# Research Papers — 2026-01 · bundle page 165 of 189

> Full-text markdown of papers #8201–8250 of 9403 published in 2026-01, newest first. AI-readable mirrors of the full 9403-paper 2026-01 archive (50 papers per file).

Canonical page: https://aipod.dev/papers/archive/2026-01/page/165
[← Previous bundle](https://aipod.dev/papers/archive/2026-01/page/164.md)
[Next bundle →](https://aipod.dev/papers/archive/2026-01/page/166.md)

### AnatomiX, an Anatomy-Aware Grounded Multimodal Large Language Model for Chest X-Ray Interpretation

Authors: Anees Ur Rehman Hashmi, Numan Saeed, Christoph Lippert
Published: 2026-01-06
Categories: cs.CV, cs.AI, cs.LG
arXiv: https://arxiv.org/abs/2601.03191v3

Multimodal medical large language models have shown substantial progress in chest X-ray interpretation but continue to face challenges in spatial reasoning and anatomical understanding. Although existing grounding techniques improve overall performance, they often fail to establish a true anatomical correspondence, resulting in incorrect anatomical understanding in the medical domain. To address this gap, we introduce AnatomiX, a multitask multimodal large language model for anatomically grounded chest X-ray interpretation. Inspired by the radiological workflow, AnatomiX adopts a two stage approach: first, it identifies anatomical structures and extracts their features, and then leverages a large language model to perform diverse downstream tasks such as phrase grounding, report generation, visual question answering, and image understanding. Extensive experiments across multiple benchmarks demonstrate that AnatomiX achieves superior anatomical reasoning and delivers over 25% improvement in performance on anatomy grounding, phrase grounding, grounded diagnosis and grounded captioning tasks compared to existing approaches. Code and pretrained model are available at https://aneesurhashmi.github.io/anatomix

---

### AnatomiX, an Anatomy-Aware Grounded Multimodal Large Language Model for Chest X-Ray Interpretation

Authors: Anees Ur Rehman Hashmi, Numan Saeed, Christoph Lippert
Published: 2026-01-06
Categories: cs.CV, cs.AI, cs.LG
arXiv: https://arxiv.org/abs/2601.03191v1

Multimodal medical large language models have shown impressive progress in chest X-ray interpretation but continue to face challenges in spatial reasoning and anatomical understanding. Although existing grounding techniques improve overall performance, they often fail to establish a true anatomical correspondence, resulting in incorrect anatomical understanding in the medical domain. To address this gap, we introduce AnatomiX, a multitask multimodal large language model explicitly designed for anatomically grounded chest X-ray interpretation. Inspired by the radiological workflow, AnatomiX adopts a two stage approach: first, it identifies anatomical structures and extracts their features, and then leverages a large language model to perform diverse downstream tasks such as phrase grounding, report generation, visual question answering, and image understanding. Extensive experiments across multiple benchmarks demonstrate that AnatomiX achieves superior anatomical reasoning and delivers over 25% improvement in performance on anatomy grounding, phrase grounding, grounded diagnosis and grounded captioning tasks compared to existing approaches. Code and pretrained model are available at https://github.com/aneesurhashmi/anatomix

---

### Maximizing Local Entropy Where It Matters: Prefix-Aware Localized LLM Unlearning

Authors: Naixin Zhai, Pengyang Shao, Binbin Zheng, Yonghui Yang, Fei Shen, Long Bai, Xun Yang
Published: 2026-01-06
Categories: cs.CL
arXiv: https://arxiv.org/abs/2601.03190v4

Machine unlearning aims to forget sensitive knowledge from Large Language Models (LLMs) while maintaining general utility. However, existing approaches typically treat all tokens in a response indiscriminately and enforce uncertainty over the entire vocabulary. This global treatment results in unnecessary utility degradation and extends optimization to content-agnostic regions. To address these limitations, we propose PALU (Prefix-Aware Localized Unlearning), a framework driven by a local entropy maximization objective across both temporal and vocabulary dimensions. PALU reveals that (i) suppressing the sensitive prefix alone is sufficient to sever the causal generation link, and (ii) flattening only the top-$k$ logits is adequate to maximize uncertainty in the critical subspace. These findings allow PALU to alleviate redundant optimization across the full vocabulary and parameter space while minimizing collateral damage to general model performance. Comprehensive evaluations validate that PALU achieves superior forgetting efficacy and utility preservation compared to state-of-the-art baselines. Our code is available at https://github.com/nxZhai/PALU.

---

### Maximizing Local Entropy Where It Matters: Prefix-Aware Localized LLM Unlearning

Authors: Naixin Zhai, Pengyang Shao, Binbin Zheng, Fei Shen, Long Bai, Xun Yang
Published: 2026-01-06
Categories: cs.CL
arXiv: https://arxiv.org/abs/2601.03190v1

Machine unlearning aims to forget sensitive knowledge from Large Language Models (LLMs) while maintaining general utility. However, existing approaches typically treat all tokens in a response indiscriminately and enforce uncertainty over the entire vocabulary. This global treatment results in unnecessary utility degradation and extends optimization to content-agnostic regions. To address these limitations, we propose PALU (Prefix-Aware Localized Unlearning), a framework driven by a local entropy maximization objective across both temporal and vocabulary dimensions. PALU reveals that (i) suppressing the sensitive prefix alone is sufficient to sever the causal generation link, and (ii) flattening only the top-$k$ logits is adequate to maximize uncertainty in the critical subspace. These findings allow PALU to avoid redundant optimization across the full vocabulary and parameter space while minimizing collateral damage to general model performance. Extensive experiments validate that PALU achieves superior forgetting efficacy and utility preservation compared to state-of-the-art baselines.

---

### Decentralized Autoregressive Generation

Authors: Stepan Maschan, Haoxuan Qu, Jun Liu
Published: 2026-01-06
Categories: cs.LG, cs.AI
arXiv: https://arxiv.org/abs/2601.03184v3

The decentralization of autoregressive generation has attracted considerable attention in recent years as a solution to scaling bottlenecks. However, despite promising empirical results, this paradigm currently lacks rigorous theoretical justification. In this work, we formally establish the theoretical equivalence between decentralized and centralized training. To achieve this, we adapt the Discrete Flow Matching framework for autoregressive generation, leveraging its inherent properties to demonstrate that global models naturally decompose into independent experts. Finally, we conduct extensive experiments across diverse multimodal benchmarks, empirically validating that decentralized training maintains competitive parity with standard centralized architectures.

---

### Decentralized Autoregressive Generation

Authors: Stepan Maschan, Haoxuan Qu, Jun Liu
Published: 2026-01-06
Categories: cs.LG, cs.AI
arXiv: https://arxiv.org/abs/2601.03184v1

We present a theoretical analysis of decentralization of autoregressive generation. We define the Decentralized Discrete Flow Matching objective, by expressing probability generating velocity as a linear combination of expert flows. We also conduct experiments demonstrat- ing the equivalence between decentralized and centralized training settings for multimodal language models across diverse set of benchmarks. Specifically, we compare two distinct paradigms: LLaVA and InternVL 2.5-1B, which uses a fixed CLIP vision encoder and per- forms full-parameter fine-tuning (ViT+MLP+LLM) during the instruction tuning stage.

---

### Multi-Modal Data-Enhanced Foundation Models for Prediction and Control in Wireless Networks: A Survey

Authors: Han Zhang, Mohammad Farzanullah, Mohammad Ghassemi, Akram Bin Sediq, Ali Afana, Melike Erol-Kantarci
Published: 2026-01-06
Categories: cs.NI, cs.AI, cs.CL, cs.CV
arXiv: https://arxiv.org/abs/2601.03181v1

Foundation models (FMs) are recognized as a transformative breakthrough that has started to reshape the future of artificial intelligence (AI) across both academia and industry. The integration of FMs into wireless networks is expected to enable the development of general-purpose AI agents capable of handling diverse network management requests and highly complex wireless-related tasks involving multi-modal data. Inspired by these ideas, this work discusses the utilization of FMs, especially multi-modal FMs in wireless networks. We focus on two important types of tasks in wireless network management: prediction tasks and control tasks. In particular, we first discuss FMs-enabled multi-modal contextual information understanding in wireless networks. Then, we explain how FMs can be applied to prediction and control tasks, respectively. Following this, we introduce the development of wireless-specific FMs from two perspectives: available datasets for development and the methodologies used. Finally, we conclude with a discussion of the challenges and future directions for FM-enhanced wireless networks.

