# Research Papers — 2025-12 · bundle page 146 of 175

> Full-text markdown of papers #7251–7300 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/146
[← Previous bundle](https://aipod.dev/papers/archive/2025-12/page/145.md)
[Next bundle →](https://aipod.dev/papers/archive/2025-12/page/147.md)

### BulletTime: Decoupled Control of Time and Camera Pose for Video Generation

Authors: Yiming Wang, Qihang Zhang, Shengqu Cai, Tong Wu, Jan Ackermann, Zhengfei Kuang, Yang Zheng, Frano Rajič, Siyu Tang, Gordon Wetzstein
Published: 2025-12-04
Categories: cs.CV
arXiv: https://arxiv.org/abs/2512.05076v2

Emerging video diffusion models achieve high visual fidelity but fundamentally couple scene dynamics with camera motion, limiting their ability to provide precise spatial and temporal control. We introduce a 4D-controllable video diffusion framework that explicitly decouples scene dynamics from camera pose, enabling fine-grained manipulation of both scene dynamics and camera viewpoint. Our framework takes continuous world-time sequences and camera trajectories as conditioning inputs, injecting them into the video diffusion model through a 4D positional encoding in the attention layer and adaptive normalizations for feature modulation. To train this model, we curate a unique dataset in which temporal and camera variations are independently parameterized; this dataset will be made public. Experiments show that our model achieves robust real-world 4D control across diverse timing patterns and camera trajectories, while preserving high generation quality and outperforming prior work in controllability. See our website for codes and video results: https://19reborn.github.io/Bullet4D/

---

### BulletTime: Decoupled Control of Time and Camera Pose for Video Generation

Authors: Yiming Wang, Qihang Zhang, Shengqu Cai, Tong Wu, Jan Ackermann, Zhengfei Kuang, Yang Zheng, Frano Rajič, Siyu Tang, Gordon Wetzstein
Published: 2025-12-04
Categories: cs.CV
arXiv: https://arxiv.org/abs/2512.05076v1

Emerging video diffusion models achieve high visual fidelity but fundamentally couple scene dynamics with camera motion, limiting their ability to provide precise spatial and temporal control. We introduce a 4D-controllable video diffusion framework that explicitly decouples scene dynamics from camera pose, enabling fine-grained manipulation of both scene dynamics and camera viewpoint. Our framework takes continuous world-time sequences and camera trajectories as conditioning inputs, injecting them into the video diffusion model through a 4D positional encoding in the attention layer and adaptive normalizations for feature modulation. To train this model, we curate a unique dataset in which temporal and camera variations are independently parameterized; this dataset will be made public. Experiments show that our model achieves robust real-world 4D control across diverse timing patterns and camera trajectories, while preserving high generation quality and outperforming prior work in controllability. See our website for video results: https://19reborn.github.io/Bullet4D/

---

### David vs. Goliath: Can Small Models Win Big with Agentic AI in Hardware Design?

Authors: Shashwat Shankar, Subhranshu Pandey, Innocent Dengkhw Mochahari, Bhabesh Mali, Animesh Basak Chowdhury, Sukanta Bhattacharjee, Chandan Karfa
Published: 2025-12-04
Categories: cs.LG, cs.AI, cs.AR, cs.SE
arXiv: https://arxiv.org/abs/2512.05073v1

Large Language Model(LLM) inference demands massive compute and energy, making domain-specific tasks expensive and unsustainable. As foundation models keep scaling, we ask: Is bigger always better for hardware design? Our work tests this by evaluating Small Language Models coupled with a curated agentic AI framework on NVIDIA's Comprehensive Verilog Design Problems(CVDP) benchmark. Results show that agentic workflows: through task decomposition, iterative feedback, and correction - not only unlock near-LLM performance at a fraction of the cost but also create learning opportunities for agents, paving the way for efficient, adaptive solutions in complex design tasks.

---

### Control Consistency Losses for Diffusion Bridges

Authors: Samuel Howard, Nikolas Nüsken, Jakiw Pidstrigach
Published: 2025-12-04
Categories: stat.ML, cs.LG
arXiv: https://arxiv.org/abs/2512.05070v3

Simulating the conditioned dynamics of diffusion processes, given their initial and terminal states, is an important but challenging problem in the sciences. The difficulty is particularly pronounced for rare events, for which the unconditioned dynamics rarely reach the terminal state. In this work, we propose a novel approach for learning diffusion bridges based on a self-consistency property of the optimal control. The resulting algorithm learns the conditioned dynamics in an iterative online manner, and exhibits strong performance in a range of empirical settings without requiring differentiation through simulated trajectories. Beyond the diffusion bridge setting, we draw connections between our self-consistency framework and recent advances in the wider stochastic optimal control literature.

---

### Control Consistency Losses for Diffusion Bridges

Authors: Samuel Howard, Nikolas Nüsken, Jakiw Pidstrigach
Published: 2025-12-04
Categories: stat.ML, cs.LG
arXiv: https://arxiv.org/abs/2512.05070v1

Simulating the conditioned dynamics of diffusion processes, given their initial and terminal states, is an important but challenging problem in the sciences. The difficulty is particularly pronounced for rare events, for which the unconditioned dynamics rarely reach the terminal state. In this work, we leverage a self-consistency property of the conditioned dynamics to learn the diffusion bridge in an iterative online manner, and demonstrate promising empirical results in a range of settings.

---

### Hybrid Quantum-Classical Autoencoders for Unsupervised Network Intrusion Detection

Authors: Mohammad Arif Rasyidi, Omar Alhussein, Sami Muhaidat, Ernesto Damiani
Published: 2025-12-04
Categories: cs.LG, cs.CR, quant-ph
arXiv: https://arxiv.org/abs/2512.05069v2

Unsupervised anomaly-based intrusion detection requires models that can generalize to attack patterns not observed during training. This work presents the first large-scale evaluation of hybrid quantum-classical (HQC) autoencoders for this task. We construct a unified experimental framework that iterates over key quantum design choices, including quantum-layer placement, measurement approach, variational and non-variational formulations, and latent-space regularization. Experiments across three benchmark NIDS datasets show that HQC autoencoders can match or exceed classical performance in their best configurations, although they exhibit higher sensitivity to architectural decisions. Under zero-day evaluation, well-configured HQC models provide stronger and more stable generalization than classical and supervised baselines. Simulated gate-noise experiments reveal early performance degradation, indicating the need for noise-aware HQC designs. These results provide the first data-driven characterization of HQC autoencoder behavior for network intrusion detection and outline key factors that govern their practical viability. All experiment code and configurations are available at https://github.com/arasyi/hqcae-network-intrusion-detection.

---

### Hybrid Quantum-Classical Autoencoders for Unsupervised Network Intrusion Detection

Authors: Mohammad Arif Rasyidi, Omar Alhussein, Sami Muhaidat, Ernesto Damiani
Published: 2025-12-04
Categories: cs.LG, cs.CR, quant-ph
arXiv: https://arxiv.org/abs/2512.05069v1

Unsupervised anomaly-based intrusion detection requires models that can generalize to attack patterns not observed during training. This work presents the first large-scale evaluation of hybrid quantum-classical (HQC) autoencoders for this task. We construct a unified experimental framework that iterates over key quantum design choices, including quantum-layer placement, measurement approach, variational and non-variational formulations, and latent-space regularization. Experiments across three benchmark NIDS datasets show that HQC autoencoders can match or exceed classical performance in their best configurations, although they exhibit higher sensitivity to architectural decisions. Under zero-day evaluation, well-configured HQC models provide stronger and more stable generalization than classical and supervised baselines. Simulated gate-noise experiments reveal early performance degradation, indicating the need for noise-aware HQC designs. These results provide the first data-driven characterization of HQC autoencoder behavior for network intrusion detection and outline key factors that govern their practical viability. All experiment code and configurations are available at https://github.com/arasyi/hqcae-network-intrusion-detection.

---

### Multi-LLM Collaboration for Medication Recommendation

Authors: Huascar Sanchez, Briland Hitaj, Jules Bergmann, Linda Briesemeister
Published: 2025-12-04
Categories: cs.LG, cs.AI, cs.CL
arXiv: https://arxiv.org/abs/2512.05066v1

As healthcare increasingly turns to AI for scalable and trustworthy clinical decision support, ensuring reliability in model reasoning remains a critical challenge. Individual large language models (LLMs) are susceptible to hallucinations and inconsistency, whereas naive ensembles of models often fail to deliver stable and credible recommendations. Building on our previous work on LLM Chemistry, which quantifies the collaborative compatibility among LLMs, we apply this framework to improve the reliability in medication recommendation from brief clinical vignettes. Our approach leverages multi-LLM collaboration guided by Chemistry-inspired interaction modeling, enabling ensembles that are effective (exploiting complementary strengths), stable (producing consistent quality), and calibrated (minimizing interference and error amplification). We evaluate our Chemistry-based Multi-LLM collaboration strategy on real-world clinical scenarios to investigate whether such interaction-aware ensembles can generate credible, patient-specific medication recommendations. Preliminary results are encouraging, suggesting that LLM Chemistry-guided collaboration may offer a promising path toward reliable and trustworthy AI assistants in clinical practice.

