AV-FlowEdit is a training‑free, inversion‑free framework that enables prompt‑driven audio‑visual editing using joint audio‑video diffusion models. It leverages source‑to‑target flow differences to guide edits while preserving non‑target structure and cross‑modal correspondence. The method includes a Cross‑Modal Event‑Aware Spectral Trajectory Preservation module that suppresses drift by combining flow saliency, aligned audio activity, multiscale low‑frequency displacement, and diffusion‑stage scheduling, achieving strong prompt alignment, temporal consistency, and competitive audio quality.
By Shaohua Pan, Xiao Liu
The paper introduces HF‑WM, an action‑conditioned world model that learns from large‑scale egocentric human videos and incorporates failure dynamics through post‑training. It bridges the embodiment gap with a dual‑branch framework that couples video generation to stabilized pointflow prediction, unifying motion representation across humans and robots while reducing egocentric camera motion. Failure modeling is enhanced by Group Relative Policy Optimization on perturbed actions, guided by action alignment and physical plausibility rewards, enabling the model to anticipate unsuccessful executions. Experiments show HF‑WM outperforms existing methods in action‑conditioned video generation and is useful for evaluating policies in downstream robotic tasks.
By Peiyan Tu, Cong Wang, Hanxin Zhu, Xin Jin, Zhibo Chen
Linear Multi-granular Body Modeling for Text-Driven Motion Generation proposes HiSTF Mamba, a linear-time framework that combines Dual‑Spatial Mamba, Bi‑Temporal Mamba, and a Dynamic Spatiotemporal Fusion Module to model both part‑level and whole‑body dynamics. The approach captures local joint dynamics and global coordination in parallel, encodes short‑term and long‑term motion details bidirectionally, and removes redundant features to produce an expressive spatiotemporal representation. Experiments on HumanML3D and KIT‑ML show that HiSTF Mamba achieves higher generation fidelity and semantic alignment while using far fewer parameters than existing segment‑based architectures.
By Xingzu Zhan, Genpei Zhang, Chen Xie, Honghang Chen, Haoran Sun, Xiaochun Mai
The paper introduces Multi-Modal Typography, a systematic investigation into how perturbations applied to speech and on-screen text can compromise audio‑visual reasoning in large language models. Experiments on the WorldSense benchmark show that such attacks can raise the attack success rate against Gemini 3.7 Flash from 9.53% to 43.41%, and fine‑tuning can reduce the success rate for Qwen2.5‑Omni‑7B from 60.6% to 52.0%. The study highlights that cross‑modal typography attacks are a significant, yet underexplored, vulnerability in multi‑modal reasoning systems.
By Tianle Chen, Deepti Ghadiyaram
The paper investigates how sequence design influences temporal synchronization in unified decoder-only transformers for video‑text‑to‑speech synthesis. By training a Visatronic model on VoxCeleb2, the authors compare positional encoding schemes and modality ordering, finding that global sequential indexing and co‑temporal ordered indexing yield similar in‑domain performance, while text‑first co‑temporal indexing best transfers under domain shift. They also introduce TimeSync, a phoneme‑level metric that uncovers misalignments missed by frame‑level measures, and demonstrate that text and video provide complementary cues for intelligibility and temporal/emotional expressiveness.
By Akshita Gupta, Tatiana Likhomanenko, Karren Dai Yang, Richard He Bai, Zakaria Aldeneh, Navdeep Jaitly
The paper argues that linear models can still perform well in complex, changing environments because the world’s complexity often exceeds what can be reliably learned from finite, partially observed data. It explains that flexible, nonlinear models may capture idiosyncratic patterns specific to a particular setting—such as a hospital or policy period—leading to poor generalization when conditions shift. The authors formalize this intuition with a squared‑error proposition and discuss how factors like bias‑variance trade‑offs, regularization, measurement error, and distribution shift influence the choice between linear and nonlinear approaches.