---

### DiffBench Meets DiffAgent: End-to-End LLM-Driven Diffusion Acceleration Code Generation

Authors: Jiajun jiao, Haowei Zhu, Puyuan Yang, Jianghui Wang, Ji Liu, Ziqiong Liu, Dong Li, Yuejian Fang, Junhai Yong, Bin Wang, Emad Barsoum
Published: 2026-01-06
Categories: cs.CV
arXiv: https://arxiv.org/abs/2601.03178v1

Diffusion models have achieved remarkable success in image and video generation. However, their inherently multiple step inference process imposes substantial computational overhead, hindering real-world deployment. Accelerating diffusion models is therefore essential, yet determining how to combine multiple model acceleration techniques remains a significant challenge. To address this issue, we introduce a framework driven by large language models (LLMs) for automated acceleration code generation and evaluation. First, we present DiffBench, a comprehensive benchmark that implements a three stage automated evaluation pipeline across diverse diffusion architectures, optimization combinations and deployment scenarios. Second, we propose DiffAgent, an agent that generates optimal acceleration strategies and codes for arbitrary diffusion models. DiffAgent employs a closed-loop workflow in which a planning component and a debugging component iteratively refine the output of a code generation component, while a genetic algorithm extracts performance feedback from the execution environment to guide subsequent code refinements. We provide a detailed explanation of the DiffBench construction and the design principles underlying DiffAgent. Extensive experiments show that DiffBench offers a thorough evaluation of generated codes and that DiffAgent significantly outperforms existing LLMs in producing effective diffusion acceleration strategies.

---

### Predicting Time Pressure of Powered Two-Wheeler Riders for Proactive Safety Interventions

Authors: Sumit S. Shevtekar, Chandresh K. Maurya, Gourab Sil
Published: 2026-01-06
Categories: cs.LG, cs.HC
arXiv: https://arxiv.org/abs/2601.03173v4

Time pressure critically influences risky maneuvers and crash proneness among powered two-wheeler riders, yet its prediction remains underexplored in intelligent transportation systems. To address this gap, we propose MotoTimePressure (MTPS), a deep learning model combining convolutional preprocessing, dual-stage temporal attention, and Squeeze-and-Excitation feature recalibration, achieving 91.53% accuracy and 98.93% ROC AUC, outperforming six baselines, with only 172K parameters, 0.66 MB model size, and 0.21 ms inference on CPU. To validate and benchmark MTPS, we present a dataset of 129,209 feature windows from 153 simulator sessions by 51 experienced male PTW riders under No, Low, and High Time Pressure conditions. Each sequence captures 63 features spanning vehicle kinematics, control inputs, behavioral violations, and environmental context. Our empirical analysis shows High Time Pressure induces 48% higher speeds, 36.4% greater speed variability, 58% more risky turns at intersections, 36% more sudden braking, and 50% higher rear brake forces versus No Time Pressure. Since time pressure cannot be directly measured in real time, we demonstrate its utility in collision prediction and threshold determination. Using MTPS-predicted time pressure as a feature improves collision risk accuracy for both Informer (91.25% to 93.51%) and TimesNet (92.10% to 93.90%), approaching oracle performance (93.72% and 94.06%, respectively). Thresholded time pressure states capture rider cognitive stress and enable proactive ITS interventions, including adaptive alerts, haptic feedback, V2I signaling, and speed guidance, supporting safer two-wheeler mobility under the Safe System Approach.

---

### Predicting Time Pressure of Powered Two-Wheeler Riders for Proactive Safety Interventions

Authors: Sumit S. Shevtekar, Chandresh K. Maurya, Gourab Sil, Subasish Das
Published: 2026-01-06
Categories: cs.LG, cs.HC
arXiv: https://arxiv.org/abs/2601.03173v1

Time pressure critically influences risky maneuvers and crash proneness among powered two-wheeler riders, yet its prediction remains underexplored in intelligent transportation systems. We present a large-scale dataset of 129,000+ labeled multivariate time-series sequences from 153 rides by 51 participants under No, Low, and High Time Pressure conditions. Each sequence captures 63 features spanning vehicle kinematics, control inputs, behavioral violations, and environmental context. Our empirical analysis shows High Time Pressure induces 48% higher speeds, 36.4% greater speed variability, 58% more risky turns at intersections, 36% more sudden braking, and 50% higher rear brake forces versus No Time Pressure. To benchmark this dataset, we propose MotoTimePressure, a deep learning model combining convolutional preprocessing, dual-stage temporal attention, and Squeeze-and-Excitation feature recalibration, achieving 91.53% accuracy and 98.93% ROC AUC, outperforming eight baselines. Since time pressure cannot be directly measured in real time, we demonstrate its utility in collision prediction and threshold determination. Using MTPS-predicted time pressure as features, improves Informer-based collision risk accuracy from 91.25% to 93.51%, approaching oracle performance (93.72%). Thresholded time pressure states capture rider cognitive stress and enable proactive ITS interventions, including adaptive alerts, haptic feedback, V2I signaling, and speed guidance, supporting safer two-wheeler mobility under the Safe System Approach.

---

### Can Embedding Similarity Predict Cross-Lingual Transfer? A Systematic Study on African Languages

Authors: Tewodros Kederalah Idris, Prasenjit Mitra, Roald Eiselen
Published: 2026-01-06
Categories: cs.CL, cs.LG
arXiv: https://arxiv.org/abs/2601.03168v1

Cross-lingual transfer is essential for building NLP systems for low-resource African languages, but practitioners lack reliable methods for selecting source languages. We systematically evaluate five embedding similarity metrics across 816 transfer experiments spanning three NLP tasks, three African-centric multilingual models, and 12 languages from four language families. We find that cosine gap and retrieval-based metrics (P@1, CSLS) reliably predict transfer success ($ρ= 0.4-0.6$), while CKA shows negligible predictive power ($ρ\approx 0.1$). Critically, correlation signs reverse when pooling across models (Simpson's Paradox), so practitioners must validate per-model. Embedding metrics achieve comparable predictive power to URIEL linguistic typology. Our results provide concrete guidance for source language selection and highlight the importance of model-specific analysis.

---

### Dynamic Hyperparameter Importance for Efficient Multi-Objective Optimization

Authors: Daphne Theodorakopoulos, Marcel Wever, Marius Lindauer
Published: 2026-01-06
Categories: cs.LG
arXiv: https://arxiv.org/abs/2601.03166v2

Choosing a suitable ML model is a complex task that can depend on several objectives, e.g., accuracy, fairness, or energy consumption. In practice, this requires trading off multiple, often competing, objectives through multi-objective optimization (MOO). However, existing MOO methods typically treat all hyperparameters as equally important, disregarding that hyperparameter importance (HPI) can vary significantly across objectives. We propose a novel dynamic optimization approach that prioritizes the most influential hyperparameters based on varying objective trade-offs during the search, thereby accelerating empirical convergence. We advance prior work on HPI for MOO from post-analysis to direct, dynamic integration within the optimization, using the recent HPI method HyperSHAP. For this, we leverage the objective weightings naturally produced by the MOO algorithm ParEGO and reduce the configuration space by fixing the unimportant hyperparameters, allowing the search to focus on the important ones. Eventually, we evaluate our method on diverse tasks from PyMOO and YAHPO-Gym. For HPO, integrating HPI yields up to 24% improvement in final Pareto front quality, while on synthetic data, integrating HPI achieves 2x better final results.

---

### Dynamic Hyperparameter Importance for Efficient Multi-Objective Optimization

Authors: Daphne Theodorakopoulos, Marcel Wever, Marius Lindauer
Published: 2026-01-06
Categories: cs.LG
arXiv: https://arxiv.org/abs/2601.03166v1

Choosing a suitable ML model is a complex task that can depend on several objectives, e.g., accuracy, model size, fairness, inference time, or energy consumption. In practice, this requires trading off multiple, often competing, objectives through multi-objective optimization (MOO). However, existing MOO methods typically treat all hyperparameters as equally important, overlooking that hyperparameter importance (HPI) can vary significantly depending on the trade-off between objectives. We propose a novel dynamic optimization approach that prioritizes the most influential hyperparameters based on varying objective trade-offs during the search process, which accelerates empirical convergence and leads to better solutions. Building on prior work on HPI for MOO post-analysis, we now integrate HPI, calculated with HyperSHAP, into the optimization. For this, we leverage the objective weightings naturally produced by the MOO algorithm ParEGO and adapt the configuration space by fixing the unimportant hyperparameters, allowing the search to focus on the important ones. Eventually, we validate our method with diverse tasks from PyMOO and YAHPO-Gym. Empirical results demonstrate improvements in convergence speed and Pareto front quality compared to baselines.