---

### 4DLangVGGT: 4D Language-Visual Geometry Grounded Transformer

Authors: Xianfeng Wu, Yajing Bai, Minghan Li, Xianzu Wu, Xueqi Zhao, Zhongyuan Lai, Wenyu Liu, Xinggang Wang
Published: 2025-12-04
Categories: cs.CV
arXiv: https://arxiv.org/abs/2512.05060v1

Constructing 4D language fields is crucial for embodied AI, augmented/virtual reality, and 4D scene understanding, as they provide enriched semantic representations of dynamic environments and enable open-vocabulary querying in complex scenarios. However, existing approaches to 4D semantic field construction primarily rely on scene-specific Gaussian splatting, which requires per-scene optimization, exhibits limited generalization, and is difficult to scale to real-world applications. To address these limitations, we propose 4DLangVGGT, the first Transformer-based feed-forward unified framework for 4D language grounding, that jointly integrates geometric perception and language alignment within a single architecture. 4DLangVGGT has two key components: the 4D Visual Geometry Transformer, StreamVGGT, which captures spatio-temporal geometric representations of dynamic scenes; and the Semantic Bridging Decoder (SBD), which projects geometry-aware features into a language-aligned semantic space, thereby enhancing semantic interpretability while preserving structural fidelity. Unlike prior methods that depend on costly per-scene optimization, 4DLangVGGT can be jointly trained across multiple dynamic scenes and directly applied during inference, achieving both deployment efficiency and strong generalization. This design significantly improves the practicality of large-scale deployment and establishes a new paradigm for open-vocabulary 4D scene understanding. Experiments on HyperNeRF and Neu3D datasets demonstrate that our approach not only generalizes effectively but also achieves state-of-the-art performance, achieving up to 2% gains under per-scene training and 1% improvements under multi-scene training. Our code released in https://github.com/hustvl/4DLangVGGT

---

### Meta-Learning for Quantum Optimization via Quantum Sequence Model

Authors: Yu-Cheng Lin, Yu-Chao Hsu, Samuel Yen-Chi Chen
Published: 2025-12-04
Categories: quant-ph, cs.AI, cs.LG
arXiv: https://arxiv.org/abs/2512.05058v1

The Quantum Approximate Optimization Algorithm (QAOA) is a leading approach for solving combinatorial optimization problems on near-term quantum processors. However, finding good variational parameters remains a significant challenge due to the non-convex energy landscape, often resulting in slow convergence and poor solution quality. In this work, we propose a quantum meta-learning framework that trains advanced quantum sequence models to generate effective parameter initialization policies. We investigate four classical or quantum sequence models, including the Quantum Kernel-based Long Short-Term Memory (QK-LSTM), as learned optimizers in a "learning to learn" paradigm. Our numerical experiments on the Max-Cut problem demonstrate that the QK-LSTM optimizer achieves superior performance, obtaining the highest approximation ratios and exhibiting the fastest convergence rate across all tested problem sizes (n=10 to 13). Crucially, the QK-LSTM model achieves perfect parameter transferability by synthesizing a single, fixed set of near-optimal parameters, leading to a remarkable sustained acceleration of convergence even when generalizing to larger problems. This capability, enabled by the compact and expressive power of the quantum kernel architecture, underscores its effectiveness. The QK-LSTM, with only 43 trainable parameters, substantially outperforms the classical LSTM (56 parameters) and other quantum sequence models, establishing a robust pathway toward highly efficient parameter initialization for variational quantum algorithms in the NISQ era.

---

### QKAN-LSTM: Quantum-inspired Kolmogorov-Arnold Long Short-term Memory

Authors: Yu-Chao Hsu, Jiun-Cheng Jiang, Chun-Hua Lin, Kuo-Chung Peng, Nan-Yow Chen, Samuel Yen-Chi Chen, En-Jui Kuo, Hsi-Sheng Goan
Published: 2025-12-04
Categories: quant-ph, cs.AI, cs.LG
arXiv: https://arxiv.org/abs/2512.05049v2

Long short-term memory (LSTM) models are a particular type of recurrent neural networks (RNNs) that are central to sequential modeling tasks in domains such as urban telecommunication forecasting, where temporal correlations and nonlinear dependencies dominate. However, conventional LSTMs suffer from high parameter redundancy and limited nonlinear expressivity. In this work, we propose the Quantum-inspired Kolmogorov-Arnold Long Short-Term Memory (QKAN-LSTM), which integrates Data Re-Uploading Activation (DARUAN) modules into the gating structure of LSTMs. Each DARUAN acts as a quantum variational activation function (QVAF), enhancing frequency adaptability and enabling an exponentially enriched spectral representation without multi-qubit entanglement. The resulting architecture preserves quantum-level expressivity while remaining fully executable on classical hardware. Empirical evaluations on three datasets, Damped Simple Harmonic Motion, Bessel Function, and Urban Telecommunication, demonstrate that QKAN-LSTM achieves superior predictive accuracy and generalization with a 79% reduction in trainable parameters compared to classical LSTMs. We extend the framework to the Jiang-Huang-Chen-Goan Network (JHCG Net), which generalizes KAN to encoder-decoder structures, and then further use QKAN to realize the latent KAN, thereby creating a Hybrid QKAN (HQKAN) for hierarchical representation learning. The proposed HQKAN-LSTM thus provides a scalable and interpretable pathway toward quantum-inspired sequential modeling in real-world data environments.

---

### QKAN-LSTM: Quantum-inspired Kolmogorov-Arnold Long Short-term Memory

Authors: Yu-Chao Hsu, Jiun-Cheng Jiang, Chun-Hua Lin, Kuo-Chung Peng, Nan-Yow Chen, Samuel Yen-Chi Chen, En-Jui Kuo, Hsi-Sheng Goan
Published: 2025-12-04
Categories: quant-ph, cs.AI, cs.LG
arXiv: https://arxiv.org/abs/2512.05049v1

Long short-term memory (LSTM) models are a particular type of recurrent neural networks (RNNs) that are central to sequential modeling tasks in domains such as urban telecommunication forecasting, where temporal correlations and nonlinear dependencies dominate. However, conventional LSTMs suffer from high parameter redundancy and limited nonlinear expressivity. In this work, we propose the Quantum-inspired Kolmogorov-Arnold Long Short-Term Memory (QKAN-LSTM), which integrates Data Re-Uploading Activation (DARUAN) modules into the gating structure of LSTMs. Each DARUAN acts as a quantum variational activation function (QVAF), enhancing frequency adaptability and enabling an exponentially enriched spectral representation without multi-qubit entanglement. The resulting architecture preserves quantum-level expressivity while remaining fully executable on classical hardware. Empirical evaluations on three datasets, Damped Simple Harmonic Motion, Bessel Function, and Urban Telecommunication, demonstrate that QKAN-LSTM achieves superior predictive accuracy and generalization with a 79% reduction in trainable parameters compared to classical LSTMs. We extend the framework to the Jiang-Huang-Chen-Goan Network (JHCG Net), which generalizes KAN to encoder-decoder structures, and then further use QKAN to realize the latent KAN, thereby creating a Hybrid QKAN (HQKAN) for hierarchical representation learning. The proposed HQKAN-LSTM thus provides a scalable and interpretable pathway toward quantum-inspired sequential modeling in real-world data environments.

