# Research Papers — 2026-01 · bundle page 89 of 189

> Full-text markdown of papers #4401–4450 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/89
[← Previous bundle](https://aipod.dev/papers/archive/2026-01/page/88.md)
[Next bundle →](https://aipod.dev/papers/archive/2026-01/page/90.md)

### Agentic Conversational Search with Contextualized Reasoning via Reinforcement Learning

Authors: Fengran Mo, Yifan Gao, Sha Li, Hansi Zeng, Xin Liu, Zhaoxuan Tan, Xian Li, Jianshu Chen, Dakuo Wang, Meng Jiang
Published: 2026-01-19
Categories: cs.CL, cs.IR
arXiv: https://arxiv.org/abs/2601.13115v2

Large Language Models (LLMs) have become a popular interface for human-AI interaction, supporting information seeking and task assistance through natural, multi-turn dialogue. To respond to users within multi-turn dialogues, the context-dependent user intent evolves across interactions, requiring contextual interpretation, query reformulation, and dynamic coordination between retrieval and generation. Existing studies usually follow static rewrite, retrieve, and generate pipelines, which optimize different procedures separately and overlook the mixed-initiative action optimization simultaneously. Although the recent developments in deep search agents demonstrate the effectiveness in jointly optimizing retrieval and generation via reasoning, these approaches focus on single-turn scenarios, which might lack the ability to handle multi-turn interactions. We introduce a conversational agent that interleaves search and reasoning across turns, enabling exploratory and adaptive behaviors learned through reinforcement learning (RL) training with tailored rewards towards evolving user goals. The experimental results across four widely used conversational benchmarks demonstrate the effectiveness of our methods by surpassing several existing strong baselines.

---

### IntAgent: NWDAF-Based Intent LLM Agent Towards Advanced Next Generation Networks

Authors: Abdelrahman Soliman, Ahmed Refaey, Aiman Erbad, Amr Mohamed
Published: 2026-01-19
Categories: cs.NI, cs.AI, eess.SY
arXiv: https://arxiv.org/abs/2601.13114v1

Intent-based networks (IBNs) are gaining prominence as an innovative technology that automates network operations through high-level request statements, defining what the network should achieve. In this work, we introduce IntAgent, an intelligent intent LLM agent that integrates NWDAF analytics and tools to fulfill the network operator's intents. Unlike previous approaches, we develop an intent tools engine directly within the NWDAF analytics engine, allowing our agent to utilize live network analytics to inform its reasoning and tool selection. We offer an enriched, 3GPP-compliant data source that enhances the dynamic, context-aware fulfillment of network operator goals, along with an MCP tools server for scheduling, monitoring, and analytics tools. We demonstrate the efficacy of our framework through two practical use cases: ML-based traffic prediction and scheduled policy enforcement, which validate IntAgent's ability to autonomously fulfill complex network intents.

---

### CORE-T: COherent REtrieval of Tables for Text-to-SQL

Authors: Hassan Soliman, Vivek Gupta, Dan Roth, Iryna Gurevych
Published: 2026-01-19
Categories: cs.CL, cs.AI, cs.IR
arXiv: https://arxiv.org/abs/2601.13111v3

Realistic text-to-SQL workflows often require joining multiple tables. As a result, accurately retrieving the relevant set of tables becomes a key bottleneck for end-to-end performance. We study an open-book setting where queries must be answered over large, heterogeneous table collections pooled from many sources, without clean scoping signals such as database identifiers. Here, dense retrieval (DR) achieves high recall but returns many distractors, while join-aware alternatives often rely on extra assumptions and/or incur high inference overhead. We propose CORE-T, a scalable, training-free framework that enriches tables with LLM-generated purpose metadata and pre-computes a lightweight table-compatibility cache. At inference time, DR returns top-K candidates; a single LLM call selects a coherent, joinable subset, and a two-step additive adjustment stage restores strongly compatible tables. Across Bird, Spider, MMQA, and Beaver, CORE-T improves over DR by up to 22.7 points in table-selection F1 while returning up to 40% fewer tables, and by up to 24.4 points in multi-table execution accuracy, and uses 1.64-4.20x fewer total selection tokens than LLM-intensive baselines.

---

### Leveraging Lora Fine-Tuning and Knowledge Bases for Construction Identification

Authors: Liu Kaipeng, Wu Ling
Published: 2026-01-19
Categories: cs.CL
arXiv: https://arxiv.org/abs/2601.13105v1

This study investigates the automatic identification of the English ditransitive construction by integrating LoRA-based fine-tuning of a large language model with a Retrieval-Augmented Generation (RAG) framework.A binary classification task was conducted on annotated data from the British National Corpus. Results demonstrate that a LoRA-fine-tuned Qwen3-8B model significantly outperformed both a native Qwen3-MAX model and a theory-only RAG system. Detailed error analysis reveals that fine-tuning shifts the model's judgment from a surface-form pattern matching towards a more semantically grounded understanding based.

---

### Approximate full conformal prediction in an RKHS

Authors: Davidson Lova Razafindrakoto, Alain Celisse, Jérôme Lacaille
Published: 2026-01-19
Categories: stat.ML, cs.LG, math.ST
arXiv: https://arxiv.org/abs/2601.13102v3

Full conformal prediction is a framework that implicitly formulates distribution-free confidence prediction regions for a wide range of estimators. However, a classical limitation of the full conformal framework is the computation of the confidence prediction regions, which is usually impossible since it requires training infinitely many estimators (for real-valued prediction for instance). The main purpose of the present work is to describe a generic strategy for designing a tight approximation to the full conformal prediction region that can be efficiently computed. Along with this approximate confidence region, a theoretical quantification of the tightness of this approximation is developed, depending on the smoothness assumptions on the loss and score functions. The new notion of thickness is introduced for quantifying the discrepancy between the approximate confidence region and the full conformal one.

---

### Recursive Meta-Distillation: An Axiomatic Framework for Iterative Knowledge Refinement

Authors: Aaron R. Flouro, Shawn P. Chadwick
Published: 2026-01-19
Categories: cs.LG
arXiv: https://arxiv.org/abs/2601.13100v1

Recent work in probability-domain knowledge distillation has established axiomatic frameworks for temperature scaling, multi-teacher aggregation, and bias-variance trade-offs in single-stage settings. However, the mathematical behavior of recursive or multi-generation distillation remains poorly understood, with prior approaches relying primarily on empirical heuristics. In this work, we introduce an axiomatic and operator-theoretic framework for recursive meta-distillation, formalizing iterative knowledge distillation as a sequence of probability-distribution operators with explicit anchoring to base teachers. We define structural axioms for valid meta-teacher construction and prove the existence of non-trivial operator families satisfying these axioms without specifying particular algorithms or loss functions. Under mild realizability and convexity assumptions, we show that anchored recursive distillation induces contraction in KL divergence, yielding geometric convergence to base teacher distributions and a unique, globally attractive fixed point. The contribution is foundational rather than algorithmic: the framework characterizes when recursive distillation is mathematically well-posed and convergent rather than error-accumulating, independent of model architecture, optimization details, or specific operator instantiations. These results provide a theoretical basis for understanding stability, bias-variance behavior, and failure modes in iterative and multi-teacher distillation under capacity constraints.

---

### Alexandria: A Multi-Domain Dialectal Arabic Machine Translation Dataset for Culturally Inclusive and Linguistically Diverse LLMs

Authors: Abdellah El Mekki, Samar M. Magdy, Houdaifa Atou, Ruwa AbuHweidi, Baraah Qawasmeh, Omer Nacar, Thikra Al-hibiri, Razan Saadie, Hamzah Alsayadi, Nadia Ghezaiel Hammouda, Alshima Alkhazimi, Aya Hamod et al.
Published: 2026-01-19
Categories: cs.CL
arXiv: https://arxiv.org/abs/2601.13099v2

Arabic is a highly diglossic language where most daily communication occurs in regional dialects rather than Modern Standard Arabic (MSA). Despite this, machine translation (MT) systems often generalize poorly to dialectal input, limiting their utility for millions of speakers. We introduce Alexandria, a large-scale, community-driven, human-translated dataset designed to bridge this gap. Alexandria covers 13 Arab countries and 11 high-impact domains, including health, education, and agriculture. Unlike previous resources, Alexandria provides unprecedented granularity by associating contributions with city-of-origin metadata, capturing authentic local varieties beyond coarse regional labels. The dataset consists of parallel English-Dialectal Arabic multi-turn conversational scenarios annotated with speaker-addressee gender configurations, enabling the study of gender-conditioned variation in dialectal use. Comprising 107K total turns, Alexandria serves as both a training resource and as a rigorous benchmark for evaluating MT and Large Language Models (LLMs). Our automatic and human evaluation benchmarks the current capabilities of Arabic-aware LLMs in translating across diverse Arabic dialects and sub-dialects while exposing significant persistent challenges. The Alexandria dataset, the creation prompts, the translation and revision guidelines, and the evaluation code are publicly available in the following repository: https://github.com/UBC-NLP/Alexandria

