# Research Papers — 2025-12 · bundle page 63 of 175 > Full-text markdown of papers #3101–3150 of 8703 published in 2025-12, newest first. AI-readable mirrors of the full 8703-paper 2025-12 archive (50 papers per file). Canonical page: https://aipod.dev/papers/archive/2025-12/page/63 [← Previous bundle](https://aipod.dev/papers/archive/2025-12/page/62.md) [Next bundle →](https://aipod.dev/papers/archive/2025-12/page/64.md) ### REGLUE Your Latents with Global and Local Semantics for Entangled Diffusion Authors: Giorgos Petsangourakis, Christos Sgouropoulos, Bill Psomas, Theodoros Giannakopoulos, Giorgos Sfikas, Ioannis Kakogeorgiou Published: 2025-12-18 Categories: cs.CV arXiv: https://arxiv.org/abs/2512.16636v2 Latent diffusion models (LDMs) achieve state-of-the-art image synthesis, yet their reconstruction-style denoising objective provides only indirect semantic supervision: high-level semantics emerge slowly, requiring longer training and limiting sample quality. Recent works inject semantics from Vision Foundation Models (VFMs) either externally via representation alignment or internally by jointly modeling only a narrow slice of VFM features inside the diffusion process, under-utilizing the rich, nonlinear, multi-layer spatial semantics available. We introduce REGLUE (Representation Entanglement with Global-Local Unified Encoding), a unified latent diffusion framework that jointly models (i) VAE image latents, (ii) compact local (patch-level) VFM semantics, and (iii) a global (image-level) (CLS) token within a single SiT backbone. A lightweight convolutional semantic compressor nonlinearly aggregates multi-layer VFM features into a low-dimensional, spatially structured representation, which is entangled with the VAE latents in the diffusion process. An external alignment loss further regularizes internal representations toward frozen VFM targets. On ImageNet 256x256, REGLUE consistently improves FID and accelerates convergence over SiT-B/2 and SiT-XL/2 baselines, as well as over REPA, ReDi, and REG. Extensive experiments show that (a) spatial VFM semantics are crucial, (b) non-linear compression is key to unlocking their full benefit, and (c) global tokens and external alignment act as complementary, lightweight enhancements within our global-local-latent joint modeling framework. The code is available at https://github.com/giorgospets/reglue . --- ### REGLUE Your Latents with Global and Local Semantics for Entangled Diffusion Authors: Giorgos Petsangourakis, Christos Sgouropoulos, Bill Psomas, Theodoros Giannakopoulos, Giorgos Sfikas, Ioannis Kakogeorgiou Published: 2025-12-18 Categories: cs.CV arXiv: https://arxiv.org/abs/2512.16636v1 Latent diffusion models (LDMs) achieve state-of-the-art image synthesis, yet their reconstruction-style denoising objective provides only indirect semantic supervision: high-level semantics emerge slowly, requiring longer training and limiting sample quality. Recent works inject semantics from Vision Foundation Models (VFMs) either externally via representation alignment or internally by jointly modeling only a narrow slice of VFM features inside the diffusion process, under-utilizing the rich, nonlinear, multi-layer spatial semantics available. We introduce REGLUE (Representation Entanglement with Global-Local Unified Encoding), a unified latent diffusion framework that jointly models (i) VAE image latents, (ii) compact local (patch-level) VFM semantics, and (iii) a global (image-level) (CLS) token within a single SiT backbone. A lightweight convolutional semantic compressor nonlinearly aggregates multi-layer VFM features into a low-dimensional, spatially structured representation, which is entangled with the VAE latents in the diffusion process. An external alignment loss further regularizes internal representations toward frozen VFM targets. On ImageNet 256x256, REGLUE consistently improves FID and accelerates convergence over SiT-B/2 and SiT-XL/2 baselines, as well as over REPA, ReDi, and REG. Extensive experiments show that (a) spatial VFM semantics are crucial, (b) non-linear compression is key to unlocking their full benefit, and (c) global tokens and external alignment act as complementary, lightweight enhancements within our global-local-latent joint modeling framework. The code is available at https://github.com/giorgospets/reglue . --- ### SARMAE: Masked Autoencoder for SAR Representation Learning Authors: Danxu Liu, Di Wang, Hebaixu Wang, Haoyang Chen, Wentao Jiang, Yilin Cheng, Haonan Guo, Wei Cui, Jing Zhang Published: 2025-12-18 Categories: cs.CV, cs.LG arXiv: https://arxiv.org/abs/2512.16635v2 Synthetic Aperture Radar (SAR) imagery plays a critical role in all-weather, day-and-night remote sensing applications. However, existing SAR-oriented deep learning is constrained by data scarcity, while the physically grounded speckle noise in SAR imagery further hampers fine-grained semantic representation learning. To address these challenges, we propose SARMAE, a Noise-Aware Masked Autoencoder for self-supervised SAR representation learning. Specifically, we construct SAR-1M, the first million-scale SAR dataset, with additional paired optical images, to enable large-scale pre-training. Building upon this, we design Speckle-Aware Representation Enhancement (SARE), which injects SAR-specific speckle noise into masked autoencoders to facilitate noise-aware and robust representation learning. Furthermore, we introduce Semantic Anchor Representation Constraint (SARC), which leverages paired optical priors to align SAR features and ensure semantic consistency. Extensive experiments across multiple SAR datasets demonstrate that SARMAE achieves state-of-the-art performance on classification, detection, and segmentation tasks. Code and models will be available at https://github.com/MiliLab/SARMAE. --- ### SARMAE: Masked Autoencoder for SAR Representation Learning Authors: Danxu Liu, Di Wang, Hebaixu Wang, Haoyang Chen, Wentao Jiang, Yilin Cheng, Haonan Guo, Wei Cui, Jing Zhang Published: 2025-12-18 Categories: cs.CV, cs.LG arXiv: https://arxiv.org/abs/2512.16635v1 Synthetic Aperture Radar (SAR) imagery plays a critical role in all-weather, day-and-night remote sensing applications. However, existing SAR-oriented deep learning is constrained by data scarcity, while the physically grounded speckle noise in SAR imagery further hampers fine-grained semantic representation learning. To address these challenges, we propose SARMAE, a Noise-Aware Masked Autoencoder for self-supervised SAR representation learning. Specifically, we construct SAR-1M, the first million-scale SAR dataset, with additional paired optical images, to enable large-scale pre-training. Building upon this, we design Speckle-Aware Representation Enhancement (SARE), which injects SAR-specific speckle noise into masked autoencoders to facilitate noise-aware and robust representation learning. Furthermore, we introduce Semantic Anchor Representation Constraint (SARC), which leverages paired optical priors to align SAR features and ensure semantic consistency. Extensive experiments across multiple SAR datasets demonstrate that SARMAE achieves state-of-the-art performance on classification, detection, and segmentation tasks. Code and models will be available at https://github.com/MiliLab/SARMAE. --- ### Stackelberg Learning from Human Feedback: Preference Optimization as a Sequential Game Authors: Barna Pásztor, Thomas Kleine Buening, Andreas Krause Published: 2025-12-18 Categories: cs.LG, cs.AI, cs.GT, cs.MA, stat.ML arXiv: https://arxiv.org/abs/2512.16626v1 We introduce Stackelberg Learning from Human Feedback (SLHF), a new framework for preference optimization. SLHF frames the alignment problem as a sequential-move game between two policies: a Leader, which commits to an action, and a Follower, which responds conditionally on the Leader's action. This approach decomposes preference optimization into a refinement problem for the Follower and an optimization problem against an adversary for the Leader. Unlike Reinforcement Learning from Human Feedback (RLHF), which assigns scalar rewards to actions, or Nash Learning from Human Feedback (NLHF), which seeks a simultaneous-move equilibrium, SLHF leverages the asymmetry of sequential play to capture richer preference structures. The sequential design of SLHF naturally enables inference-time refinement, as the Follower learns to improve the Leader's actions, and these refinements can be leveraged through iterative sampling. We compare the solution concepts of SLHF, RLHF, and NLHF, and lay out key advantages in consistency, data sensitivity, and robustness to intransitive preferences. Experiments on large language models demonstrate that SLHF achieves strong alignment across diverse preference datasets, scales from 0.5B to 8B parameters, and yields inference-time refinements that transfer across model families without further fine-tuning. --- ### DeContext as Defense: Safe Image Editing in Diffusion Transformers Authors: Linghui Shen, Mingyue Cui, Xingyi Yang Published: 2025-12-18 Categories: cs.CV arXiv: https://arxiv.org/abs/2512.16625v2 In-context diffusion models allow users to modify images with remarkable ease and realism. However, the same power raises serious privacy concerns: personal images can be easily manipulated for identity impersonation, misinformation, or other malicious uses, all without the owner's consent. While prior work has explored input perturbations to protect against misuse in personalized text-to-image generation, the robustness of modern, large-scale in-context DiT-based models remains largely unexamined. In this paper, we propose DeContext, a new method to safeguard input images from unauthorized in-context editing. Our key insight is that contextual information from the source image propagates to the output primarily through multimodal attention layers. By injecting small, targeted perturbations that weaken these cross-attention pathways, DeContext breaks this flow, effectively decouples the link between input and output. This simple defense is both efficient and robust. We further show that early denoising steps and specific transformer blocks