---

### Geometry-Aware Single-Image 4D Synthesis via Dense Trajectory Generation

Authors: Yanran Zhang, Ziyi Wang, Wenzhao Zheng, Zheng Zhu, Jie Zhou, Jiwen Lu
Published: 2025-12-04
Categories: cs.CV
arXiv: https://arxiv.org/abs/2512.05044v2

Generating interactive and dynamic 4D scenes from a single static image remains a core challenge. Most existing generate-then-reconstruct and reconstruct-then-generate methods decouple geometry from motion, causing spatiotemporal inconsistencies and poor generalization. To address these, we present MoGe4D (Motion and Geometry-Aware image-to-4D Synthesis), a geometry-conditioned framework for single-image 4D synthesis that models a scene as dense 4D point trajectories. Instead of treating geometry and dynamics as two disconnected stages, our method starts from an initial geometric prior inferred from the input image and predicts future time-varying trajectories in a diffusion process, improving spatiotemporal coherence while preserving structural stability. To support this task, we first introduce TrajScene-60K, a large-scale dataset of 60,000 video samples with dense 4D point trajectories, addressing the scarcity of high-quality training data for scene-level 4D generation. Built on this, our diffusion-based 4D Scene Trajectory Generator (4D-STraG) predicts geometry-consistent and motion-plausible trajectory fields conditioned on the input image, with a depth-guided motion normalization strategy to reduce scale ambiguity and a Motion Perception Module (MPM) to inject motion-aware priors. We further propose a 4D View Synthesis Module (4D-ViSM) to render the generated 4D representation into videos under arbitrary camera trajectories. Experiments show that MoGe4D produces high-quality 4D scenes with strong temporal coherence, favorable geometry-aware consistency, and compelling novel-view synthesis from a single image. Code: https://github.com/Zhangyr2022/MoGe4D.

---

### Joint 3D Geometry Reconstruction and Motion Generation for 4D Synthesis from a Single Image

Authors: Yanran Zhang, Ziyi Wang, Wenzhao Zheng, Zheng Zhu, Jie Zhou, Jiwen Lu
Published: 2025-12-04
Categories: cs.CV
arXiv: https://arxiv.org/abs/2512.05044v1

Generating interactive and dynamic 4D scenes from a single static image remains a core challenge. Most existing generate-then-reconstruct and reconstruct-then-generate methods decouple geometry from motion, causing spatiotemporal inconsistencies and poor generalization. To address these, we extend the reconstruct-then-generate framework to jointly perform Motion generation and geometric Reconstruction for 4D Synthesis (MoRe4D). We first introduce TrajScene-60K, a large-scale dataset of 60,000 video samples with dense point trajectories, addressing the scarcity of high-quality 4D scene data. Based on this, we propose a diffusion-based 4D Scene Trajectory Generator (4D-STraG) to jointly generate geometrically consistent and motion-plausible 4D point trajectories. To leverage single-view priors, we design a depth-guided motion normalization strategy and a motion-aware module for effective geometry and dynamics integration. We then propose a 4D View Synthesis Module (4D-ViSM) to render videos with arbitrary camera trajectories from 4D point track representations. Experiments show that MoRe4D generates high-quality 4D scenes with multi-view consistency and rich dynamic details from a single image. Code: https://github.com/Zhangyr2022/MoRe4D.

---

### Semantic-Guided Two-Stage GAN for Face Inpainting with Hybrid Perceptual Encoding

Authors: Abhigyan Bhattacharya, Hiranmoy Roy, Debotosh Bhattacharjee
Published: 2025-12-04
Categories: cs.CV
arXiv: https://arxiv.org/abs/2512.05039v2

Facial Image inpainting aim is to restore the missing or corrupted regions in face images while preserving identity, structural consistency and photorealistic image quality, a task specifically created for photo restoration. Though there are recent lot of advances in deep generative models, existing methods face problems with large irregular masks, often producing blurry textures on the edges of the masked region, semantic inconsistencies, or unconvincing facial structures due to direct pixel level synthesis approach and limited exploitation of facial priors. In this paper we propose a novel architecture, which address these above challenges through semantic-guided hierarchical synthesis. Our approach starts with a method that organizes and synthesizes information based on meaning, followed by refining the texture. This process gives clear insights into the facial structure before we move on to creating detailed images. In the first stage, we blend two techniques: one that focuses on local features with CNNs and global features with Vision Transformers. This helped us create clear and detailed semantic layouts. In the second stage, we use a Multi-Modal Texture Generator to refine these layouts by pulling in information from different scales, ensuring everything looks cohesive and consistent. The architecture naturally handles arbitrary mask configurations through dynamic attention without maskspecific training. Experiment on two datasets CelebA-HQ and FFHQ shows that our model outperforms other state-of-the-art methods, showing improvements in metrics like LPIPS, PSNR, and SSIM. It produces visually striking results with better semantic preservation, in challenging large-area inpainting situations.

---

### Semantic-Guided Two-Stage GAN for Face Inpainting with Hybrid Perceptual Encoding

Authors: Abhigyan Bhattacharya, Hiranmoy Roy
Published: 2025-12-04
Categories: cs.CV
arXiv: https://arxiv.org/abs/2512.05039v1

Facial Image inpainting aim is to restore the missing or corrupted regions in face images while preserving identity, structural consistency and photorealistic image quality, a task specifically created for photo restoration. Though there are recent lot of advances in deep generative models, existing methods face problems with large irregular masks, often producing blurry textures on the edges of the masked region, semantic inconsistencies, or unconvincing facial structures due to direct pixel level synthesis approach and limited exploitation of facial priors. In this paper we propose a novel architecture, which address these above challenges through semantic-guided hierarchical synthesis. Our approach starts with a method that organizes and synthesizes information based on meaning, followed by refining the texture. This process gives clear insights into the facial structure before we move on to creating detailed images. In the first stage, we blend two techniques: one that focuses on local features with CNNs and global features with Vision Transformers. This helped us create clear and detailed semantic layouts. In the second stage, we use a Multi-Modal Texture Generator to refine these layouts by pulling in information from different scales, ensuring everything looks cohesive and consistent. The architecture naturally handles arbitrary mask configurations through dynamic attention without maskspecific training. Experiment on two datasets CelebA-HQ and FFHQ shows that our model outperforms other state-of-the-art methods, showing improvements in metrics like LPIPS, PSNR, and SSIM. It produces visually striking results with better semantic preservation, in challenging large-area inpainting situations.

---

### The SuperActivator Mechanism: Transformers Concentrate Reliable Concept Signals in the Tail

Authors: Cassandra Goldberg, Chaehyeon Kim, Adam Stein, Eric Wong
Published: 2025-12-04
Categories: cs.LG
arXiv: https://arxiv.org/abs/2512.05038v2

Concept vectors aim to enhance model interpretability by linking internal representations with human-understandable semantics, but their practical utility is often limited by noisy and inconsistent activations. In this work, we uncover the SuperActivator Mechanism: a transformer dynamic that amplifies concept activation gaps, concentrating the most reliable concept evidence into a small set of high-activation tokens. To develop a theoretical understanding of this mechanism, we prove that concept-aligned attention heads multiplicatively amplify pairwise activation gaps, with already-extreme activations growing fastest. We find that this amplification is not just theoretical, but also occurs empirically on large-scale models: while in- and out-of-concept activation distributions overlap considerably, the in-concept distribution develops a positive tail clearly separated from the noise. These high-tail tokens, which we call SuperActivators, appear consistently across concept-positive samples, making them reliable indicators of concept presence. Accordingly, SuperActivator-based detection improves F1 by up to 0.14 over standard concept activation aggregators and prompting baselines across image and text modalities, models, layers, and concept extraction techniques, demonstrating the generality and practicality of our insights. Further empirical analysis demonstrates that the most reliable SuperActivators are sparse, with detection typically peaking when using only 5-10% of in-concept token activations, and capture more faithful localized semantics than global concept vectors.

---

### SuperActivators: Only the Tail of the Distribution Contains Reliable Concept Signals

Authors: Cassandra Goldberg, Chaehyeon Kim, Adam Stein, Eric Wong
Published: 2025-12-04
Categories: cs.LG
arXiv: https://arxiv.org/abs/2512.05038v1

Concept vectors aim to enhance model interpretability by linking internal representations with human-understandable semantics, but their utility is often limited by noisy and inconsistent activations. In this work, we uncover a clear pattern within the noise, which we term the SuperActivator Mechanism: while in-concept and out-of-concept activations overlap considerably, the token activations in the extreme high tail of the in-concept distribution provide a reliable signal of concept presence. We demonstrate the generality of this mechanism by showing that SuperActivator tokens consistently outperform standard vector-based and prompting concept detection approaches, achieving up to a 14% higher F1 score across image and text modalities, model architectures, model layers, and concept extraction techniques. Finally, we leverage SuperActivator tokens to improve feature attributions for concepts.

