Diffusion
-
Beyond Model Weights: Sampling-Dependent Fairness in Diffusion Language Models
2026-10-20 03:49:20Koshiro Saito, Diasuke Oba, Danushka Bollegala, Naoaki OkazakiAbstract
Diffusion language models (DLMs) are emerging as a practical alternative to autoregressive language models (ARLMs), yet their fairness remains largely uncharacterized. DLMs introduce sampling controls with no direct autoregressive counterpart, such as block length and remasking strategy, that directly set their speed--quality operating point. This raises a basic question: does a DLM's measured social bias remain stable across these settings? Across LLaDA, LLaDA-1.5, and Dream on Open-BBQ, we find that all three models exhibit social bias, but both the magnitude of measured bias and the relative ordering of models by bias are highly sensitive to the sampling configuration. This sampling dependence also changes what successful debiasing means. While mitigation strategies established for ARLMs transfer effectively to DLMs and reduce bias without degrading downstream performance, sampling dependence persists even after mitigation. Our results show that fairness in DLMs cannot be characterized independently of the sampling configuration, even after debiasing. Project page is available at this https URL.
Abstract (translated)
URL
https://arxiv.org/abs/2610.27584
PDF
-
TeCoP: Accelerating Diffusion Large Language Model Decoding via Tentative Context Pipelining
2026-10-20 03:48:01Jui-Chien Tsou, Hsuan-Chi Liu, Wei-Chin Wang, Yu-Syuan Xu, Chun-Yi LeeAbstract
Diffusion large language models (dLLMs) predict multiple token positions per forward pass, yet block-wise decoders serialize generation and repeatedly recompute stable predictions. We attribute this inefficiency to coupling a prediction's readiness for irreversible commitment with its utility as context. We introduce Tentative Context Pipelining (TeCoP), a training-free decoder that decouples these roles through staggered block-parallel pipelining and draft-and-verify. TeCoP reuses stable but uncommitted predictions as provisional cross-block context, with commitment governed by confidence and stability checks. Across four LLaDA and Dream dLLMs on GSM8K, Minerva MATH, and MBPP, TeCoP reduces number of forward evaluations by up to 87% and achieves up to 7.50-fold throughput without KV caching and 11.77-fold with caching over the uncached original decoder. It outperforms Fast-dLLM and DSB in throughput across all 12 model-benchmark settings, with and without caching, while keeping average cached accuracy within one percentage point of the original decoder.
Abstract (translated)
URL
https://arxiv.org/abs/2610.27583
PDF
-
Factual Access States in Diffusion Language Models: From Mechanism to Knowledge Editing and Retrieval
2026-10-20 03:45:45Yuchen Guo, Zhuoyuan Mao, Yiwen GuoAbstract
Factual recall is a lookup: some representation inside the network serves as the key, and what that key is made of decides which queries can retrieve the fact, what an edit must cover, and what a retriever can read. We call it the factual access state. In autoregressive (AR) models the subject position enriches attributes largely independently of the relation, and relation-specific selection happens later, at the readout position. In masked diffusion language models (DLMs) we find the state is assembled differently, and much earlier: in the first attention layer the wording of the relation is routed into the position of whichever entity is visible and the early feed-forward layers look the fact up on this fused representation. The state therefore varies with the relation's surface form and with the query direction, so the queries about one fact occupy a region of state space rather than a point, and an operation on the fact succeeds only for queries that enter it. We treat the state as an interface to a DLM's factual knowledge, and put it to two uses. Writing to the state gives a knowledge editor: from one declarative statement, Multi-View Edit fits one shared value per direction across that region and installs it in the early-MLP block, preserving source-prompt reliability while substantially improving paraphrase and reverse question answering, where AR editors gain near zero. Reading it suggests a complementary use: the state at each entity position can serve as a relation-specific retrieval vector with no training, returning both pieces of evidence for two-entity queries on which pooled, prompted and lexical retrievers almost never return both.