---

### RM -RF: Reward Model for Run-Free Unit Test Evaluation

Authors: Elena Bruches, Daniil Grebenkin, Mikhail Klementev, Vadim Alperovich, Roman Derunets, Dari Baturova, Georgy Mkrtchyan, Oleg Sedukhin, Ivan Bondarenko, Nikolay Bushkov, Stanislav Moiseev
Published: 2026-01-19
Categories: cs.SE, cs.LG
arXiv: https://arxiv.org/abs/2601.13097v1

We present RM-RF, a lightweight reward model for run-free evaluation of automatically generated unit tests. Instead of repeatedly compiling and executing candidate tests, RM-RF predicts - from source and test code alone - three execution-derived signals: (1) whether the augmented test suite compiles and runs successfully, (2) whether the generated test cases increase code coverage, and (3) whether the generated test cases improve the mutation kill rate. To train and evaluate RM-RF we assemble a multilingual dataset (Java, Python, Go) of focal files, test files, and candidate test additions labeled by an execution-based pipeline, and we release an associated dataset and methodology for comparative evaluation. We tested multiple model families and tuning regimes (zero-shot, full fine-tuning, and PEFT via LoRA), achieving an average F1 of 0.69 across the three targets. Compared to conventional compile-and-run instruments, RM-RF provides substantially lower latency and infrastructure cost while delivering competitive predictive fidelity, enabling fast, scalable feedback for large-scale test generation and RL-based code optimization.

---

### LLM-VLM Fusion Framework for Autonomous Maritime Port Inspection using a Heterogeneous UAV-USV System

Authors: Muhayy Ud Din, Waseem Akram, Ahsan B. Bakht, Irfan Hussain
Published: 2026-01-19
Categories: cs.RO, cs.CV
arXiv: https://arxiv.org/abs/2601.13096v1

Maritime port inspection plays a critical role in ensuring safety, regulatory compliance, and operational efficiency in complex maritime environments. However, existing inspection methods often rely on manual operations and conventional computer vision techniques that lack scalability and contextual understanding. This study introduces a novel integrated engineering framework that utilizes the synergy between Large Language Models (LLMs) and Vision Language Models (VLMs) to enable autonomous maritime port inspection using cooperative aerial and surface robotic platforms. The proposed framework replaces traditional state-machine mission planners with LLM-driven symbolic planning and improved perception pipelines through VLM-based semantic inspection, enabling context-aware and adaptive monitoring. The LLM module translates natural language mission instructions into executable symbolic plans with dependency graphs that encode operational constraints and ensure safe UAV-USV coordination. Meanwhile, the VLM module performs real-time semantic inspection and compliance assessment, generating structured reports with contextual reasoning. The framework was validated using the extended MBZIRC Maritime Simulator with realistic port infrastructure and further assessed through real-world robotic inspection trials. The lightweight on-board design ensures suitability for resource-constrained maritime platforms, advancing the development of intelligent, autonomous inspection systems. Project resources (code and videos) can be found here: https://github.com/Muhayyuddin/llm-vlm-fusion-port-inspection

---

### Patient-Conditioned Adaptive Offsets for Reliable Diagnosis across Subgroups

Authors: Gelei Xu, Yuying Duan, Jun Xia, Ruining Deng, Wei Jin, Yiyu Shi
Published: 2026-01-19
Categories: cs.CV
arXiv: https://arxiv.org/abs/2601.13094v1

AI models for medical diagnosis often exhibit uneven performance across patient populations due to heterogeneity in disease prevalence, imaging appearance, and clinical risk profiles. Existing algorithmic fairness approaches typically seek to reduce such disparities by suppressing sensitive attributes. However, in medical settings these attributes often carry essential diagnostic information, and removing them can degrade accuracy and reliability, particularly in high-stakes applications. In contrast, clinical decision making explicitly incorporates patient context when interpreting diagnostic evidence, suggesting a different design direction for subgroup-aware models. In this paper, we introduce HyperAdapt, a patient-conditioned adaptation framework that improves subgroup reliability while maintaining a shared diagnostic model. Clinically relevant attributes such as age and sex are encoded into a compact embedding and used to condition a hypernetwork-style module, which generates small residual modulation parameters for selected layers of a shared backbone. This design preserves the general medical knowledge learned by the backbone while enabling targeted adjustments that reflect patient-specific variability. To ensure efficiency and robustness, adaptations are constrained through low-rank and bottlenecked parameterizations, limiting both model complexity and computational overhead. Experiments across multiple public medical imaging benchmarks demonstrate that the proposed approach consistently improves subgroup-level performance without sacrificing overall accuracy. On the PAD-UFES-20 dataset, our method outperforms the strongest competing baseline by 4.1% in recall and 4.4% in F1 score, with larger gains observed for underrepresented patient populations.

---

### Adversarial News and Lost Profits: Manipulating Headlines in LLM-Driven Algorithmic Trading

Authors: Advije Rizvani, Giovanni Apruzzese, Pavel Laskov
Published: 2026-01-19
Categories: cs.CR, cs.LG
arXiv: https://arxiv.org/abs/2601.13082v1

Large Language Models (LLMs) are increasingly adopted in the financial domain. Their exceptional capabilities to analyse textual data make them well-suited for inferring the sentiment of finance-related news. Such feedback can be leveraged by algorithmic trading systems (ATS) to guide buy/sell decisions. However, this practice bears the risk that a threat actor may craft "adversarial news" intended to mislead an LLM. In particular, the news headline may include "malicious" content that remains invisible to human readers but which is still ingested by the LLM. Although prior work has studied textual adversarial examples, their system-wide impact on LLM-supported ATS has not yet been quantified in terms of monetary risk. To address this threat, we consider an adversary with no direct access to an ATS but able to alter stock-related news headlines on a single day. We evaluate two human-imperceptible manipulations in a financial context: Unicode homoglyph substitutions that misroute models during stock-name recognition, and hidden-text clauses that alter the sentiment of the news headline. We implement a realistic ATS in Backtrader that fuses an LSTM-based price forecast with LLM-derived sentiment (FinBERT, FinGPT, FinLLaMA, and six general-purpose LLMs), and quantify monetary impact using portfolio metrics. Experiments on real-world data show that manipulating a one-day attack over 14 months can reliably mislead LLMs and reduce annual returns by up to 17.7 percentage points. To assess real-world feasibility, we analyze popular scraping libraries and trading platforms and survey 27 FinTech practitioners, confirming our hypotheses. We notified trading platform owners of this security issue.

---

### Polychronous Wave Computing: Timing-Native Address Selection in Spiking Networks

Authors: Natalila G. Berloff
Published: 2026-01-19
Categories: cond-mat.dis-nn, cs.LG, cs.NE, physics.optics
arXiv: https://arxiv.org/abs/2601.13079v1

Spike timing offers a combinatorial address space, suggesting that timing-based spiking inference can be executed as lookup and routing rather than as dense multiply--accumulate. Yet most neuromorphic and photonic systems still digitize events into timestamps, bins, or rates and then perform selection in clocked logic. We introduce Polychronous Wave Computing (PWC), a timing-native address-selection primitive that maps relative spike latencies directly to a discrete output route in the wave domain. Spike times are phase-encoded in a rotating frame and processed by a programmable multiport interferometer that evaluates K template correlations in parallel; a driven--dissipative winner-take-all stage then performs a physical argmax, emitting a one-hot output port. We derive the operating envelope imposed by phase wrapping and mutual coherence, and collapse timing jitter, static phase mismatch, and dephasing into a single effective phase-noise budget whose induced winner--runner-up margin predicts boundary-first failures and provides an intensity-only calibration target. Simulations show that nonlinear competition improves routing fidelity compared with noisy linear intensity readout, and that hardware-in-the-loop phase tuning rescues a temporal-order gate from 55.9% to 97.2% accuracy under strong static mismatch. PWC provides a fast routing coprocessor for LUT-style spiking networks and sparse top-1 gates (e.g., mixture-of-experts routing) across polaritonic, photonic, and oscillator platforms.