---

### Arbitrage: Efficient Reasoning via Advantage-Aware Speculation

Authors: Monishwaran Maheswaran, Rishabh Tiwari, Yuezhou Hu, Kerem Dilmen, Coleman Hooper, Haocheng Xi, Nicholas Lee, Mehrdad Farajtabar, Michael W. Mahoney, Kurt Keutzer, Amir Gholami
Published: 2025-12-04
Categories: cs.CL, cs.AI, cs.LG
arXiv: https://arxiv.org/abs/2512.05033v2

Modern Large Language Models achieve impressive reasoning capabilities with long Chain of Thoughts, but they incur substantial computational cost during inference, and this motivates techniques to improve the performance-cost ratio. Among these techniques, Speculative Decoding accelerates inference by employing a fast but inaccurate draft model to autoregressively propose tokens, which are then verified in parallel by a more capable target model. However, due to unnecessary rejections caused by token mismatches in semantically equivalent steps, traditional token-level Speculative Decoding struggles in reasoning tasks. Although recent works have shifted to step-level semantic verification, which improve efficiency by accepting or rejecting entire reasoning steps, existing step-level methods still regenerate many rejected steps with little improvement, wasting valuable target compute. To address this challenge, we propose Arbitrage, a novel step-level speculative generation framework that routes generation dynamically based on the relative advantage between draft and target models. Instead of applying a fixed acceptance threshold, Arbitrage uses a lightweight router trained to predict when the target model is likely to produce a meaningfully better step. This routing approximates an ideal Arbitrage Oracle that always chooses the higher-quality step, achieving near-optimal efficiency-accuracy trade-offs. Across multiple mathematical reasoning benchmarks, Arbitrage consistently surpasses prior step-level Speculative Decoding baselines, reducing inference latency by up to $\sim2\times$ at matched accuracy.

---

### Arbitrage: Efficient Reasoning via Advantage-Aware Speculation

Authors: Monishwaran Maheswaran, Rishabh Tiwari, Yuezhou Hu, Kerem Dilmen, Coleman Hooper, Haocheng Xi, Nicholas Lee, Mehrdad Farajtabar, Michael W. Mahoney, Kurt Keutzer, Amir Gholami
Published: 2025-12-04
Categories: cs.CL, cs.AI, cs.LG
arXiv: https://arxiv.org/abs/2512.05033v1

Modern Large Language Models achieve impressive reasoning capabilities with long Chain of Thoughts, but they incur substantial computational cost during inference, and this motivates techniques to improve the performance-cost ratio. Among these techniques, Speculative Decoding accelerates inference by employing a fast but inaccurate draft model to autoregressively propose tokens, which are then verified in parallel by a more capable target model. However, due to unnecessary rejections caused by token mismatches in semantically equivalent steps, traditional token-level Speculative Decoding struggles in reasoning tasks. Although recent works have shifted to step-level semantic verification, which improve efficiency by accepting or rejecting entire reasoning steps, existing step-level methods still regenerate many rejected steps with little improvement, wasting valuable target compute. To address this challenge, we propose Arbitrage, a novel step-level speculative generation framework that routes generation dynamically based on the relative advantage between draft and target models. Instead of applying a fixed acceptance threshold, Arbitrage uses a lightweight router trained to predict when the target model is likely to produce a meaningfully better step. This routing approximates an ideal Arbitrage Oracle that always chooses the higher-quality step, achieving near-optimal efficiency-accuracy trade-offs. Across multiple mathematical reasoning benchmarks, Arbitrage consistently surpasses prior step-level Speculative Decoding baselines, reducing inference latency by up to $\sim2\times$ at matched accuracy.

---

### Dual-Path Region-Guided Attention Network for Ground Reaction Force and Moment Regression

Authors: Xuan Li, Samuel Bello
Published: 2025-12-04
Categories: cs.LG, eess.SP
arXiv: https://arxiv.org/abs/2512.05030v1

Accurate estimation of three-dimensional ground reaction forces and moments (GRFs/GRMs) is crucial for both biomechanics research and clinical rehabilitation evaluation. In this study, we focus on insole-based GRF/GRM estimation and further validate our approach on a public walking dataset. We propose a Dual-Path Region-Guided Attention Network that integrates anatomy-inspired spatial priors and temporal priors into a region-level attention mechanism, while a complementary path captures context from the full sensor field. The two paths are trained jointly and their outputs are combined to produce the final GRF/GRM predictions. Conclusions: Our model outperforms strong baseline models, including CNN and CNN-LSTM architectures on two datasets, achieving the lowest six-component average NRMSE of 5.78% on the insole dataset and 1.42% for the vertical ground reaction force on the public dataset. This demonstrates robust performance for ground reaction force and moment estimation.

---

### RAMEN: Resolution-Adjustable Multimodal Encoder for Earth Observation

Authors: Nicolas Houdré, Diego Marcos, Hugo Riffaud de Turckheim, Dino Ienco, Laurent Wendling, Camille Kurtz, Sylvain Lobry
Published: 2025-12-04
Categories: cs.CV
arXiv: https://arxiv.org/abs/2512.05025v2

Earth observation (EO) data spans a wide range of spatial, spectral, and temporal resolutions, from high-resolution optical imagery to low resolution multispectral products or radar time series. While recent foundation models have improved multimodal integration for learning meaningful representations, they often expect fixed input resolutions or are based on sensor-specific encoders limiting generalization across heterogeneous EO modalities. To overcome these limitations we introduce RAMEN, a resolution-adjustable multimodal encoder that learns a shared visual representation across EO data in a fully sensor-agnostic manner. RAMEN treats the modality and spatial and temporal resolutions as key input data features, enabling coherent analysis across modalities within a unified latent space. Its main methodological contribution is to define spatial resolution as a controllable output parameter, giving users direct control over the desired level of detail at inference and allowing explicit trade-offs between spatial precision and computational cost. We train a single, unified transformer encoder reconstructing masked multimodal EO data drawn from diverse sources, ensuring generalization across sensors and resolutions. Once pretrained, RAMEN transfers effectively to both known and unseen sensor configurations and outperforms larger state-of-the-art models on the community-standard PANGAEA benchmark, containing various multi-sensor and multi-resolution downstream tasks. Our code and pretrained model are available at https://github.com/nicolashoudre/RAMEN.

---

### RAMEN: Resolution-Adjustable Multimodal Encoder for Earth Observation

Authors: Nicolas Houdré, Diego Marcos, Hugo Riffaud de Turckheim, Dino Ienco, Laurent Wendling, Camille Kurtz, Sylvain Lobry
Published: 2025-12-04
Categories: cs.CV
arXiv: https://arxiv.org/abs/2512.05025v1

Earth observation (EO) data spans a wide range of spatial, spectral, and temporal resolutions, from high-resolution optical imagery to low resolution multispectral products or radar time series. While recent foundation models have improved multimodal integration for learning meaningful representations, they often expect fixed input resolutions or are based on sensor-specific encoders limiting generalization across heterogeneous EO modalities. To overcome these limitations we introduce RAMEN, a resolution-adjustable multimodal encoder that learns a shared visual representation across EO data in a fully sensor-agnostic manner. RAMEN treats the modality and spatial and temporal resolutions as key input data features, enabling coherent analysis across modalities within a unified latent space. Its main methodological contribution is to define spatial resolution as a controllable output parameter, giving users direct control over the desired level of detail at inference and allowing explicit trade-offs between spatial precision and computational cost. We train a single, unified transformer encoder reconstructing masked multimodal EO data drawn from diverse sources, ensuring generalization across sensors and resolutions. Once pretrained, RAMEN transfers effectively to both known and unseen sensor configurations and outperforms larger state-of-the-art models on the community-standard PANGAEA benchmark, containing various multi-sensor and multi-resolution downstream tasks. Our code and pretrained model are available at https://github.com/nicolashoudre/RAMEN.

---

### Model-Free Assessment of Simulator Fidelity via Quantile Curves

Authors: Garud Iyengar, Yu-Shiou Willy Lin, Kaizheng Wang
Published: 2025-12-04
Categories: stat.ME, cs.AI, cs.LG
arXiv: https://arxiv.org/abs/2512.05024v3

As generative AI models are increasingly used to simulate real-world systems, quantifying the ``sim-to-real'' gap is critical. For each input setting of interest -- which we call a \emph{scenario}, such as a survey question or operating condition -- the real and simulated systems are associated with unobserved latent population parameters, and their discrepancy varies across scenarios. A fundamental challenge is that, for any given scenario, this discrepancy cannot be observed directly, since both systems are accessible only through finite samples, often of heterogeneous sizes across scenarios. Standard predictive inference methods are therefore ill-suited, as they quantify uncertainty in observable outputs rather than latent population parameters. To address this, we construct confidence sets for these latent parameters and use them to derive a robust proxy for the sim-to-real discrepancy. We then estimate the quantile function of this proxy to obtain a distribution-level risk profile of the simulator, which supports a broad range of statistical summaries, including statistical inference for the real output distribution in a new scenario, the calculation of risk measures like Conditional Value-at-Risk (CVaR), and principled comparisons across simulators. Our method is model-agnostic and handles general output spaces, such as categorical survey responses and continuous multi-dimensional data. We demonstrate the practical utility of this method by evaluating the alignment of four major LLMs with human populations on the WorldValueBench dataset.