Abstract (translated)
URL
https://arxiv.org/abs/2610.27580
PDF
-
CityPaint: A Free Lunch from 2D Diffusion for City-Scale 3D Texturing
Abstract
Generating high-fidelity, city-scale 3D environments plays a crucial role in applications such as virtual reality, gaming, and urban digital twins. However, the scarcity of real-world city-scale 3D training data forces existing methods to either stitch independently generated assets or learn from synthetic urban corpora, often sacrificing realism and global consistency. In contrast, 2D image data and their corresponding generative models are abundant and encode rich visual priors. To bridge this gap, we propose CityPaint, a novel framework that leverages 2D diffusion models to generate globally consistent city-scale 3D textures without relying on city-scale 3D training data. Specifically, CityPaint integrates geometric priors with the generative power of 2D diffusion models, introducing a robust top-$K$ camera baking strategy for iterative texture projection. To address the artifacts of 2D-to-3D projection, we further propose a decoupled refinement mechanism that separately optimizes terrain and building textures. Extensive experiments show that CityPaint produces globally coherent textures with fine local details and achieves favorable results against strong baselines in both visual quality and cross-view consistency.
Abstract (translated)
URL
https://arxiv.org/abs/2610.27564
PDF
-
Diffusion Subgoal Planning for Long-Horizon Offline Goal-Conditioned Reinforcement Learning
2026-10-20 00:56:00Hengrui Zhang, Yuhu Cheng, C. L. Philip Chen, Xuesong WangAbstract
Offline goal-conditioned reinforcement learning (GCRL) learns goal-directed policies from reward-free data, but in long-horizon tasks, goal-conditioned value functions often provide unstable guidance due to sparse rewards and discounting. Hierarchical methods partially mitigate this issue via subgoal decomposition; however, high-level decision-making still relies on noise-sensitive value estimates, leading to unstable behavior in complex environments. We address this limitation by proposing \textbf{D}iffusion \textbf{S}ubgoal \textbf{P}lanning (\textbf{DSP}), a diffusion-based framework for high-level subgoal generation. DSP casts high-level planning as guided generative inference over goal-conditioned subgoals and learns both conditional and unconditional flows, enabling classifier-free guidance to introduce a goal-directed bias at inference time. By removing explicit value-based guidance from high-level planning, DSP generates reachable and goal-directed subgoals through a generative model while retaining hierarchical execution. Experiments on offline GCRL benchmarks demonstrate that DSP outperforms prior methods on a range of navigation and manipulation tasks, with particularly strong performance in maze environments that require multi-step subgoal planning.
Abstract (translated)
URL
https://arxiv.org/abs/2609.34575
PDF
-
LoRASpace: A Pool-Wide Shared Substrate for Static and Dynamic Multi-LoRA Composition
2026-10-20 00:46:19Thamirawaran Sathiyalogeswaran, Navindu De Silva, Chamikara Siriwardane, W. K. Alwis, Jason Mars, Udaya Sampath K. Perera Miriya Thanthrige, Ranga RodrigoAbstract
Low-rank adaptation (LoRA) attaches small low-rank matrices to a frozen backbone, and public repositories host many LoRAs per backbone, each for a different concept or task. Combining several LoRAs in one generation forces a trade-off. Merging their weights into one update needs a single forward pass, but LoRAs trained for different concepts interfere and quality drops. Running each LoRA in its own forward pass (composite) preserves quality, but inference time grows linearly with the number of active LoRAs. We study a property of LoRA pools overlooked by prior multi-LoRA work. In real LoRA pools, the stacked low-rank factors need far fewer directions than their summed ranks. Using singular value decomposition, we find the minimal subspace that keeps 99% of the factor energy. On three diffusion pools, only 46% to 57% of the summed LoRA ranks is needed, and the subspace grows sub-linearly with pool size. A 48-LoRA sub-pool of LoRA-Hub on FLAN-T5 tasks shows comparable overlap (62%), while the diversity-curated LoRA-Retriever benchmark sits near the no-overlap baseline (96%), so pool composition rather than modality predicts where compression is large. We build on this overlap with LoRASpace, a training-free method that computes the subspace once per pool and reuses it in two inference modes, a static mode at merge cost and a dynamic mode for input-dependent routing. On the diffusion pools, static LoRASpace matches composite quality at merge cost, and dynamic LoRASpace is more than 6x faster than composite with six active LoRAs on a 17-LoRA pool, mainly because it runs one backbone pass instead of six. In an ablation, static variants at the same rank match LoRASpace, so the quality gain does not depend on the pool-wide basis. On the near-orthogonal LoRA-Retriever benchmark, compression is small and gains come from coupling with external retrieval.