dominate context propagation, which allows us to concentrate perturbations where they matter most. Experiments on Flux Kontext and Step1X-Edit show that DeContext consistently blocks unwanted image edits while preserving visual quality. These results highlight the effectiveness of attention-based perturbations as a powerful defense against image manipulation. Code is available at https://github.com/LinghuiiShen/DeContext. --- ### DeContext as Defense: Safe Image Editing in Diffusion Transformers Authors: Linghui Shen, Mingyue Cui, Xingyi Yang Published: 2025-12-18 Categories: cs.CV arXiv: https://arxiv.org/abs/2512.16625v1 In-context diffusion models allow users to modify images with remarkable ease and realism. However, the same power raises serious privacy concerns: personal images can be easily manipulated for identity impersonation, misinformation, or other malicious uses, all without the owner's consent. While prior work has explored input perturbations to protect against misuse in personalized text-to-image generation, the robustness of modern, large-scale in-context DiT-based models remains largely unexamined. In this paper, we propose DeContext, a new method to safeguard input images from unauthorized in-context editing. Our key insight is that contextual information from the source image propagates to the output primarily through multimodal attention layers. By injecting small, targeted perturbations that weaken these cross-attention pathways, DeContext breaks this flow, effectively decouples the link between input and output. This simple defense is both efficient and robust. We further show that early denoising steps and specific transformer blocks dominate context propagation, which allows us to concentrate perturbations where they matter most. Experiments on Flux Kontext and Step1X-Edit show that DeContext consistently blocks unwanted image edits while preserving visual quality. These results highlight the effectiveness of attention-based perturbations as a powerful defense against image manipulation. --- ### Plug to Place: Indoor Multimedia Geolocation from Electrical Sockets for Digital Investigation Authors: Kanwal Aftab, Graham Adams, Mark Scanlon Published: 2025-12-18 Categories: cs.CV arXiv: https://arxiv.org/abs/2512.16620v1 Computer vision is a rapidly evolving field, giving rise to powerful new tools and techniques in digital forensic investigation, and shows great promise for novel digital forensic applications. One such application, indoor multimedia geolocation, has the potential to become a crucial aid for law enforcement in the fight against human trafficking, child exploitation, and other serious crimes. While outdoor multimedia geolocation has been widely explored, its indoor counterpart remains underdeveloped due to challenges such as similar room layouts, frequent renovations, visual ambiguity, indoor lighting variability, unreliable GPS signals, and limited datasets in sensitive domains. This paper introduces a pipeline that uses electric sockets as consistent indoor markers for geolocation, since plug socket types are standardised by country or region. The three-stage deep learning pipeline detects plug sockets (YOLOv11, mAP@0.5 = 0.843), classifies them into one of 12 plug socket types (Xception, accuracy = 0.912), and maps the detected socket types to countries (accuracy = 0.96 at >90% threshold confidence). To address data scarcity, two dedicated datasets were created: socket detection dataset of 2,328 annotated images expanded to 4,072 through augmentation, and a classification dataset of 3,187 images across 12 plug socket classes. The pipeline was evaluated on the Hotels-50K dataset, focusing on the TraffickCam subset of crowd-sourced hotel images, which capture real-world conditions such as poor lighting and amateur angles. This dataset provides a more realistic evaluation than using professional, well-lit, often wide-angle images from travel websites. This framework demonstrates a practical step toward real-world digital forensic applications. The code, trained models, and the data for this paper are available open source. --- ### Trainable Log-linear Sparse Attention for Efficient Diffusion Transformers Authors: Yifan Zhou, Zeqi Xiao, Tianyi Wei, Shuai Yang, Xingang Pan Published: 2025-12-18 Categories: cs.CV arXiv: https://arxiv.org/abs/2512.16615v2 Diffusion Transformers (DiTs) set the state of the art in visual generation, yet their quadratic self-attention cost fundamentally limits scaling to long token sequences. Recent Top-K sparse attention approaches reduce the computation of DiTs by compressing tokens into block-wise representation and selecting a small set of relevant key blocks, but still suffer from (i) quadratic selection cost on compressed tokens and (ii) increasing K required to maintain model quality as sequences grow. We identify that their inefficiency is due to the single-level design, as a single coarse level is insufficient to represent the global structure. In this paper, we introduce Log-linear Sparse Attention (LLSA), a trainable sparse attention mechanism for extremely long token sequences that reduces both selection and attention costs from quadratic to log-linear complexity by utilizing a hierarchical structure. LLSA performs hierarchical Top-K selection, progressively adopting sparse Top-K selection with the indices found at the previous level, and introduces a Hierarchical KV Enrichment mechanism that preserves global context while using fewer tokens of different granularity during attention computation. To support efficient training, we develop a high-performance GPU implementation that uses only sparse indices for both the forward and backward passes, eliminating the need for dense attention masks. We evaluate LLSA on high-resolution pixel-space image generation without using patchification and VAE encoding. LLSA accelerates attention inference by 28.27x and DiT training by 6.09x on 256x256 pixel token sequences, while maintaining generation quality. The results demonstrate that LLSA offers a promising direction for training long-sequence DiTs efficiently. Code is available at: https://github.com/SingleZombie/LLSA --- ### Trainable Log-linear Sparse Attention for Efficient Diffusion Transformers Authors: Yifan Zhou, Zeqi Xiao, Tianyi Wei, Shuai Yang, Xingang Pan Published: 2025-12-18 Categories: cs.CV arXiv: https://arxiv.org/abs/2512.16615v1 Diffusion Transformers (DiTs) set the state of the art in visual generation, yet their quadratic self-attention cost fundamentally limits scaling to long token sequences. Recent Top-K sparse attention approaches reduce the computation of DiTs by compressing tokens into block-wise representation and selecting a small set of relevant key blocks, but still suffer from (i) quadratic selection cost on compressed tokens and (ii) increasing K required to maintain model quality as sequences grow. We identify that their inefficiency is due to the single-level design, as a single coarse level is insufficient to represent the global structure. In this paper, we introduce Log-linear Sparse Attention (LLSA), a trainable sparse attention mechanism for extremely long token sequences that reduces both selection and attention costs from quadratic to log-linear complexity by utilizing a hierarchical structure. LLSA performs hierarchical Top-K selection, progressively adopting sparse Top-K selection with the indices found at the previous level, and introduces a Hierarchical KV Enrichment mechanism that preserves global context while using fewer tokens of different granularity during attention computation. To support efficient training, we develop a high-performance GPU implementation that uses only sparse indices for both the forward and backward passes, eliminating the need for dense attention masks. We evaluate LLSA on high-resolution pixel-space image generation without using patchification and VAE encoding. LLSA accelerates attention inference by 28.27x and DiT training by 6.09x on 256x256 pixel token sequences, while maintaining generation quality. The results demonstrate that LLSA offers a promising direction for training long-sequence DiTs efficiently. Code is available at: https://github.com/SingleZombie/LLSA --- ### Don't Guess, Escalate: Towards Explainable Uncertainty-Calibrated AI Forensic Agents Authors: Giulia Boato, Andrea Montibeller, Edward Delp, Luisa Verdoliva, Daniele Miorandi Published: 2025-12-18 Categories: cs.MA, cs.AI, cs.CV, cs.MM arXiv: https://arxiv.org/abs/2512.16614v1 AI is reshaping the landscape of multimedia forensics. We propose AI forensic agents: reliable orchestrators that select and combine forensic detectors, identify provenance and context, and provide uncertainty-aware assessments. We highlight pitfalls in current solutions and introduce a unified framework to improve the authenticity verification process. --- ### Hazedefy: A Lightweight Real-Time Image and Video Dehazing Pipeline for Practical Deployment Authors: Ayush Bhavsar Published: 2025-12-18 Categories: cs.CV arXiv: https://arxiv.org/abs/2512.16609v1 This paper introduces Hazedefy, a lightweight and application-focused dehazing pipeline intended for real-time video and live camera feed enhancement. Hazedefy prioritizes computational simplicity and practical deployability on consumer-grade hardware, building upon the Dark Channel Prior (DCP) concept and the atmospheric scattering model. Key elements include gamma-adaptive reconstruction, a fast transmission approximation with lower bounds for numerical stability, a stabilized atmospheric light estimator based on fractional top-pixel averaging, and an optional color balance stage. The pipeline is suitable for mobile and embedded applications, as experimental demonstrations on real-world images and videos show improved visibility and contrast without requiring GPU acceleration. --- ### Boltzmann generators for amorphous particle systems Authors: Louis Grenioux, Leonardo Galliano, Ludovic Berthier, Giulio Biroli, Marylou Gabrié Published: 2025-12-18 Categories: stat.ML, cond-mat.stat-mech, cs.LG, physics.comp-ph arXiv: https://arxiv.org/abs/2512.16607v4 Sampling configurations in