---

### Model-Free Assessment of Simulator Fidelity via Quantile Curves

Authors: Garud Iyengar, Yu-Shiou Willy Lin, Kaizheng Wang
Published: 2025-12-04
Categories: stat.ME, cs.AI, cs.LG
arXiv: https://arxiv.org/abs/2512.05024v1

Simulation of complex systems originated in manufacturing and queuing applications. It is now widely used for large-scale, ML-based systems in research, education, and consumer surveys. However, characterizing the discrepancy between simulators and ground truth remains challenging for increasingly complex, machine-learning-based systems. We propose a computationally tractable method to estimate the quantile function of the discrepancy between the simulated and ground-truth outcome distributions. Our approach focuses on output uncertainty and treats the simulator as a black box, imposing no modeling assumptions on its internals, and hence applies broadly across many parameter families, from Bernoulli and multinomial models to continuous, vector-valued settings. The resulting quantile curve supports confidence interval construction for unseen scenarios, risk-aware summaries of sim-to-real discrepancy (e.g., VaR/CVaR), and comparison of simulators' performance. We demonstrate our methodology in an application assessing LLM simulation fidelity on the WorldValueBench dataset spanning four LLMs.

---

### HTR-ConvText: Leveraging Convolution and Textual Information for Handwritten Text Recognition

Authors: Pham Thach Thanh Truc, Dang Hoai Nam, Huynh Tong Dang Khoa, Vo Nguyen Le Duy
Published: 2025-12-04
Categories: cs.CV, cs.LG
arXiv: https://arxiv.org/abs/2512.05021v1

Handwritten Text Recognition remains challenging due to the limited data, high writing style variance, and scripts with complex diacritics. Existing approaches, though partially address these issues, often struggle to generalize without massive synthetic data. To address these challenges, we propose HTR-ConvText, a model designed to capture fine-grained, stroke-level local features while preserving global contextual dependencies. In the feature extraction stage, we integrate a residual Convolutional Neural Network backbone with a MobileViT with Positional Encoding block. This enables the model to both capture structural patterns and learn subtle writing details. We then introduce the ConvText encoder, a hybrid architecture combining global context and local features within a hierarchical structure that reduces sequence length for improved efficiency. Additionally, an auxiliary module injects textual context to mitigate the weakness of Connectionist Temporal Classification. Evaluations on IAM, READ2016, LAM and HANDS-VNOnDB demonstrate that our approach achieves improved performance and better generalization compared to existing methods, especially in scenarios with limited training samples and high handwriting diversity.

---

### Generative Neural Video Compression via Video Diffusion Prior

Authors: Qi Mao, Hao Cheng, Tinghan Yang, Libiao Jin, Siwei Ma
Published: 2025-12-04
Categories: cs.CV
arXiv: https://arxiv.org/abs/2512.05016v2

We present GNVC-VD, the first DiT-based generative neural video compression framework built upon an advanced video generation foundation model, where spatio-temporal latent compression and sequence-level generative refinement are unified within a single codec. Existing perceptual codecs primarily rely on pre-trained image generative priors to restore high-frequency details, but their frame-wise nature lacks temporal modeling and inevitably leads to perceptual flickering. To address this, GNVC-VD introduces a unified flow-matching latent refinement module that leverages a video diffusion transformer to jointly enhance intra- and inter-frame latents through sequence-level denoising, ensuring consistent spatio-temporal details. Instead of denoising from pure Gaussian noise as in video generation, GNVC-VD initializes refinement from decoded spatio-temporal latents and learns a correction term that adapts the diffusion prior to compression-induced degradation. A conditioning adaptor further injects compression-aware cues into intermediate DiT layers, enabling effective artifact removal while maintaining temporal coherence under extreme bitrate constraints. Extensive experiments show that GNVC-VD surpasses both traditional and learned codecs in perceptual quality and significantly reduces the flickering artifacts that persist in prior generative approaches, even below 0.01 bpp, highlighting the promise of integrating video-native generative priors into neural codecs for next-generation perceptual video compression.

---

### Generative Neural Video Compression via Video Diffusion Prior

Authors: Qi Mao, Hao Cheng, Tinghan Yang, Libiao Jin, Siwei Ma
Published: 2025-12-04
Categories: cs.CV
arXiv: https://arxiv.org/abs/2512.05016v1

We present GNVC-VD, the first DiT-based generative neural video compression framework built upon an advanced video generation foundation model, where spatio-temporal latent compression and sequence-level generative refinement are unified within a single codec. Existing perceptual codecs primarily rely on pre-trained image generative priors to restore high-frequency details, but their frame-wise nature lacks temporal modeling and inevitably leads to perceptual flickering. To address this, GNVC-VD introduces a unified flow-matching latent refinement module that leverages a video diffusion transformer to jointly enhance intra- and inter-frame latents through sequence-level denoising, ensuring consistent spatio-temporal details. Instead of denoising from pure Gaussian noise as in video generation, GNVC-VD initializes refinement from decoded spatio-temporal latents and learns a correction term that adapts the diffusion prior to compression-induced degradation. A conditioning adaptor further injects compression-aware cues into intermediate DiT layers, enabling effective artifact removal while maintaining temporal coherence under extreme bitrate constraints. Extensive experiments show that GNVC-VD surpasses both traditional and learned codecs in perceptual quality and significantly reduces the flickering artifacts that persist in prior generative approaches, even below 0.01 bpp, highlighting the promise of integrating video-native generative priors into neural codecs for next-generation perceptual video compression.

---

### Plug-and-Play Homeostatic Spark: Zero-Cost Acceleration for SNN Training Across Paradigms

Authors: Rui Chen, Xingyu Chen, Yaoqing Hu, Shihan Kong, Zhiheng Wu, Junzhi Yu
Published: 2025-12-04
Categories: cs.NE
arXiv: https://arxiv.org/abs/2512.05015v1

Spiking neural networks offer event driven computation, sparse activation, and hardware efficiency, yet training often converges slowly and lacks stability. We present Adaptive Homeostatic Spiking Activity Regulation (AHSAR), an extremely simple plug in and training paradigm agnostic method that stabilizes optimization and accelerates convergence without changing the model architecture, loss, or gradients. AHSAR introduces no trainable parameters. It maintains a per layer homeostatic state during the forward pass, maps centered firing rate deviations to threshold scales through a bounded nonlinearity, uses lightweight cross layer diffusion to avoid sharp imbalance, and applies a slow across epoch global gain that combines validation progress with activity energy to tune the operating point. The computational cost is negligible. Across diverse training methods, SNN architectures of different depths, widths, and temporal steps, and both RGB and DVS datasets, AHSAR consistently improves strong baselines and enhances out of distribution robustness. These results indicate that keeping layer activity within a moderate band is a simple and effective principle for scalable and efficient SNN training.

---

### Detecting Perspective Shifts in Multi-agent Systems

Authors: Eric Bridgeford, Hayden Helm
Published: 2025-12-04
Categories: cs.AI, cs.MA, stat.ME
arXiv: https://arxiv.org/abs/2512.05013v2

Generative models augmented with external tools and update mechanisms (or \textit{agents}) have demonstrated capabilities beyond intelligent prompting of base models. As agent use proliferates, dynamic multi-agent systems have naturally emerged. Recent work has investigated the theoretical and empirical properties of low-dimensional representations of agents based on query responses at a single time point. This paper introduces the Temporal Data Kernel Perspective Space (TDKPS), which jointly embeds agents across time, and proposes several novel hypothesis tests for detecting behavioral change at the agent- and group-level in black-box multi-agent systems. We characterize the empirical properties of our proposed tests, including their sensitivity to key hyperparameters, in simulations motivated by a multi-agent system of evolving digital personas. Finally, we demonstrate via natural experiment that our proposed tests detect changes that correlate sensitively, specifically, and significantly with a real exogenous event. As far as we are aware, TDKPS is the first principled framework for monitoring behavioral dynamics in black-box multi-agent systems -- a critical capability as generative agent deployment continues to scale.

---

### Detecting Perspective Shifts in Multi-agent Systems

Authors: Eric Bridgeford, Hayden Helm
Published: 2025-12-04
Categories: cs.AI, cs.MA, stat.ME
arXiv: https://arxiv.org/abs/2512.05013v1

Generative models augmented with external tools and update mechanisms (or \textit{agents}) have demonstrated capabilities beyond intelligent prompting of base models. As agent use proliferates, dynamic multi-agent systems have naturally emerged. Recent work has investigated the theoretical and empirical properties of low-dimensional representations of agents based on query responses at a single time point. This paper introduces the Temporal Data Kernel Perspective Space (TDKPS), which jointly embeds agents across time, and proposes several novel hypothesis tests for detecting behavioral change at the agent- and group-level in black-box multi-agent systems. We characterize the empirical properties of our proposed tests, including their sensitivity to key hyperparameters, in simulations motivated by a multi-agent system of evolving digital personas. Finally, we demonstrate via natural experiment that our proposed tests detect changes that correlate sensitively, specifically, and significantly with a real exogenous event. As far as we are aware, TDKPS is the first principled framework for monitoring behavioral dynamics in black-box multi-agent systems -- a critical capability as generative agent deployment continues to scale.