By Fangzhou Lin, Dengzhe Hou, Yinxing Li, Kazunori D Yamada
FedPCD introduces a priority-constrained federated optimization framework that decouples the global aggregation objective from client protection. By preserving the standard federated aggregate and applying only minimal corrections to satisfy client performance constraints, FedPCD achieves higher global accuracy and improved worst‑10% client accuracy across heterogeneous benchmarks. The method also maintains competitive trade‑offs between global accuracy and client variability while converging faster with fewer communication rounds.
By Yousef Irshaid, Dara Varam, Mohamed I. AlHajri
arXiv:2610. 13631v1 Announce Type: new Abstract: Fault diagnosis in asset-intensive industries requires multiclass models trained on noisy, sensitive data collected across geographically distributed client sites.
By Andrea P. Hernandez, Michael Ibrahim, Nagi Gebraeel
The paper presents a method for exact machine unlearning in federated socially fair k‑means clustering, aiming to maintain fairness across demographic groups while allowing data deletion. It introduces separate local summaries per group so that unaffected institutions can keep valid global fairness weights, and demonstrates that reusing cached computations—especially when an entire client withdraws—can dramatically speed up retraining. Experiments on three datasets show significant speedups for initialization and moderate gains for individual record deletions, highlighting when exact unlearning is most beneficial.
By Aravind Reddy, Deeraj S K, Mohd Wasif Raza, Saurav Prakash
The paper argues that neural networks can exhibit modular organization even when individual neurons are not specialized. It introduces a nested hierarchy of modularity—unit-level, orthogonal, and linear—and shows that many vision, language, and RNN models lack unit-level modularity but still display clear functional modularity at the subspace level. The authors demonstrate that these subspaces can be causally validated through ablation and align with brain data, revealing distinct content/style or syntax/semantics separations.
By Shreya Saha, Benjamin Bergen, Meenakshi Khosla
The paper investigates discrete graph diffusion models (GDMs) and demonstrates that they can memorize training data when conditioned on corrupted training graphs. By analyzing reverse diffusion trajectories on two‑block stochastic block models, the authors find that the model almost perfectly reconstructs the corrupted training graph, a phenomenon they call conditional memorization. This memorization coincides with a growing generalization gap and precedes the emergence of community structure in generated graphs, occurring across all tested node sizes.
By Adarsh Jamadandi, Nicolas Keriven
The paper proposes a holistic architectural and algorithmic framework for sustainable cloud‑fog‑edge systems in smart buildings, emphasizing real‑time responsiveness and reduced environmental impact. It validates resilient data acquisition using affordable Arduino hardware and secure persistence solutions, introduces the lightweight KOptim orchestration framework for dynamic SLA management, and adapts the Generalized Hebbian Algorithm for PCA on low‑power microcontrollers with an iterative learning strategy that preserves data diversity. Experimental results on high‑dimensional sensor data demonstrate that this approach converges to quality comparable to full‑memory processing while operating within strict memory constraints.
By Mamadou Sow, Christophe Cerin
CAST (Chunk-aware credit Assignment with Selective Token Anchoring) is a new framework for cross-tokenizer on-policy distillation that tackles two key problems: credit alignment and unaligned token positions. It redistributes chunk-level advantage signals across student tokens using predictive uncertainty and teacher-signal smoothness, while anchoring unaligned tokens to the initial policy via a KL penalty. Experiments on six mathematics, science, and code benchmarks show that CAST improves average performance over the strongest baseline by up to 2.45 points in cross-tokenizer transfer scenarios.
By Jingru Yang, Jinhang Wu, Minjing Shi, Shengfeng He, Xuyan Li, Man Tan, Biye Li
The paper introduces an entropy-based estimator for Kullback-Leibler divergence that uses normalized importance weights. It establishes a universal finite-sample bound, proving that the estimator’s expectation can exceed the true target‑proposal KL divergence by no more than λ2, regardless of sample size. Additionally, it shows that the KL divergence between the distribution produced by importance‑weighted resampling and the proposal is bounded by both the estimator’s expectation and the target‑proposal KL divergence, without requiring bounded weights or finite higher moments. The authors apply these theoretical results to test‑time alignment of generative models via Soft Best‑of‑N sampling and to relative entropy coding, yielding finite‑sample policy‑divergence guarantees and moment‑free bounds on coding cost.
