arXiv:2607.04634v1 Announce Type: new Abstract: In heterogeneous robot swarms, caste reassignment (rebinding a robot to a new capability-bound role) is a high-frequency runtime event driven by battery, payload, and priority changes. Existing approaches treat it as an internal allocation algorithm and do not expose the reassignment to external authority. We argue that for regulated embodied deployments a caste change that elevates a robot's privilege envelope is a governance event that must be auditable and externally authorised. We propose an asymmetric-trust protocol: auto-tightening reassignments (to safer, lower-privilege castes) are admitted automatically, while bounded relaxation (to higher-privilege castes) requires an operator countersignature against a per-axis budget. Each transition carries a signed cause-chain, committed to a hash-chained Merkle audit log that an offline auditor verifies from an operator-signed identity manifest alone. We evaluate a reference implementation with real Ed25519 signatures over fleets up to 100 robots: auto-tightening completes in single-digit to low-double-digit milliseconds, and the governed protocol refuses four explicit attacks (caste laundering, repeated-relaxation escalation, operator impersonation, cause-chain forgery) by construction, with a partially-governed baseline isolating which gate stops which attack and a randomized fuzz adversary finding no admission. A distributed audit layer replicates the log across N per-member replicas with quorum-committed total order and cryptographic fork exclusion; we prove agreement and fork exclusion and validate them both in simulation and as a real multi-process deployment over TCP sockets (up to 100 real processes) with a Byzantine equivocator, on which every honest replica agrees, detects the equivocation, and commits no fork. The construction generalises a single-agent persona-mutation governance gate to swarm-level caste governance.
Science Journals
arXiv:2510.08807v3 Announce Type: replace Abstract: From loco-motion to dextrous manipulation, humanoid robots have made remarkable strides in demonstrating complex full-body capabilities. However, the majority of current robot learning datasets and benchmarks mainly focus on stationary robot arms, and the few existing humanoid datasets are either confined to fixed environments or limited in task diversity, often lacking human-humanoid interaction and lower-body locomotion. Moreover, there are a few standardized evaluation platforms for benchmarking learning-based policies on humanoid data. In this work, we present Humanoid Everyday, a large-scale and diverse humanoid manipulation dataset characterized by extensive task variety involving dextrous object manipulation, human-humanoid interaction, locomotion-integrated actions, and more. Leveraging a highly efficient human-supervised teleoperation pipeline, Humanoid Everyday aggregates high-quality multimodal sensory data, including RGB, depth, LiDAR, and tactile inputs, together with natural language annotations, comprising 10.3k trajectories and over 3 million frames of data across 260 tasks across 7 broad categories. In addition, we conduct an analysis of representative policy learning methods on our dataset, providing insights into their strengths and limitations across different task categories. For standardized evaluation, we introduce a cloud-based evaluation platform that allows researchers to seamlessly deploy their policies in our controlled setting and receive performance feedback. By releasing Humanoid Everyday along with our policy learning analysis and a standardized cloud-based evaluation platform, we intend to advance research in general-purpose humanoid manipulation and lay the groundwork for more capable and embodied robotic agents in real-world scenarios. Our dataset, data collection code, and cloud evaluation website are made publicly available on our project website.
arXiv:2512.11308v2 Announce Type: replace Abstract: This paper proposes the framework of an efficient gig-work management system. A gig-work management system recommends one-off tasks with information about task hours and wages to gig-workers. To enable effective management, this paper develops a model of gig-workers' decision-making. Then, based on the model, we formulate an optimization problem to determine the optimal task hours and wages. The formulated problem belongs to the class of chance-constrained model predictive control (CC-MPC) problems. To efficiently solve the CC-MPC problem, we develop an approximate solution algorithm with guaranteed confidence levels. Finally, we develop gig-worker models based on data collected through crowdsourcing.
arXiv:2607.05134v1 Announce Type: new Abstract: We present PDEFlow, an autonomous agentic framework that turns user-level ODE and PDE descriptions into solver-backed neural-operator pipelines. The workflow links problem specification, data generation, operator training, and checkpoint-based inference. A stateful input graph converts multi-turn natural-language input and user edits into validated problem specifications. The data-generation module then samples parameters, solves the configured governing-equation with FEniCSx finite-element backend, and stores the solutions as operator-ready tensors. The training and inference stages use a registry-based interface, allowing different neural operators to be trained and deployed without changing the surrounding pipeline. In the current implementation, we instantiate this interface with a multi-branch Bayesian DeepONet. Experiments on benchmark ODE and PDE tasks show that PDEFlow can construct valid specifications, generate solver-backed datasets, train neural operators across steady and transient problem classes, and provide solver-free predictions from saved checkpoints. The framework is designed for repeatable scientific and engineering workflows where many related physics configurations must be specified, simulated, learned, and queried with minimal manual intervention.