thermodynamic equilibrium is a long-standing challenge in statistical physics. Boltzmann generators address this problem by employing generative models to propose independent configurations, which are then reweighted via importance sampling using exact likelihood evaluations. Recent Boltzmann Generators based on continuous normalizing flows and flow matching have achieved significant success for particle systems and biomolecules. However, these approaches have not been extended to amorphous materials (glasses), for which equilibrium sampling is notoriously slow. Because of their disordered structure, the invariances and geometrical constraints of amorphous materials differ from those of crystals and biomolecules, preventing the direct use of existing generative models. Here, we develop Boltzmann Generators tailored to amorphous materials by building the required equivariances directly into Riemannian stochastic interpolants. Our framework incorporates periodic boundary conditions and particle symmetries using equivariant graph neural networks. Numerical experiments demonstrate that enforcing physical symmetries significantly improves the accuracy of Boltzmann Generators, but also reveal an intrinsic limitation of the continuous-flow formulation: accumulated numerical errors during likelihood integration break time-reversibility, compromising exact thermodynamic reweighting. These results reveal a fundamental challenge for continuous-flow generative models in statistical mechanics and call for alternative approaches that preserve exact thermodynamic consistency. --- ### Riemannian Stochastic Interpolants for Amorphous Particle Systems Authors: Louis Grenioux, Leonardo Galliano, Ludovic Berthier, Giulio Biroli, Marylou Gabrié Published: 2025-12-18 Categories: stat.ML, cond-mat.stat-mech, cs.LG, physics.comp-ph arXiv: https://arxiv.org/abs/2512.16607v1 Modern generative models hold great promise for accelerating diverse tasks involving the simulation of physical systems, but they must be adapted to the specific constraints of each domain. Significant progress has been made for biomolecules and crystalline materials. Here, we address amorphous materials (glasses), which are disordered particle systems lacking atomic periodicity. Sampling equilibrium configurations of glass-forming materials is a notoriously slow and difficult task. This obstacle could be overcome by developing a generative framework capable of producing equilibrium configurations with well-defined likelihoods. In this work, we address this challenge by leveraging an equivariant Riemannian stochastic interpolation framework which combines Riemannian stochastic interpolant and equivariant flow matching. Our method rigorously incorporates periodic boundary conditions and the symmetries of multi-component particle systems, adapting an equivariant graph neural network to operate directly on the torus. Our numerical experiments on model amorphous systems demonstrate that enforcing geometric and symmetry constraints significantly improves generative performance. --- ### Refusal Steering: Fine-grained Control over LLM Refusal Behaviour for Sensitive Topics Authors: Iker García-Ferrero, David Montero, Roman Orus Published: 2025-12-18 Categories: cs.CL, cs.AI arXiv: https://arxiv.org/abs/2512.16602v3 We introduce Refusal Steering, an inference-time method to exercise fine-grained control over Large Language Models refusal behaviour on politically sensitive topics without retraining. We replace fragile pattern-based refusal detection with an LLM-as-a-judge that assigns refusal confidence scores and we propose a ridge-regularized variant to compute steering vectors that better isolate the refusal--compliance direction. On Qwen3-Next-80B-A3B-Thinking, our method removes the refusal behaviour of the model around politically sensitive topics while maintaining safety on JailbreakBench and near-baseline performance on general benchmarks. The approach generalizes across 4B and 80B models and can also induce targeted refusals when desired. We analize the steering vectors and show that refusal signals concentrate in deeper layers of the transformer and are distributed across many dimensions. Together, these results demonstrate that activation steering can remove political refusal behaviour while retaining safety alignment for harmful content, offering a practical path to controllable, transparent moderation at inference time. --- ### Refusal Steering: Fine-grained Control over LLM Refusal Behaviour for Sensitive Topics Authors: Iker García-Ferrero, David Montero, Roman Orus Published: 2025-12-18 Categories: cs.CL, cs.AI arXiv: https://arxiv.org/abs/2512.16602v1 We introduce Refusal Steering, an inference-time method to exercise fine-grained control over Large Language Models refusal behaviour on politically sensitive topics without retraining. We replace fragile pattern-based refusal detection with an LLM-as-a-judge that assigns refusal confidence scores and we propose a ridge-regularized variant to compute steering vectors that better isolate the refusal--compliance direction. On Qwen3-Next-80B-A3B-Thinking, our method removes the refusal behaviour of the model around politically sensitive topics while maintaining safety on JailbreakBench and near-baseline performance on general benchmarks. The approach generalizes across 4B and 80B models and can also induce targeted refusals when desired. We analize the steering vectors and show that refusal signals concentrate in deeper layers of the transformer and are distributed across many dimensions. Together, these results demonstrate that activation steering can remove political refusal behaviour while retaining safety alignment for harmful content, offering a practical path to controllable, transparent moderation at inference time. --- ### Muon is Provably Faster with Momentum Variance Reduction Authors: Xun Qian, Hussein Rammal, Dmitry Kovalev, Peter Richtárik Published: 2025-12-18 Categories: math.OC, cs.LG arXiv: https://arxiv.org/abs/2512.16598v1 Recent empirical research has demonstrated that deep learning optimizers based on the linear minimization oracle (LMO) over specifically chosen Non-Euclidean norm balls, such as Muon and Scion, outperform Adam-type methods in the training of large language models. In this work, we show that such optimizers can be provably improved by replacing their vanilla momentum by momentum variance reduction (MVR). Instead of proposing and analyzing MVR variants of Muon and Scion separately, we incorporate MVR into the recently proposed Gluon framework, which captures Muon, Scion and other specific Non-Euclidean LMO-based methods as special cases, and at the same time works with a more general smoothness assumption which better captures the layer-wise structure of neural networks. In the non-convex case, we incorporate MVR into Gluon in three different ways. All of them improve the convergence rate from ${\cal O} (\frac{1}{K^{1/4}})$ to ${\cal O} (\frac{1}{K^{1/3}})$. Additionally, we provide improved rates in the star-convex case. Finally, we conduct several numerical experiments that verify the superior performance of our proposed algorithms in terms of iteration complexity. --- ### Yuan-TecSwin: A text conditioned Diffusion model with Swin-transformer blocks Authors: Shaohua Wu, Tong Yu, Shenling Wang, Xudong Zhao Published: 2025-12-18 Categories: cs.CV, cs.AI arXiv: https://arxiv.org/abs/2512.16586v1 Diffusion models have shown remarkable capacity in image synthesis based on their U-shaped architecture and convolutional neural networks (CNN) as basic blocks. The locality of the convolution operation in CNN may limit the model's ability to understand long-range semantic information. To address this issue, we propose Yuan-TecSwin, a text-conditioned diffusion model with Swin-transformer in this work. The Swin-transformer blocks take the place of CNN blocks in the encoder and decoder, to improve the non-local modeling ability in feature extraction and image restoration. The text-image alignment is improved with a well-chosen text encoder, effective utilization of text embedding, and careful design in the incorporation of text condition. Using an adapted time step to search in different diffusion stages, inference performance is further improved by 10%. Yuan-TecSwin achieves the state-of-the-art FID score of 1.37 on ImageNet generation benchmark, without any additional models at different denoising stages. In a side-by-side comparison, we find it difficult for human interviewees to tell the model-generated images from the human-painted ones. --- ### Sketch-in-Latents: Eliciting Unified Reasoning in MLLMs Authors: Jintao Tong, Jiaqi Gu, Yujing Lou, Lubin Fan, Yixiong Zou, Yue Wu, Jieping Ye, Ruixuan Li Published: 2025-12-18 Categories: cs.CV arXiv: https://arxiv.org/abs/2512.16584v1 While Multimodal Large Language Models (MLLMs) excel at visual understanding tasks through text reasoning, they often fall short in scenarios requiring visual imagination. Unlike current works that take predefined external toolkits or generate images during thinking, however, humans can form flexible visual-text imagination and interactions during thinking without predefined toolkits, where one important reason is that humans construct the visual-text thinking process in a unified space inside the brain. Inspired by this capability, given that current MLLMs already encode visual and text information in the same feature space, we hold that visual tokens can be seamlessly inserted into the reasoning process carried by text tokens, where ideally, all visual imagination processes can be encoded by the latent features. To achieve this goal, we propose Sketch-in-Latents (SkiLa), a novel paradigm for unified multi-modal reasoning that expands the auto-regressive capabilities of MLLMs to natively generate continuous visual embeddings, termed latent sketch tokens, as visual thoughts. During multi-step reasoning, the model dynamically alternates between textual thinking mode for generating textual think tokens and visual sketching mode for generating latent sketch tokens. A latent visual semantics reconstruction mechanism is proposed