---

### METIS: Mentoring Engine for Thoughtful Inquiry & Solutions

Authors: Abhinav Rajeev Kumar, Dhruv Trehan, Paras Chopra
Published: 2026-01-19
Categories: cs.LG, cs.AI
arXiv: https://arxiv.org/abs/2601.13075v1

Many students lack access to expert research mentorship. We ask whether an AI mentor can move undergraduates from an idea to a paper. We build METIS, a tool-augmented, stage-aware assistant with literature search, curated guidelines, methodology checks, and memory. We evaluate METIS against GPT-5 and Claude Sonnet 4.5 across six writing stages using LLM-as-a-judge pairwise preferences, student-persona rubrics, short multi-turn tutoring, and evidence/compliance checks. On 90 single-turn prompts, LLM judges preferred METIS to Claude Sonnet 4.5 in 71% and to GPT-5 in 54%. Student scores (clarity/actionability/constraint-fit; 90 prompts x 3 judges) are higher across stages. In multi-turn sessions (five scenarios/agent), METIS yields slightly higher final quality than GPT-5. Gains concentrate in document-grounded stages (D-F), consistent with stage-aware routing and groundings failure modes include premature tool routing, shallow grounding, and occasional stage misclassification.

---

### MagicGUI-RMS: A Multi-Agent Reward Model System for Self-Evolving GUI Agents via Automated Feedback Reflux

Authors: Zecheng Li, Zhihui Cao, Wenke Huang, Yudong Zhang, Keying Qi, Rui Wang, Zeyu Zheng, Jian Zhao, Hao Zhu, Hengxin Wu, Yuran Wang, Guitao Fan et al.
Published: 2026-01-19
Categories: cs.AI
arXiv: https://arxiv.org/abs/2601.13060v1

Graphical user interface (GUI) agents are rapidly progressing toward autonomous interaction and reliable task execution across diverse applications. However, two central challenges remain unresolved: automating the evaluation of agent trajectories and generating high-quality training data at scale to enable continual improvement. Existing approaches often depend on manual annotation or static rule-based verification, which restricts scalability and limits adaptability in dynamic environments. We present MagicGUI-RMS, a multi-agent reward model system that delivers adaptive trajectory evaluation, corrective feedback, and self-evolving learning capabilities. MagicGUI-RMS integrates a Domain-Specific Reward Model (DS-RM) with a General-Purpose Reward Model (GP-RM), enabling fine-grained action assessment and robust generalization across heterogeneous GUI tasks. To support reward learning at scale, we design a structured data construction pipeline that automatically produces balanced and diverse reward datasets, effectively reducing annotation costs while maintaining sample fidelity. During execution, the reward model system identifies erroneous actions, proposes refined alternatives, and continuously enhances agent behavior through an automated data-reflux mechanism. Extensive experiments demonstrate that MagicGUI-RMS yields substantial gains in task accuracy, behavioral robustness. These results establish MagicGUI-RMS as a principled and effective foundation for building self-improving GUI agents driven by reward-based adaptation.

---

### Prototype Learning-Based Few-Shot Segmentation for Low-Light Crack on Concrete Structures

Authors: Yulun Guo
Published: 2026-01-19
Categories: cs.CV
arXiv: https://arxiv.org/abs/2601.13059v1

Crack detection is critical for concrete infrastructure safety, but real-world cracks often appear in low-light environments like tunnels and bridge undersides, degrading computer vision segmentation accuracy. Pixel-level annotation of low-light crack images is extremely time-consuming, yet most deep learning methods require large, well-illuminated datasets. We propose a dual-branch prototype learning network integrating Retinex theory with few-shot learning for low-light crack segmentation. Retinex-based reflectance components guide illumination-invariant global representation learning, while metric learning reduces dependence on large annotated datasets. We introduce a cross-similarity prior mask generation module that computes high-dimensional similarities between query and support features to capture crack location and structure, and a multi-scale feature enhancement module that fuses multi-scale features with the prior mask to alleviate spatial inconsistency. Extensive experiments on multiple benchmarks demonstrate consistent state-of-the-art performance under low-light conditions. Code: https://github.com/YulunGuo/CrackFSS.

---

### TinyML-Enabled IoT for Sustainable Precision Irrigation

Authors: Kamogelo Taueatsoala, Caitlyn Daniels, Angelina J. Ramsunar, Petrus Bronkhorst, Absalom E. Ezugwu
Published: 2026-01-19
Categories: cs.LG, cs.AI
arXiv: https://arxiv.org/abs/2601.13054v1

Small-scale farming communities are disproportionately affected by water scarcity, erratic climate patterns, and a lack of access to advanced, affordable agricultural technologies. To address these challenges, this paper presents a novel, edge-first IoT framework that integrates Tiny Machine Learning (TinyML) for intelligent, offline-capable precision irrigation. The proposed four-layer architecture leverages low-cost hardware, an ESP32 microcontroller as an edge inference node, and a Raspberry Pi as a local edge server to enable autonomous decision-making without cloud dependency. The system utilizes capacitive soil moisture, temperature, humidity, pH, and ambient light sensors for environmental monitoring. A rigorous comparative analysis of ensemble models identified gradient boosting as superior, achieving an R^2 score of 0.9973 and a Mean Absolute Percentage Error (MAPE) of 0.99%, outperforming a random forest model (R^2 = 0.9916, MAPE = 1.81%). This optimized model was converted and deployed as a lightweight TinyML inference engine on the ESP32 and predicts irrigation needs with exceptional accuracy (MAPE < 1%). Local communication is facilitated by an MQTT-based LAN protocol, ensuring reliable operation in areas with limited or no internet connectivity. Experimental validation in a controlled environment demonstrated a significant reduction in water usage compared to traditional methods, while the system's low-power design and offline functionality confirm its viability for sustainable, scalable deployment in resource-constrained rural settings. This work provides a practical, cost-effective blueprint for bridging the technological divide in agriculture and enhancing water-use efficiency through on-device artificial intelligence.

---

### GridNet-HD: A High-Resolution Multi-Modal Dataset for LiDAR-Image Fusion on Power Line Infrastructure

Authors: Antoine Carreaud, Shanci Li, Malo De Lacour, Digre Frinde, Jan Skaloud, Adrien Gressin
Published: 2026-01-19
Categories: cs.CV
arXiv: https://arxiv.org/abs/2601.13052v1

This paper presents GridNet-HD, a multi-modal dataset for 3D semantic segmentation of overhead electrical infrastructures, pairing high-density LiDAR with high-resolution oblique imagery. The dataset comprises 7,694 images and 2.5 billion points annotated into 11 classes, with predefined splits and mIoU metrics. Unimodal (LiDAR-only, image-only) and multi-modal fusion baselines are provided. On GridNet-HD, fusion models outperform the best unimodal baseline by +5.55 mIoU, highlighting the complementarity of geometry and appearance. As reviewed in Sec. 2, no public dataset jointly provides high-density LiDAR and high-resolution oblique imagery with 3D semantic labels for power-line assets. Dataset, baselines, and codes are available: https://huggingface.co/collections/heig-vd-geo/gridnet-hd.

---

### Profiling German Text Simplification with Interpretable Model-Fingerprints

Authors: Lars Klöser, Mika Beele, Bodo Kraft
Published: 2026-01-19
Categories: cs.CL
arXiv: https://arxiv.org/abs/2601.13050v1

While Large Language Models (LLMs) produce highly nuanced text simplifications, developers currently lack tools for a holistic, efficient, and reproducible diagnosis of their behavior. This paper introduces the Simplification Profiler, a diagnostic toolkit that generates a multidimensional, interpretable fingerprint of simplified texts. Multiple aggregated simplifications of a model result in a model's fingerprint. This novel evaluation paradigm is particularly vital for languages, where the data scarcity problem is magnified when creating flexible models for diverse target groups rather than a single, fixed simplification style. We propose that measuring a model's unique behavioral signature is more relevant in this context as an alternative to correlating metrics with human preferences. We operationalize this with a practical meta-evaluation of our fingerprints' descriptive power, which bypasses the need for large, human-rated datasets. This test measures if a simple linear classifier can reliably identify various model configurations by their created simplifications, confirming that our metrics are sensitive to a model's specific characteristics. The Profiler can distinguish high-level behavioral variations between prompting strategies and fine-grained changes from prompt engineering, including few-shot examples. Our complete feature set achieves classification F1-scores up to 71.9 %, improving upon simple baselines by over 48 percentage points. The Simplification Profiler thus offers developers a granular, actionable analysis to build more effective and truly adaptive text simplification systems.

---

### Analysis of Long Range Dependency Understanding in State Space Models

Authors: Srividya Ravikumar, Abhinav Anand, Shweta Verma, Mira Mezini
Published: 2026-01-19
Categories: cs.LG, cs.AI
arXiv: https://arxiv.org/abs/2601.13048v1

Although state-space models (SSMs) have demonstrated strong performance on long-sequence benchmarks, most research has emphasized predictive accuracy rather than interpretability. In this work, we present the first systematic kernel interpretability study of the diagonalized state-space model (S4D) trained on a real-world task (vulnerability detection in source code). Through time and frequency domain analysis of the S4D kernel, we show that the long-range modeling capability of S4D varies significantly under different model architectures, affecting model performance. For instance, we show that the depending on the architecture, S4D kernel can behave as low-pass, band-pass or high-pass filter. The insights from our analysis can guide future work in designing better S4D-based models.

