# Research Papers — 2026-01 · bundle page 94 of 189

> Full-text markdown of papers #4651–4700 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/94
[← Previous bundle](https://aipod.dev/papers/archive/2026-01/page/93.md)
[Next bundle →](https://aipod.dev/papers/archive/2026-01/page/95.md)

### TrojanPraise: Jailbreak LLMs via Benign Fine-Tuning

Authors: Zhixin Xie, Xurui Song, Jun Luo
Published: 2026-01-18
Categories: cs.CR, cs.LG
arXiv: https://arxiv.org/abs/2601.12460v1

The demand of customized large language models (LLMs) has led to commercial LLMs offering black-box fine-tuning APIs, yet this convenience introduces a critical security loophole: attackers could jailbreak the LLMs by fine-tuning them with malicious data. Though this security issue has recently been exposed, the feasibility of such attacks is questionable as malicious training dataset is believed to be detectable by moderation models such as Llama-Guard-3. In this paper, we propose TrojanPraise, a novel finetuning-based attack exploiting benign and thus filter-approved data. Basically, TrojanPraise fine-tunes the model to associate a crafted word (e.g., "bruaf") with harmless connotations, then uses this word to praise harmful concepts, subtly shifting the LLM from refusal to compliance. To explain the attack, we decouple the LLM's internal representation of a query into two dimensions of knowledge and attitude. We demonstrate that successful jailbreak requires shifting the attitude while avoiding knowledge shift, a distortion in the model's understanding of the concept. To validate this attack, we conduct experiments on five opensource LLMs and two commercial LLMs under strict black-box settings. Results show that TrojanPraise achieves a maximum attack success rate of 95.88% while evading moderation.

---

### AgenTRIM: Tool Risk Mitigation for Agentic AI

Authors: Roy Betser, Amit Giloni, Shamik Bose, Sindhu Padakandla, Chiara Picardi, Lidor Erez, Roman Vainshtein
Published: 2026-01-18
Categories: cs.CR, cs.AI
arXiv: https://arxiv.org/abs/2601.12449v2

AI agents are autonomous systems that combine LLMs with external tools to solve complex tasks. While such tools extend capability, improper tool permissions introduce security risks such as indirect prompt injection and tool misuse. We characterize these failures as unbalanced tool-driven agency. Agents may retain unnecessary permissions (excessive agency) or fail to invoke required tools (insufficient agency), amplifying the attack surface and reducing performance. We introduce AgenTRIM, a framework for detecting and mitigating tool-driven agency risks without altering an agent's internal reasoning. AgenTRIM addresses these risks through complementary offline and online phases. Offline, AgenTRIM reconstructs and verifies the agent's tool interface from code and execution traces. At runtime, it enforces per-step least-privilege tool access through adaptive filtering and status-aware validation of tool calls. Evaluating on the AgentDojo benchmark, AgenTRIM substantially reduces attack success while maintaining high task performance. Additional experiments show robustness to description-based attacks and effective enforcement of explicit safety policies. Together, these results show that AgenTRIM provides a practical, capability-preserving approach to safer tool use in LLM-based agents.

---

### Privacy-Preserving Federated Learning with Verifiable Fairness Guarantees

Authors: Mohammed Himayath Ali, Mohammed Aqib Abdullah, Syed Muneer Hussain, Mohammed Mudassir Uddin, Shahnawaz Alam
Published: 2026-01-18
Categories: cs.CR, cs.CL, cs.CV
arXiv: https://arxiv.org/abs/2601.12447v2