to ensure these latent sketch tokens are semantically grounded. Extensive experiments demonstrate that SkiLa achieves superior performance on vision-centric tasks while exhibiting strong generalization to diverse general multi-modal benchmarks. Codes will be released at https://github.com/TungChintao/SkiLa. --- ### Abacus: Self-Supervised Event Counting-Aligned Distributional Pretraining for Sequential User Modeling Authors: Sullivan Castro, Artem Betlei, Thomas Di Martino, Nadir El Manouzi Published: 2025-12-18 Categories: cs.LG, cs.IR arXiv: https://arxiv.org/abs/2512.16581v1 Modeling user purchase behavior is a critical challenge in display advertising systems, necessary for real-time bidding. The difficulty arises from the sparsity of positive user events and the stochasticity of user actions, leading to severe class imbalance and irregular event timing. Predictive systems usually rely on hand-crafted "counter" features, overlooking the fine-grained temporal evolution of user intent. Meanwhile, current sequential models extract direct sequential signal, missing useful event-counting statistics. We enhance deep sequential models with self-supervised pretraining strategies for display advertising. Especially, we introduce Abacus, a novel approach of predicting the empirical frequency distribution of user events. We further propose a hybrid objective unifying Abacus with sequential learning objectives, combining stability of aggregated statistics with the sequence modeling sensitivity. Experiments on two real-world datasets show that Abacus pretraining outperforms existing methods accelerating downstream task convergence, while hybrid approach yields up to +6.1% AUC compared to the baselines. --- ### CRONOS: Continuous Time Reconstruction for 4D Medical Longitudinal Series Authors: Nico Albert Disch, Saikat Roy, Constantin Ulrich, Yannick Kirchhoff, Maximilian Rokuss, Robin Peretzke, David Zimmerer, Klaus Maier-Hein Published: 2025-12-18 Categories: cs.CV arXiv: https://arxiv.org/abs/2512.16577v1 Forecasting how 3D medical scans evolve over time is important for disease progression, treatment planning, and developmental assessment. Yet existing models either rely on a single prior scan, fixed grid times, or target global labels, which limits voxel-level forecasting under irregular sampling. We present CRONOS, a unified framework for many-to-one prediction from multiple past scans that supports both discrete (grid-based) and continuous (real-valued) timestamps in one model, to the best of our knowledge the first to achieve continuous sequence-to-image forecasting for 3D medical data. CRONOS learns a spatio-temporal velocity field that transports context volumes toward a target volume at an arbitrary time, while operating directly in 3D voxel space. Across three public datasets spanning Cine-MRI, perfusion CT, and longitudinal MRI, CRONOS outperforms other baselines, while remaining computationally competitive. We will release code and evaluation protocols to enable reproducible, multi-dataset benchmarking of multi-context, continuous-time forecasting. --- ### Causal-Tune: Mining Causal Factors from Vision Foundation Models for Domain Generalized Semantic Segmentation Authors: Yin Zhang, Yongqiang Zhang, Yaoyue Zheng, Bogdan Raducanu, Dan Liu Published: 2025-12-18 Categories: cs.CV arXiv: https://arxiv.org/abs/2512.16567v1 Fine-tuning Vision Foundation Models (VFMs) with a small number of parameters has shown remarkable performance in Domain Generalized Semantic Segmentation (DGSS). Most existing works either train lightweight adapters or refine intermediate features to achieve better generalization on unseen domains. However, they both overlook the fact that long-term pre-trained VFMs often exhibit artifacts, which hinder the utilization of valuable representations and ultimately degrade DGSS performance. Inspired by causal mechanisms, we observe that these artifacts are associated with non-causal factors, which usually reside in the low- and high-frequency components of the VFM spectrum. In this paper, we explicitly examine the causal and non-causal factors of features within VFMs for DGSS, and propose a simple yet effective method to identify and disentangle them, enabling more robust domain generalization. Specifically, we propose Causal-Tune, a novel fine-tuning strategy designed to extract causal factors and suppress non-causal ones from the features of VFMs. First, we extract the frequency spectrum of features from each layer using the Discrete Cosine Transform (DCT). A Gaussian band-pass filter is then applied to separate the spectrum into causal and non-causal components. To further refine the causal components, we introduce a set of causal-aware learnable tokens that operate in the frequency domain, while the non-causal components are discarded. Finally, refined features are transformed back into the spatial domain via inverse DCT and passed to the next layer. Extensive experiments conducted on various cross-domain tasks demonstrate the effectiveness of Causal-Tune. In particular, our method achieves superior performance under adverse weather conditions, improving +4.8% mIoU over the baseline in snow conditions. --- ### Non-Asymptotic Global Convergence of PPO-Clip Authors: Qiming Dai, Yin Liu, Junyu Zhang, Zaiwen Wen Published: 2025-12-18 Categories: math.OC, cs.LG arXiv: https://arxiv.org/abs/2512.16565v2 Reinforcement learning has gained attention for modern Large Language Model post-training. The actor-only variants of Proximal Policy Optimization (PPO) are widely applied for their efficiency. These algorithms incorporate a clipping mechanism to improve stability. Besides, a regularization term, such as the reverse KL-divergence or a more general \(f\)-divergence, is introduced to control excessive deviation from a reference policy. Despite their empirical success, a rigorous theoretical understanding of the problem and the algorithm's properties is limited. This paper advances the theoretical foundations of the PPO-Clip algorithm by analyzing a deterministic actor-only PPO algorithm within the general RL setting with \(f\)-divergence regularization under the softmax policy parameterization. We derive a non-uniform Lipschitz smoothness condition and a Łojasiewicz inequality for the considered problem. Based on these properties, we establish non-asymptotic global linear convergence in value gap for the forward KL regularizer. For the reverse KL regularizer, we derive global linear convergence from any finite softmax initialization in both value gap and squared policy distance. --- ### Non-Asymptotic Global Convergence of PPO-Clip Authors: Yin Liu, Qiming Dai, Junyu Zhang, Zaiwen Wen Published: 2025-12-18 Categories: math.OC, cs.LG arXiv: https://arxiv.org/abs/2512.16565v1 Reinforcement learning (RL) has gained attention for aligning large language models (LLMs) via reinforcement learning from human feedback (RLHF). The actor-only variants of Proximal Policy Optimization (PPO) are widely applied for their efficiency. These algorithms incorporate a clipping mechanism to improve stability. Besides, a regularization term, such as the reverse KL-divergence or a more general \(f\)-divergence, is introduced to prevent policy drift. Despite their empirical success, a rigorous theoretical understanding of the problem and the algorithm's properties is limited. This paper advances the theoretical foundations of the PPO-Clip algorithm by analyzing a deterministic actor-only PPO algorithm within the general RL setting with \(f\)-divergence regularization under the softmax policy parameterization. We derive a non-uniform Lipschitz smoothness condition and a Łojasiewicz inequality for the considered problem. Based on these, a non-asymptotic linear convergence rate to the globally optimal policy is established for the forward KL-regularizer. Furthermore, stationary convergence and local linear convergence are derived for the reverse KL-regularizer. --- ### 4D Primitive-Mâché: Glueing Primitives for Persistent 4D Scene Reconstruction Authors: Kirill Mazur, Marwan Taher, Andrew J. Davison Published: 2025-12-18 Categories: cs.CV arXiv: https://arxiv.org/abs/2512.16564v1 We present a dynamic reconstruction system that receives a casual monocular RGB video as input, and outputs a complete and persistent reconstruction of the scene. In other words, we reconstruct not only the the currently visible parts of the scene, but also all previously viewed parts, which enables replaying the complete reconstruction across all timesteps. Our method decomposes the scene into a set of rigid 3D primitives, which are assumed to be moving throughout the scene. Using estimated dense 2D correspondences, we jointly infer the rigid motion of these primitives through an optimisation pipeline, yielding a 4D reconstruction of the scene, i.e. providing 3D geometry dynamically moving through time. To achieve this, we also introduce a mechanism to extrapolate motion for objects that become invisible, employing motion-grouping techniques to maintain continuity. The resulting system enables 4D spatio-temporal awareness, offering capabilities such as replayable 3D reconstructions of articulated objects through time, multi-object scanning, and object permanence. On object scanning and multi-object datasets, our system significantly outperforms existing methods both quantitatively and qualitatively. --- ### N3D-VLM: Native 3D Grounding Enables Accurate Spatial Reasoning in Vision-Language Models Authors: Yuxin Wang, Lei Ke, Boqiang Zhang, Tianyuan Qu, Hanxun Yu, Zhenpeng Huang, Meng Yu, Dan Xu, Dong Yu Published: 2025-12-18 Categories: cs.CV arXiv: https://arxiv.org/abs/2512.16561v1 While current multimodal models can answer questions based on 2D images, they lack intrinsic 3D object perception, limiting their ability to comprehend spatial relationships and depth cues in 3D scenes. In this work, we propose N3D-VLM, a novel unified framework that seamlessly integrates native 3D object perception with 3D-aware visual reasoning, enabling both precise 3D grounding and interpretable spatial understanding. Unlike conventional end-to-end models that directly