---

### WebAnchor: Anchoring Agent Planning to Stabilize Long-Horizon Web Reasoning

Authors: Xinmiao Yu, Liwen Zhang, Xiaocheng Feng, Yong Jiang, Bing Qin, Pengjun Xie, Jingren Zhou
Published: 2026-01-06
Categories: cs.CL
arXiv: https://arxiv.org/abs/2601.03164v2

Large Language Model(LLM)-based agents have shown strong capabilities in web information seeking, with reinforcement learning (RL) becoming a key optimization paradigm. However, planning remains a bottleneck, as existing methods struggle with long-horizon strategies. Our analysis reveals a critical phenomenon, plan anchor, where the first reasoning step disproportionately impacts downstream behavior in long-horizon web reasoning tasks. Current RL algorithms, fail to account for this by uniformly distributing rewards across the trajectory. To address this, we propose Anchor-GRPO, a two-stage RL framework that decouples planning and execution. In Stage 1, the agent optimizes its first-step planning using fine-grained rubrics derived from self-play experiences and human calibration. In Stage 2, execution is aligned with the initial plan through sparse rewards, ensuring stable and efficient tool usage. We evaluate Anchor-GRPO on four benchmarks: BrowseComp, BrowseComp-Zh, GAIA, and XBench-DeepSearch. Across models from 3B to 30B, Anchor-GRPO outperforms baseline GRPO and First-step GRPO, improving task success and tool efficiency. Notably, WebAnchor-30B achieves 46.0% pass@1 on BrowseComp and 76.4% on GAIA. Anchor-GRPO also demonstrates strong scalability, getting higher accuracy as model size and context length increase.

---

### WebAnchor: Anchoring Agent Planning to Stabilize Long-Horizon Web Reasoning

Authors: Yu Xinmiao, Zhang Liwen, Feng Xiaocheng, Jiang Yong, Qin Bing, Xie Pengjun, Zhou Jingren
Published: 2026-01-06
Categories: cs.CL
arXiv: https://arxiv.org/abs/2601.03164v1

Large Language Model(LLM)-based agents have shown strong capabilities in web information seeking, with reinforcement learning (RL) becoming a key optimization paradigm. However, planning remains a bottleneck, as existing methods struggle with long-horizon strategies. Our analysis reveals a critical phenomenon, plan anchor, where the first reasoning step disproportionately impacts downstream behavior in long-horizon web reasoning tasks. Current RL algorithms, fail to account for this by uniformly distributing rewards across the trajectory. To address this, we propose Anchor-GRPO, a two-stage RL framework that decouples planning and execution. In Stage 1, the agent optimizes its first-step planning using fine-grained rubrics derived from self-play experiences and human calibration. In Stage 2, execution is aligned with the initial plan through sparse rewards, ensuring stable and efficient tool usage. We evaluate Anchor-GRPO on four benchmarks: BrowseComp, BrowseComp-Zh, GAIA, and XBench-DeepSearch. Across models from 3B to 30B, Anchor-GRPO outperforms baseline GRPO and First-step GRPO, improving task success and tool efficiency. Notably, WebAnchor-30B achieves 46.0% pass@1 on BrowseComp and 76.4% on GAIA. Anchor-GRPO also demonstrates strong scalability, getting higher accuracy as model size and context length increase.

---

### LSP-DETR: Efficient and Scalable Nuclei Segmentation in Whole-Slide Images

Authors: Matěj Pekár, Vít Musil, Rudolf Nenutil, Petr Holub, Tomáš Brázdil
Published: 2026-01-06
Categories: cs.CV
arXiv: https://arxiv.org/abs/2601.03163v2

Background and Objective: Precise and scalable instance segmentation of cell nuclei is a fundamental prerequisite for computational pathology, yet gigapixel whole-slide images (WSIs) pose significant computational challenges. While patch-based processing is standard during training, existing methods are often limited to small tile sizes during inference due to architectural bottlenecks or reliance on computationally expensive post-processing for instance separation. We introduce a faster, scalable, and end-to-end framework capable of processing large-scale image tiles while accurately modeling biologically realistic overlapping nuclei. Methods: We propose LSP-DETR (Local Star Polygon DEtection TRansformer). The model represents nuclei as star-convex polygons and employs a lightweight transformer with linear complexity, enabling the processing of high-resolution images in a single forward pass. A novel radial distance loss accommodates annotation uncertainty, allowing the segmentation of overlapping nuclei to emerge naturally without explicit overlap labels. Results: LSP-DETR achieves state-of-the-art efficiency, with an inference time of 0.45 s/mm^2, a 3.2x speedup over StarDist, the next-fastest method. On PanNuke, the model achieves competitive accuracy (67.5 bPQ), while yielding an F$_1$-score of 0.964 in polygon overlap when evaluated against consensus annotations from two expert pathologists. Furthermore, it outperforms larger models such as LKCell in generalization robustness, reaching an F1-score of 85.0 on MoNuSeg. Conclusions: LSP-DETR bridges the gap between high-fidelity segmentation and practical clinical requirements by eliminating heuristic post-processing. By providing a scalable, linear-complexity solution that naturally handles overlaps between nuclei, this framework sets a new direction for efficient high-throughput WSI analysis in digital pathology.

---

### LSP-DETR: Efficient and Scalable Nuclei Segmentation in Whole Slide Images

Authors: Matěj Pekár, Vít Musil, Rudolf Nenutil, Petr Holub, Tomáš Brázdil
Published: 2026-01-06
Categories: cs.CV
arXiv: https://arxiv.org/abs/2601.03163v1

Precise and scalable instance segmentation of cell nuclei is essential for computational pathology, yet gigapixel Whole-Slide Images pose major computational challenges. Existing approaches rely on patch-based processing and costly post-processing for instance separation, sacrificing context and efficiency. We introduce LSP-DETR (Local Star Polygon DEtection TRansformer), a fully end-to-end framework that uses a lightweight transformer with linear complexity to process substantially larger images without additional computational cost. Nuclei are represented as star-convex polygons, and a novel radial distance loss function allows the segmentation of overlapping nuclei to emerge naturally, without requiring explicit overlap annotations or handcrafted post-processing. Evaluations on PanNuke and MoNuSeg show strong generalization across tissues and state-of-the-art efficiency, with LSP-DETR being over five times faster than the next-fastest leading method. Code and models are available at https://github.com/RationAI/lsp-detr.

---

### On the Convergence Behavior of Preconditioned Gradient Descent Toward the Rich Learning Regime

Authors: Shuai Jiang, Alexey Voronin, Eric Cyr, Ben Southworth
Published: 2026-01-06
Categories: cs.LG
arXiv: https://arxiv.org/abs/2601.03162v2

Spectral bias, the tendency of neural networks to learn low frequencies first, can be both a blessing and a curse. While it enhances the generalization capabilities by suppressing high-frequency noise, it can be a limitation in scientific tasks that require capturing fine-scale structures. The delayed generalization phenomenon known as grokking is another barrier to rapid training of neural networks. Grokking has been hypothesized to arise as learning transitions from the NTK to the feature-rich regime. This paper explores the impact of preconditioned gradient descent (PGD), such as Gauss-Newton, on spectral bias and grokking phenomena. We demonstrate through theoretical and empirical results how PGD can mitigate issues associated with spectral bias. Additionally, building on the rich learning regime grokking hypothesis, we study how PGD can be used to reduce delays associated with grokking. Our conjecture is that PGD, without the impediment of spectral bias, enables uniform exploration of the parameter space in the NTK regime. Our experimental results confirm this prediction, providing strong evidence that grokking represents a transitional behavior between the lazy regime characterized by the NTK and the rich regime. These findings deepen our understanding of the interplay between optimization dynamics, spectral bias, and the phases of neural network learning.

---

### On the Convergence Behavior of Preconditioned Gradient Descent Toward the Rich Learning Regime

Authors: Shuai Jiang, Alexey Voronin, Eric Cyr, Ben Southworth
Published: 2026-01-06
Categories: cs.LG
arXiv: https://arxiv.org/abs/2601.03162v1

Spectral bias, the tendency of neural networks to learn low frequencies first, can be both a blessing and a curse. While it enhances the generalization capabilities by suppressing high-frequency noise, it can be a limitation in scientific tasks that require capturing fine-scale structures. The delayed generalization phenomenon known as grokking is another barrier to rapid training of neural networks. Grokking has been hypothesized to arise as learning transitions from the NTK to the feature-rich regime. This paper explores the impact of preconditioned gradient descent (PGD), such as Gauss-Newton, on spectral bias and grokking phenomena. We demonstrate through theoretical and empirical results how PGD can mitigate issues associated with spectral bias. Additionally, building on the rich learning regime grokking hypothesis, we study how PGD can be used to reduce delays associated with grokking. Our conjecture is that PGD, without the impediment of spectral bias, enables uniform exploration of the parameter space in the NTK regime. Our experimental results confirm this prediction, providing strong evidence that grokking represents a transitional behavior between the lazy regime characterized by the NTK and the rich regime. These findings deepen our understanding of the interplay between optimization dynamics, spectral bias, and the phases of neural network learning.