---

### Typhoon ASR Real-time: FastConformer-Transducer for Thai Automatic Speech Recognition

Authors: Warit Sirichotedumrong, Adisai Na-Thalang, Potsawee Manakul, Pittawat Taveekitworachai, Sittipong Sripaisarnmongkol, Kunat Pipatanakul
Published: 2026-01-19
Categories: cs.CL
arXiv: https://arxiv.org/abs/2601.13044v1

Large encoder-decoder models like Whisper achieve strong offline transcription but remain impractical for streaming applications due to high latency. However, due to the accessibility of pre-trained checkpoints, the open Thai ASR landscape remains dominated by these offline architectures, leaving a critical gap in efficient streaming solutions. We present Typhoon ASR Real-time, a 115M-parameter FastConformer-Transducer model for low-latency Thai speech recognition. We demonstrate that rigorous text normalization can match the impact of model scaling: our compact model achieves a 45x reduction in computational cost compared to Whisper Large-v3 while delivering comparable accuracy. Our normalization pipeline resolves systemic ambiguities in Thai transcription --including context-dependent number verbalization and repetition markers (mai yamok) --creating consistent training targets. We further introduce a two-stage curriculum learning approach for Isan (north-eastern) dialect adaptation that preserves Central Thai performance. To address reproducibility challenges in Thai ASR, we release the Typhoon ASR Benchmark, a gold-standard human-labeled datasets with transcriptions following established Thai linguistic conventions, providing standardized evaluation protocols for the research community.

---

### SASA: Semantic-Aware Contrastive Learning Framework with Separated Attention for Triple Classification

Authors: Xu Xiaodan, Hu Xiaolin
Published: 2026-01-19
Categories: cs.CL, cs.LG
arXiv: https://arxiv.org/abs/2601.13035v1

Knowledge Graphs~(KGs) often suffer from unreliable knowledge, which restricts their utility. Triple Classification~(TC) aims to determine the validity of triples from KGs. Recently, text-based methods learn entity and relation representations from natural language descriptions, significantly improving the generalization capabilities of TC models and setting new benchmarks in performance. However, there are still two critical challenges. First, existing methods often ignore the effective semantic interaction among different KG components. Second, most approaches adopt single binary classification training objective, leading to insufficient semantic representation learning. To address these challenges, we propose \textbf{SASA}, a novel framework designed to enhance TC models via separated attention mechanism and semantic-aware contrastive learning~(CL). Specifically, we first propose separated attention mechanism to encode triples into decoupled contextual representations and then fuse them through a more effective interactive way. Then, we introduce semantic-aware hierarchical CL as auxiliary training objective to guide models in improving their discriminative capabilities and achieving sufficient semantic learning, considering both local level and global level CL. Experimental results across two benchmark datasets demonstrate that SASA significantly outperforms state-of-the-art methods. In terms of accuracy, we advance the state-of-the-art by +5.9\% on FB15k-237 and +3.4\% on YAGO3-10.

---

### Think3D: Thinking with Space for Spatial Reasoning

Authors: Zaibin Zhang, Yuhan Wu, Lianjie Jia, Yifan Wang, Zhongbo Zhang, Yijiang Li, Binghao Ran, Fuxi Zhang, Zhuohan Sun, Yizhuang Peng, Zhenfei Yin, Lijun Wang et al.
Published: 2026-01-19
Categories: cs.CV
arXiv: https://arxiv.org/abs/2601.13029v4

While Vision-Language Models (VLMs) excel at 2D visual understanding, they remain constrained by 2D-centric paradigm that severely limits genuine 3D spatial reasoning. To bridge this gap, we introduce Think3D, a novel framework that equips VLM agents with interactive, 3D chain-of-thought reasoning capabilities. By integrating a suite of 3D manipulation tools, Think3D transforms perception into active spatial exploration, mirroring human geometric reasoning. Think3D consistently improves proprietary models, including GPT-4.1 and Gemini 2.5 Pro, across BLINK Multi-view, MindCube-1K, and VSI-Bench-Tiny. We further propose Think3D-RL to teach smaller open-weight models how to manipulate 3D space effectively. Using only final-answer rewards, without process supervision or handcrafted exploration trajectories, Think3D-RL enables Qwen3-VL-4B to autonomously learn effective 3D exploration strategies. After training, the model exhibits tool-use patterns similar to those of stronger proprietary models, while shifting the effect of 3D tool use on MindCube-1K from a performance drop to a substantial improvement. These results show that active exploration in 3D space provides an effective and general paradigm for improving spatial reasoning in multimodal agents. Code, models, and data are available at https://github.com/zhangzaibin/spagent.

---

### Tears or Cheers? Benchmarking LLMs via Culturally Elicited Distinct Affective Responses

Authors: Chongyuan Dai, Yaling Shen, Zihan Gao, Jia Li, Yishun Jiang, Yaxiong Wang, Liu Liu, Zongyuan Ge, Jinpeng Hu
Published: 2026-01-19
Categories: cs.CL
arXiv: https://arxiv.org/abs/2601.13024v2

Culture serves as a fundamental determinant of human affective processing and profoundly shapes how individuals perceive and interpret emotional stimuli. Despite this intrinsic link extant evaluations regarding cultural alignment within Large Language Models primarily prioritize declarative knowledge such as geographical facts or established societal customs. These benchmarks remain insufficient to capture the subjective interpretative variance inherent to diverse sociocultural lenses. To address this limitation, we introduce CEDAR, a multimodal benchmark constructed entirely from scenarios capturing Culturally \underline{\textsc{E}}licited \underline{\textsc{D}}istinct \underline{\textsc{A}}ffective \underline{\textsc{R}}esponses. To construct CEDAR, we implement a novel pipeline that leverages LLM-generated provisional labels to isolate instances yielding cross-cultural emotional distinctions, and subsequently derives reliable ground-truth annotations through rigorous human evaluation. The resulting benchmark comprises 10,962 instances across seven languages and 14 fine-grained emotion categories, with each language including 400 multimodal and 1,166 text-only samples. Comprehensive evaluations of 17 representative multilingual models reveal a dissociation between language consistency and cultural alignment, demonstrating that culturally grounded affective understanding remains a significant challenge for current models.

---

### Enhancing Generalization in Sickle Cell Disease Diagnosis through Ensemble Methods and Feature Importance Analysis

Authors: Nataša Petrović, Gabriel Moyà-Alcover, Antoni Jaume-i-Capó, Jose Maria Buades Rubio
Published: 2026-01-19
Categories: cs.LG
arXiv: https://arxiv.org/abs/2601.13021v1

This work presents a novel approach for selecting the optimal ensemble-based classification method and features with a primarly focus on achieving generalization, based on the state-of-the-art, to provide diagnostic support for Sickle Cell Disease using peripheral blood smear images of red blood cells. We pre-processed and segmented the microscopic images to ensure the extraction of high-quality features. To ensure the reliability of our proposed system, we conducted an in-depth analysis of interpretability. Leveraging techniques established in the literature, we extracted features from blood cells and employed ensemble machine learning methods to classify their morphology. Furthermore, we have devised a methodology to identify the most critical features for classification, aimed at reducing complexity and training time and enhancing interpretability in opaque models. Lastly, we validated our results using a new dataset, where our model overperformed state-of-the-art models in terms of generalization. The results of classifier ensembled of Random Forest and Extra Trees classifier achieved an harmonic mean of precision and recall (F1-score) of 90.71\% and a Sickle Cell Disease diagnosis support score (SDS-score) of 93.33\%. These results demonstrate notable enhancement from previous ones with Gradient Boosting classifier (F1-score 87.32\% and SDS-score 89.51\%). To foster scientific progress, we have made available the parameters for each model, the implemented code library, and the confusion matrices with the raw data.

---

### PASs-MoE: Mitigating Misaligned Co-drift among Router and Experts via Pathway Activation Subspaces for Continual Learning

Authors: Zhiyan Hou, Haiyun Guo, Haokai Ma, Yandu Sun, Yonghui Yang, Jinqiao Wang
Published: 2026-01-19
Categories: cs.LG, cs.AI
arXiv: https://arxiv.org/abs/2601.13020v2

Continual instruction tuning (CIT) requires multimodal large language models (MLLMs) to adapt to a stream of tasks without forgetting prior capabilities. A common strategy is to isolate updates by routing inputs to different LoRA experts. However, existing LoRA-based Mixture-of-Experts (MoE) methods often jointly update the router and experts in an indiscriminate way, causing the router's preferences to co-drift with experts' adaptation pathways and gradually deviate from early-stage input--expert specialization. We term this as Misaligned Co-drift, which blurs expert responsibilities and exacerbates forgetting. To address this, we introduce the pathway activation subspace (PASs), a LoRA-induced subspace that reflects which low-rank pathway directions an input activates in each expert, providing a capability-aligned coordinate system for routing and preservation. Based on PASs, we propose a fixed-capacity PASs-based MoE--LoRA method with two components: PAS-guided Reweighting, which calibrates routing using each expert's pathway activation signals, and PAS-aware Rank Stabilization, which selectively stabilizes rank directions important to previous tasks. Experiments on a CIT benchmark show that our approach consistently outperforms a range of conventional continual learning baselines and MoE--LoRA variants in both accuracy and resistance to forgetting, without increasing model parameters. Our code is publicly available at https://github.com/yueluoshuangtian/PASs-MoE.