predict answers from RGB/RGB-D inputs, our approach equips the model with native 3D object perception capabilities, enabling it to directly localize objects in 3D space based on textual descriptions. Building upon accurate 3D object localization, the model further performs explicit reasoning in 3D, achieving more interpretable and structured spatial understanding. To support robust training for these capabilities, we develop a scalable data construction pipeline that leverages depth estimation to lift large-scale 2D annotations into 3D space, significantly increasing the diversity and coverage for 3D object grounding data, yielding over six times larger than the largest existing single-image 3D detection dataset. Moreover, the pipeline generates spatial question-answering datasets that target chain-of-thought (CoT) reasoning in 3D, facilitating joint training for both 3D object localization and 3D spatial reasoning. Experimental results demonstrate that our unified framework not only achieves state-of-the-art performance on 3D grounding tasks, but also consistently surpasses existing methods in 3D spatial reasoning in vision-language model. --- ### Persistent Multiscale Density-based Clustering Authors: Daniël Bot, Leland McInnes, Jan Aerts Published: 2025-12-18 Categories: cs.LG arXiv: https://arxiv.org/abs/2512.16558v3 Clustering is a cornerstone of modern data analysis. Detecting clusters in exploratory data analyses (EDA) requires algorithms that make few assumptions about the data. Density-based clustering algorithms are particularly well-suited for EDA because they describe high-density regions, assuming only that a density exists. Applying density-based clustering algorithms in practice, however, requires selecting appropriate hyperparameters, which is difficult without prior knowledge of the data distribution. For example, DBSCAN requires selecting a density threshold, and HDBSCAN* relies on a minimum cluster size parameter. In this work, we propose Persistent Leaves Spatial Clustering for Applications with Noise (PLSCAN), a multiscale density-based clustering algorithm that replaces HDBSCAN*'s fixed minimum cluster size pruning of a mutual-reachability linkage hierarchy with a persistence-based cluster selection procedure. Effectively, PLSCAN identifies all minimum cluster sizes for which HDBSCAN* produces stable (leaf) clusters. In concept, PLSCAN applies scale-space clustering principles and is equivalent to persistent homology on a novel metric space. We compare its performance to HDBSCAN* on several real-world datasets, demonstrating that it achieves a higher median ARI, is less sensitive to changes in the number of mutual reachability neighbours, and has higher stability under resampling. Additionally, we compare PLSCAN's computational costs to $k$-Means++, demonstrating competitive run-times on low-dimensional datasets. At higher dimensions, run times scale more similarly to HDBSCAN*. --- ### Persistent Multiscale Density-based Clustering Authors: Daniël Bot, Leland McInnes, Jan Aerts Published: 2025-12-18 Categories: cs.LG arXiv: https://arxiv.org/abs/2512.16558v1 Clustering is a cornerstone of modern data analysis. Detecting clusters in exploratory data analyses (EDA) requires algorithms that make few assumptions about the data. Density-based clustering algorithms are particularly well-suited for EDA because they describe high-density regions, assuming only that a density exists. Applying density-based clustering algorithms in practice, however, requires selecting appropriate hyperparameters, which is difficult without prior knowledge of the data distribution. For example, DBSCAN requires selecting a density threshold, and HDBSCAN* relies on a minimum cluster size parameter. In this work, we propose Persistent Leaves Spatial Clustering for Applications with Noise (PLSCAN). This novel density-based clustering algorithm efficiently identifies all minimum cluster sizes for which HDBSCAN* produces stable (leaf) clusters. PLSCAN applies scale-space clustering principles and is equivalent to persistent homology on a novel metric space. We compare its performance to HDBSCAN* on several real-world datasets, demonstrating that it achieves a higher average ARI and is less sensitive to changes in the number of mutual reachability neighbours. Additionally, we compare PLSCAN's computational costs to k-Means, demonstrating competitive run-times on low-dimensional datasets. At higher dimensions, run times scale more similarly to HDBSCAN*. --- ### Needle in the Web: A Benchmark for Retrieving Targeted Web Pages in the Wild Authors: Yumeng Wang, Tianyu Fan, Lingrui Xu, Chao Huang Published: 2025-12-18 Categories: cs.AI, cs.CL arXiv: https://arxiv.org/abs/2512.16553v1 Large Language Models (LLMs) have evolved from simple chatbots into sophisticated agents capable of automating complex real-world tasks, where browsing and reasoning over live web content is key to assessing retrieval and cognitive skills. Existing benchmarks like BrowseComp and xBench-DeepSearch emphasize complex reasoning searches requiring multi-hop synthesis but neglect Fuzzy Exploratory Search, namely queries that are vague and multifaceted, where users seek the most relevant webpage rather than a single factual answer. To address this gap, we introduce Needle in the Web, a novel benchmark specifically designed to evaluate modern search agents and LLM-based systems on their ability to retrieve and reason over real-world web content in response to ambiguous, exploratory queries under varying levels of difficulty. Needle in the Web comprises 663 questions spanning seven distinct domains. To ensure high query quality and answer uniqueness, we employ a flexible methodology that reliably generates queries of controllable difficulty based on factual claims of web contents. We benchmark three leading LLMs and three agent-based search systems on Needle in the Web, finding that most models struggle: many achieve below 35% accuracy, and none consistently excel across domains or difficulty levels. These findings reveal that Needle in the Web presents a significant challenge for current search systems and highlights the open problem of effective fuzzy retrieval under semantic ambiguity. --- ### Predictive Inorganic Synthesis based on Machine Learning using Small Data sets: a case study of Hydrodynamic Diameter-controlled Cu Nanoparticles Authors: Brent Motmans, Digvijay Ghogare, Thijs G. I. van Wijk, Joren Van Herck, Saba Heidarian, Pieter De Meyer, Berend Smit, An Hardy, Danny E. P. Vanpoucke Published: 2025-12-18 Categories: cond-mat.mtrl-sci, cs.LG arXiv: https://arxiv.org/abs/2512.16545v3 Cu NPs have a broad applicability, yet their synthesis is sensitive to subtle changes in reaction parameters. This sensitivity, combined with the time- and resource-intensive nature of experimental optimization, poses a major challenge in achieving reproducible and size-controlled synthesis. While ML shows promise in materials research, its application is often limited by scarcity of large high-quality experimental data sets. This study explores ML to predict the DLS-derived hydrodynamic diameter of Cu NPs using a small data set of 25 syntheses. Latin Hypercube Sampling is used to efficiently cover the parameter space while creating the experimental data set. Ensemble regression models successfully predict hydrodynamic diameters with good predictive performance given the limited dataset. Since quantitative regression requires a unique DLS-derived hydrodynamic diameter, the regression model is restricted to mono-modal DLS distributions, while a complementary classification model identifies synthesis conditions for which quantitative prediction is applicable. Using equivalent out-of-sample validation, the ML and DoE models showed comparable generalization. The final ensemble model achieved an R2=0.74 compared to 0.60 for the DoE model, while retaining the complete synthesis parameter space, making it better suited for synthesis guidance. Additionally, classification models using both random forests and LLMs are evaluated to distinguish between large and small particles. These classification models exhibited only modest predictive performance, indicating that this small dataset is insufficient to fully exploit the capabilities of complex LLMs. Overall, this study demonstrates that carefully curated small data sets, paired with robust classical ML, can effectively support the synthesis of Cu NPs and highlights that for lab-scale studies, complex models like LLMs may offer limited benefits. --- ### Predictive Inorganic Synthesis based on Machine Learning using Small Data sets: a case study of size-controlled Cu Nanoparticles Authors: Brent Motmans, Digvijay Ghogare, Thijs G. I. van Wijk, An Hardy, Danny E. P. Vanpoucke Published: 2025-12-18 Categories: cond-mat.mtrl-sci, cs.LG arXiv: https://arxiv.org/abs/2512.16545v1 Copper nanoparticles (Cu NPs) have a broad applicability, yet their synthesis is sensitive to subtle changes in reaction parameters. This sensitivity, combined with the time- and resource-intensive nature of experimental optimization, poses a major challenge in achieving reproducible and size-controlled synthesis. While Machine Learning (ML) shows promise in materials research, its application is often limited by scarcity of large high-quality experimental data sets. This study explores ML to predict the size of Cu NPs from microwave-assisted polyol synthesis using a small data set of 25 in-house performed syntheses. Latin Hypercube Sampling is used to efficiently cover the parameter space while creating the experimental data set. Ensemble regression models, built with the AMADEUS framework, successfully predict particle sizes with high accuracy ($R^2 = 0.74$), outperforming classical statistical approaches ($R^2 = 0.60$). Overall, this study highlights that, for lab-scale synthesis optimization, high-quality small datasets combined with classical, interpretable ML models outperform traditional statistical methods and are fully sufficient for quantitative synthesis prediction. This approach provides a sustainable and experimentally realistic pathway toward data-driven