Abstract (translated)
URL
https://arxiv.org/abs/2610.27444
PDF
-
DStill: Trajectory Consistency Distillation for Accelerated One-Step Drafting in Block Diffusion Speculative Decoding
2026-10-19 23:05:16Shreyas Kulkarni, Uttam Gurram, Madhav Kanda, Krishna Teja Chitty Venkata, Shichu Chen, Kartik Balasubramaniam, Supriyo Chakraborty, Sambit SahuAbstract
Block-diffusion speculative decoding accelerates large language model inference by drafting multiple tokens in parallel. However, one-step drafting in discrete diffusion is limited by token factorization error: each token is predicted from the same masked block, without conditioning on the tokens selected at other positions in that step. This can produce inconsistent drafts and reduce acceptance length. Iterative refinement mitigates this limitation by conditioning later predictions on previously committed tokens, though additional passes increase drafting cost. To address this trade-off between draft fidelity and inference latency, we introduce DStill, a two-stage framework that distills a target-conditioned drafter's own multi-step refinement into one-step drafting. We first train the drafter to complete partially revealed blocks, then use its refinement trajectories to train a student to predict the final completion in one pass. Using a curriculum that gradually transitions from partial to fully masked inputs, we successfully transfer the multi-step refinement capabilities into one step. Across reasoning, math, code benchmarks on three model families DStill improves acceptance length by up to 28% and speedup by up to 25% over one-step DFlash, achieving up to 6.45x speedup over auto-regressive decoding.
Abstract (translated)
URL
https://arxiv.org/abs/2610.27379
PDF
-
Learning to Be Different: Component Specialization for Compositional Diffusion Policies
2026-10-19 19:29:58Alec Laprevotte, Haonan Chen, Xiaoshen Han, Yilun DuAbstract
Diffusion policies with linearly composed scores have emerged as a promising approach to multitask robotic learning by combining separate factors to enable modular, task-wise adaptation, compositional generalization, and interpretable functional decomposition. However, current training objectives do not encourage factor specialization, often resulting in uninterpretable denoisers whose composed prediction is an average over identical copies, undermining their architectural premise. To address this, we propose a drop-in score-decorrelation auxiliary loss, motivated by projective composition theory. Our approach encourages specialization by penalizing overlap directly between component noise predictions, with little architectural change or computational overhead. We demonstrate across multiple benchmarks that our decorrelation loss consistently reduces measurable double-counting between components, improves task decomposition, and produces behaviorally meaningful specialization without compromising multitask performance. Website: this https URL.
Abstract (translated)
URL
https://arxiv.org/abs/2610.27194
PDF
-
EDISCO: Equivariant DIScrete Diffusion for Euclidean Combinatorial Optimization
2026-10-19 18:44:19Ruogu Chen, Jie HanAbstract
Euclidean combinatorial optimization problems (ECOPs), such as the Traveling Salesman Problem (TSP) and Capacitated Vehicle Routing Problem (CVRP), possess inherent symmetries under the two-dimensional Euclidean group E(2), including rotations, reflections, and translations. Existing learning-based methods, including recent diffusion-based methods, rely on data augmentation or regularization to approximate E(2)-equivariance. This paper presents EDISCO, the first discrete diffusion model for ECOPs with exact E(2)-invariant generative distributions over node-index solutions. EDISCO introduces an E(2)-equivariant edge-score network coupled with a categorical continuous-time Markov chain over discrete edge variables, and exact posterior sampling provides efficient multi-step inference. This design gives EDISCO a local geometric inductive bias: edge neighborhoods with the same relative geometry and combinatorial context are represented consistently regardless of absolute position or orientation, making learning more efficient and inference more robust than non-equivariant methods. EDISCO outperforms previous learning-based state-of-the-art solvers on synthetic TSP from 100 to 10000 nodes and CVRP from 50 to 2000 customers, while using only 33-50% of the training instances. Trained only on uniform synthetic data, EDISCO also outperforms competing learning-based baselines under spatial distribution shift and CVRP constraint-tightness shift. Code is available at this https URL.