Federated learning enables collaborative model training across distributed institutions without centralizing sensitive data; however, ensuring algorithmic fairness across heterogeneous data distributions while preserving privacy remains fundamentally unresolved. This paper introduces CryptoFair-FL, a novel cryptographic framework providing the first verifiable fairness guarantees for federated learning systems under formal security definitions. The proposed approach combines additively homomorphic encryption with secure multi-party computation to enable privacy-preserving verification of demographic parity and equalized odds metrics without revealing protected attribute distributions or individual predictions. A novel batched verification protocol reduces computational complexity from BigO(n^2) to BigO(n \log n) while maintaining (\dparam, \deltap)-differential privacy with dparam = 0.5 and deltap = 10^{-6}. Theoretical analysis establishes information-theoretic lower bounds on the privacy cost of fairness verification, demonstrating that the proposed protocol achieves near-optimal privacy-fairness tradeoffs. Comprehensive experiments across four benchmark datasets (MIMIC-IV healthcare records, Adult Income, CelebA, and a novel FedFair-100 benchmark) demonstrate that CryptoFair-FL reduces fairness violations from 0.231 to 0.031 demographic parity difference while incurring only 2.3 times computational overhead compared to standard federated averaging. The framework successfully defends against attribute inference attacks, maintaining adversarial success probability below 0.05 across all tested configurations. These results establish a practical pathway for deploying fairness-aware federated learning in regulated industries requiring both privacy protection and algorithmic accountability.

---

### Large Language Model for OWL Proofs

Authors: Hui Yang, Jiaoyan Chen, Uli Sattler
Published: 2026-01-18
Categories: cs.AI, cs.LO
arXiv: https://arxiv.org/abs/2601.12444v1

The ability of Large Language Models (LLMs) to perform reasoning tasks such as deduction has been widely investigated in recent years. Yet, their capacity to generate proofs-faithful, human-readable explanations of why conclusions follow-remains largely under explored. In this work, we study proof generation in the context of OWL ontologies, which are widely adopted for representing and reasoning over complex knowledge, by developing an automated dataset construction and evaluation framework. Our evaluation encompassing three sequential tasks for complete proving: Extraction, Simplification, and Explanation, as well as an additional task of assessing Logic Completeness of the premise. Through extensive experiments on widely used reasoning LLMs, we achieve important findings including: (1) Some models achieve overall strong results but remain limited on complex cases; (2) Logical complexity, rather than representation format (formal logic language versus natural language), is the dominant factor shaping LLM performance; and (3) Noise and incompleteness in input data substantially diminish LLMs' performance. Together, these results underscore both the promise of LLMs for explanation with rigorous logics and the gap of supporting resilient reasoning under complex or imperfect conditions. Code and data are available at https://github.com/HuiYang1997/LLMOwlR.

---

### Adversarial Defense in Vision-Language Models: An Overview

Authors: Xiaowei Fu, Lei Zhang
Published: 2026-01-18
Categories: cs.CV, cs.AI
arXiv: https://arxiv.org/abs/2601.12443v1

The widespread use of Vision Language Models (VLMs, e.g. CLIP) has raised concerns about their vulnerability to sophisticated and imperceptible adversarial attacks. These attacks could compromise model performance and system security in cross-modal tasks. To address this challenge, three main defense paradigms have been proposed: Training-time Defense, Test-time Adaptation Defense, and Training-free Defense. Training-time Defense involves modifying the training process, typically through adversarial fine-tuning to improve the robustness to adversarial examples. While effective, this approach requires substantial computational resources and may not generalize across all adversarial attacks. Test-time Adaptation Defense focuses on adapting the model at inference time by updating its parameters to handle unlabeled adversarial examples, offering flexibility but often at the cost of increased complexity and computational overhead. Training-free Defense avoids modifying the model itself, instead focusing on altering the adversarial inputs or their feature embeddings, which enforces input perturbations to mitigate the impact of attacks without additional training. This survey reviews the latest advancements in adversarial defense strategies for VLMs, highlighting the strengths and limitations of such approaches and discussing ongoing challenges in enhancing the robustness of VLMs.