inorganic synthesis design. --- ### UM_FHS at the CLEF 2025 SimpleText Track: Comparing No-Context and Fine-Tune Approaches for GPT-4.1 Models in Sentence and Document-Level Text Simplification Authors: Primoz Kocbek, Gregor Stiglic Published: 2025-12-18 Categories: cs.CL arXiv: https://arxiv.org/abs/2512.16541v1 This work describes our submission to the CLEF 2025 SimpleText track Task 1, addressing both sentenceand document-level simplification of scientific texts. The methodology centered on using the gpt-4.1, gpt-4.1mini, and gpt-4.1-nano models from OpenAI. Two distinct approaches were compared: a no-context method relying on prompt engineering and a fine-tuned (FT) method across models. The gpt-4.1-mini model with no-context demonstrated robust performance at both levels of simplification, while the fine-tuned models showed mixed results, highlighting the complexities of simplifying text at different granularities, where gpt-4.1-nano-ft performance stands out at document-level simplification in one case. --- ### A Systematic Study of Code Obfuscation Against LLM-based Vulnerability Detection Authors: Xiao Li, Yue Li, Hao Wu, Yue Zhang, Yechao Zhang, Fengyuan Xu, Sheng Zhong Published: 2025-12-18 Categories: cs.CR, cs.LG arXiv: https://arxiv.org/abs/2512.16538v1 As large language models (LLMs) are increasingly adopted for code vulnerability detection, their reliability and robustness across diverse vulnerability types have become a pressing concern. In traditional adversarial settings, code obfuscation has long been used as a general strategy to bypass auditing tools, preserving exploitability without tampering with the tools themselves. Numerous efforts have explored obfuscation methods and tools, yet their capabilities differ in terms of supported techniques, granularity, and programming languages, making it difficult to systematically assess their impact on LLM-based vulnerability detection. To address this gap, we provide a structured systematization of obfuscation techniques and evaluate them under a unified framework. Specifically, we categorize existing obfuscation methods into three major classes (layout, data flow, and control flow) covering 11 subcategories and 19 concrete techniques. We implement these techniques across four programming languages (Solidity, C, C++, and Python) using a consistent LLM-driven approach, and evaluate their effects on 15 LLMs spanning four model families (DeepSeek, OpenAI, Qwen, and LLaMA), as well as on two coding agents (GitHub Copilot and Codex). Our findings reveal both positive and negative impacts of code obfuscation on LLM-based vulnerability detection, highlighting conditions under which obfuscation leads to performance improvements or degradations. We further analyze these outcomes with respect to vulnerability characteristics, code properties, and model attributes. Finally, we outline several open problems and propose future directions to enhance the robustness of LLMs for real-world vulnerability detection. --- ### From Personalization to Prejudice: Bias and Discrimination in Memory-Enhanced AI Agents for Recruitment Authors: Himanshu Gharat, Himanshi Agrawal, Gourab K. Patro Published: 2025-12-18 Categories: cs.AI, cs.IR arXiv: https://arxiv.org/abs/2512.16532v1 Large Language Models (LLMs) have empowered AI agents with advanced capabilities for understanding, reasoning, and interacting across diverse tasks. The addition of memory further enhances them by enabling continuity across interactions, learning from past experiences, and improving the relevance of actions and responses over time; termed as memory-enhanced personalization. Although such personalization through memory offers clear benefits, it also introduces risks of bias. While several previous studies have highlighted bias in ML and LLMs, bias due to memory-enhanced personalized agents is largely unexplored. Using recruitment as an example use case, we simulate the behavior of a memory-enhanced personalized agent, and study whether and how bias is introduced and amplified in and across various stages of operation. Our experiments on agents using safety-trained LLMs reveal that bias is systematically introduced and reinforced through personalization, emphasizing the need for additional protective measures or agent guardrails in memory-enhanced LLM-based AI agents. --- ### Scaling Laws for Energy Efficiency of Local LLMs Authors: Ander Alvarez, Alessandro Genuardi, Nilotpal Sinha, Antonio Tiene, Mikail Okyay, Bakbergen Ryskulov, David Montero, Samuel Mugel, Román Orús Published: 2025-12-18 Categories: cs.AI arXiv: https://arxiv.org/abs/2512.16531v4 Deploying local large language models and vision-language models on edge devices requires balancing accuracy with constrained computational and energy budgets. Although graphics processors dominate modern artificial-intelligence deployment, most consumer hardware--including laptops, desktops, industrial controllers, and embedded systems--relies on central processing units. Despite this, the computational laws governing central-processing-unit-only inference for local language and vision-language workloads remain largely unexplored. We systematically benchmark large language and vision-language models on two representative central-processing-unit tiers widely used for local inference: a MacBook Pro M2, reflecting mainstream laptop-class deployment, and a Raspberry Pi 5, representing constrained, low-power embedded settings. Using a unified methodology based on continuous sampling of processor and memory usage together with area-under-curve integration, we characterize how computational load scales with input text length for language models and with image resolution for vision-language models. We uncover two empirical scaling laws: (1) computational cost for language-model inference scales approximately linearly with token length; and (2) vision-language models exhibit a preprocessing-driven "resolution knee", where compute remains constant above an internal resolution clamp and decreases sharply below it. Beyond these laws, we show that quantum-inspired compression reduces processor and memory usage by up to 71.9% and energy consumption by up to 62%, while preserving or improving semantic accuracy. These results provide a systematic quantification of multimodal central-processing-unit-only scaling for local language and vision-language workloads, and they identify model compression and input-resolution preprocessing as effective, low-cost levers for sustainable edge inference. --- ### Scaling Laws for Energy Efficiency of Local LLMs Authors: Ander Alvarez, Alessandro Genuardi, Nilotpal Sinha, Antonio Tiene, Samuel Mugel, Román Orús Published: 2025-12-18 Categories: cs.AI arXiv: https://arxiv.org/abs/2512.16531v1 Deploying local large language models and vision-language models on edge devices requires balancing accuracy with constrained computational and energy budgets. Although graphics processors dominate modern artificial-intelligence deployment, most consumer hardware--including laptops, desktops, industrial controllers, and embedded systems--relies on central processing units. Despite this, the computational laws governing central-processing-unit-only inference for local language and vision-language workloads remain largely unexplored. We systematically benchmark large language and vision-language models on two representative central-processing-unit tiers widely used for local inference: a MacBook Pro M2, reflecting mainstream laptop-class deployment, and a Raspberry Pi 5, representing constrained, low-power embedded settings. Using a unified methodology based on continuous sampling of processor and memory usage together with area-under-curve integration, we characterize how computational load scales with input text length for language models and with image resolution for vision-language models. We uncover two empirical scaling laws: (1) computational cost for language-model inference scales approximately linearly with token length; and (2) vision-language models exhibit a preprocessing-driven "resolution knee", where compute remains constant above an internal resolution clamp and decreases sharply below it. Beyond these laws, we show that quantum-inspired compression reduces processor and memory usage by up to 71.9% and energy consumption by up to 62%, while preserving or improving semantic accuracy. These results provide a systematic quantification of multimodal central-processing-unit-only scaling for local language and vision-language workloads, and they identify model compression and input-resolution preprocessing as effective, low-cost levers for sustainable edge inference. --- ### Plain language adaptations of biomedical text using LLMs: Comparision of evaluation metrics Authors: Primoz Kocbek, Leon Kopitar, Gregor Stiglic Published: 2025-12-18 Categories: cs.CL, cs.AI arXiv: https://arxiv.org/abs/2512.16530v1 This study investigated the application of Large Language Models (LLMs) for simplifying biomedical texts to enhance health literacy. Using a public dataset, which included plain language adaptations of biomedical abstracts, we developed and evaluated several approaches, specifically a baseline approach using a prompt template, a two AI agent approach, and a fine-tuning approach. We selected OpenAI gpt-4o and gpt-4o mini models as baselines for further research. We evaluated our approaches with quantitative metrics, such as Flesch-Kincaid grade level, SMOG Index, SARI, and BERTScore, G-Eval, as well as with qualitative metric, more precisely 5-point Likert scales for simplicity, accuracy, completeness, brevity. Results showed a superior performance of gpt-4o-mini and an underperformance of FT approaches. G-Eval, a LLM based quantitative metric, showed promising results, ranking the approaches similarly as the qualitative metric. --- ### ParamExplorer: A framework for exploring parameters in generative art Authors: Julien Gachadoat, Guillaume Lagarde Published: 2025-12-18 Categories: cs.AI, cs.HC, cs.SE arXiv: https://arxiv.org/abs/2512.16529v2 Generative art systems often involve high-dimensional and complex parameter spaces in which