relative to matched firms with similar pre-event trajectories. Together, these findings suggest that corporate AI advantage depends not only on how much talent firms accumulate, but also on where they sit in its circulation.

In a previous paper, we investigate the control problem of discrete event systems under cyber attacks, where CA-controllability and CA-observability are proposed. We prove that a discrete event system can be controlled to generate a specification language K if and only if K is CA-controllable and CA-observable. While CA controllability can be "converted" to (conventional) controllability, CA-observability cannot be "converted" to (conventional) observability. In this paper, we further investigate CA-observability. We develop a method and an algorithm to check CA-observability. The method involves constructing an augmented automaton whose states are pairs of the current state and state estimate of the supervised system. We prove that K is CA-observable if and only if the language generated by the augmented automaton is equal to K. An algorithm is then developed to check CA-observability. We also investigate the properties of CA-observability. We show that CA-observability is not closed under (set) in tersection. Note that (conventional) observability is closed under intersection. We further show that CA-observability is not closed under union. This is similar to observability. If K is not CA-controllable and CA-observable, we pro pose a method to calculate a CA-controllable and CA observable sublanguage. We first calculate the supremal CA-controllable sublanguage for K which can be described by a sub-automaton of the automaton for K. We then propose a state-estimate-based supervisor based on the subautomaton, which ensures the supervised system stays within K.

Can authentication make memory, rather than computational hardness, the attacker's bottleneck? Contextual Chain is a lightweight continuity protocol for intermittently connected devices that share evolving physical or operational context. An honest device follows one realized history, updating a compact accumulator and fixed hash-based readiness lanes; outages cause pause or bounded rollback, not branch search. After the epoch is frozen, a fresh challenge selects one lane under a short deadline. An outsider that missed context may therefore need to prepare for many mature histories before learning which one will be tested. In the standard random-oracle model, a causal counting theorem lower-bounds the deadline-accessible retained state required for a target success probability against arbitrary nonlinear preselection encoding and adaptive post-selection queries, accounting for sequential depth, candidate queries, and cross-target protected information obtained online. Honest readiness memory remains fixed and independent of the number of plausible histories. Contextual Chain thus converts shared-experience uncertainty into a tunable preparation requirement without transferring combinatorial complexity to lightweight devices.

Auditing the claimed outcomes of large language model (LLM) training is challenging when model weights and training data are private, while cryptographically proving the full training process is prohibitively expensive at Transformer scale. We present zkLLMPoT, a zero-knowledge framework that certifies auditor-defined properties of a trained checkpoint through forward evaluation rather than verification of its optimization trajectory. zkLLMPoT includes 2 phases: 1) The trainer fixes the architecture and the model weights are committed. Then the auditor selects challenge sequences, preventing the trainer from modifying the checkpoint in response to the audit data. 2) Then the trainer proves the objective value attained by the committed model on those sequences. This formulation makes the certification cost independent of the number of training iterations, without revealing model weights or requiring access to private training data. We build on sumcheck- and lookup-based arguments to certify Transformer computations, while supporting next-token loss and task-specific audit objectives. Across four model families, operator-level benchmarks yield proving times of 41-59 seconds for 1.1-1.5B-parameter models and 131 seconds at 13B for the covered operators, with verification below half a second at a sequence length of 512.