---

### Bi-Attention HateXplain : Taking into account the sequential aspect of data during explainability in a multi-task context

Authors: Ghislain Dorian Tchuente Mondjo
Published: 2026-01-19
Categories: cs.CL, cs.AI
arXiv: https://arxiv.org/abs/2601.13018v1

Technological advances in the Internet and online social networks have brought many benefits to humanity. At the same time, this growth has led to an increase in hate speech, the main global threat. To improve the reliability of black-box models used for hate speech detection, post-hoc approaches such as LIME, SHAP, and LRP provide the explanation after training the classification model. In contrast, multi-task approaches based on the HateXplain benchmark learn to explain and classify simultaneously. However, results from HateXplain-based algorithms show that predicted attention varies considerably when it should be constant. This attention variability can lead to inconsistent interpretations, instability of predictions, and learning difficulties. To solve this problem, we propose the BiAtt-BiRNN-HateXplain (Bidirectional Attention BiRNN HateXplain) model which is easier to explain compared to LLMs which are more complex in view of the need for transparency, and will take into account the sequential aspect of the input data during explainability thanks to a BiRNN layer. Thus, if the explanation is correctly estimated, thanks to multi-task learning (explainability and classification task), the model could classify better and commit fewer unintentional bias errors related to communities. The experimental results on HateXplain data show a clear improvement in detection performance, explainability and a reduction in unintentional bias.

---

### HT-GNN: Hyper-Temporal Graph Neural Network for Customer Lifetime Value Prediction in Baidu Ads

Authors: Xiaohui Zhao, Xinjian Zhao, Jiahui Zhang, Guoyu Liu, Houzhi Wang, Shu Wu
Published: 2026-01-19
Categories: cs.LG, cs.AI
arXiv: https://arxiv.org/abs/2601.13013v1

Lifetime value (LTV) prediction is crucial for news feed advertising, enabling platforms to optimize bidding and budget allocation for long-term revenue growth. However, it faces two major challenges: (1) demographic-based targeting creates segment-specific LTV distributions with large value variations across user groups; and (2) dynamic marketing strategies generate irregular behavioral sequences where engagement patterns evolve rapidly. We propose a Hyper-Temporal Graph Neural Network (HT-GNN), which jointly models demographic heterogeneity and temporal dynamics through three key components: (i) a hypergraph-supervised module capturing inter-segment relationships; (ii) a transformer-based temporal encoder with adaptive weighting; and (iii) a task-adaptive mixture-of-experts with dynamic prediction towers for multi-horizon LTV forecasting. Experiments on \textit{Baidu Ads} with 15 million users demonstrate that HT-GNN consistently outperforms state-of-the-art methods across all metrics and prediction horizons.

---

### ArchAgent: Scalable Legacy Software Architecture Recovery with LLMs

Authors: Rusheng Pan, Bingcheng Mao, Tianyi Ma, Zhenhua Ling
Published: 2026-01-19
Categories: cs.SE, cs.AI
arXiv: https://arxiv.org/abs/2601.13007v1

Recovering accurate architecture from large-scale legacy software is hindered by architectural drift, missing relations, and the limited context of Large Language Models (LLMs). We present ArchAgent, a scalable agent-based framework that combines static analysis, adaptive code segmentation, and LLM-powered synthesis to reconstruct multiview, business-aligned architectures from cross-repository codebases. ArchAgent introduces scalable diagram generation with contextual pruning and integrates cross-repository data to identify business-critical modules. Evaluations of typical large-scale GitHub projects show significant improvements over existing benchmarks. An ablation study confirms that dependency context improves the accuracy of generated architectures of production-level repositories, and a real-world case study demonstrates effective recovery of critical business logics from legacy projects. The dataset is available at https://github.com/panrusheng/arch-eval-benchmark.

---

### OFA-MAS: One-for-All Multi-Agent System Topology Design based on Mixture-of-Experts Graph Generative Models

Authors: Shiyuan Li, Yixin Liu, Yu Zheng, Mei Li, Quoc Viet Hung Nguyen, Shirui Pan
Published: 2026-01-19
Categories: cs.MA, cs.LG
arXiv: https://arxiv.org/abs/2601.12996v1

Multi-Agent Systems (MAS) offer a powerful paradigm for solving complex problems, yet their performance is critically dependent on the design of their underlying collaboration topology. As MAS become increasingly deployed in web services (e.g., search engines), designing adaptive topologies for diverse cross-domain user queries becomes essential. Current graph learning-based design methodologies often adhere to a "one-for-one" paradigm, where a specialized model is trained for each specific task domain. This approach suffers from poor generalization to unseen domains and fails to leverage shared structural knowledge across different tasks. To address this, we propose OFA-TAD, a one-for-all framework that generates adaptive collaboration graphs for any task described in natural language through a single universal model. Our approach integrates a Task-Aware Graph State Encoder (TAGSE) that filters task-relevant node information via sparse gating, and a Mixture-of-Experts (MoE) architecture that dynamically selects specialized sub-networks to drive node and edge prediction. We employ a three-stage training strategy: unconditional pre-training on canonical topologies for structural priors, large-scale conditional pre-training on LLM-generated datasets for task-topology mappings, and supervised fine-tuning on empirically validated graphs. Experiments across six diverse benchmarks show that OFA-TAD significantly outperforms specialized one-for-one models, generating highly adaptive MAS topologies. Code: https://github.com/Shiy-Li/OFA-MAS.

---

### Graph Reasoning Paradigm: Structured and Symbolic Reasoning with Topology-Aware Reinforcement Learning for Large Language Models

Authors: Runxuan Liu, Xianhao Ou, Xinyan Ma, Jiyuan Wang, Jiafeng Liang, Jiaqi Li, Tao He, Zheng Chu, Rongchuan Mu, Zekun Wang, Baoxin Wang, Dayong Wu et al.
Published: 2026-01-19
Categories: cs.CL
arXiv: https://arxiv.org/abs/2601.12995v1

Long Chain-of-Thought (LCoT), achieved by Reinforcement Learning with Verifiable Rewards (RLVR), has proven effective in enhancing the reasoning capabilities of Large Language Models (LLMs). However, reasoning in current LLMs is primarily generated as plain text, where performing semantic evaluation on such unstructured data creates a computational bottleneck during training. Despite RLVR-based optimization, existing methods still suffer from coarse-grained supervision, reward hacking, high training costs, and poor generalization. To address these issues, we propose the Graph Reasoning Paradigm (GRP), which realizes structured and symbolic reasoning, implemented via graph-structured representations with step-level cognitive labels. Building upon GRP, we further design Process-Aware Stratified Clipping Group Relative Policy Optimization (PASC-GRPO), which leverages structured evaluation to replace semantic evaluation, achieves process-aware verification through graph-structured outcome rewards, and mitigates reward hacking via stratified clipping advantage estimation. Experiments demonstrate significant improvements across mathematical reasoning and code generation tasks. Data, models, and code will be released later.

---

### AsyncBEV: Cross-modal Flow Alignment in Asynchronous 3D Object Detection

Authors: Shiming Wang, Holger Caesar, Liangliang Nan, Julian F. P. Kooij
Published: 2026-01-19
Categories: cs.CV
arXiv: https://arxiv.org/abs/2601.12994v1

In autonomous driving, multi-modal perception tasks like 3D object detection typically rely on well-synchronized sensors, both at training and inference. However, despite the use of hardware- or software-based synchronization algorithms, perfect synchrony is rarely guaranteed: Sensors may operate at different frequencies, and real-world factors such as network latency, hardware failures, or processing bottlenecks often introduce time offsets between sensors. Such asynchrony degrades perception performance, especially for dynamic objects. To address this challenge, we propose AsyncBEV, a trainable lightweight and generic module to improve the robustness of 3D Birds' Eye View (BEV) object detection models against sensor asynchrony. Inspired by scene flow estimation, AsyncBEV first estimates the 2D flow from the BEV features of two different sensor modalities, taking into account the known time offset between these sensor measurements. The predicted feature flow is then used to warp and spatially align the feature maps, which we show can easily be integrated into different current BEV detector architectures (e.g., BEV grid-based and token-based). Extensive experiments demonstrate AsyncBEV improves robustness against both small and large asynchrony between LiDAR or camera sensors in both the token-based CMT and grid-based UniBEV, especially for dynamic objects. We significantly outperform the ego motion compensated CMT and UniBEV baselines, notably by $16.6$ % and $11.9$ % NDS on dynamic objects in the worst-case scenario of a $0.5 s$ time offset. Code will be released upon acceptance.