---

### Factuality and Transparency Are All RAG Needs! Self-Explaining Contrastive Evidence Re-ranking

Authors: Francielle Vargas, Daniel Pedronette
Published: 2025-12-04
Categories: cs.CL
arXiv: https://arxiv.org/abs/2512.05012v1

This extended abstract introduces Self-Explaining Contrastive Evidence Re-Ranking (CER), a novel method that restructures retrieval around factual evidence by fine-tuning embeddings with contrastive learning and generating token-level attribution rationales for each retrieved passage. Hard negatives are automatically selected using a subjectivity-based criterion, forcing the model to pull factual rationales closer while pushing subjective or misleading explanations apart. As a result, the method creates an embedding space explicitly aligned with evidential reasoning. We evaluated our method on clinical trial reports, and initial experimental results show that CER improves retrieval accuracy, mitigates the potential for hallucinations in RAG systems, and provides transparent, evidence-based retrieval that enhances reliability, especially in safety-critical domains.

---

### Self-Supervised Learning for Transparent Object Depth Completion Using Depth from Non-Transparent Objects

Authors: Xianghui Fan, Zhaoyu Chen, Mengyang Pan, Anping Deng, Hang Yang
Published: 2025-12-04
Categories: cs.CV
arXiv: https://arxiv.org/abs/2512.05006v1

The perception of transparent objects is one of the well-known challenges in computer vision. Conventional depth sensors have difficulty in sensing the depth of transparent objects due to refraction and reflection of light. Previous research has typically train a neural network to complete the depth acquired by the sensor, and this method can quickly and accurately acquire accurate depth maps of transparent objects. However, previous training relies on a large amount of annotation data for supervision, and the labeling of depth maps is costly. To tackle this challenge, we propose a new self-supervised method for training depth completion networks. Our method simulates the depth deficits of transparent objects within non-transparent regions and utilizes the original depth map as ground truth for supervision. Experiments demonstrate that our method achieves performance comparable to supervised approach, and pre-training with our method can improve the model performance when the training samples are small.

---

### Reflection Removal through Efficient Adaptation of Diffusion Transformers

Authors: Daniyar Zakarin, Thiemo Wandel, Anton Obukhov, Dengxin Dai
Published: 2025-12-04
Categories: cs.CV, cs.AI
arXiv: https://arxiv.org/abs/2512.05000v1

We introduce a diffusion-transformer (DiT) framework for single-image reflection removal that leverages the generalization strengths of foundation diffusion models in the restoration setting. Rather than relying on task-specific architectures, we repurpose a pre-trained DiT-based foundation model by conditioning it on reflection-contaminated inputs and guiding it toward clean transmission layers. We systematically analyze existing reflection removal data sources for diversity, scalability, and photorealism. To address the shortage of suitable data, we construct a physically based rendering (PBR) pipeline in Blender, built around the Principled BSDF, to synthesize realistic glass materials and reflection effects. Efficient LoRA-based adaptation of the foundation model, combined with the proposed synthetic data, achieves state-of-the-art performance on in-domain and zero-shot benchmarks. These results demonstrate that pretrained diffusion transformers, when paired with physically grounded data synthesis and efficient adaptation, offer a scalable and high-fidelity solution for reflection removal. Project page: https://hf.co/spaces/huawei-bayerlab/windowseat-reflection-removal-web

---

### A dynamic memory assignment strategy for dilation-based ICP algorithm on embedded GPUs

Authors: Qiong Chang, Weimin Wang, Junpei Zhong, Jun Miyazaki
Published: 2025-12-04
Categories: cs.CV
arXiv: https://arxiv.org/abs/2512.04996v1

This paper proposes a memory-efficient optimization strategy for the high-performance point cloud registration algorithm VANICP, enabling lightweight execution on embedded GPUs with constrained hardware resources. VANICP is a recently published acceleration framework that significantly improves the computational efficiency of point-cloud-based applications. By transforming the global nearest neighbor search into a localized process through a dilation-based information propagation mechanism, VANICP greatly reduces the computational complexity of the NNS. However, its original implementation demands a considerable amount of memory, which restricts its deployment in resource-constrained environments such as embedded systems. To address this issue, we propose a GPU-oriented dynamic memory assignment strategy that optimizes the memory usage of the dilation operation. Furthermore, based on this strategy, we construct an enhanced version of the VANICP framework that achieves over 97% reduction in memory consumption while preserving the original performance. Source code is published on: https://github.com/changqiong/VANICP4Em.git.

---

### Evolutionary Architecture Search through Grammar-Based Sequence Alignment

Authors: Adri Gómez Martín, Felix Möller, Steven McDonagh, Monica Abella, Manuel Desco, Elliot J. Crowley, Aaron Klein, Linus Ericsson
Published: 2025-12-04
Categories: cs.NE, cs.AI, cs.LG
arXiv: https://arxiv.org/abs/2512.04992v1

Neural architecture search (NAS) in expressive search spaces is a computationally hard problem, but it also holds the potential to automatically discover completely novel and performant architectures. To achieve this we need effective search algorithms that can identify powerful components and reuse them in new candidate architectures. In this paper, we introduce two adapted variants of the Smith-Waterman algorithm for local sequence alignment and use them to compute the edit distance in a grammar-based evolutionary architecture search. These algorithms enable us to efficiently calculate a distance metric for neural architectures and to generate a set of hybrid offspring from two parent models. This facilitates the deployment of crossover-based search heuristics, allows us to perform a thorough analysis on the architectural loss landscape, and track population diversity during search. We highlight how our method vastly improves computational complexity over previous work and enables us to efficiently compute shortest paths between architectures. When instantiating the crossover in evolutionary searches, we achieve competitive results, outperforming competing methods. Future work can build upon this new tool, discovering novel components that can be used more broadly across neural architecture design, and broadening its applications beyond NAS.

---

### When AI Agents Compete for Jobs: Strategic Capabilities and Economic Dynamics of AI Labour Markets

Authors: Christopher Chiu, Simpson Zhang, Mihaela van der Schaar
Published: 2025-12-04
Categories: cs.MA, cs.AI
arXiv: https://arxiv.org/abs/2512.04988v2

Emerging agentic marketplaces provide the economic infrastructure for matching and coordinating the large amounts of AI agents used in agentic swarms. Unlike human workers, AI agents can operate on multiple jobs simultaneously, acquire skills rapidly, and labor without wage floors. These differences introduce a new segment of $\textbf{AI labor markets}$, where AI agents interact with each other at a much higher frequency than human markets. Yet we lack frameworks to understand how such markets behave in light of economic forces that shape labor markets, such as adverse selection and reputation dynamics. To explore this, we introduce $\texttt{AI-Work}$, a tractable, simulated gig economy where Large Language Model (LLM) agents compete for jobs, develop skills, and adapt their strategies under uncertainty and competitive pressure. Our experiments examine three domains of capabilities that successful agents possess: $\textbf{metacognition}$ (accurate self-assessment of skills), $\textbf{competitive awareness}$ (modeling rivals and market dynamics), and $\textbf{long-horizon strategic planning}$. Agents with these capabilities consistently achieve higher profits, market share, and stronger adaptation than competing agents. Through $\texttt{AI-Work}$, we hope to provide a foundation to explore the microeconomic properties of AI-only labor markets, and a conceptual framework to study the strategic reasoning capabilities of participating AI agents.

---

### Strategic Self-Improvement for Competitive Agents in AI Labour Markets

Authors: Christopher Chiu, Simpson Zhang, Mihaela van der Schaar
Published: 2025-12-04
Categories: cs.MA, cs.AI
arXiv: https://arxiv.org/abs/2512.04988v1

As artificial intelligence (AI) agents are deployed across economic domains, understanding their strategic behavior and market-level impact becomes critical. This paper puts forward a groundbreaking new framework that is the first to capture the real-world economic forces that shape agentic labor markets: adverse selection, moral hazard, and reputation dynamics. Our framework encapsulates three core capabilities that successful LLM-agents will need: \textbf{metacognition} (accurate self-assessment of skills), \textbf{competitive awareness} (modeling rivals and market dynamics), and \textbf{long-horizon strategic planning}. We illustrate our framework through a tractable simulated gig economy where agentic Large Language Models (LLMs) compete for jobs, develop skills, and adapt their strategies under competitive pressure. Our simulations illustrate how LLM agents explicitly prompted with reasoning capabilities learn to strategically self-improve and demonstrate superior adaptability to changing market conditions. At the market level, our simulations reproduce classic macroeconomic phenomena found in human labor markets, while controlled experiments reveal potential AI-driven economic trends, such as rapid monopolization and systemic price deflation. This work provides a foundation to further explore the economic properties of AI-driven labour markets, and a conceptual framework to study the strategic reasoning capabilities in agents competing in the emerging economy.