---

### Rapid Augmentations for Time Series (RATS): A High-Performance Library for Time Series Augmentation

Authors: Wadie Skaf, Felix Kern, Aryamaan Basu Roy, Tejas Pradhan, Roman Kalkreuth, Holger Hoos
Published: 2026-01-06
Categories: cs.LG, cs.AI, cs.PF
arXiv: https://arxiv.org/abs/2601.03159v1

Time series augmentation is critical for training robust deep learning models, particularly in domains where labelled data is scarce and expensive to obtain. However, existing augmentation libraries for time series, mainly written in Python, suffer from performance bottlenecks, where running time grows exponentially as dataset sizes increase -- an aspect limiting their applicability in large-scale, production-grade systems. We introduce RATS (Rapid Augmentations for Time Series), a high-performance library for time series augmentation written in Rust with Python bindings (RATSpy). RATS implements multiple augmentation methods spanning basic transformations, frequency-domain operations and time warping techniques, all accessible through a unified pipeline interface with built-in parallelisation. Comprehensive benchmarking of RATSpy versus a commonly used library (tasug) on 143 datasets demonstrates that RATSpy achieves an average speedup of 74.5\% over tsaug (up to 94.8\% on large datasets), with up to 47.9\% less peak memory usage.

---

### Prompt-Counterfactual Explanations for Generative AI System Behavior

Authors: Sofie Goethals, Foster Provost, João Sedoc
Published: 2026-01-06
Categories: cs.LG, cs.AI, cs.CL, cs.CY
arXiv: https://arxiv.org/abs/2601.03156v2

As generative AI systems become integrated into real-world applications, organizations increasingly need to be able to understand and interpret their behavior. In particular, decision-makers need to understand what causes generative AI systems to exhibit specific output characteristics. Within this general topic, this paper examines a key question: what is it about the input -- the prompt -- that causes an LLM-based generative AI system to produce output that exhibits specific characteristics, such as toxicity, negative sentiment, or political bias. To examine this question, we adapt a common technique from the Explainable AI literature: counterfactual explanations. We explain why traditional counterfactual explanations cannot be applied directly to generative AI systems, due to several differences in how generative AI systems function. We then propose a flexible framework that adapts counterfactual explanations to non-deterministic, generative AI systems in scenarios where downstream classifiers can reveal key characteristics of their outputs. Based on this framework, we introduce an algorithm for generating prompt-counterfactual explanations (PCEs). Finally, we demonstrate the production of counterfactual explanations for generative AI systems with three case studies, examining different output characteristics (viz., political leaning, toxicity, and sentiment). The case studies further show that PCEs can streamline prompt engineering to suppress undesirable output characteristics and can enhance red-teaming efforts to uncover additional prompts that elicit undesirable outputs. Ultimately, this work lays a foundation for prompt-focused interpretability in generative AI: a capability that will become indispensable as these models are entrusted with higher-stakes tasks and subject to emerging regulatory requirements for transparency and accountability.

---

### Prompt-Counterfactual Explanations for Generative AI System Behavior

Authors: Sofie Goethals, Foster Provost, João Sedoc
Published: 2026-01-06
Categories: cs.LG, cs.AI, cs.CL, cs.CY
arXiv: https://arxiv.org/abs/2601.03156v1

As generative AI systems become integrated into real-world applications, organizations increasingly need to be able to understand and interpret their behavior. In particular, decision-makers need to understand what causes generative AI systems to exhibit specific output characteristics. Within this general topic, this paper examines a key question: what is it about the input -the prompt- that causes an LLM-based generative AI system to produce output that exhibits specific characteristics, such as toxicity, negative sentiment, or political bias. To examine this question, we adapt a common technique from the Explainable AI literature: counterfactual explanations. We explain why traditional counterfactual explanations cannot be applied directly to generative AI systems, due to several differences in how generative AI systems function. We then propose a flexible framework that adapts counterfactual explanations to non-deterministic, generative AI systems in scenarios where downstream classifiers can reveal key characteristics of their outputs. Based on this framework, we introduce an algorithm for generating prompt-counterfactual explanations (PCEs). Finally, we demonstrate the production of counterfactual explanations for generative AI systems with three case studies, examining different output characteristics (viz., political leaning, toxicity, and sentiment). The case studies further show that PCEs can streamline prompt engineering to suppress undesirable output characteristics and can enhance red-teaming efforts to uncover additional prompts that elicit undesirable outputs. Ultimately, this work lays a foundation for prompt-focused interpretability in generative AI: a capability that will become indispensable as these models are entrusted with higher-stakes tasks and subject to emerging regulatory requirements for transparency and accountability.

---

### Decoupling the Effect of Chain-of-Thought Reasoning: A Human Label Variation Perspective

Authors: Beiduo Chen, Tiancheng Hu, Caiqi Zhang, Robert Litschko, Anna Korhonen, Barbara Plank
Published: 2026-01-06
Categories: cs.CL
arXiv: https://arxiv.org/abs/2601.03154v2

Reasoning-tuned LLMs utilizing long Chain-of-Thought (CoT) excel at single-answer tasks, yet their ability to model Human Label Variation--which requires capturing probabilistic ambiguity rather than resolving it--remains underexplored. We investigate this through systematic disentanglement experiments on distribution-based tasks, employing Cross-CoT experiments to isolate the effect of reasoning text from intrinsic model priors. We observe a distinct "decoupled mechanism": while CoT improves distributional alignment, final accuracy is dictated by CoT content (99% variance contribution), whereas distributional ranking is governed by model priors (over 80%). Step-wise analysis further shows that while CoT's influence on accuracy grows monotonically during the reasoning process, distributional structure is largely determined by LLM's intrinsic priors. These findings suggest that long CoT serves as a decisive LLM decision-maker for the top option but fails to function as a granular distribution calibrator for ambiguous tasks.

---

### Decoupling the Effect of Chain-of-Thought Reasoning: A Human Label Variation Perspective

Authors: Beiduo Chen, Tiancheng Hu, Caiqi Zhang, Robert Litschko, Anna Korhonen, Barbara Plank
Published: 2026-01-06
Categories: cs.CL
arXiv: https://arxiv.org/abs/2601.03154v1

Reasoning-tuned LLMs utilizing long Chain-of-Thought (CoT) excel at single-answer tasks, yet their ability to model Human Label Variation--which requires capturing probabilistic ambiguity rather than resolving it--remains underexplored. We investigate this through systematic disentanglement experiments on distribution-based tasks, employing Cross-CoT experiments to isolate the effect of reasoning text from intrinsic model priors. We observe a distinct "decoupled mechanism": while CoT improves distributional alignment, final accuracy is dictated by CoT content (99% variance contribution), whereas distributional ranking is governed by model priors (over 80%). Step-wise analysis further shows that while CoT's influence on accuracy grows monotonically during the reasoning process, distributional structure is largely determined by LLM's intrinsic priors. These findings suggest that long CoT serves as a decisive LLM decision-maker for the top option but fails to function as a granular distribution calibrator for ambiguous tasks.

---

### PersonaLedger: Generating Realistic Financial Transactions with Persona Conditioned LLMs and Rule Grounded Feedback

Authors: Dehao Yuan, Tyler Farnan, Stefan Tesliuc, Doron L Bergman, Yulun Wu, Xiaoyu Liu, Minghui Liu, James Montgomery, Nam H Nguyen, C. Bayan Bruss, Furong Huang
Published: 2026-01-06
Categories: cs.LG
arXiv: https://arxiv.org/abs/2601.03149v2

Strict privacy regulations limit access to real transaction data, slowing open research in financial AI. Synthetic data can bridge this gap, but existing generators do not jointly achieve behavioral diversity and logical groundedness. Rule-driven simulators rely on hand-crafted workflows and shallow stochasticity, which miss the richness of human behavior. Learning-based generators such as GANs capture correlations yet often violate hard financial constraints and still require training on private data. We introduce PersonaLedger, a generation engine that uses a large language model conditioned on rich user personas to produce diverse transaction streams, coupled with an expert configurable programmatic engine that maintains correctness. The LLM and engine interact in a closed loop: after each event, the engine updates the user state, enforces financial rules, and returns a context aware "nextprompt" that guides the LLM toward feasible next actions. With this engine, we create a public dataset of 30 million transactions from 23,000 users and a benchmark suite with two tasks, illiquidity classification and identity theft segmentation. PersonaLedger offers a realistic, privacy preserving resource that supports rigorous evaluation of forecasting and anomaly detection models. PersonaLedger offers the community a rich, realistic, and privacy preserving resource -- complete with code, rules, and generation logs -- to accelerate innovation in financial AI and enable rigorous, reproducible evaluation.