Abstract (translated)
URL
https://arxiv.org/abs/2610.04953
PDF
-
Representing short-term movement to reveal animal identity and learning progress
2026-10-19 18:35:06Ping Zhang, Quang-Huy Nguyen, Yingtie Lei, Sameera Shridhar, Zachary Leveroni, Ellie Quattrocchi, Jai Y Yu, Wei-Lun ChaoAbstract
How learning progresses over time and varies across individuals is a longstanding question in cognitive and behavioral science. However, a major challenge is identifying observable behavioral measures that faithfully capture learning progression. Existing approaches commonly rely on predefined behavioral features or task-level performance, while fine-scale posture and movement in freely behaving animals remain less explored. In this paper, we ask: can short-term movement trajectories reveal learning progress and individual identity? We introduce a self-supervised representation learning framework built on a simple hypothesis: behavioral characteristics associated with identity and learning remain relatively stable over nearby time intervals. Using temporal proximity as supervision, our framework learns compact representations directly from movement trajectories without manually defining behavioral states or kinematic features. Using behavior data from freely behaving rats, we show that the learned representations not only improve prediction of identity and learning-related readouts over existing methods, but also provide a tool for investigating broader scientific questions about when and how these signals appear in behavior. Specifically, trajectories as short as one second contain informative learning and identity signals. Egocentric pose dynamics are more predictive of identity, whereas allocentric motion is more predictive of learning-related variation, and both signals are strongest at decision junctions in the alternation cohort. We further use a representation-conditioned diffusion model to inspect movement patterns associated with identity and learning progress. Our findings establish short-term movement as an informative behavioral "fingerprint" for studying learning and individual differences. Our code is available at this https URL.
Abstract (translated)
URL
https://arxiv.org/abs/2610.27140
PDF
-
Beyond Two Tracks: Speaker-Anchored Full-Duplex Dialogue Recovery
2026-10-19 18:09:47Yuxuan Chen, Yang Xiao, Jack Sun, Chak Sun, Brick Li, SamAbstract
Spoken interaction is moving from alternating turns to full-duplex conversation, where an agent listens while speaking. Learning this coordination requires natural dialogue with persistent speaker-resolved tracks, a data requirement that remains difficult to satisfy at scale. Generative recovery offers a route from ordinary recordings, but recovering two voices is not enough: their track assignments can drift as the conversation unfolds. We address this missing link with speaker-anchored routing before diffusion: a 1.84M-parameter module matches the separator's coarse streams to fixed enrollment identities and orders them before a frozen generator reconstructs the speech, without retraining the generator or altering the reconstructed audio. On 25-minute remote conversations, routing cuts wrong-track time from 43.9% to 4.1%, below the 6.2% of matched-enrollment waveform re-identification, with its largest advantage when one participant holds the floor (17.8% versus 28.6%), and raises fixed-track activity F1 from 0.46 to 0.84 while per-chunk reconstruction quality is unchanged. The same router transfers to unseen face-to-face and telephone recordings, and a margin-gated verification step brings their wrong-track time to 0.9-1.7% while embedding 20-36% of the output, where full re-identification embeds all of it. Recovered voices thus become the participant-consistent tracks that full-duplex dialogue modeling needs. Audio examples: this https URL
Abstract (translated)
URL
https://arxiv.org/abs/2610.27109
PDF
-
Breaking Diffusion: Parallel Reasoning with Draft Refinement
2026-10-19 15:02:45Jaesang Yu, Seoyeon Byun, Yun Wang, Yan Lu, Juho KimAbstract
Drafting an answer allows a reasoning model to reconsider each decision in the context of the others. Yet repeated diffusion denoising does not ensure that individually plausible token updates form a better answer together. We introduce Draft Refinement Language Models (DRLMs), which use a single denoiser to propose parallel revisions and evaluate a candidate draft. We train the denoiser to reconstruct every token from uniformly corrupted drafts, teaching both preservation and repair. Its predictions specify revisions directly, without a diffusion reverse transition. Evaluation requires a different use of these predictions: confidence in an existing token reflects both contextual support and evidence from observing the token itself. Using the known corruption process, we correct for this self-token evidence to estimate how well each token is supported by the prompt and the remaining draft. Aggregating these estimates yields a cavity energy computed in one forward pass, without an additional scoring model. Each round proposes a revision, reads the completed candidate in its revised context, and retains it only if its energy is lower. Across Countdown, Sudoku-Extreme, and Maze-Hard, our method outperforms autoregressive and discrete diffusion baselines. With 6M parameters, DRLM reaches 44.4% on Countdown-5 and 93.9% on Sudoku-Extreme. The results establish complete-draft parallel refinement as an effective alternative to reverse diffusion for complex reasoning.