arXiv:2607.04860v1 Announce Type: new Abstract: Gaze target estimation, the task of predicting where a person is looking in a scene, is crucial to understanding human attention and intent. It is a challenging task that combines high-level understanding of global scene semantics and precise spatial reasoning using human appearance (e.g. pose, eye orientation). As a result, human-level performance remains elusive for existing models, limiting their practical application. To this end, we propose PaGE (Practical Gaze Estimator), a gaze estimation model that explicitly models the complex interaction between scene and head features. Using a PaGE model with a large ViT-H+ backbone as the teacher, we further distill student models with lighter backbones on a much larger and more diverse unlabeled dataset. The architectural improvements and novel training recipe allow PaGE to achieve state-of-the-art performance on several gaze estimation tasks, outperforming humans in 7 out of 9 metrics while reducing the human-AI gap by at least 60% in the remaining 2. The distilled student models retain most of the teacher's performance while being lightweight enough for practical deployment on robots and consumer devices. The code and model checkpoints are available at our project page.
arXiv:2507.06169v3 Announce Type: replace-cross Abstract: In recent years, there has been significant interest in characterizing the induced subgraph obstructions to bounded treewidth and pathwidth. While this has recently been resolved for pathwidth, the case of treewidth remains open, and prior work has reduced the problem to understanding the layered-wheel-like obstructions -- graphs that contain large complete minor models with each branching set inducing a path; exclude large walls as induced minors; exclude large complete bipartite graphs as induced minors; and exclude large complete subgraphs. There are various constructions of such graphs, but they are all rather involved. In this paper, we present a simple construction of layered-wheel-like graphs with arbitrarily large treewidth. Three notable features of our construction are: (a) the vertices of degree at least four can be made to be arbitrarily far apart; (b) the girth can be made to be arbitrarily large; and (c) every outerstring induced subgraph of the graphs from our construction has treewidth bounded by an absolute constant. In contrast, among several previously known constructions of layered wheels, none achieves (a); at most one satisfies either (b) or (c); and none satisfies both (b) and (c) simultaneously. In particular, this is related to a former conjecture of Trotignon, that every graph with large enough treewidth, excluding large walls and large complete bipartite graphs as induced minors, and large complete subgraphs, must contain an outerstring induced subgraph of large treewidth. Our construction provides the first counterexample to this conjecture that can also be made to have arbitrarily large girth.
arXiv:2507.08922v3 Announce Type: replace-cross Abstract: Continual learning is an online paradigm where a learner continually accumulates knowledge from different tasks encountered over sequential time steps. Importantly, the learner is required to extend and update its knowledge without forgetting about the learning experience acquired from the past, and while avoiding the need to retrain from scratch. Given its sequential nature and its resemblance to the way humans think, continual learning offers an opportunity to address several challenges which currently stand in the way of widening the range of applicability of deep models to further real-world problems. The continual need to update the learner with data arriving sequentially strikes inherent congruence between continual learning and Bayesian inference which provides a principal platform to keep updating the prior beliefs of a model given new data, without completely forgetting the knowledge acquired from the old data. This survey inspects different settings of Bayesian continual learning, namely task-incremental learning and class-incremental learning. We begin by discussing definitions of continual learning along with its Bayesian setting, as well as the links with related fields, such as domain adaptation, transfer learning and meta-learning. Afterwards, we introduce a taxonomy offering a comprehensive categorization of algorithms belonging to the Bayesian continual learning paradigm. Meanwhile, we analyze the state-of-the-art while zooming in on some of the most prominent Bayesian continual learning algorithms to date. Furthermore, we shed some light on links between continual learning and developmental psychology, and correspondingly introduce analogies between both fields. We follow that with a discussion of current challenges, and finally conclude with potential areas for future research on Bayesian continual learning.