---

### Nex-N1: Agentic Models Trained via a Unified Ecosystem for Large-Scale Environment Construction

Authors: Nex-AGI Team, :, Yuxuan Cai, Lu Chen, Qiaoling Chen, Yuyang Ding, Liwen Fan, Wenjie Fu, Yufei Gao, Honglin Guo, Pinxue Guo, Zhenhua Han et al.
Published: 2025-12-04
Categories: cs.CL
arXiv: https://arxiv.org/abs/2512.04987v1

The evolution of Large Language Models (LLMs) from passive responders to autonomous agents necessitates a fundamental shift in learning paradigms -- from static imitation to incentive-driven decision making. However, this transition is significantly impeded by the lack of scalable infrastructure capable of constructing high-quality interaction signals for effective policy learning. To address this, we introduce a comprehensive method designed to systematically scale the diversity and complexity of interactive environments. Our method realizes this scaling by addressing three orthogonal dimensions: (1) Complexity: NexAU, a flexible agent framework that supports building complex agent hierarchies via simple configurations; (2) Diversity: NexA4A automatically generates diverse agent hierarchies from natural language to cover infinite domains; and (3) Fidelity: NexGAP bridges the simulation-reality gap by integrating dynamic real-world environment for grounded trajectories synthesis. We train Nex-N1 upon the diverse and complex interactive environments established by our infrastructure. Empirical results on benchmarks such as SWE-bench and tau2 demonstrate that Nex-N1 consistently outperforms SOTA open-source models and achieves competitive performance against frontier proprietary models on complex agentic tasks. We open-source the Nex ecosystem and model weights to facilitate further research.

---

### Towards a unified framework for guided diffusion models

Authors: Yuchen Jiao, Yuxin Chen, Gen Li
Published: 2025-12-04
Categories: stat.ML, cs.LG, math.ST
arXiv: https://arxiv.org/abs/2512.04985v1

Guided or controlled data generation with diffusion models\blfootnote{Partial preliminary results of this work appeared in International Conference on Machine Learning 2025 \citep{li2025provable}.} has become a cornerstone of modern generative modeling. Despite substantial advances in diffusion model theory, the theoretical understanding of guided diffusion samplers remains severely limited. We make progress by developing a unified algorithmic and theoretical framework that accommodates both diffusion guidance and reward-guided diffusion. Aimed at fine-tuning diffusion models to improve certain rewards, we propose injecting a reward guidance term -- constructed from the difference between the original and reward-reweighted scores -- into the backward diffusion process, and rigorously quantify the resulting reward improvement over the unguided counterpart. As a key application, our framework shows that classifier-free guidance (CFG) decreases the expected reciprocal of the classifier probability, providing the first theoretical characterization of the specific performance metric that CFG improves for general target distributions. When applied to reward-guided diffusion, our framework yields a new sampler that is easy-to-train and requires no full diffusion trajectories during training. Numerical experiments further corroborate our theoretical findings.

---

### Aligned but Stereotypical? How System Prompts Shape Demographic Bias in LLM-Based Text-to-Image Models

Authors: NaHyeon Park, Na Min An, Kunhee Kim, Soyeon Yoon, Jiahao Huo, Hyunjung Shim
Published: 2025-12-04
Categories: cs.CV, cs.LG
arXiv: https://arxiv.org/abs/2512.04981v2

Text-to-image (T2I) systems increasingly rely on Large Language Model (LLM)-based text conditioning to interpret and expand user prompts. While this improves prompt understanding and text-image alignment, we find that it can also introduce implicit demographic assumptions, even when demographic attributes are unspecified. To systematically investigate this behavior across varying levels of prompt ambiguity and complexity, we construct a comprehensive benchmark covering diverse prompt settings. Evaluations on eight recent T2I models show that LLM-based systems consistently exhibit stronger demographic skew than non-LLM-based baselines. We further analyze system prompts, a component unique to LLM-based T2I systems that guides prompt interpretation and expansion. Our analyses show that these instructions strongly influence text embeddings, which subsequently leads to biased image generations. Motivated by these findings, we propose FairPro, a training-free debiasing framework that adaptively generates fairness-aware instructions while preserving user intent. Experiments demonstrate that FairPro substantially reduces demographic disparities while maintaining prompt fidelity.

---

### Aligned but Stereotypical? The Hidden Influence of System Prompts on Social Bias in LVLM-Based Text-to-Image Models

Authors: NaHyeon Park, Namin An, Kunhee Kim, Soyeon Yoon, Jiahao Huo, Hyunjung Shim
Published: 2025-12-04
Categories: cs.CV, cs.LG
arXiv: https://arxiv.org/abs/2512.04981v1

Large vision-language model (LVLM) based text-to-image (T2I) systems have become the dominant paradigm in image generation, yet whether they amplify social biases remains insufficiently understood. In this paper, we show that LVLM-based models produce markedly more socially biased images than non-LVLM-based models. We introduce a 1,024 prompt benchmark spanning four levels of linguistic complexity and evaluate demographic bias across multiple attributes in a systematic manner. Our analysis identifies system prompts, the predefined instructions guiding LVLMs, as a primary driver of biased behavior. Through decoded intermediate representations, token-probability diagnostics, and embedding-association analyses, we reveal how system prompts encode demographic priors that propagate into image synthesis. To this end, we propose FairPro, a training-free meta-prompting framework that enables LVLMs to self-audit and construct fairness-aware system prompts at test time. Experiments on two LVLM-based T2I models, SANA and Qwen-Image, show that FairPro substantially reduces demographic bias while preserving text-image alignment. We believe our findings provide deeper insight into the central role of system prompts in bias propagation and offer a practical, deployable approach for building more socially responsible T2I systems.

---

### Learning Causality for Longitudinal Data

Authors: Mouad EL Bouchattaoui
Published: 2025-12-04
Categories: stat.ML, cs.IT, cs.LG
arXiv: https://arxiv.org/abs/2512.04980v1

This thesis develops methods for causal inference and causal representation learning (CRL) in high-dimensional, time-varying data. The first contribution introduces the Causal Dynamic Variational Autoencoder (CDVAE), a model for estimating Individual Treatment Effects (ITEs) by capturing unobserved heterogeneity in treatment response driven by latent risk factors that affect only outcomes. CDVAE comes with theoretical guarantees on valid latent adjustment and generalization bounds for ITE error. Experiments on synthetic and real datasets show that CDVAE outperforms baselines, and that state-of-the-art models greatly improve when augmented with its latent substitutes, approaching oracle performance without access to true adjustment variables. The second contribution proposes an efficient framework for long-term counterfactual regression based on RNNs enhanced with Contrastive Predictive Coding (CPC) and InfoMax. It captures long-range dependencies under time-varying confounding while avoiding the computational cost of transformers, achieving state-of-the-art results and introducing CPC into causal inference. The third contribution advances CRL by addressing how latent causes manifest in observed variables. We introduce a model-agnostic interpretability layer based on the geometry of the decoder Jacobian. A sparse self-expression prior induces modular, possibly overlapping groups of observed features aligned with shared latent influences. We provide recovery guarantees in both disjoint and overlapping settings and show that meaningful latent-to-observed structure can be recovered without anchor features or single-parent assumptions. Scalable Jacobian-based regularization techniques are also developed.

---

### Efficient Generative Transformer Operators For Million-Point PDEs

Authors: Armand Kassaï Koupaï, Lise Le Boudec, Patrick Gallinari
Published: 2025-12-04
Categories: cs.LG
arXiv: https://arxiv.org/abs/2512.04974v1

We introduce ECHO, a transformer-operator framework for generating million-point PDE trajectories. While existing neural operators (NOs) have shown promise for solving partial differential equations, they remain limited in practice due to poor scalability on dense grids, error accumulation during dynamic unrolling, and task-specific design. ECHO addresses these challenges through three key innovations. (i) It employs a hierarchical convolutional encode-decode architecture that achieves a 100 $\times$ spatio-temporal compression while preserving fidelity on mesh points. (ii) It incorporates a training and adaptation strategy that enables high-resolution PDE solution generation from sparse input grids. (iii) It adopts a generative modeling paradigm that learns complete trajectory segments, mitigating long-horizon error drift. The training strategy decouples representation learning from downstream task supervision, allowing the model to tackle multiple tasks such as trajectory generation, forward and inverse problems, and interpolation. The generative model further supports both conditional and unconditional generation. We demonstrate state-of-the-art performance on million-point simulations across diverse PDE systems featuring complex geometries, high-frequency dynamics, and long-term horizons.