By Buu Phan, Lester Mackey, David Alvarez-Melis
X-AVDD is a cross‑attention‑based audio‑visual dataset distillation method that explicitly models inter‑modal relationships during synthesis. By decoupling cross‑modal alignment from joint optimization, it reduces memory usage and end‑to‑end GPU‑hour cost by about eight times while producing highly aligned synthetic data. Empirical results show significant performance gains over prior methods, improving top‑1 test accuracy on VGGSound from 8.2 % to 24.0 % at 10 samples‑per‑class and enabling efficient scaling to complex architectures such as AudioCLIP.
By Saumyaranjan Mohanty, Aravind Reddy, Konda Reddy Mopuri
The paper proposes a fairness‑guided deep Q‑network (DQN) to control the transmission opportunity (TXOP) of NR‑U in unlicensed spectrum, aiming to balance coexistence with Wi‑Fi. Two reward designs, Q1 and Q2, target different fairness regimes: Q1 enforces strict per‑node throughput balance, while Q2 allows a broader fairness band and rewards throughput gains within it. The DQN policies are benchmarked against 3GPP LBT, a multi‑armed‑bandit approach, utility‑fairness tabular Q‑learning, and a fixed‑TXOP baseline, with experiments across four priority classes and asymmetric populations measuring aggregate throughput, Jain fairness, and concave throughput utility.
By Po-Heng Chou, Yi-Fang Yu, Shou-Yu Chen, Chiapin Wang
The paper introduces Kernelized Parameter Sensitivity Atlases (KPSA), a method that uses a block decomposition of the empirical neural tangent kernel to estimate input‑specific parameter sensitivities efficiently. KPSA embeds parameter groups in a query space and applies a lightweight kernel to predict sensitivities for unseen inputs, achieving a high fraction of the perturbation susceptibility of full back‑propagation while dramatically reducing computation and memory usage. Experiments across vision, forecasting, and protein‑modeling tasks show that KPSA identifies sparse, depth‑spanning sensitivity circuits and improves throughput and memory efficiency in activation‑attribution problems.
By Davide Wiest, Ahanaf Hasan Ariq, Abdul Monaf Chowdhury
arXiv:2610. 13789v1 Announce Type: new Abstract: Momentum methods such as STORM and second-order momentum track the gradient by adding a transport correction to momentum, at the cost of an extra oracle call per iteration.
By El Mahdi Chayti, Martin Jaggi
The paper introduces a framework for constructing verified quadratic characterizations of scalar relations in two-dimensional space, using convex quadratic programs and sum-of-squares certificates to generate locally valid inequalities that are globally verified. These constraints provide sound overapproximations compatible with existing QC and IQC analysis tools, and can be embedded in semidefinite programs for reachability and safety analysis of feedforward neural networks. The method improves reachability results for smooth activations like tanh, reduces conservatism for ReLU networks by exploiting neuron dependencies, and extends beyond neural networks to other nonlinear systems such as saturation.
By Elias Khalife, Mazen Farhood, Pierre-Loic Garoche
The paper examines whether deep ensembles effectively capture uncertainty in message‑passing graph neural networks. Across eight diverse datasets, ensembles only marginally outperform single models, mainly by reducing optimization noise rather than improving uncertainty estimates. The authors identify an epistemic collapse—independently trained networks converge to similar predictions—due to the smoothing bias of message passing, and provide a theoretical bound linking this collapse to graph topology and Dirichlet energy.
By Pedro C. Vieira, Viacheslav Borovitskiy, Pedro Ribeiro