arXiv:2607.05396v1 Announce Type: new Abstract: Real-world robot deployment rarely maintains the training-stage camera setup, where cameras often experience repositioning or remounting depending on actual scenarios. Existing view-robust Vision-Language-Action (VLA) policies tolerate such camera variations only when the camera extrinsics are explicitly provided, making them fragile and hard to use especially when view robustness is critical. We argue that the policy should not be told where the camera is, but rather figure it out by itself. To this end, we introduce Camera-Centric VLA (CamVLA), a new VLA model that decouples manipulation controls from camera geometry by predicting (i) a camera-centric end-effector action expressed in the local camera frame, and (ii) a 6-DoF hand-eye matrix relating cameras to the robot base. A deterministic geometric transformation composes the two predictions into a robot base-frame action. This disentangles how I should move in pose-independent camera-centric action generation from where I am looking from in camera-perspective geometric grounding. The resulting policy is calibration-free, depth-free, and single-view, requiring only a single monocular RGB image as the visual observation and task instruction at deployment. Evaluations in both simulation and real-world robot data show that CamVLA consistently improves success rates across diverse unseen viewpoints. Project page: https://alibaba-damo-academy.github.io/CamVLA/.
arXiv:2208.00657v2 Announce Type: cross Abstract: Building detection and change detection using remote sensing images can help urban and rescue planning. Moreover, they can be used for building damage assessment after natural disasters. Currently, most of the existing models for building detection use only one image (pre-disaster image) to detect buildings. This is based on the idea that post-disaster images reduce the model's performance because of presence of destroyed buildings. In this paper, we propose a siamese model, called SiamixFormer, which uses pre- and post-disaster images as input. Our model has two encoders and has a hierarchical transformer architecture. The output of each stage in both encoders is given to a temporal transformer for feature fusion in a way that query is generated from pre-disaster images and (key, value) is generated from post-disaster images. To this end, temporal features are also considered in feature fusion. Another advantage of using temporal transformers in feature fusion is that they can better maintain large receptive fields generated by transformer encoders compared with CNNs. Finally, the output of the temporal transformer is given to a simple MLP decoder at each stage. The SiamixFormer model is evaluated on xBD, and WHU datasets, for building detection and on LEVIR-CD and CDD datasets for change detection and could outperform the state-of-the-art.
arXiv:2603.17418v4 Announce Type: replace Abstract: Distribution grid analyses include tasks such as network information retrieval, power-flow analysis, hosting-capacity assessment, DER planning, and state estimation. Completing these tasks often requires long-horizon, stateful workflows in which an engineer retrieves data, loads a feeder, runs simulations, evaluates results, and exports outputs. The growing volume of these analyses is outpacing the limited engineering workforce, causing suboptimal outcomes and delays. Large Language Model (LLM)-orchestrated agents can help, but they often struggle for two reasons: (i) they lack algorithms to determine the right context for an unseen grid task, and (ii) they cannot verify proposed actions against the environment state beforehand and instead rely on feedback after execution. We propose PowerDAG, an agentic artificial intelligence (AI) system that formalizes workflows as directed acyclic graphs (DAGs) and addresses current gaps in this formalism through two mechanisms, adaptive retrieval and Just-in-Time supervision. To dynamically retrieve relevant context, it curates and ranks expert exemplars using an adaptive score-decay cutoff that matches the query complexity. For supervision, it evaluates prerequisites before every tool call. If an agent proposes an invalid action, the supervisor blocks execution, preserves the environment, and returns a corrective advisory. We evaluate PowerDAG on 150 held-out queries from a 200-record expert-verified benchmark that covers 10 of the most commonly performed distribution-grid analyses, comparing 6 agentic systems across 10 LLMs for a total of 9,000 runs. PowerDAG reaches a success rate of 98.0% with GPT-5.5, 97.3% with Gemini 3.1 Pro, and 92.7% with Qwen3.6-27B, improving success rates by 6 to 50 percentage points over baselines.
arXiv:2510.21692v3 Announce Type: replace-cross Abstract: This paper lays out the principles of how Bose-Einstein condensates can modify radioactive decay. We highlight the challenges of many modes and short coherence times due to the $\approx$ MeV energies of the emitted radiation. Recent proposals for gamma ray and neutrino lasers claim that using a Bose-Einstein condensate as a source would solve these issues. We show that this is not the case, and the proposed experiments would have a gain of only $10^{-16}$ or smaller. We also analyze proposals for gamma ray lasers based on stimulated annihilation of positronium Bose-Einstein condensates.