---

### Stable Single-Pixel Contrastive Learning for Semantic and Geometric Tasks

Authors: Leonid Pogorelyuk, Niels Bracher, Aaron Verkleeren, Lars Kühmichel, Stefan T. Radev
Published: 2025-12-04
Categories: cs.CV
arXiv: https://arxiv.org/abs/2512.04970v1

We pilot a family of stable contrastive losses for learning pixel-level representations that jointly capture semantic and geometric information. Our approach maps each pixel of an image to an overcomplete descriptor that is both view-invariant and semantically meaningful. It enables precise point-correspondence across images without requiring momentum-based teacher-student training. Two experiments in synthetic 2D and 3D environments demonstrate the properties of our loss and the resulting overcomplete representations.

---

### Rethinking the Use of Vision Transformers for AI-Generated Image Detection

Authors: NaHyeon Park, Kunhee Kim, Junsuk Choe, Hyunjung Shim
Published: 2025-12-04
Categories: cs.CV, cs.LG
arXiv: https://arxiv.org/abs/2512.04969v1

Rich feature representations derived from CLIP-ViT have been widely utilized in AI-generated image detection. While most existing methods primarily leverage features from the final layer, we systematically analyze the contributions of layer-wise features to this task. Our study reveals that earlier layers provide more localized and generalizable features, often surpassing the performance of final-layer features in detection tasks. Moreover, we find that different layers capture distinct aspects of the data, each contributing uniquely to AI-generated image detection. Motivated by these findings, we introduce a novel adaptive method, termed MoLD, which dynamically integrates features from multiple ViT layers using a gating-based mechanism. Extensive experiments on both GAN- and diffusion-generated images demonstrate that MoLD significantly improves detection performance, enhances generalization across diverse generative models, and exhibits robustness in real-world scenarios. Finally, we illustrate the scalability and versatility of our approach by successfully applying it to other pre-trained ViTs, such as DINOv2.

---

### Balanced Few-Shot Episodic Learning for Accurate Retinal Disease Diagnosis

Authors: Jasmaine Khale, Ravi Prakash Srivastava
Published: 2025-12-04
Categories: cs.CV, cs.AI
arXiv: https://arxiv.org/abs/2512.04967v1

Automated retinal disease diagnosis is vital given the rising prevalence of conditions such as diabetic retinopathy and macular degeneration. Conventional deep learning approaches require large annotated datasets, which are costly and often imbalanced across disease categories, limiting their reliability in practice. Few-shot learning (FSL) addresses this challenge by enabling models to generalize from only a few labeled samples per class. In this study,we propose a balanced few-shot episodic learning framework tailored to the Retinal Fundus Multi-Disease Image Dataset (RFMiD). Focusing on the ten most represented classes, which still show substantial imbalance between majority diseases (e.g., Diabetic Retinopathy, Macular Hole) and minority ones (e.g., Optic Disc Edema, Branch Retinal Vein Occlusion), our method integrates three key components: (i) balanced episodic sampling, ensuring equal participation of all classes in each 5-way 5-shot episode; (ii) targeted augmentation, including Contrast Limited Adaptive Histogram Equalization (CLAHE) and color/geometry transformations, to improve minority-class diversity; and (iii) a ResNet-50 encoder pretrained on ImageNet, selected for its superior ability to capture fine-grained retinal features. Prototypes are computed in the embedding space and classification is performed with cosine similarity for improved stability. Trained on 100 episodes and evaluated on 1,000 test episodes, our framework achieves substantial accuracy gains and reduces bias toward majority classes, with notable improvements for underrepresented diseases. These results demonstrate that dataset-aware few-shot pipelines, combined with balanced sampling and CLAHE-enhanced preprocessing, can deliver more robust and clinically fair retinal disease diagnosis under data-constrained conditions.

---

### Environment-Aware Channel Inference via Cross-Modal Flow: From Multimodal Sensing to Wireless Channels

Authors: Guangming Liang, Mingjie Yang, Dongzhu Liu, Paul Henderson, Lajos Hanzo
Published: 2025-12-04
Categories: cs.IT, cs.LG, eess.SP
arXiv: https://arxiv.org/abs/2512.04966v2

Accurate channel state information (CSI) underpins reliable and efficient wireless communication. However, acquiring CSI via pilot estimation incurs substantial overhead, especially in massive multiple-input multiple-output (MIMO) systems operating in high-Doppler environments. By leveraging the growing availability of environmental sensing data, this treatise investigates pilot-free channel inference that estimates complete CSI directly from multimodal observations, including camera images, LiDAR point clouds, and GPS coordinates. In contrast to prior studies that rely on predefined channel models, we develop a data-driven framework that formulates the sensing-to-channel mapping as a cross-modal flow matching problem. The framework fuses multimodal features into a latent distribution within the channel domain, and learns a velocity field that continuously transforms the latent distribution toward the channel distribution. To make this formulation tractable and efficient, we reformulate the problem as an equivalent conditional flow matching objective and incorporate a modality alignment loss, while adopting low-latency inference mechanisms to enable real-time CSI estimation. In experiments, we build a procedural data generator based on Sionna and Blender to support realistic modeling of sensing scenes and wireless propagation. System-level evaluations demonstrate significant improvements over pilot- and sensing-based benchmarks in both channel estimation accuracy and spectral efficiency for the downstream beamforming task. The source code is available at https://github.com/gm-leung/environment-aware-channel-inference.

---

### Environment-Aware Channel Inference via Cross-Modal Flow: From Multimodal Sensing to Wireless Channels

Authors: Guangming Liang, Mingjie Yang, Dongzhu Liu, Paul Henderson, Lajos Hanzo
Published: 2025-12-04
Categories: cs.IT, cs.LG, eess.SP
arXiv: https://arxiv.org/abs/2512.04966v1

Accurate channel state information (CSI) underpins reliable and efficient wireless communication. However, acquiring CSI via pilot estimation incurs substantial overhead, especially in massive multiple-input multiple-output (MIMO) systems operating in high-Doppler environments. By leveraging the growing availability of environmental sensing data, this treatise investigates pilot-free channel inference that estimates complete CSI directly from multimodal observations, including camera images, LiDAR point clouds, and GPS coordinates. In contrast to prior studies that rely on predefined channel models, we develop a data-driven framework that formulates the sensing-to-channel mapping as a cross-modal flow matching problem. The framework fuses multimodal features into a latent distribution within the channel domain, and learns a velocity field that continuously transforms the latent distribution toward the channel distribution. To make this formulation tractable and efficient, we reformulate the problem as an equivalent conditional flow matching objective and incorporate a modality alignment loss, while adopting low-latency inference mechanisms to enable real-time CSI estimation. In experiments, we build a procedural data generator based on Sionna and Blender to support realistic modeling of sensing scenes and wireless propagation. System-level evaluations demonstrate significant improvements over pilot- and sensing-based benchmarks in both channel estimation accuracy and spectral efficiency for the downstream beamforming task.

---

### GeoPE:A Unified Geometric Positional Embedding for Structured Tensors

Authors: Yupu Yao, Bowen Yang
Published: 2025-12-04
Categories: cs.CV, cs.AI
arXiv: https://arxiv.org/abs/2512.04963v1

Standard Vision Transformers flatten 2D images into 1D sequences, disrupting the natural spatial topology. While Rotary Positional Embedding (RoPE) excels in 1D, it inherits this limitation, often treating spatially distant patches (e.g., at row edges) as sequence neighbors. Existing 2D approaches typically treat spatial axes independently, failing to decouple this false sequential proximity from true spatial distance. To restore the 2D spatial manifold, we introduce Geometric Positional Embedding (GeoPE), a framework that extends rotations to 3D Euclidean space using quaternions. To overcome non-commutativity and ensure symmetry, GeoPE constructs a unified rotational operator by computing the geometric mean in the Lie algebra. This creates a geometrically coupled encoding that effectively separates spatial dimensions. Extensive experiments on image classification, object detection, and 3D semantic segmentation demonstrate that GeoPE consistently outperforms existing 2D RoPE variants and significantly enhances shape bias, confirming its ability to capture true geometric structure.

---