---

### Constraint-Aware Neurosymbolic Uncertainty Quantification with Bayesian Deep Learning for Scientific Discovery

Authors: Shahnawaz Alam, Mohammed Mudassir Uddin, Mohammed Kaif Pasha
Published: 2026-01-18
Categories: cs.LG, cs.AI, cs.CV
arXiv: https://arxiv.org/abs/2601.12442v1

Scientific Artificial Intelligence (AI) applications require models that deliver trustworthy uncertainty estimates while respecting domain constraints. Existing uncertainty quantification methods lack mechanisms to incorporate symbolic scientific knowledge, while neurosymbolic approaches operate deterministically without principled uncertainty modeling. We introduce the Constraint-Aware Neurosymbolic Uncertainty Framework (CANUF), unifying Bayesian deep learning with differentiable symbolic reasoning. The architecture comprises three components: automated constraint extraction from scientific literature, probabilistic neural backbone with variational inference, and differentiable constraint satisfaction layer ensuring physical consistency. Experiments on Materials Project (140,000+ materials), QM9 molecular properties, and climate benchmarks show CANUF reduces Expected Calibration Error by 34.7% versus Bayesian neural networks while maintaining 99.2% constraint satisfaction. Ablations reveal constraint-guided recalibration contributes 18.3% performance gain, with constraint extraction achieving 91.4% precision. CANUF provides the first end-to-end differentiable pipeline simultaneously addressing uncertainty quantification, constraint satisfaction, and interpretable explanations for scientific predictions.

---

### Purification Before Fusion: Toward Mask-Free Speech Enhancement for Robust Audio-Visual Speech Recognition

Authors: Linzhi Wu, Xingyu Zhang, Hao Yuan, Yakun Zhang, Changyan Zheng, Liang Xie, Tiejun Liu, Erwei Yin
Published: 2026-01-18
Categories: eess.AS, cs.AI, cs.LG, cs.MM, cs.SD
arXiv: https://arxiv.org/abs/2601.12436v2

Audio-visual speech recognition (AVSR) typically improves recognition accuracy in noisy environments by integrating noise-immune visual cues with audio signals. Nevertheless, high-noise audio inputs are prone to introducing adverse interference into the feature fusion process. To mitigate this, recent AVSR methods often adopt mask-based strategies to filter audio noise during feature interaction and fusion, yet such methods risk discarding semantically relevant information alongside noise. In this work, we propose an end-to-end noise-robust AVSR framework coupled with speech enhancement, eliminating the need for explicit noise mask generation. This framework leverages a Conformer-based bottleneck fusion module to implicitly refine noisy audio features with video assistance. By reducing modality redundancy and enhancing inter-modal interactions, our method preserves speech semantic integrity to achieve robust recognition performance. Experimental evaluations on the public LRS3 benchmark suggest that our method outperforms prior advanced mask-based baselines under noisy conditions.

---

### Temporal Data and Short-Time Averages Improve Multiphase Mass Flow Metering

Authors: Amanda Nyholm, Yessica Arellano, Jinyu Liu, Damian Krakowiak, Pierluigi Salvo Rossi
Published: 2026-01-18
Categories: eess.SP, cs.LG
arXiv: https://arxiv.org/abs/2601.12433v1

Reliable flow measurements are essential in many industries, but current instruments often fail to accurately estimate multiphase flows, which are frequently encountered in real-world operations. Combining machine learning (ML) algorithms with accurate single-phase flowmeters has therefore received extensive research attention in recent years. The Coriolis mass flowmeter is a widely used single-phase meter that provides direct mass flow measurements, which ML models can be trained to correct, thereby reducing measurement errors in multiphase conditions. This paper demonstrates that preserving temporal information significantly improves model performance in such scenarios. We compare a multilayer perceptron, a windowed multilayer perceptron, and a convolutional neural network (CNN) on three-phase air-water-oil flow data from 342 experiments. Whereas prior work typically compresses each experiment into a single averaged sample, we instead compute short-time averages from within each experiment and train models that preserve temporal information at several downsampling intervals. The CNN performed best at 0.25 Hz with approximately 95 % of relative errors below 13 %, a normalized root mean squared error of 0.03, and a mean absolute percentage error of approximately 4.3 %, clearly outperforming the best single-averaged model and demonstrating that short-time averaging within individual experiments is preferable. Results are consistent across multiple data splits and random seeds, demonstrating robustness.