aesthetically compelling outputs occupy only small, fragmented regions. Because of this combinatorial explosion, artists typically rely on extensive manual trial-and-error, leaving many potentially interesting configurations undiscovered. In this work we make two contributions. First, we introduce ParamExplorer, an interactive and modular framework inspired by reinforcement learning that helps the exploration of parameter spaces in generative art algorithms, guided by human-in-the-loop or even automated feedback. The framework also integrates seamlessly with existing p5js projects. Second, within this framework we implement and evaluate several exploration strategies, referred to as agents. --- ### ParamExplorer: A framework for exploring parameters in generative art Authors: Julien Gachadoat, Guillaume Lagarde Published: 2025-12-18 Categories: cs.AI, cs.HC, cs.SE arXiv: https://arxiv.org/abs/2512.16529v1 Generative art systems often involve high-dimensional and complex parameter spaces in which aesthetically compelling outputs occupy only small, fragmented regions. Because of this combinatorial explosion, artists typically rely on extensive manual trial-and-error, leaving many potentially interesting configurations undiscovered. In this work we make two contributions. First, we introduce ParamExplorer, an interactive and modular framework inspired by reinforcement learning that helps the exploration of parameter spaces in generative art algorithms, guided by human-in-the-loop or even automated feedback. The framework also integrates seamlessly with existing p5.js projects. Second, within this framework we implement and evaluate several exploration strategies, referred to as agents. --- ### TTP: Test-Time Padding for Adversarial Detection and Robust Adaptation on Vision-Language Models Authors: Zhiwei Li, Yitian Pang, Weining Wang, Zhenan Sun, Qi Li Published: 2025-12-18 Categories: cs.CV, cs.AI arXiv: https://arxiv.org/abs/2512.16523v2 Vision-Language Models (VLMs), such as CLIP, have achieved impressive zero-shot recognition performance but remain highly susceptible to adversarial perturbations, posing significant risks in safety-critical scenarios. Previous training-time defenses rely on adversarial fine-tuning, which requires labeled data and costly retraining, while existing test-time strategies fail to reliably distinguish between clean and adversarial inputs, thereby preventing both adversarial robustness and clean accuracy from reaching their optimum. To address these limitations, we propose Test-Time Padding (TTP), a lightweight defense framework that performs adversarial detection followed by targeted adaptation at inference. TTP identifies adversarial inputs via the cosine similarity shift between CLIP feature embeddings computed before and after spatial padding, yielding a universal threshold for reliable detection across architectures and datasets. For detected adversarial cases, TTP employs trainable padding to restore disrupted attention patterns, coupled with a similarity-aware ensemble strategy for a more robust final prediction. For clean inputs, TTP leaves them unchanged by default or optionally integrates existing test-time adaptation techniques for further accuracy gains. Comprehensive experiments on diverse CLIP backbones and fine-grained benchmarks show that TTP consistently surpasses state-of-the-art test-time defenses, delivering substantial improvements in adversarial robustness without compromising clean accuracy. The code for this paper will be released soon. --- ### TTP: Test-Time Padding for Adversarial Detection and Robust Adaptation on Vision-Language Models Authors: Zhiwei Li, Yitian Pang, Weining Wang, Zhenan Sun, Qi Li Published: 2025-12-18 Categories: cs.CV, cs.AI arXiv: https://arxiv.org/abs/2512.16523v1 Vision-Language Models (VLMs), such as CLIP, have achieved impressive zero-shot recognition performance but remain highly susceptible to adversarial perturbations, posing significant risks in safety-critical scenarios. Previous training-time defenses rely on adversarial fine-tuning, which requires labeled data and costly retraining, while existing test-time strategies fail to reliably distinguish between clean and adversarial inputs, thereby preventing both adversarial robustness and clean accuracy from reaching their optimum. To address these limitations, we propose Test-Time Padding (TTP), a lightweight defense framework that performs adversarial detection followed by targeted adaptation at inference. TTP identifies adversarial inputs via the cosine similarity shift between CLIP feature embeddings computed before and after spatial padding, yielding a universal threshold for reliable detection across architectures and datasets. For detected adversarial cases, TTP employs trainable padding to restore disrupted attention patterns, coupled with a similarity-aware ensemble strategy for a more robust final prediction. For clean inputs, TTP leaves them unchanged by default or optionally integrates existing test-time adaptation techniques for further accuracy gains. Comprehensive experiments on diverse CLIP backbones and fine-grained benchmarks show that TTP consistently surpasses state-of-the-art test-time defenses, delivering substantial improvements in adversarial robustness without compromising clean accuracy. The code for this paper will be released soon. --- ### Pseudo-Cepstrum: Pitch Modification for Mel-Based Neural Vocoders Authors: Nikolaos Ellinas, Alexandra Vioni, Panos Kakoulidis, Georgios Vamvoukakis, Myrsini Christidou, Konstantinos Markopoulos, Junkwang Oh, Gunu Jho, Inchul Hwang, Aimilios Chalamandaris, Pirros Tsiakoulis Published: 2025-12-18 Categories: cs.SD, cs.LG, eess.AS arXiv: https://arxiv.org/abs/2512.16519v1 This paper introduces a cepstrum-based pitch modification method that can be applied to any mel-spectrogram representation. As a result, this method is compatible with any mel-based vocoder without requiring any additional training or changes to the model. This is achieved by directly modifying the cepstrum feature space in order to shift the harmonic structure to the desired target. The spectrogram magnitude is computed via the pseudo-inverse mel transform, then converted to the cepstrum by applying DCT. In this domain, the cepstral peak is shifted without having to estimate its position and the modified mel is recomputed by applying IDCT and mel-filterbank. These pitch-shifted mel-spectrogram features can be converted to speech with any compatible vocoder. The proposed method is validated experimentally with objective and subjective metrics on various state-of-the-art neural vocoders as well as in comparison with traditional pitch modification methods. --- ### The Universe Learning Itself: On the Evolution of Dynamics from the Big Bang to Machine Intelligence Authors: Pradeep Singh, Mudasani Rushikesh, Bezawada Sri Sai Anurag, Balasubramanian Raman Published: 2025-12-18 Categories: nlin.AO, cs.AI arXiv: https://arxiv.org/abs/2512.16515v2 We develop a unified, dynamical-systems narrative of the universe that traces a continuous chain of structure formation from the Big Bang to contemporary human societies and their artificial learning systems. Rather than treating cosmology, astrophysics, geophysics, biology, cognition, and machine intelligence as disjoint domains, we view each as successive regimes of dynamics on ever-richer state spaces, stitched together by phase transitions, symmetry-breaking events, and emergent attractors. Starting from inflationary field dynamics and the growth of primordial perturbations, we describe how gravitational instability sculpts the cosmic web, how dissipative collapse in baryonic matter yields stars and planets, and how planetary-scale geochemical cycles define long-lived nonequilibrium attractors. Within these attractors, we frame the origin of life as the emergence of self-maintaining reaction networks, evolutionary biology as flow on high-dimensional genotype-phenotype-environment manifolds, and brains as adaptive dynamical systems operating near critical surfaces. Human culture and technology-including modern machine learning and artificial intelligence-are then interpreted as symbolic and institutional dynamics that implement and refine engineered learning flows which recursively reshape their own phase space. Throughout, we emphasize recurring mathematical motifs-instability, bifurcation, multiscale coupling, and constrained flows on measure-zero subsets of the accessible state space. Our aim is not to present any new cosmological or biological model, but a cross-scale, theoretical perspective: a way of reading the universe's history as the evolution of dynamics itself, culminating (so far) in biological and artificial systems capable of modeling, predicting, and deliberately perturbing their own future trajectories. --- ### The Universe Learning Itself: On the Evolution of Dynamics from the Big Bang to Machine Intelligence Authors: Pradeep Singh, Mudasani Rushikesh, Bezawada Sri Sai Anurag, Balasubramanian Raman Published: 2025-12-18 Categories: nlin.AO, cs.AI arXiv: https://arxiv.org/abs/2512.16515v1 We develop a unified, dynamical-systems narrative of the universe that traces a continuous chain of structure formation from the Big Bang to contemporary human societies and their artificial learning systems. Rather than treating cosmology, astrophysics, geophysics, biology, cognition, and machine intelligence as disjoint domains, we view each as successive regimes of dynamics on ever-richer state spaces, stitched together by phase transitions, symmetry-breaking events, and emergent attractors. Starting from inflationary field dynamics and the growth of primordial perturbations, we describe how gravitational instability sculpts the cosmic web, how dissipative collapse in baryonic matter yields stars and planets, and how planetary-scale geochemical cycles define long-lived nonequilibrium attractors. Within these attractors, we frame the origin of life as the emergence of self-maintaining reaction networks, evolutionary biology as flow on high-dimensional genotype-phenotype-environment manifolds, and brains as adaptive