---

### PersonaLedger: Generating Realistic Financial Transactions with Persona Conditioned LLMs and Rule Grounded Feedback

Authors: Dehao Yuan, Tyler Farnan, Stefan Tesliuc, Doron L Bergman, Yulun Wu, Xiaoyu Liu, Minghui Liu, James Montgomery, Nam H Nguyen, C. Bayan Bruss, Furong Huang
Published: 2026-01-06
Categories: cs.LG
arXiv: https://arxiv.org/abs/2601.03149v1

Strict privacy regulations limit access to real transaction data, slowing open research in financial AI. Synthetic data can bridge this gap, but existing generators do not jointly achieve behavioral diversity and logical groundedness. Rule-driven simulators rely on hand-crafted workflows and shallow stochasticity, which miss the richness of human behavior. Learning-based generators such as GANs capture correlations yet often violate hard financial constraints and still require training on private data. We introduce PersonaLedger, a generation engine that uses a large language model conditioned on rich user personas to produce diverse transaction streams, coupled with an expert configurable programmatic engine that maintains correctness. The LLM and engine interact in a closed loop: after each event, the engine updates the user state, enforces financial rules, and returns a context aware "nextprompt" that guides the LLM toward feasible next actions. With this engine, we create a public dataset of 30 million transactions from 23,000 users and a benchmark suite with two tasks, illiquidity classification and identity theft segmentation. PersonaLedger offers a realistic, privacy preserving resource that supports rigorous evaluation of forecasting and anomaly detection models. PersonaLedger offers the community a rich, realistic, and privacy preserving resource -- complete with code, rules, and generation logs -- to accelerate innovation in financial AI and enable rigorous, reproducible evaluation.

---

### Self-Verification is All You Need To Pass The Japanese Bar Examination

Authors: Andrew Shin
Published: 2026-01-06
Categories: cs.CL, cs.AI
arXiv: https://arxiv.org/abs/2601.03144v1

Despite rapid advances in large language models (LLMs), achieving reliable performance on highly professional and structured examinations remains a significant challenge. The Japanese bar examination is a particularly demanding benchmark, requiring not only advanced legal reasoning but also strict adherence to complex answer formats that involve joint evaluation of multiple propositions. While recent studies have reported improvements by decomposing such questions into simpler true--false judgments, these approaches have not been systematically evaluated under the original exam format and scoring scheme, leaving open the question of whether they truly capture exam-level competence. In this paper, we present a self-verification model trained on a newly constructed dataset that faithfully replicates the authentic format and evaluation scale of the exam. Our model is able to exceed the official passing score when evaluated on the actual exam scale, marking the first demonstration, to our knowledge, of an LLM passing the Japanese bar examination without altering its original question structure or scoring rules. We further conduct extensive comparisons with alternative strategies, including multi-agent inference and decomposition-based supervision, and find that these methods fail to achieve comparable performance. Our results highlight the importance of format-faithful supervision and consistency verification, and suggest that carefully designed single-model approaches can outperform more complex systems in high-stakes professional reasoning tasks. Our dataset and codes are publicly available.

---

### Accurate Table Question Answering with Accessible LLMs

Authors: Yangfan Jiang, Fei Wei, Ergute Bao, Yaliang Li, Bolin Ding, Yin Yang, Xiaokui Xiao
Published: 2026-01-06
Categories: cs.DB, cs.CL
arXiv: https://arxiv.org/abs/2601.03137v1

Given a table T in a database and a question Q in natural language, the table question answering (TQA) task aims to return an accurate answer to Q based on the content of T. Recent state-of-the-art solutions leverage large language models (LLMs) to obtain high-quality answers. However, most rely on proprietary, large-scale LLMs with costly API access, posing a significant financial barrier. This paper instead focuses on TQA with smaller, open-weight LLMs that can run on a desktop or laptop. This setting is challenging, as such LLMs typically have weaker capabilities than large proprietary models, leading to substantial performance degradation with existing methods. We observe that a key reason for this degradation is that prior approaches often require the LLM to solve a highly sophisticated task using long, complex prompts, which exceed the capabilities of small open-weight LLMs. Motivated by this observation, we present Orchestra, a multi-agent approach that unlocks the potential of accessible LLMs for high-quality, cost-effective TQA. Orchestra coordinates a group of LLM agents, each responsible for a relatively simple task, through a structured, layered workflow to solve complex TQA problems -- akin to an orchestra. By reducing the prompt complexity faced by each agent, Orchestra significantly improves output reliability. We implement Orchestra on top of AgentScope, an open-source multi-agent framework, and evaluate it on multiple TQA benchmarks using a wide range of open-weight LLMs. Experimental results show that Orchestra achieves strong performance even with small- to medium-sized models. For example, with Qwen2.5-14B, Orchestra reaches 72.1% accuracy on WikiTQ, approaching the best prior result of 75.3% achieved with GPT-4; with larger Qwen, Llama, or DeepSeek models, Orchestra outperforms all prior methods and establishes new state-of-the-art results across all benchmarks.

---

### Limited Linguistic Diversity in Embodied AI Datasets

Authors: Selma Wanna, Agnes Luhtaru, Jonathan Salfity, Ryan Barron, Juston Moore, Cynthia Matuszek, Mitch Pryor
Published: 2026-01-06
Categories: cs.CL, cs.AI, cs.RO
arXiv: https://arxiv.org/abs/2601.03136v2

Language plays a critical role in Vision-Language-Action (VLA) models, yet the linguistic characteristics of the datasets used to train and evaluate these systems remain poorly documented. In this work, we present a systematic dataset audit of several widely used VLA corpora, aiming to characterize what kinds of instructions these datasets actually contain and how much linguistic variety they provide. We quantify instruction language along complementary dimensions--including lexical variety, duplication and overlap, semantic similarity, and syntactic complexity. Our analysis shows that many datasets rely on highly repetitive, template-like commands with limited structural variation, yielding a narrow distribution of instruction forms. We position these findings as descriptive documentation of the language signal available in current VLA training and evaluation data, intended to support more detailed dataset reporting, more principled dataset selection, and targeted curation or augmentation strategies that broaden language coverage.

---

### Limited Linguistic Diversity in Embodied AI Datasets

Authors: Selma Wanna, Agnes Luhtaru, Jonathan Salfity, Ryan Barron, Juston Moore, Cynthia Matuszek, Mitch Pryor
Published: 2026-01-06
Categories: cs.CL, cs.AI, cs.RO
arXiv: https://arxiv.org/abs/2601.03136v1

Language plays a critical role in Vision-Language-Action (VLA) models, yet the linguistic characteristics of the datasets used to train and evaluate these systems remain poorly documented. In this work, we present a systematic dataset audit of several widely used VLA corpora, aiming to characterize what kinds of instructions these datasets actually contain and how much linguistic variety they provide. We quantify instruction language along complementary dimensions-including lexical variety, duplication and overlap, semantic similarity, and syntactic complexity. Our analysis shows that many datasets rely on highly repetitive, template-like commands with limited structural variation, yielding a narrow distribution of instruction forms. We position these findings as descriptive documentation of the language signal available in current VLA training and evaluation data, intended to support more detailed dataset reporting, more principled dataset selection, and targeted curation or augmentation strategies that broaden language coverage.

---

### Improving Indigenous Language Machine Translation with Synthetic Data and Language-Specific Preprocessing

Authors: Aashish Dhawan, Christopher Driggers-Ellis, Christan Grant, Daisy Zhe Wang
Published: 2026-01-06
Categories: cs.CL
arXiv: https://arxiv.org/abs/2601.03135v2

Low-resource indigenous languages often lack the parallel corpora required for effective neural machine translation (NMT). Synthetic data generation offers a practical strategy for mitigating this limitation in data-scarce settings. In this work, we augment curated parallel datasets for indigenous languages of the Americas with synthetic sentence pairs generated using a high-capacity multilingual translation model. We fine-tune a multilingual mBART model on curated-only and synthetically augmented data and evaluate translation quality using chrF++, the primary metric used in recent AmericasNLP shared tasks for agglutinative languages. We further apply language-specific preprocessing, including orthographic normalization and noise-aware filtering, to reduce corpus artifacts. Experiments on Guarani-Spanish and Quechua-Spanish translation show consistent chrF++ improvements from synthetic data augmentation, while diagnostic experiments on Aymara highlight the limitations of generic preprocessing for highly agglutinative languages.