arXiv:2510.22270v2 Announce Type: replace-cross Abstract: This paper aims to investigate the distributed stochastic optimization problems on compact embedded submanifolds (in the Euclidean space) where the local cost functions are weakly-convex. To address the manifold structure, we propose a distributed Riemannian stochastic proximal algorithm framework by utilizing the retraction and Riemannian consensus protocol, and analyze three specific algorithms: the distributed Riemannian stochastic subgradient, proximal point, and prox-linear algorithms. When the initial points satisfy certain conditions, we show that the iterates generated by this framework converge to a nearly stationary point in expectation while achieving consensus. We further establish the convergence rate of the algorithm framework as $\mathcal{O}(\frac{1+\kappa_g}{\sqrt{k}})$ where $k$ denotes the number of iterations and $\kappa_g$ shows the impact of manifold geometry on the algorithm performance. Finally, numerical experiments are implemented to demonstrate the theoretical results and show the empirical performance.
arXiv:2601.22652v2 Announce Type: replace-cross Abstract: Spectral gradient methods, such as the Muon optimizer, modify gradient updates by preserving directional information while discarding scale, and have shown strong empirical performance in deep learning. We investigate the mechanisms underlying these gains through a dynamical analysis of a nonlinear phase retrieval model with anisotropic Gaussian inputs, equivalent to training a two-layer neural network with the quadratic activation and fixed second-layer weights. Focusing on a spiked covariance setting where the dominant variance direction is orthogonal to the signal, we show that gradient descent (GD) suffers from a variance-induced misalignment: during the early escaping stage, the high-variance but uninformative spike direction is multiplicatively amplified, degrading alignment with the true signal under strong anisotropy. In contrast, spectral gradient descent (SpecGD) removes this spike amplification effect, leading to stable alignment and accelerated noise contraction. Numerical experiments confirm the theory and show that these phenomena persist under broader anisotropic covariances.
arXiv:2512.16396v2 Announce Type: replace-cross Abstract: We establish $L^p$-universal approximation theorems for general path-dependent and non-anticipative functionals on suitable rough path spaces, showing that linear functionals acting on signatures of time-extended rough paths are dense with respect to the $L^p$-distance. To that end, we derive global universal approximation theorems for weighted rough path spaces. We demonstrate that these $L^p$-universal approximation theorems apply to Gaussian processes, in particular, to fractional Brownian motion. As a consequence, linear functionals on the signature of the time-extended Brownian motion can approximate any $p$-integrable stochastic process adapted to the Brownian filtration, including solutions to stochastic differential equations.
arXiv:2505.16713v3 Announce Type: replace-cross Abstract: We examine the concentration of uniform generalization errors around their expectation in binary linear classification problems via an isoperimetric argument. In particular, we establish Poincar\'{e} and log-Sobolev inequalities for the joint distribution of the output labels and the label-weighted input vectors, which we apply to derive concentration bounds. The derived results improve upon existing bounds obtained from general unbounded empirical processes, as well as that tailored specifically to logistic regression. In asymptotic analysis, we also show that almost sure convergence of uniform generalization errors to their expectation occurs in very broad settings, such as proportionally high-dimensional regimes. Using this convergence, we establish uniform laws of large numbers under dimension-free conditions.
arXiv:2602.02190v2 Announce Type: replace-cross Abstract: A common approach to perform PCA on probability measures is to embed them into a Hilbert space where standard functional PCA techniques apply. While convergence rates for estimating the embedding of a single measure from $m$ samples are well understood, the literature has not addressed the setting involving multiple measures. In this paper, we study PCA in a double asymptotic regime where $n$ probability measures are observed, each through $m$ samples. We derive convergence rates of the form $n^{-1/2} + m^{-\alpha}$ for the empirical covariance operator and the PCA excess risk, where $\alpha>0$ depends on the chosen embedding. This characterizes the relationship between the number $n$ of measures and the number $m$ of samples per measure, revealing a sparse (small $m$) to dense (large $m$) transition in the convergence behavior. Moreover, we prove that the dense-regime rate is minimax optimal for the empirical covariance error. Our numerical experiments validate these theoretical rates and demonstrate that appropriate subsampling preserves PCA accuracy while reducing computational cost.
arXiv:2607.05154v1 Announce Type: new Abstract: Phase-separated liquid droplets organize molecules in cells, but the underlying physical principles differ from abiotic mixing and quantitative rules in living systems remain poorly understood. The pyrenoid -- a liquid-like organelle that enhances photosynthetic carbon fixation in algae and hornworts -- provides an unusually tractable model system. Here, we review recent advances in our understanding of pyrenoids from the perspective of biophysics. We highlight how reaction-diffusion models connect compartment architecture to catalytic performance, how soft matter theories link molecular interactions to condensate assembly, and how modern experimental methods enable these predictions to be tested quantitatively. Recent studies suggest that pyrenoid function may be described by a small number of effective transport and reaction processes, while condensate assembly can be understood through molecular design parameters and thermodynamic constraints. Together, these findings establish the pyrenoid as a powerful system for investigating catalytic compartmentalization, biomolecular self-organization and the emergence of effective physical descriptions in living systems.