---

### RAGExplorer: A Visual Analytics System for the Comparative Diagnosis of RAG Systems

Authors: Haoyu Tian, Yingchaojie Feng, Zhen Wen, Haoxuan Li, Minfeng Zhu, Wei Chen
Published: 2026-01-19
Categories: cs.HC, cs.CL
arXiv: https://arxiv.org/abs/2601.12991v2

The advent of Retrieval-Augmented Generation (RAG) has significantly enhanced the ability of Large Language Models (LLMs) to produce factually accurate and up-to-date responses. However, the performance of a RAG system is not determined by a single component but emerges from a complex interplay of modular choices, such as embedding models and retrieval algorithms. This creates a vast and often opaque configuration space, making it challenging for developers to understand performance trade-offs and identify optimal designs. To address this challenge, we present RAGExplorer, a visual analytics system for the systematic comparison and diagnosis of RAG configurations. RAGExplorer guides users through a seamless macro-to-micro analytical workflow. Initially, it empowers developers to survey the performance landscape across numerous configurations, allowing for a high-level understanding of which design choices are most effective. For a deeper analysis, the system enables users to drill down into individual failure cases, investigate how differences in retrieved information contribute to errors, and interactively test hypotheses by manipulating the provided context to observe the resulting impact on the generated answer. We demonstrate the effectiveness of RAGExplorer through detailed case studies and user studies, validating its ability to empower developers in navigating the complex RAG design space. Our code and user guide are publicly available at https://github.com/Thymezzz/RAGExplorer.

---

### Beyond Visual Realism: Toward Reliable Financial Time Series Generation

Authors: Fan Zhang, Jiabin Luo, Zheng Zhang, Shuanghong Huang, Zhipeng Liu, Yu Chen
Published: 2026-01-19
Categories: q-fin.ST, cs.LG
arXiv: https://arxiv.org/abs/2601.12990v1

Generative models for financial time series often create data that look realistic and even reproduce stylized facts such as fat tails or volatility clustering. However, these apparent successes break down under trading backtests: models like GANs or WGAN-GP frequently collapse, yielding extreme and unrealistic results that make the synthetic data unusable in practice. We identify the root cause in the neglect of financial asymmetry and rare tail events, which strongly affect market risk but are often overlooked by objectives focusing on distribution matching. To address this, we introduce the Stylized Facts Alignment GAN (SFAG), which converts key stylized facts into differentiable structural constraints and jointly optimizes them with adversarial loss. This multi-constraint design ensures that generated series remain aligned with market dynamics not only in plots but also in backtesting. Experiments on the Shanghai Composite Index (2004--2024) show that while baseline GANs produce unstable and implausible trading outcomes, SFAG generates synthetic data that preserve stylized facts and support robust momentum strategy performance. Our results highlight that structure-preserving objectives are essential to bridge the gap between superficial realism and practical usability in financial generative modeling.

---

### PaperGuide: Making Small Language-Model Paper-Reading Agents More Efficient

Authors: Zijian Wang, Tiancheng Huang, Hanqi Li, Da Ma, Lu Chen, Kai Yu
Published: 2026-01-19
Categories: cs.LG
arXiv: https://arxiv.org/abs/2601.12988v1

The accelerating growth of the scientific literature makes it increasingly difficult for researchers to track new advances through manual reading alone. Recent progress in large language models (LLMs) has therefore spurred interest in autonomous agents that can read scientific papers and extract task-relevant information. However, most existing approaches rely either on heavily engineered prompting or on a conventional SFT-RL training pipeline, both of which often lead to excessive and low-yield exploration. Drawing inspiration from cognitive science, we propose PaperCompass, a framework that mitigates these issues by separating high-level planning from fine-grained execution. PaperCompass first drafts an explicit plan that outlines the intended sequence of actions, and then performs detailed reasoning to instantiate each step by selecting the parameters for the corresponding function calls. To train such behavior, we introduce Draft-and-Follow Policy Optimization (DFPO), a tailored RL method that jointly optimizes both the draft plan and the final solution. DFPO can be viewed as a lightweight form of hierarchical reinforcement learning, aimed at narrowing the `knowing-doing' gap in LLMs. We provide a theoretical analysis that establishes DFPO's favorable optimization properties, supporting a stable and reliable training process. Experiments on paper-based question answering (Paper-QA) benchmarks show that PaperCompass improves efficiency over strong baselines without sacrificing performance, achieving results comparable to much larger models.

---

### ChartAttack: Testing the Vulnerability of LLMs to Malicious Prompting in Chart Generation

Authors: Jesus-German Ortiz-Barajas, Jonathan Tonglet, Vivek Gupta, Iryna Gurevych
Published: 2026-01-19
Categories: cs.CL
arXiv: https://arxiv.org/abs/2601.12983v4

Multimodal large language models (MLLMs) are increasingly used to automate chart generation from data tables, improving efficiency but introducing new misuse risks. We present ChartAttack, a framework for evaluating how MLLMs use design misleaders to generate charts that induce incorrect interpretations. We also introduce AttackViz, a chart question-answering (QA) dataset labeled with effective misleaders and their induced incorrect answers. ChartAttack reduces MLLM QA accuracy by 17.2 points in-domain and 11.9 points cross-domain. Conditional deception rates show targeted effects: correct answers shift to attacker-intended answers 11.2\% of the time in-domain and 11.7-14.9\% cross-domain, while originally incorrect answers rarely change. A controlled human study shows that ChartAttack-generated charts also reduce human QA performance. Finally, fine-tuning on AttackViz improves in-domain MLLM robustness to misleading charts. Our findings highlight the need for secure, robust MLLM chart generation. Code and data are publicly available on the project website.

---

### Early Prediction of Type 2 Diabetes Using Multimodal data and Tabular Transformers

Authors: Sulaiman Khan, Md. Rafiul Biswas, Zubair Shah
Published: 2026-01-19
Categories: cs.CV, cs.LG
arXiv: https://arxiv.org/abs/2601.12981v1

This study introduces a novel approach for early Type 2 Diabetes Mellitus (T2DM) risk prediction using a tabular transformer (TabTrans) architecture to analyze longitudinal patient data. By processing patients` longitudinal health records and bone-related tabular data, our model captures complex, long-range dependencies in disease progression that conventional methods often overlook. We validated our TabTrans model on a retrospective Qatar BioBank (QBB) cohort of 1,382 subjects, comprising 725 men (146 diabetic, 579 healthy) and 657 women (133 diabetic, 524 healthy). The study integrated electronic health records (EHR) with dual-energy X-ray absorptiometry (DXA) data. To address class imbalance, we employed SMOTE and SMOTE-ENN resampling techniques. The proposed model`s performance is evaluated against conventional machine learning (ML) and generative AI models, including Claude 3.5 Sonnet (Anthropic`s constitutional AI), GPT-4 (OpenAI`s generative pre-trained transformer), and Gemini Pro (Google`s multimodal language model). Our TabTrans model demonstrated superior predictive performance, achieving ROC AUC $\geq$ 79.7 % for T2DM prediction compared to both generative AI models and conventional ML approaches. Feature interpretation analysis identified key risk indicators, with visceral adipose tissue (VAT) mass and volume, ward bone mineral density (BMD) and bone mineral content (BMC), T and Z-scores, and L1-L4 scores emerging as the most important predictors associated with diabetes development in Qatari adults. These findings demonstrate the significant potential of TabTrans for analyzing complex tabular healthcare data, providing a powerful tool for proactive T2DM management and personalized clinical interventions in the Qatari population. Index Terms: tabular transformers, multimodal data, DXA data, diabetes, T2DM, feature interpretation, tabular data

---

### The Bitter Lesson of Diffusion Language Models for Agentic Workflows: A Comprehensive Reality Check

Authors: Qingyu Lu, Liang Ding, Kanjian Zhang, Jinxia Zhang, Dacheng Tao
Published: 2026-01-19
Categories: cs.CL
arXiv: https://arxiv.org/abs/2601.12979v3

The pursuit of real-time agentic interaction has driven interest in Diffusion-based Large Language Models (dLLMs) as alternatives to auto-regressive backbones, promising to break the sequential latency bottleneck. However, does such efficiency gains translate into effective agentic behavior? In this work, we present a comprehensive evaluation of dLLMs (e.g., LLaDA, Dream) across two distinct agentic paradigms: Embodied Agents (requiring long-horizon planning) and Tool-Calling Agents (requiring precise formatting). Contrary to the efficiency hype, our results on Agentboard and BFCL reveal a "bitter lesson": current dLLMs fail to serve as reliable agentic backbones, frequently leading to systematically failure. (1) In Embodied settings, dLLMs suffer repeated attempts, failing to branch under temporal feedback. (2) In Tool-Calling settings, dLLMs fail to maintain symbolic precision (e.g. strict JSON schemas) under diffusion noise. To assess the potential of dLLMs in agentic workflows, we introduce DiffuAgent, a multi-agent evaluation framework that integrates dLLMs as plug-and-play cognitive cores. Our analysis shows that dLLMs are effective in non-causal roles (e.g., memory summarization and tool selection) but require the incorporation of causal, precise, and logically grounded reasoning mechanisms into the denoising process to be viable for agentic tasks.