---

### Improving Indigenous Language Machine Translation with Synthetic Data and Language-Specific Preprocessing

Authors: Aashish Dhawan, Christopher Driggers-Ellis, Christan Grant, Daisy Zhe Wang
Published: 2026-01-06
Categories: cs.CL
arXiv: https://arxiv.org/abs/2601.03135v1

Low-resource indigenous languages often lack the parallel corpora required for effective neural machine translation (NMT). Synthetic data generation offers a practical strategy for mitigating this limitation in data-scarce settings. In this work, we augment curated parallel datasets for indigenous languages of the Americas with synthetic sentence pairs generated using a high-capacity multilingual translation model. We fine-tune a multilingual mBART model on curated-only and synthetically augmented data and evaluate translation quality using chrF++, the primary metric used in recent AmericasNLP shared tasks for agglutinative languages. We further apply language-specific preprocessing, including orthographic normalization and noise-aware filtering, to reduce corpus artifacts. Experiments on Guarani--Spanish and Quechua--Spanish translation show consistent chrF++ improvements from synthetic data augmentation, while diagnostic experiments on Aymara highlight the limitations of generic preprocessing for highly agglutinative languages.

---

### The Anatomy of Conversational Scams: A Topic-Based Red Teaming Analysis of Multi-Turn Interactions in LLMs

Authors: Xiangzhe Yuan, Zhenhao Zhang, Haoming Tang, Siying Hu
Published: 2026-01-06
Categories: cs.CL
arXiv: https://arxiv.org/abs/2601.03134v2

As LLMs gain persuasive capabilities through extended dialogues, they create new opportunities for studying adversarial conversational behavior in extended interaction settings that traditional single-turn safety evaluations fail to capture. We systematically study these interactional dynamics using a controlled LLM-to-LLM simulation framework for automated red-teaming across bilingual social engineering scenarios. Evaluating eight state-of-the-art models in English and Chinese, we analyze dialogue-level outcomes, annotate attacker and defender strategy families, and model interaction dynamics between them. Results show that multi-turn adversarial dialogues follow recurrent escalation patterns, while defensive responses frequently rely on verification, delay, and channel control. We further find statistically significant cross-model and cross-lingual differences in outcome distributions, and transition analysis reveals systematic structural variation in how defender strategies respond to attacker tactics across languages. These findings highlight the importance of studying interactional structure in multi-turn adversarial dialogue settings and demonstrate how controlled LLM-to-LLM simulations can support mechanistic analysis of adversarial conversational dynamics.

---

### The Anatomy of Conversational Scams: A Topic-Based Red Teaming Analysis of Multi-Turn Interactions in LLMs

Authors: Xiangzhe Yuan, Zhenhao Zhang, Haoming Tang, Siying Hu
Published: 2026-01-06
Categories: cs.CL
arXiv: https://arxiv.org/abs/2601.03134v1

As LLMs gain persuasive agentic capabilities through extended dialogues, they introduce novel risks in multi-turn conversational scams that single-turn safety evaluations fail to capture. We systematically study these risks using a controlled LLM-to-LLM simulation framework across multi-turn scam scenarios. Evaluating eight state-of-the-art models in English and Chinese, we analyze dialogue outcomes and qualitatively annotate attacker strategies, defensive responses, and failure modes. Results reveal that scam interactions follow recurrent escalation patterns, while defenses employ verification and delay mechanisms. Furthermore, interactional failures frequently stem from safety guardrail activation and role instability. Our findings highlight multi-turn interactional safety as a critical, distinct dimension of LLM behavior.

---

### Finite Memory Belief Approximation for Optimal Control in Partially Observable Markov Decision Processes

Authors: Mintae Kim
Published: 2026-01-06
Categories: eess.SY, cs.IT, cs.LG
arXiv: https://arxiv.org/abs/2601.03132v1

We study finite memory belief approximation for partially observable (PO) stochastic optimal control (SOC) problems. While belief states are sufficient for SOC in partially observable Markov decision processes (POMDPs), they are generally infinite-dimensional and impractical. We interpret truncated input-output (IO) histories as inducing a belief approximation and develop a metric-based theory that directly relates information loss to control performance. Using the Wasserstein metric, we derive policy-conditional performance bounds that quantify value degradation induced by finite memory along typical closed-loop trajectories. Our analysis proceeds via a fixed-policy comparison: we evaluate two cost functionals under the same closed-loop execution and isolate the effect of replacing the true belief by its finite memory approximation inside the belief-level cost. For linear quadratic Gaussian (LQG) systems, we provide closed-form belief mismatch evaluation and empirically validate the predicted mechanism, demonstrating that belief mismatch decays approximately exponentially with memory length and that the induced performance mismatch scales accordingly. Together, these results provide a metric-aware characterization of what finite memory belief approximation can and cannot achieve in PO settings.

---

### Automatic Prompt Engineering with No Task Cues and No Tuning

Authors: Faisal Chowdhury, Nandana Mihindukulasooriya, Niharika S D'Souza, Horst Samulowitz, Neeru Gupta, Tomasz Hanusiak, Michal Kapitonow
Published: 2026-01-06
Categories: cs.AI, cs.CL
arXiv: https://arxiv.org/abs/2601.03130v1

This paper presents a system for automatic prompt engineering that is much simpler in both design and application and yet as effective as the existing approaches. It requires no tuning and no explicit clues about the task. We evaluated our approach on cryptic column name expansion (CNE) in database tables, a task which is critical for tabular data search, access, and understanding and yet there has been very little existing work. We evaluated on datasets in two languages, English and German. This is the first work to report on the application of automatic prompt engineering for the CNE task. To the best of our knowledge, this is also the first work on the application of automatic prompt engineering for a language other than English.

---

### Unified Thinker: A General Reasoning Modular Core for Image Generation

Authors: Sashuai Zhou, Qiang Zhou, Jijin Hu, Hanqing Yang, Yue Cao, Junpeng Ma, Yinchao Ma, Jun Song, Tiezheng Ge, Cheng Yu, Bo Zheng, Zhou Zhao
Published: 2026-01-06
Categories: cs.CV, cs.AI
arXiv: https://arxiv.org/abs/2601.03127v2

Despite impressive progress in high-fidelity image synthesis, generative models still struggle with logic-intensive instruction following, exposing a persistent reasoning--execution gap. Meanwhile, closed-source systems (e.g., Nano Banana) have demonstrated strong reasoning-driven image generation, highlighting a substantial gap to current open-source models. We argue that closing this gap requires not merely better visual generators, but executable reasoning: decomposing high-level intents into grounded, verifiable plans that directly steer the generative process. To this end, we propose Unified Thinker, a task-agnostic reasoning architecture for general image generation, designed as a unified planning core that can plug into diverse generators and workflows. Unified Thinker decouples a dedicated Thinker from the image Generator, enabling modular upgrades of reasoning without retraining the entire generative model. We further introduce a two-stage training paradigm: we first build a structured planning interface for the Thinker, then apply reinforcement learning to ground its policy in pixel-level feedback, encouraging plans that optimize visual correctness over textual plausibility. Extensive experiments on text-to-image generation and image editing show that Unified Thinker substantially improves image reasoning and generation quality.

---

### Unified Thinker: A General Reasoning Modular Core for Image Generation

Authors: Sashuai Zhou, Qiang Zhou, Jijin Hu, Hanqing Yang, Yue Cao, Junpeng Ma, Yinchao Ma, Jun Song, Tiezheng Ge, Cheng Yu, Bo Zheng, Zhou Zhao
Published: 2026-01-06
Categories: cs.CV, cs.AI
arXiv: https://arxiv.org/abs/2601.03127v1

Despite impressive progress in high-fidelity image synthesis, generative models still struggle with logic-intensive instruction following, exposing a persistent reasoning--execution gap. Meanwhile, closed-source systems (e.g., Nano Banana) have demonstrated strong reasoning-driven image generation, highlighting a substantial gap to current open-source models. We argue that closing this gap requires not merely better visual generators, but executable reasoning: decomposing high-level intents into grounded, verifiable plans that directly steer the generative process. To this end, we propose Unified Thinker, a task-agnostic reasoning architecture for general image generation, designed as a unified planning core that can plug into diverse generators and workflows. Unified Thinker decouples a dedicated Thinker from the image Generator, enabling modular upgrades of reasoning without retraining the entire generative model. We further introduce a two-stage training paradigm: we first build a structured planning interface for the Thinker, then apply reinforcement learning to ground its policy in pixel-level feedback, encouraging plans that optimize visual correctness over textual plausibility. Extensive experiments on text-to-image generation and image editing show that Unified Thinker substantially improves image reasoning and generation quality.