arXiv:2606.30988v3 Announce Type: replace Abstract: Robot manipulation often relies on sensory feedback beyond vision, particularly in contact-rich settings where force, tactile, or audio signals reveal interaction states that are not directly observable from images. However, these modalities are often hardware- and task-specific, and large-scale multisensory robot datasets remain scarce. As a result, it is impractical to pretrain policies with every sensor they may encounter. We study multisensory continual learning: adapting a pretrained robot policy to new tasks with newly introduced modalities while preserving performance under the original sensor suite. We propose MultiSensory World Model (MuSe), which incorporates limited multisensory data into pretrained vision-only policies through multi-stage fusion, multisensory future prediction, and experience replay over pretraining data. We instantiate MuSe by augmenting a pretrained vision-only policy with force-torque sensing and evaluate it on real-world manipulation tasks. Our experiments show that MuSe performs strongly on contact-rich finetuning tasks while preserving, and in some cases improving, performance on the original pretraining tasks. These results suggest that a modest multisensory dataset can improve general robot capabilities beyond the finetuning distribution. Project website: https://jadenvc.github.io/multisensory-continual-learning/
arXiv:2607.03161v1 Announce Type: cross Abstract: In selective deployment, practitioners act only on a model-chosen subset of individuals based on predicted conditional average treatment effects, but marginal conformal guarantees need not control reliability on that selected subset. We study reliable selection for black-box CATE predictors: selecting candidates whose CATE errors are below a tolerance while controlling the false discovery rate (FDR). Since CATE errors are unobservable, we construct doubly robust proxy errors from pseudo-outcomes; however, naive proxies can lose power under heteroskedasticity because variance overwhelms the reliability signal. We propose Denoised Conformal Alignment, which subtracts an estimated conditional variance component and combines conformal calibration with Benjamini--Hochberg selection. Our analysis shows that validity is governed by stability of proxy/oracle threshold labels, rather than pointwise perfection of the variance estimator. Experiments show substantially improved power while maintaining FDR control across challenging settings.
arXiv:2605.06890v4 Announce Type: replace Abstract: AI agents are promising for high-stakes enterprise workflows, but dependable deployment remains limited because these tool-use decisions are difficult to diagnose and control. Agents may skip required tool calls, invoke tools unnecessarily, or take actions whose consequences become visible only after execution. Existing observability methods are external: prompts reveal correlations, evaluations score outputs, and logs arrive only after the model has already acted. In long-horizon settings, these failures are costly because an early tool mistake can alter the rest of the execution trajectory, increase token consumption, and create downstream safety and security risk. We introduce a mechanistic-interpretability toolkit built on Sparse Autoencoders (SAEs), which decompose activations into sparse internal features, and linear probes, lightweight classifiers that read signals from those features. The framework reads model states before each action and infers whether a tool is needed and how risky the next tool action is. It identifies the model layers and features most associated with tool decisions and tests their functional importance through feature ablation. We train the probes on multi-step agent execution traces from the NVIDIA Nemotron function-calling dataset and apply the same workflow to GPT-OSS 20B and Gemma 3 27B models. The goal is not to replace external evaluation, but to add a missing layer: visibility into what the model signaled internally before action. This helps surface deeper causes of agent failure, especially in long-horizon runs where an early mistake can impact subsequent agent behavior. More broadly, the paper shows how mechanistic interpretability can support internal observability for monitoring tool calls and risk in agent systems.
arXiv:2511.04805v2 Announce Type: replace Abstract: Mixture-of-Experts (MoE) models have shown strong potential in scaling language models efficiently by activating only a small subset of experts per input. However, their widespread deployment remains limited due to the high memory overhead associated with storing all expert parameters, particularly as the number of experts increases. To address this challenge, prior works have explored expert dropping and merging strategies, yet they often suffer from performance drop at high compression ratios. In this paper, we introduce PuzzleMoE, a training-free MoE compression method that achieves both high accuracy and efficient inference through two key innovations: First, PuzzleMoE performs sparse expert merging by identifying element-wise weight redundancy and specialization. It uses a dual-mask to capture both shared and expert-specific parameters. Second, to avoid the overhead of storing binary masks and signs, PuzzleMoE introduces a bit-packed encoding scheme that reuses underutilized exponent bits, enabling efficient MoE inference on GPUs. Extensive experiments demonstrate that PuzzleMoE can compress MoE models by up to 50% while maintaining accuracy across various tasks. Specifically, it outperforms prior MoE compression methods by up to 16.7% on MMLU at 50% compression ratio, and achieves up to 1.28\times inference speedup.