Abstract (translated)
URL
https://arxiv.org/abs/2610.26542
PDF
-
CaMeWM: A Causal Medical World Model for Generative Tumor Evolution and Treatment Planning
2026-10-17 11:53:00Huanlong Gao, Xuelei He, Fengjun Zhao, Xiaowei HeAbstract
Treatment planning is intrinsically counterfactual: clinicians must compare how a patient's tumor and survival would evolve under competing regimens, whereas a retrospective cohort records only the regimen received and one follow-up scan. A treatment-conditioned generator trained on such data can therefore reproduce historical treatment allocation rather than simulate the interventional disease dynamics needed to rank regimens. We introduce \textbf{CaMeWM}, a causal medical world model for supported post-TACE tumor evolution and treatment planning. CaMeWM implements one intervention-to-decision loop rather than three independent prediction tasks: it first admits a queried regimen only for clinically comparable histories with empirical treatment support; it then generates a distribution of lesion-local follow-up worlds through an action-conditioned response descriptor, deterministic anchor, and residual 3D diffusion transition; and it finally maps sampled worlds to restricted mean survival time (RMST) with a response-Cox utility model. Under pre-treatment adjustment and positivity, this loop estimates a prespecified, support-restricted g-formula target; all selection and nuisance estimation are fold-local, and the induced policy is evaluated by cross-fitted, censoring-adjusted doubly robust estimation. Across 79 paired out-of-fold episodes, CaMeWM achieves 0.618 SSIM and 0.528 change-map SSIM for follow-up CT transition, while its generated-world survival pathway attains a 0.615 C-index; factual-only AIPW estimates a 9.60-week RMST gain (95\% CI 0.61--18.13) for the frozen policy. CaMeWM provides an inspectable basis for retrospective, support-restricted treatment planning.
Abstract (translated)
URL
https://arxiv.org/abs/2610.26923
PDF
-
Enhancing analytical photon transport in turbid media through physics-guided neural network corrections
2026-10-17 07:28:34Guojun Tang, Paul D. Teal, Alexander DoroninAbstract
Analytical methods provide rapid and computationally efficient predictions of light transport in turbid media, making them attractive for real-time and inverse applications. However, their accuracy is limited by the approximations and assumptions used to obtain tractable solutions. On the other hand, Monte Carlo (MC) simulation is widely regarded as the "gold standard" for modeling light transport in turbid media, but its high computational cost limits its use in applications requiring repeated forward evaluations. Here, we introduce a physics-guided residual learning framework that combines the computational efficiency of analytical modeling with the accuracy of MC simulation. As a representative case study, we consider the diffusion approximation (DA), which can exhibit systematic errors under relatively strong absorption, weak scattering, and imperfect boundary treatment. Given the absorption coefficient, scattering coefficient, and anisotropy factor, the corresponding DA prediction is incorporated as a physical prior, while the network learns only the residual between the analytical prediction and the MC reference. Across the evaluated optical-property range, the proposed framework reduces scalar-reflectance RMSE and spatially resolved reflectance RMSLE by factors of approximately 62 and 14, respectively, relative to the diffusion approximation. With only 1\% of the available training data, it achieves approximately 87\% and 93\% lower corresponding errors than a direct neural-network surrogate. These results demonstrate that learning corrections to an analytical physical baseline improves both prediction accuracy and data efficiency.
Abstract (translated)
URL
https://arxiv.org/abs/2610.26917
PDF
-
Rethinking the design space of Vanilla Pixel Diffusion Transformers
2026-10-16 17:59:30Soumik Mukhopadhyay, Long Mai, Abhinav Shrivastava, Aniruddha MahapatraAbstract
Pixel-space diffusion models have recently re-emerged as a compelling alternative to latent diffusion, avoiding the lossy latent bottleneck and separately trained autoencoder. JiT shows that a vanilla diffusion transformer can generate directly in pixel space by replacing the commonly used velocity prediction ($\boldsymbol{u}$-prediction) with clean image prediction ($\boldsymbol{x_0}$-prediction). In this work, we revisit this choice and introduce MFlow, an adaptive parameterization that learns to interpolate between $\boldsymbol{x_0}$- and $\boldsymbol{u}$-prediction across principal directions of the data, learned end-to-end with diffusion training. We find that the preferred parameterization varies systematically with the data representation and model capacity -- notably, increasing model capacity shifts preference toward velocity prediction. We evaluate MFlow across ImageNet, UCF-101 video, and large-scale text-to-image generation, demonstrating competitive or improved generative performance across diverse settings. In addition, we find that vanilla pixel diffusion transformers, irrespective of parameterization, retain a persistent high-frequency deficit relative to real images. To address this, we introduce ConvRefiner, a lightweight convolutional module trained post-hoc with an adversarial loss while keeping the diffusion backbone frozen. With only a small amount of additional training, ConvRefiner restores the high-frequency statistics of generated images while preserving FID. Project webpage: this https URL.