---

### LeafLife: An Explainable Deep Learning Framework with Robustness for Grape Leaf Disease Recognition

Authors: B. M. Shahria Alam, Md. Nasim Ahmed
Published: 2026-01-06
Categories: cs.CV, cs.AI, cs.LG
arXiv: https://arxiv.org/abs/2601.03124v1

Plant disease diagnosis is essential to farmers' management choices because plant diseases frequently lower crop yield and product quality. For harvests to flourish and agricultural productivity to boost, grape leaf disease detection is important. The plant disease dataset contains grape leaf diseases total of 9,032 images of four classes, among them three classes are leaf diseases, and the other one is healthy leaves. After rigorous pre-processing dataset was split (70% training, 20% validation, 10% testing), and two pre-trained models were deployed: InceptionV3 and Xception. Xception shows a promising result of 96.23% accuracy, which is remarkable than InceptionV3. Adversarial Training is used for robustness, along with more transparency. Grad-CAM is integrated to confirm the leaf disease. Finally deployed a web application using Streamlit with a heatmap visualization and prediction with confidence level for robust grape leaf disease classification.

---

### Gradient descent reliably finds depth- and gate-optimal circuits for generic unitaries

Authors: Janani Gomathi, Alex Meiburg
Published: 2026-01-06
Categories: quant-ph, cs.LG
arXiv: https://arxiv.org/abs/2601.03123v2

When the gate set has continuous parameters, synthesizing a unitary operator as a quantum circuit is, in principle, always possible using exact methods. However, efficiently finding depth- and gate-minimal circuits remains a major challenge. The landscape is very different for compiled unitaries, which arise from programming and typically have short circuits, as compared with generic unitaries, which use all parameters and typically require circuits of maximal size. Previous approaches based on random combinatorial search indicate a low success rate even when the circuit ansatz is nominally adequately parameterized, motivating the use of heavily overparameterized circuits. In this work, we present a gradient-based optimization framework that enables the synthesis of depth- and gate-optimal circuits for generic unitaries without overparameterization, even under restricted hardware connectivity. We prescribe parameter-optimal circuit skeletons and eliminate the need for random combinatorial search. We further show that the poor performance of earlier random-search approaches can be attributed to the inadvertent selection of parameter-deficient circuit topologies. By systematically avoiding such skeletons, our approach achieves reliable convergence while maintaining parameter efficiency.

---

### Gradient descent reliably finds depth- and gate-optimal circuits for generic unitaries

Authors: Janani Gomathi, Alex Meiburg
Published: 2026-01-06
Categories: quant-ph, cs.LG
arXiv: https://arxiv.org/abs/2601.03123v1

When the gate set has continuous parameters, synthesizing a unitary operator as a quantum circuit is always possible using exact methods, but finding minimal circuits efficiently remains a challenging problem. The landscape is very different for compiled unitaries, which arise from programming and typically have short circuits, as compared with generic unitaries, which use all parameters and typically require circuits of maximal size. We show that simple gradient descent reliably finds depth- and gate-optimal circuits for generic unitaries, including in the presence of restricted chip connectivity. This runs counter to earlier evidence that optimal synthesis required combinatorial search, and we show that this discrepancy can be explained by avoiding the random selection of certain parameter-deficient circuit skeletons.

---

### ToxiGAN: Toxic Data Augmentation via LLM-Guided Directional Adversarial Generation

Authors: Peiran Li, Jan Fillies, Adrian Paschke
Published: 2026-01-06
Categories: cs.CL, cs.AI, cs.LG
arXiv: https://arxiv.org/abs/2601.03121v1

Augmenting toxic language data in a controllable and class-specific manner is crucial for improving robustness in toxicity classification, yet remains challenging due to limited supervision and distributional skew. We propose ToxiGAN, a class-aware text augmentation framework that combines adversarial generation with semantic guidance from large language models (LLMs). To address common issues in GAN-based augmentation such as mode collapse and semantic drift, ToxiGAN introduces a two-step directional training strategy and leverages LLM-generated neutral texts as semantic ballast. Unlike prior work that treats LLMs as static generators, our approach dynamically selects neutral exemplars to provide balanced guidance. Toxic samples are explicitly optimized to diverge from these exemplars, reinforcing class-specific contrastive signals. Experiments on four hate speech benchmarks show that ToxiGAN achieves the strongest average performance in both macro-F1 and hate-F1, consistently outperforming traditional and LLM-based augmentation methods. Ablation and sensitivity analyses further confirm the benefits of semantic ballast and directional training in enhancing classifier robustness.

---

### A framework for assuring the accuracy and fidelity of an AI-enabled Digital Twin of en route UK airspace

Authors: Adam Keane, Nick Pepper, Chris Burr, Amy Hodgkin, Dewi Gould, John Korna, Marc Thomas
Published: 2026-01-06
Categories: cs.AI
arXiv: https://arxiv.org/abs/2601.03120v1

Digital Twins combine simulation, operational data and Artificial Intelligence (AI), and have the potential to bring significant benefits across the aviation industry. Project Bluebird, an industry-academic collaboration, has developed a probabilistic Digital Twin of en route UK airspace as an environment for training and testing AI Air Traffic Control (ATC) agents. There is a developing regulatory landscape for this kind of novel technology. Regulatory requirements are expected to be application specific, and may need to be tailored to each specific use case. We draw on emerging guidance for both Digital Twin development and the use of Artificial Intelligence/Machine Learning (AI/ML) in Air Traffic Management (ATM) to present an assurance framework. This framework defines actionable goals and the evidence required to demonstrate that a Digital Twin accurately represents its physical counterpart and also provides sufficient functionality across target use cases. It provides a structured approach for researchers to assess, understand and document the strengths and limitations of the Digital Twin, whilst also identifying areas where fidelity could be improved. Furthermore, it serves as a foundation for engagement with stakeholders and regulators, supporting discussions around the regulatory needs for future applications, and contributing to the emerging guidance through a concrete, working example of a Digital Twin. The framework leverages a methodology known as Trustworthy and Ethical Assurance (TEA) to develop an assurance case. An assurance case is a nested set of structured arguments that provides justified evidence for how a top-level goal has been realised. In this paper we provide an overview of each structured argument and a number of deep dives which elaborate in more detail upon particular arguments, including the required evidence, assumptions and justifications.

---

### Transformers self-organize like newborn visual systems when trained in prenatal worlds

Authors: Lalit Pandey, Samantha M. W. Wood, Justin N. Wood
Published: 2026-01-06
Categories: q-bio.NC, cs.AI, cs.CV
arXiv: https://arxiv.org/abs/2601.03117v1

Do transformers learn like brains? A key challenge in addressing this question is that transformers and brains are trained on fundamentally different data. Brains are initially "trained" on prenatal sensory experiences (e.g., retinal waves), whereas transformers are typically trained on large datasets that are not biologically plausible. We reasoned that if transformers learn like brains, then they should develop the same structure as newborn brains when exposed to the same prenatal data. To test this prediction, we simulated prenatal visual input using a retinal wave generator. Then, using self-supervised temporal learning, we trained transformers to adapt to those retinal waves. During training, the transformers spontaneously developed the same structure as newborn visual systems: (1) early layers became sensitive to edges, (2) later layers became sensitive to shapes, and (3) the models developed larger receptive fields across layers. The organization of newborn visual systems emerges spontaneously when transformers adapt to a prenatal visual world. This developmental convergence suggests that brains and transformers learn in common ways and follow the same general fitting principles.