arXiv:2603.14038v2 Announce Type: replace Abstract: We present a machine-verified proof of the correctness of Toom-Cook multiplication with generalized integer evaluation points. Toom-Cook is a class of fast multiplication algorithms parameterized by a triple $(k_x, k_y, \vec v)$ consisting of two positive integer split sizes $k_x, k_y$ and a vector $\vec v$ of distinct evaluation points. As part of our proof, we verify that for any selection of $k_x+k_y-1$ distinct integer evaluation points, we can compute a threshold function $\theta(k_x, k_y, \vec v)$ such that, if the algorithm's base-case problem size is set above this threshold, then the algorithm's termination is guaranteed regardless of the values of the operands. The threshold formula, which we derive by obtaining upper bounds on the subproblem sizes produced by the Toom-Cook recurrence, does not depend on the operands; it depends only on $k_x$, $k_y$, $\vec v$, and the base $b$ in which we operate. We write the proof in Lean 4, making use of the Mathlib library. We formalize the algorithm, our base case threshold formula, and our key lemma statements in Lean. We then use the AI theorem prover Aristotle to assist in completing the machine verification of the algorithm's correctness. This proof, through its synthesis of human input and AI assistance, demonstrates the considerable power of AI to automate the machine verification process.
arXiv:2603.22770v2 Announce Type: replace Abstract: The deployment of deep neural networks (DNNs) in safety-critical edge environments necessitates robustness against hardware-induced bit-flip errors. While empirical studies indicate that reducing numerical precision can improve fault tolerance, the theoretical basis of this phenomenon remains underexplored. In this work, we study resilience as a structural property of neural architectures rather than solely as a property of a dataset-specific trained solution. By deriving the expected squared error (MSE) under independent parameter bit flips across multiple numerical formats and layer primitives, we show that lower precision, higher sparsity, bounded activations, and shallow depth are consistently favored under this corruption model. We then argue that logic and lookup-based neural networks realize the joint limit of these design trends. Through ablation studies on the MLPerf Tiny benchmark suite, we show that the observed empirical trends are consistent with the theoretical predictions, and that LUT-based models remain highly stable in corruption regimes where standard floating-point models fail sharply. Furthermore, we identify a novel even-layer recovery effect unique to logic-based architectures and analyze the structural conditions under which it emerges. Overall, our results suggest that shifting from continuous arithmetic weights to discrete Boolean lookups can provide a favorable accuracy-resilience trade-off for hardware fault tolerance.
arXiv:2506.18020v3 Announce Type: replace Abstract: Robust distributed learning algorithms aim to maintain reliable performance despite the presence of misbehaving workers. Such misbehaviors are commonly modeled as \textit{Byzantine failures}, allowing arbitrarily corrupted communication, or as \textit{data poisoning}, a weaker form of corruption restricted to local training data. While prior work shows similar optimization guarantees for both models, an important question remains: \textit{How do these threat models impact generalization?} We show, for the first time, a fundamental gap in generalization guarantees between the two threat models: Byzantine failures yield strictly worse rates than those achievable under data poisoning. Our findings are based upon a tight algorithmic stability analysis of robust distributed learning. Specifically, with $f$ out of $n$ workers misbehaving, we prove that: \textit{(i)} under data poisoning, the uniform algorithmic stability of a robust distributed learning algorithm
arXiv:2603.22107v3 Announce Type: replace Abstract: In this paper we propose a detectability condition for nonlinear continuous-time systems with irregular/infrequent output measurements, namely a sample-based version of incremental integral input/output-to-state stability (i-iIOSS). We provide a sufficient condition for an i-iIOSS system to be sample-based i-iIOSS. This condition is also exploited to analyze the relationship between sample-based i-iIOSS and sample-based observability for linear systems, such that previously established sampling strategies for linear systems can be used to guarantee sample-based i-iIOSS. Furthermore, we present a sample-based moving horizon estimation scheme, for which robust stability can be shown. Finally, we illustrate the applicability of the proposed estimation scheme through a biomedical simulation example.