---

### Bridging the Knowledge-Action Gap by Evaluating LLMs in Dynamic Dental Clinical Scenarios

Authors: Hongyang Ma, Tiantian Gu, Huaiyuan Sun, Huilin Zhu, Yongxin Wang, Jie Li, Wubin Sun, Zeliang Lian, Yinghong Zhou, Yi Gao, Shirui Wang, Zhihui Tang
Published: 2026-01-19
Categories: cs.CL
arXiv: https://arxiv.org/abs/2601.12974v1

The transition of Large Language Models (LLMs) from passive knowledge retrievers to autonomous clinical agents demands a shift in evaluation-from static accuracy to dynamic behavioral reliability. To explore this boundary in dentistry, a domain where high-quality AI advice uniquely empowers patient-participatory decision-making, we present the Standardized Clinical Management & Performance Evaluation (SCMPE) benchmark, which comprehensively assesses performance from knowledge-oriented evaluations (static objective tasks) to workflow-based simulations (multi-turn simulated patient interactions). Our analysis reveals that while models demonstrate high proficiency in static objective tasks, their performance precipitates in dynamic clinical dialogues, identifying that the primary bottleneck lies not in knowledge retention, but in the critical challenges of active information gathering and dynamic state tracking. Mapping "Guideline Adherence" versus "Decision Quality" reveals a prevalent "High Efficacy, Low Safety" risk in general models. Furthermore, we quantify the impact of Retrieval-Augmented Generation (RAG). While RAG mitigates hallucinations in static tasks, its efficacy in dynamic workflows is limited and heterogeneous, sometimes causing degradation. This underscores that external knowledge alone cannot bridge the reasoning gap without domain-adaptive pre-training. This study empirically charts the capability boundaries of dental LLMs, providing a roadmap for bridging the gap between standardized knowledge and safe, autonomous clinical practice.

---

### Pardon? Evaluating Conversational Repair in Large Audio-Language Models

Authors: Shuanghong Huang, Jinlei Xu, Youchao Zhou, Yanghao Zhou, Xuan Zhao, Chong Feng, Wenxuan Zhang
Published: 2026-01-19
Categories: cs.CL
arXiv: https://arxiv.org/abs/2601.12973v1

Large Audio-Language Models (LALMs) have demonstrated strong performance in spoken question answering (QA), with existing evaluations primarily focusing on answer accuracy and robustness to acoustic perturbations. However, such evaluations implicitly assume that spoken inputs remain semantically answerable, an assumption that often fails in real-world interaction when essential information is missing. In this work, we introduce a repair-aware evaluation setting that explicitly distinguishes between answerable and unanswerable audio inputs. We define answerability as a property of the input itself and construct paired evaluation conditions using a semantic-acoustic masking protocol. Based on this setting, we propose the Evaluability Awareness and Repair (EAR) score, a non-compensatory metric that jointly evaluates task competence under answerable conditions and repair behavior under unanswerable conditions. Experiments on two spoken QA benchmarks across diverse LALMs reveal a consistent gap between answer accuracy and conversational reliability: while many models perform well when inputs are answerable, most fail to recognize semantic unanswerability and initiate appropriate conversational repair. These findings expose a limitation of prevailing accuracy-centric evaluation practices and motivate reliability assessments that treat unanswerable inputs as cues for repair and continued interaction.

---

### Architecture--Optimization Co-Design for Physics-Informed Neural Networks via Layer-wise Coordinate Adaptation and Gradient Conflict Resolution

Authors: Pancheng Niu, Jun Guo, Qiaolin He, Yongming Chen, Yanchao Shi
Published: 2026-01-19
Categories: cs.LG
arXiv: https://arxiv.org/abs/2601.12971v2

Physics-informed neural networks (PINNs) can be limited by coordinate representations and conflicting gradients from heterogeneous physical constraints. We propose Architecture--Conflict-Resolved PINN (ACR-PINN), combining Layer-wise Dynamic Adaptation (LDA) and Gradient-Conflict-Resolved PINN (GCR-PINN). LDA constructs layer-specific coordinate features and fuses two encoding branches through input-conditioned, feature-wise gates. GCR-PINN treats PDE, initial-condition, and boundary-condition losses as separate tasks and conditionally projects negatively aligned gradients before aggregation, without changing the physical loss formulation. Across seven benchmark problems, ACR-PINN achieves the lowest observed mean across all metrics among four core architecture--optimizer combinations, reducing mean relative $L_2$ error by approximately 56--97\% relative to Std-PINN and improving on the better single-component variant. Capacity-controlled comparisons and gate interventions indicate that these gains reflect the learned coordinate-fusion pathway rather than parameter count alone. Gradient diagnostics suggest that LDA can improve task-gradient compatibility and reduce the correction required by GCR-PINN, with effects depending on the problem and update rule. Comparisons with representative PINN baselines and tests over sampling and temporal-horizon changes support the observed accuracy gains within the studied settings, while equal-time experiments reveal an accuracy--cost trade-off from the richer LDA representation.

---

### Lombard Speech Synthesis for Any Voice with Controllable Style Embeddings

Authors: Seymanur Akti, Alexander Waibel
Published: 2026-01-19
Categories: cs.SD, cs.CL
arXiv: https://arxiv.org/abs/2601.12966v1

The Lombard effect plays a key role in natural communication, particularly in noisy environments or when addressing hearing-impaired listeners. We present a controllable text-to-speech (TTS) system capable of synthesizing Lombard speech for any speaker without requiring explicit Lombard data during training. Our approach leverages style embeddings learned from a large, prosodically diverse dataset and analyzes their correlation with Lombard attributes using principal component analysis (PCA). By shifting the relevant PCA components, we manipulate the style embeddings and incorporate them into our TTS model to generate speech at desired Lombard levels. Evaluations demonstrate that our method preserves naturalness and speaker identity, enhances intelligibility under noise, and provides fine-grained control over prosody, offering a robust solution for controllable Lombard TTS for any speaker.

---

### Deterministic Dynamics of Sampling Processes in Score-Based Diffusion Models with Multiplicative Noise Conditioning

Authors: Doheon Kim
Published: 2026-01-19
Categories: cs.LG
arXiv: https://arxiv.org/abs/2601.12965v1

Score-based diffusion models generate new samples by learning the score function associated with a diffusion process. While the effectiveness of these models can be theoretically explained using differential equations related to the sampling process, previous work by Song and Ermon (2020) demonstrated that neural networks using multiplicative noise conditioning can still generate satisfactory samples. In this setup, the model is expressed as the product of two functions: one depending on the spatial variable and the other on the noise magnitude. This structure limits the model's ability to represent a more general relationship between the spatial variable and the noise, indicating that it cannot fully learn the correct score. Despite this limitation, the models perform well in practice. In this work, we provide a theoretical explanation for this phenomenon by studying the deterministic dynamics of the associated differential equations, offering insight into how the model operates.

---

### Cross-Scale Pretraining: Enhancing Self-Supervised Learning for Low-Resolution Satellite Imagery for Semantic Segmentation

Authors: John Waithaka, Gustave Bwirayesu, Moise Busogi
Published: 2026-01-19
Categories: cs.CV
arXiv: https://arxiv.org/abs/2601.12964v2

Self-supervised pretraining in remote sensing is mostly done using mid-spatial resolution (MR) image datasets due to their high availability. Given the release of high-resolution (HR) datasets, we ask how HR datasets can be included in self-supervised pretraining to enhance MR image representation learning and downstream segmentation performance on MR tasks. We design a spatial affinity component that can be added to existing self-supervised learning frameworks and that uses HR imagery to learn better representations of MR imagery. We test the spatial affinity component on two self-supervised learning frameworks and show that it outperforms models pretrained on HR or MR images alone.