Abstract (translated)
URL
https://arxiv.org/abs/2610.23854
PDF
-
Persistent Planar Memory for Video World Models
2026-10-16 17:21:54Yuze He, Bin Tan, Zelin Gao, Yujun Shen, Yong-Jin Liu, Nan Xue, Yinghao XuAbstract
Video world models that generate and explore environments at scale need a persistent 3D memory of what they have produced. Plane primitives offer a natural candidate: a compact set of oriented planes captures the dominant structural elements (e.g., walls, floors, facades) of indoor and outdoor scenes, while requiring only 7 parameters per primitive. We present PPMem, a video world model that adopts plane primitives as its persistent 3D memory. PPMem augments a video diffusion transformer with a planar prediction head and a Depth Token Fusion (DTF) module: each autoregressive chunk jointly produces RGB video and planar geometry, which merges into a persistent plane memory that conditions subsequent chunks via rendered depth. The resulting memory is three orders of magnitude more compact than per-pixel point clouds (~100K primitives, 5MB for a full scene), geometrically explicit (metric depth, normals, exportable surfaces), cross-view consistent (jointly predicted by the shared backbone), and adaptive (refined by the merging procedure as generation proceeds). On DL3DV and RealEstate10K, PPMem outperforms prior 3D-grounded world models in video quality while jointly producing metric-scale depth, surface normals, and exportable planar geometry, sustaining coherence across hundreds of frames under large viewpoint changes. Project page: this https URL.
Abstract (translated)
URL
https://arxiv.org/abs/2610.23795
PDF
-
Video Diffusion Avatars: Taming Video Diffusion Models into Controllable 3D Avatars
2026-10-16 17:08:33Simon Giebenhain, Tobias Kirschstein, Liam Schoneveld, Davide Davoli, Zhe Chen, Matthias Nie{\ss}nerAbstract
Video diffusion models (VDMs) trained on internet-scale data encode a rich prior on appearance, motion, and physics that is difficult to replicate with explicit 3D rendering and costly 3D data capture. We introduce Video Diffusion Avatars (VDAs), a method that creates photo-realistic, multi-view consistent avatars from a single image by fine-tuning a VDM for precise control over camera viewpoint, facial expression, and upper-body pose. Departing from 3D morphable models (3DMMs), VDAs use a learned expression space that lifts the expressiveness ceiling of 3DMMs, e.g., for the tongue and mouth interior. However, learned expression codes exhibit heavy viewpoint bias, as our DISTRACT benchmark reveals. We show that this bias can be removed through disentanglement, enabling their use as driving signals for multi-view consistent avatars. Camera and upper-body pose are controlled jointly via proxy-mesh renderings, giving independent control over expression, pose, and viewpoint. A camera-triplet training strategy makes the approach efficient: our models train for 34 hours on a single A100 using only public multi-view data. VDAs outperform state-of-the-art single-image 3D avatars on novel-view reenactment, reducing face LPIPS by 29% and improving perceptual quality (JOD) by 66% on BecomingLit. A static 3DGS reconstruction probe further confirms that our generations are the most multi-view consistent. Built on a video prior, VDAs produce effects beyond the reach of explicit renderers, including physically plausible hair dynamics, shadows, reflections, and refractions.