---

### Discovering and Causally Validating Emotion-Sensitive Neurons in Large Audio-Language Models

Authors: Xiutian Zhao, Björn Schuller, Berrak Sisman
Published: 2026-01-06
Categories: cs.CL, eess.AS
arXiv: https://arxiv.org/abs/2601.03115v2

Emotion is a central dimension of spoken communication, yet, we still lack a mechanistic account of how modern large audio-language models (LALMs) encode it internally. We present the first neuron-level interpretability study of emotion-sensitive neurons (ESNs) in LALMs and provide causal evidence supporting the existence of such units in Qwen2.5-Omni, Kimi-Audio, and Audio Flamingo 3. Across these three widely used open-source models, we compare frequency-, entropy-, mean-deviation-, and contrast-based neuron selectors on multiple emotion recognition benchmarks. Using inference-time interventions, we reveal a consistent emotion-specific signature: deactivating neurons selected for a given emotion disproportionately degrades recognition of that emotion while largely preserving other classes, whereas targeted steering amplifies these units to bias predictions toward the target emotion. These effects arise with modest amounts of identification data and scale systematically with intervention strength. We further observe that ESNs exhibit non-uniform layer-wise clustering with partial cross-dataset transfer. Taken together, our results offer a causal, neuron-level account of emotion decisions in LALMs and highlight targeted neuron interventions as an actionable handle for controllable affective behaviors.

---

### Discovering and Causally Validating Emotion-Sensitive Neurons in Large Audio-Language Models

Authors: Xiutian Zhao, Björn Schuller, Berrak Sisman
Published: 2026-01-06
Categories: cs.CL, eess.AS
arXiv: https://arxiv.org/abs/2601.03115v1

Emotion is a central dimension of spoken communication, yet, we still lack a mechanistic account of how modern large audio-language models (LALMs) encode it internally. We present the first neuron-level interpretability study of emotion-sensitive neurons (ESNs) in LALMs and provide causal evidence that such units exist in Qwen2.5-Omni, Kimi-Audio, and Audio Flamingo 3. Across these three widely used open-source models, we compare frequency-, entropy-, magnitude-, and contrast-based neuron selectors on multiple emotion recognition benchmarks. Using inference-time interventions, we reveal a consistent emotion-specific signature: ablating neurons selected for a given emotion disproportionately degrades recognition of that emotion while largely preserving other classes, whereas gain-based amplification steers predictions toward the target emotion. These effects arise with modest identification data and scale systematically with intervention strength. We further observe that ESNs exhibit non-uniform layer-wise clustering with partial cross-dataset transfer. Taken together, our results offer a causal, neuron-level account of emotion decisions in LALMs and highlight targeted neuron interventions as an actionable handle for controllable affective behaviors.

---

### DiT-JSCC: Rethinking Deep JSCC with Diffusion Transformers and Semantic Representations

Authors: Kailin Tan, Jincheng Dai, Sixian Wang, Guo Lu, Shuo Shao, Kai Niu, Wenjun Zhang, Ping Zhang
Published: 2026-01-06
Categories: eess.IV, cs.CV
arXiv: https://arxiv.org/abs/2601.03112v2

Generative joint source-channel coding (GJSCC) has emerged as a new Deep JSCC paradigm for achieving high-fidelity and robust image transmission under extreme wireless channel conditions, such as ultra-low bandwidth and low signal-to-noise ratio. Recent studies commonly adopt diffusion models as generative decoders, but they frequently produce visually realistic results with limited semantic consistency. This limitation stems from a fundamental mismatch between reconstruction-oriented JSCC encoders and generative decoders, as the former lack explicit semantic discriminability and fail to provide reliable conditional cues. In this paper, we propose DiT-JSCC, a novel GJSCC backbone that can jointly learn a semantics-prioritized representation encoder and a diffusion transformer (DiT) based generative decoder, our open-source project aims to promote the future research in GJSCC. Specifically, we design a semantics-detail dual-branch encoder that aligns naturally with a coarse-to-fine conditional DiT decoder, prioritizing semantic consistency under extreme channel conditions. Moreover, a training-free adaptive bandwidth allocation strategy inspired by Kolmogorov complexity is introduced to further improve the transmission efficiency, thereby indeed redefining the notion of information value in the era of generative decoding. Extensive experiments demonstrate that DiT-JSCC consistently outperforms existing JSCC methods in both semantic consistency and visual quality, particularly in extreme regimes.

---

### DiT-JSCC: Rethinking Deep JSCC with Diffusion Transformers and Semantic Representations

Authors: Kailin Tan, Jincheng Dai, Sixian Wang, Guo Lu, Shuo Shao, Kai Niu, Wenjun Zhang, Ping Zhang
Published: 2026-01-06
Categories: eess.IV, cs.CV
arXiv: https://arxiv.org/abs/2601.03112v1

Generative joint source-channel coding (GJSCC) has emerged as a new Deep JSCC paradigm for achieving high-fidelity and robust image transmission under extreme wireless channel conditions, such as ultra-low bandwidth and low signal-to-noise ratio. Recent studies commonly adopt diffusion models as generative decoders, but they frequently produce visually realistic results with limited semantic consistency. This limitation stems from a fundamental mismatch between reconstruction-oriented JSCC encoders and generative decoders, as the former lack explicit semantic discriminability and fail to provide reliable conditional cues. In this paper, we propose DiT-JSCC, a novel GJSCC backbone that can jointly learn a semantics-prioritized representation encoder and a diffusion transformer (DiT) based generative decoder, our open-source project aims to promote the future research in GJSCC. Specifically, we design a semantics-detail dual-branch encoder that aligns naturally with a coarse-to-fine conditional DiT decoder, prioritizing semantic consistency under extreme channel conditions. Moreover, a training-free adaptive bandwidth allocation strategy inspired by Kolmogorov complexity is introduced to further improve the transmission efficiency, thereby indeed redefining the notion of information value in the era of generative decoding. Extensive experiments demonstrate that DiT-JSCC consistently outperforms existing JSCC methods in both semantic consistency and visual quality, particularly in extreme regimes.

---

### One Sample to Rule Them All: Extreme Data Efficiency in Multidiscipline Reasoning with Reinforcement Learning

Authors: Yiyuan Li, Zhen Huang, Yanan Wu, Weixun Wang, Xuefeng Li, Yijia Luo, Wenbo Su, Bo Zheng, Pengfei Liu
Published: 2026-01-06
Categories: cs.LG, cs.CL
arXiv: https://arxiv.org/abs/2601.03111v2

The reasoning ability of large language models (LLMs) can be unleashed with reinforcement learning (RL) (OpenAI, 2024; DeepSeek-AI et al., 2025a; Zeng et al., 2025). The success of existing RL attempts in LLMs usually rely on high-quality samples of large volumes. In this paper, we challenge conventional assumptions about data requirements in RL for LLMs by demonstrating the effectiveness of one-shot reinforcement learning. Specifically, we introduce polymath learning, a framework for designing one training sample that elicits multidisciplinary reasoning improvement. We present three key findings: (1) A single, strategically selected math reasoning sample can produce significant performance improvements across multiple domains, including physics, chemistry, and biology; (2) Analysis of salient mathematical skills provides insight into the characteristics associated with effective polymath samples; and (3) An engineered synthetic sample that integrates multidisciplinary elements and broader skill coverage achieves stronger performance than naturally occurring individual samples. Across various reasoning benchmarks, polymath learning achieves stronger performance than larger datasets, demonstrating that reasoning structure and skills in samples, rather than quantity, may be the key to unlock enhanced reasoning capabilities in language models. Our results suggest a shift, dubbed as sample engineering, toward precision engineering of samples that complements simply increasing data volume.

---

### One Sample to Rule Them All: Extreme Data Efficiency in RL Scaling

Authors: Yiyuan Li, Zhen Huang, Yanan Wu, Weixun Wang, Xuefeng Li, Yijia Luo, Wenbo Su, Bo Zheng, Pengfei Liu
Published: 2026-01-06
Categories: cs.LG, cs.CL
arXiv: https://arxiv.org/abs/2601.03111v1

The reasoning ability of large language models (LLMs) can be unleashed with reinforcement learning (RL) (OpenAI, 2024; DeepSeek-AI et al., 2025a; Zeng et al., 2025). The success of existing RL attempts in LLMs usually relies on high-quality samples of thousands or beyond. In this paper, we challenge fundamental assumptions about data requirements in RL for LLMs by demonstrating the remarkable effectiveness of one-shot learning. Specifically, we introduce polymath learning, a framework for designing one training sample that elicits multidisciplinary impact. We present three key findings: (1) A single, strategically selected math reasoning sample can produce significant performance improvements across multiple domains, including physics, chemistry, and biology with RL; (2) The math skills salient to reasoning suggest the characteristics of the optimal polymath sample; and (3) An engineered synthetic sample that integrates multidiscipline elements outperforms training with individual samples that naturally occur. Our approach achieves superior performance to training with larger datasets across various reasoning benchmarks, demonstrating that sample quality and design, rather than quantity, may be the key to unlock enhanced reasoning capabilities in language models. Our results suggest a shift, dubbed as sample engineering, toward precision engineering of training samples rather than simply increasing data volume.

---