dynamical systems operating near critical surfaces. Human culture and technology-including modern machine learning and artificial intelligence-are then interpreted as symbolic and institutional dynamics that implement and refine engineered learning flows which recursively reshape their own phase space. Throughout, we emphasize recurring mathematical motifs-instability, bifurcation, multiscale coupling, and constrained flows on measure-zero subsets of the accessible state space. Our aim is not to present any new cosmological or biological model, but a cross-scale, theoretical perspective: a way of reading the universe's history as the evolution of dynamics itself, culminating (so far) in biological and artificial systems capable of modeling, predicting, and deliberately perturbing their own future trajectories. --- ### XTC, A Research Platform for Optimizing AI Workload Operators Authors: Pompougnac Hugo, Guillon Christophe, Noiry Sylvain, Dutilleul Alban, Iooss Guillaume, Rastello Fabrice Published: 2025-12-18 Categories: cs.PF, cs.AI arXiv: https://arxiv.org/abs/2512.16512v1 Achieving high efficiency on AI operators demands precise control over computation and data movement. However, existing scheduling languages are locked into specific compiler ecosystems, preventing fair comparison, reuse, and evaluation across frameworks. No unified interface currently decouples scheduling specification from code generation and measurement. We introduce XTC, a platform that unifies scheduling and performance evaluation across compilers. With its common API and reproducible measurement framework, XTC enables portable experimentation and accelerates research on optimization strategies. --- ### Multi-scale Attention-Guided Intrinsic Decomposition and Rendering Pass Prediction for Facial Images Authors: Hossein Javidnia Published: 2025-12-18 Categories: cs.CV, cs.GR arXiv: https://arxiv.org/abs/2512.16511v1 Accurate intrinsic decomposition of face images under unconstrained lighting is a prerequisite for photorealistic relighting, high-fidelity digital doubles, and augmented-reality effects. This paper introduces MAGINet, a Multi-scale Attention-Guided Intrinsics Network that predicts a $512\times512$ light-normalized diffuse albedo map from a single RGB portrait. MAGINet employs hierarchical residual encoding, spatial-and-channel attention in a bottleneck, and adaptive multi-scale feature fusion in the decoder, yielding sharper albedo boundaries and stronger lighting invariance than prior U-Net variants. The initial albedo prediction is upsampled to $1024\times1024$ and refined by a lightweight three-layer CNN (RefinementNet). Conditioned on this refined albedo, a Pix2PixHD-based translator then predicts a comprehensive set of five additional physically based rendering passes: ambient occlusion, surface normal, specular reflectance, translucency, and raw diffuse colour (with residual lighting). Together with the refined albedo, these six passes form the complete intrinsic decomposition. Trained with a combination of masked-MSE, VGG, edge, and patch-LPIPS losses on the FFHQ-UV-Intrinsics dataset, the full pipeline achieves state-of-the-art performance for diffuse albedo estimation and demonstrates significantly improved fidelity for the complete rendering stack compared to prior methods. The resulting passes enable high-quality relighting and material editing of real faces. --- ### Skeleton-Snippet Contrastive Learning with Multiscale Feature Fusion for Action Localization Authors: Qiushuo Cheng, Jingjing Liu, Catherine Morgan, Alan Whone, Majid Mirmehdi Published: 2025-12-18 Categories: cs.CV arXiv: https://arxiv.org/abs/2512.16504v3 The self-supervised pretraining paradigm has achieved great success in learning 3D action representations for skeleton-based action recognition using contrastive learning. However, learning effective representations for skeleton-based temporal action localization remains challenging and underexplored. Unlike video-level {action} recognition, detecting action boundaries requires temporally sensitive features that capture subtle differences between adjacent frames where labels change. To this end, we formulate a snippet discrimination pretext task for self-supervised pretraining, which densely projects skeleton sequences into non-overlapping segments and promotes features that distinguish them across videos via contrastive learning. Additionally, we build on strong backbones of skeleton-based action recognition models by fusing intermediate features with a U-shaped module to enhance feature resolution for frame-level localization. Our approach consistently improves existing skeleton-based contrastive learning methods for action localization on BABEL across diverse subsets and evaluation protocols. We also achieve state-of-the-art transfer learning performance on PKUMMD with pretraining on NTU RGB+D and BABEL. --- ### Skeleton-Snippet Contrastive Learning with Multiscale Feature Fusion for Action Localization Authors: Qiushuo Cheng, Jingjing Liu, Catherine Morgan, Alan Whone, Majid Mirmehdi Published: 2025-12-18 Categories: cs.CV arXiv: https://arxiv.org/abs/2512.16504v1 The self-supervised pretraining paradigm has achieved great success in learning 3D action representations for skeleton-based action recognition using contrastive learning. However, learning effective representations for skeleton-based temporal action localization remains challenging and underexplored. Unlike video-level {action} recognition, detecting action boundaries requires temporally sensitive features that capture subtle differences between adjacent frames where labels change. To this end, we formulate a snippet discrimination pretext task for self-supervised pretraining, which densely projects skeleton sequences into non-overlapping segments and promotes features that distinguish them across videos via contrastive learning. Additionally, we build on strong backbones of skeleton-based action recognition models by fusing intermediate features with a U-shaped module to enhance feature resolution for frame-level localization. Our approach consistently improves existing skeleton-based contrastive learning methods for action localization on BABEL across diverse subsets and evaluation protocols. We also achieve state-of-the-art transfer learning performance on PKUMMD with pretraining on NTU RGB+D and BABEL. --- ### VenusBench-GD: A Comprehensive Multi-Platform GUI Benchmark for Diverse Grounding Tasks Authors: Beitong Zhou, Zhexiao Huang, Yuan Guo, Zhangxuan Gu, Tianyu Xia, Zichen Luo, Fei Tang, Dehan Kong, Yanyi Shang, Suling Ou, Zhenlin Guo, Changhua Meng et al. Published: 2025-12-18 Categories: cs.CV arXiv: https://arxiv.org/abs/2512.16501v1 GUI grounding is a critical component in building capable GUI agents. However, existing grounding benchmarks suffer from significant limitations: they either provide insufficient data volume and narrow domain coverage, or focus excessively on a single platform and require highly specialized domain knowledge. In this work, we present VenusBench-GD, a comprehensive, bilingual benchmark for GUI grounding that spans multiple platforms, enabling hierarchical evaluation for real-word applications. VenusBench-GD contributes as follows: (i) we introduce a large-scale, cross-platform benchmark with extensive coverage of applications, diverse UI elements, and rich annotated data, (ii) we establish a high-quality data construction pipeline for grounding tasks, achieving higher annotation accuracy than existing benchmarks, and (iii) we extend the scope of element grounding by proposing a hierarchical task taxonomy that divides grounding into basic and advanced categories, encompassing six distinct subtasks designed to evaluate models from complementary perspectives. Our experimental findings reveal critical insights: general-purpose multimodal models now match or even surpass specialized GUI models on basic grounding tasks. In contrast, advanced tasks, still favor GUI-specialized models, though they exhibit significant overfitting and poor robustness. These results underscore the necessity of comprehensive, multi-tiered evaluation frameworks. --- ### PoseMoE: Mixture-of-Experts Network for Monocular 3D Human Pose Estimation Authors: Mengyuan Liu, Jiajie Liu, Jinyan Zhang, Wenhao Li, Junsong Yuan Published: 2025-12-18 Categories: cs.CV, cs.AI arXiv: https://arxiv.org/abs/2512.16494v1 The lifting-based methods have dominated monocular 3D human pose estimation by leveraging detected 2D poses as intermediate representations. The 2D component of the final 3D human pose benefits from the detected 2D poses, whereas its depth counterpart must be estimated from scratch. The lifting-based methods encode the detected 2D pose and unknown depth in an entangled feature space, explicitly introducing depth uncertainty to the detected 2D pose, thereby limiting overall estimation accuracy. This work reveals that the depth representation is pivotal for the estimation process. Specifically, when depth is in an initial, completely unknown state, jointly encoding depth features with 2D pose features is detrimental to the estimation process. In contrast, when depth is initially refined to a more dependable state via network-based estimation, encoding it together with 2D pose information is beneficial. To address this limitation, we present a Mixture-of-Experts network for monocular 3D pose estimation named PoseMoE. Our approach introduces: (1) A mixture-of-experts network where specialized expert modules refine the well-detected 2D pose features and learn the depth features. This mixture-of-experts design disentangles the feature encoding process for 2D pose and depth, therefore reducing the explicit influence of uncertain depth features on 2D pose features. (2) A cross-expert knowledge aggregation module is proposed to aggregate cross-expert spatio-temporal contextual information. This step enhances features through bidirectional mapping between 2D pose and depth. Extensive experiments show that our proposed PoseMoE outperforms the conventional lifting-based methods on three widely used datasets: Human3.6M, MPI-INF-3DHP, and 3DPW. ---