Abstract (translated)
URL
https://arxiv.org/abs/2610.23772
PDF
-
SemPurify: Semantics-Aware Data Purification for Black-box Backdoor Defense
2026-10-16 15:09:14Jiawei Yu, Syed Irfan Ali Meerza, Yi Wu, Amir Sadovnik, Jian LiuAbstract
Backdoor attacks have emerged as a critical security threat in machine learning, where adversaries embed triggers within poisoned training samples to manipulate a model's predictions on targeted inputs while leaving its general behavior unaffected. This stealthy nature makes backdoor attacks difficult to detect, as the model functions normally until activated by a specific malicious trigger. Current purification-based defenses use image transformations (e.g., blurring) to disrupt trigger patterns but often degrade image quality. Even with advanced restoration techniques like diffusion models, these approaches can still alter visual or semantic information, reducing utility for downstream tasks. To address this, we propose the first semantics-aware data purification framework for black-box backdoor defense, SemPurify, which explicitly incorporates multi-level semantics (i.e., low-level semantic embeddings and high-level image tags) within a text-to-image (T2I) framework to guide image restoration during purification. Using this semantic information, the T2I model subsequently refines the transformed samples, helping preserve visual details and semantic information during restoration. Extensive experiments spanning multiple backdoor scenarios, i.e., classification models, CLIP-style encoders, and generative models, demonstrate that SemPurify reliably removes backdoors while effectively preserving semantic information for downstream tasks.
Abstract (translated)
URL
https://arxiv.org/abs/2610.23589
PDF
-
LiveSync: Iterative Self-Training for Lip-Sync Generation
2026-10-16 14:46:38Huaishao Luo, Huan Zhang, Guanbo Bao, Yinglu Liu, Youzheng Wu, Xiaodong HeAbstract
Audio-driven lip synchronization aims to match target speech through localized mouth and jaw articulation while preserving subject identity, head pose, and surrounding facial context. Naturally recorded videos provide synchronized reconstruction targets but rarely offer pairs that are matched in identity and pose yet differ primarily in speech-related articulation. Mask-based training enables reconstruction from these videos, but random-reference inputs differ from the aligned self-reference used at inference. We present LiveSync, an iterative self-training framework in which a single mask-based model generates pseudo pairs and is refined for deployment. The model is first trained on a curated real-video corpus augmented with externally generated clips covering under-represented visual conditions. A frozen checkpoint then uses independently sampled audio to generate mouth-edited references from real-video frames, copying pixels outside the edit mask from each source frame. After confidence filtering, these references are paired with aligned masked inputs and original real-frame targets. The retained pseudo pairs are mixed with real training data to refine the same model from its current weights, updating training supervision without adding inference stages. Although we instantiate the framework with a single-pass GAN-based model, the self-training strategy could in principle also be adapted to diffusion-based models. Experimental results show that LiveSync achieves leading synchronization performance on HDTF and competitive performance compared with the baselines on the AIGC-LipSync subset, while maintaining low inference latency.
Abstract (translated)
URL
https://arxiv.org/abs/2610.23558
PDF
-
Edit-Your-Interest: Efficient Video Editing via Feature Most-Similar Propagation
2026-10-16 13:59:01Yi Zuo, Zitao Wang, Lingling Li, Xu Liu, Fang Liu, Licheng JiaoAbstract
Text-to-image diffusion models have advanced video editing. However, existing video editing methods are limited by their high computational overhead and memory consumption. Furthermore, these approaches often sacrifice visual fidelity, leading to temporal inconsistencies and artifacts such as blurring and mosaic-like patterns. To address these dual challenges and strike a balance between computational efficiency and visual fidelity, we propose Edit-Your-Interest, a lightweight, text-driven, zero-shot video editing method. Edit-Your-Interest introduces a Spatio-Temporal Feature Memory bank (SFM) to cache spatial-attention features from previous frames under a fixed memory budget. The SFM maintains a representative candidate pool, reducing the computational overhead of full-sequence spatio-temporal modeling. To query this memory, we propose Feature Most-Similar Propagation (FMP), a similarity-based retrieval rule that propagates relevant cached tokens to the current frame. Together, SFM and FMP form a bounded-memory retrieval-based propagation design that improves temporal consistency while reducing blurring and mosaic-like artifacts. Finally, we introduce a feature-diversity-based SFM update algorithm to keep the memory representative and non-redundant throughout the video sequence. Furthermore, to enable precise object editing, we leverage cross-attention maps to automatically extract masks for the instances of interest. These masks are seamlessly integrated into the diffusion denoising process, enabling fine-grained control over target objects and allowing Edit-Your-Interest to perform highly accurate edits while robustly preserving the background integrity. Extensive experiments demonstrate that Edit-Your-Interest outperforms state-of-the-art methods in both efficiency and visual fidelity. Project page: this https URL.
Abstract (translated)
URL
https://arxiv.org/abs/2510.13084
PDF