---

### Trustworthy Data-driven Chronological Age Estimation from Panoramic Dental Images

Authors: Ainhoa Vivel-Couso, Nicolás Vila-Blanco, María J. Carreira, Alberto Bugarín-Diz, Inmaculada Tomás, Jose M. Alonso-Moral
Published: 2026-01-19
Categories: cs.CL
arXiv: https://arxiv.org/abs/2601.12960v1

Integrating deep learning into healthcare enables personalized care but raises trust issues due to model opacity. To improve transparency, we propose a system for dental age estimation from panoramic images that combines an opaque and a transparent method within a natural language generation (NLG) module. This module produces clinician-friendly textual explanations about the age estimations, designed with dental experts through a rule-based approach. Following the best practices in the field, the quality of the generated explanations was manually validated by dental experts using a questionnaire. The results showed a strong performance, since the experts rated 4.77+/-0.12 (out of 5) on average across the five dimensions considered. We also performed a trustworthy self-assessment procedure following the ALTAI checklist, in which it scored 4.40+/-0.27 (out of 5) across seven dimensions of the AI Trustworthiness Assessment List.

---

### StyMam: A Mamba-Based Generator for Artistic Style Transfer

Authors: Zhou Hong, Ning Dong, Yicheng Di, Xiaolong Xu, Rongsheng Hu, Yihua Shao, Run Ling, Yun Wang, Juqin Wang, Zhanjie Zhang, Ao Ma
Published: 2026-01-19
Categories: cs.CV
arXiv: https://arxiv.org/abs/2601.12954v3

Image style transfer aims to integrate the visual patterns of a specific artistic style into a content image while preserving its content structure. Existing methods mainly rely on the generative adversarial network (GAN) or stable diffusion (SD). GAN-based approaches using CNNs or Transformers struggle to jointly capture local and global dependencies, leading to artifacts and disharmonious patterns. SD-based methods reduce such issues but often fail to preserve content structures and suffer from slow inference. To address these issues, we revisit GAN and propose a mamba-based generator, termed as StyMam, to produce high-quality stylized images without introducing artifacts and disharmonious patterns. Specifically, we introduce a mamba-based generator with a residual dual-path strip scanning mechanism and a channel-reweighted spatial attention module. The former efficiently captures local texture features, while the latter models global dependencies. Finally, extensive qualitative and quantitative experiments demonstrate that the proposed method outperforms state-of-the-art algorithms in both quality and speed.

---

### Beyond Accuracy: Characterizing Code Comprehension Capabilities in (Large) Language Models

Authors: Felix Mächtle, Jan-Niclas Serr, Nils Loose, Thomas Eisenbarth
Published: 2026-01-19
Categories: cs.SE, cs.AI
arXiv: https://arxiv.org/abs/2601.12951v1

Large Language Models (LLMs) are increasingly integrated into software engineering workflows, yet current benchmarks provide only coarse performance summaries that obscure the diverse capabilities and limitations of these models. This paper investigates whether LLMs' code-comprehension performance aligns with traditional human-centric software metrics or instead reflects distinct, non-human regularities. We introduce a diagnostic framework that reframes code understanding as a binary input-output consistency task, enabling the evaluation of classification and generative models. Using a large-scale dataset, we correlate model performance with traditional, human-centric complexity metrics, such as lexical size, control-flow complexity, and abstract syntax tree structure. Our analyses reveal minimal correlation between human-defined metrics and LLM success (AUROC 0.63), while shadow models achieve substantially higher predictive performance (AUROC 0.86), capturing complex, partially predictable patterns beyond traditional software measures. These findings suggest that LLM comprehension reflects model-specific regularities only partially accessible through either human-designed or learned features, emphasizing the need for benchmark methodologies that move beyond aggregate accuracy and toward instance-level diagnostics, while acknowledging fundamental limits in predicting correct outcomes.

---

### GazeD: Context-Aware Diffusion for Accurate 3D Gaze Estimation

Authors: Riccardo Catalini, Davide Di Nucci, Guido Borghi, Davide Davoli, Lorenzo Garattoni, Gianpiero Francesca, Yuki Kawana, Roberto Vezzani
Published: 2026-01-19
Categories: cs.CV
arXiv: https://arxiv.org/abs/2601.12948v2

We introduce GazeD, a new 3D gaze estimation method that jointly provides 3D gaze and human pose from a single RGB image. Leveraging the ability of diffusion models to deal with uncertainty, it generates multiple plausible 3D gaze and pose hypotheses based on the 2D context information extracted from the input image. Specifically, we condition the denoising process on the 2D pose, the surroundings of the subject, and the context of the scene. With GazeD we also introduce a novel way of representing the 3D gaze by positioning it as an additional body joint at a fixed distance from the eyes. The rationale is that the gaze is usually closely related to the pose, and thus it can benefit from being jointly denoised during the diffusion process. Evaluations across three benchmark datasets demonstrate that GazeD achieves state-of-the-art performance in 3D gaze estimation, even surpassing methods that rely on temporal information. Project details will be available at https://aimagelab.ing.unimore.it/go/gazed.

---

### AI-generated data contamination erodes pathological variability and diagnostic reliability

Authors: Hongyu He, Shaowen Xiang, Ye Zhang, Yingtao Zhu, Jin Zhang, Hao Deng, Emily Alsentzer, Yun Liu, Qingyu Chen, Kun-Hsing Yu, Andrew Marshall, Tingting Chen et al.
Published: 2026-01-19
Categories: cs.CY, cs.AI, cs.CL, cs.CV, cs.LG
arXiv: https://arxiv.org/abs/2601.12946v4

Generative artificial intelligence (AI) is rapidly populating medical records with synthetic content, creating a feedback loop where future models are increasingly at risk of training on uncurated AI-generated data. However, the clinical consequences of this AI-generated data contamination remain unexplored. Here, we show that in the absence of mandatory human verification, this self-referential cycle drives a rapid erosion of pathological variability and diagnostic reliability. By analysing more than 800,000 synthetic data points across clinical text generation, vision-language reporting, and medical image synthesis, we find that models progressively converge toward generic phenotypes regardless of the model architecture. Specifically, rare but critical findings, including pneumothorax and effusions, vanish from the synthetic content generated by AI models, while demographic representations skew heavily toward middle-aged male phenotypes. Crucially, this degradation is masked by false diagnostic confidence; models continue to issue reassuring reports while failing to detect life-threatening pathology, with false reassurance rates tripling to 40%. Blinded physician evaluation confirms that this decoupling of confidence and accuracy renders AI-generated documentation clinically useless after just two generations. We systematically evaluate three mitigation strategies, finding that while synthetic volume scaling fails to prevent collapse, mixing real data with quality-aware filtering effectively preserves diversity. Ultimately, our results suggest that without policy-mandated human oversight, the deployment of generative AI threatens to degrade the very healthcare data ecosystems it relies upon.

---

### A Component-Based Survey of Interactions between Large Language Models and Multi-Armed Bandits

Authors: Siguang Chen, Chunli Lv, Miao Xie
Published: 2026-01-19
Categories: cs.CL, cs.LG
arXiv: https://arxiv.org/abs/2601.12945v3

Large language models (LLMs) have become powerful and widely used systems for language understanding and generation, while multi-armed bandit (MAB) algorithms provide a principled framework for adaptive decision-making under uncertainty. This survey explores the potential at the intersection of these two fields. As we know, it is the first survey to systematically review the bidirectional interaction between large language models and multi-armed bandits at the component level. We highlight the bidirectional benefits: MAB algorithms address critical LLM challenges, spanning from pre-training to retrieval-augmented generation (RAG) and personalization. Conversely, LLMs enhance MAB systems by redefining core components such as arm definition and environment modeling, thereby improving decision-making in sequential tasks. We analyze existing LLM-enhanced bandit systems and bandit-enhanced LLM systems, providing insights into their design, methodologies, and performance. Key challenges and representative findings are identified to help guide future research. An accompanying GitHub repository that indexes relevant literature is available at https://github.com/bucky1119/Awesome-LLM-Bandit-Interaction.

---

### Active Inference-Driven World Modeling for Adaptive UAV Swarm Trajectory Design

Authors: Kaleem Arshid, Ali Krayani, Lucio Marcenaro, David Martin Gomez, Carlo Regazzoni
Published: 2026-01-19
Categories: cs.RO, cs.AI, eess.SP
arXiv: https://arxiv.org/abs/2601.12939v1

This paper proposes an Active Inference-based framework for autonomous trajectory design in UAV swarms. The method integrates probabilistic reasoning and self-learning to enable distributed mission allocation, route ordering, and motion planning. Expert trajectories generated using a Genetic Algorithm with Repulsion Forces (GA-RF) are employed to train a hierarchical World Model capturing swarm behavior across mission, route, and motion levels. During online operation, UAVs infer actions by minimizing divergence between current beliefs and model-predicted states, enabling adaptive responses to dynamic environments. Simulation results show faster convergence, higher stability, and safer navigation than Q-Learning, demonstrating the scalability and cognitive grounding of the proposed framework for intelligent UAV swarm control.

---