---

### SkeFi: Cross-Modal Knowledge Transfer for Wireless Skeleton-Based Action Recognition

Authors: Shunyu Huang, Yunjiao Zhou, Jianfei Yang
Published: 2026-01-18
Categories: cs.CV, cs.MM
arXiv: https://arxiv.org/abs/2601.12432v1

Skeleton-based action recognition leverages human pose keypoints to categorize human actions, which shows superior generalization and interoperability compared to regular end-to-end action recognition. Existing solutions use RGB cameras to annotate skeletal keypoints, but their performance declines in dark environments and raises privacy concerns, limiting their use in smart homes and hospitals. This paper explores non-invasive wireless sensors, i.e., LiDAR and mmWave, to mitigate these challenges as a feasible alternative. Two problems are addressed: (1) insufficient data on wireless sensor modality to train an accurate skeleton estimation model, and (2) skeletal keypoints derived from wireless sensors are noisier than RGB, causing great difficulties for subsequent action recognition models. Our work, SkeFi, overcomes these gaps through a novel cross-modal knowledge transfer method acquired from the data-rich RGB modality. We propose the enhanced Temporal Correlation Adaptive Graph Convolution (TC-AGC) with frame interactive enhancement to overcome the noise from missing or inconsecutive frames. Additionally, our research underscores the effectiveness of enhancing multiscale temporal modeling through dual temporal convolution. By integrating TC-AGC with temporal modeling for cross-modal transfer, our framework can extract accurate poses and actions from noisy wireless sensors. Experiments demonstrate that SkeFi realizes state-of-the-art performances on mmWave and LiDAR. The code is available at https://github.com/Huang0035/Skefi.

---

### System-Mediated Attention Imbalances Make Vision-Language Models Say Yes

Authors: Tsan Tsai Chan, Varsha Suresh, Anisha Saha, Michael Hahn, Vera Demberg
Published: 2026-01-18
Categories: cs.CL
arXiv: https://arxiv.org/abs/2601.12430v2

Vision-language model (VLM) hallucination is commonly linked to imbalanced allocation of attention across input modalities: system, image and text. However, existing mitigation strategies tend towards an image-centric interpretation of these imbalances, often prioritising increased image attention while giving less consideration to the roles of the other modalities. In this study, we evaluate a more holistic, system-mediated account, which attributes these imbalances to functionally redundant system weights that reduce attention to image and textual inputs. We show that this framework offers a useful empirical perspective on the yes-bias, a common form of hallucination in which VLMs indiscriminately respond `yes'. Causally redistributing attention from the system modality to image and textual inputs substantially suppresses this bias, often outperforming existing approaches. We further present evidence suggesting that system-mediated attention imbalances contribute to the yes-bias by encouraging a default reliance on coarse input representations, which are effective for some tasks but ill-suited to others. Taken together, these findings firmly establish system attention as a key factor in VLM hallucination and highlight its potential as a lever for mitigation.

---

### ReWorld: Multi-Dimensional Reward Modeling for Embodied World Models

Authors: Baorui Peng, Wenyao Zhang, Liang Xu, Zekun Qi, Jiazhao Zhang, Hongsi Liu, Wenjun Zeng, Xin Jin
Published: 2026-01-18
Categories: cs.RO, cs.CV
arXiv: https://arxiv.org/abs/2601.12428v1

Recently, video-based world models that learn to simulate the dynamics have gained increasing attention in robot learning. However, current approaches primarily emphasize visual generative quality while overlooking physical fidelity, dynamic consistency, and task logic, especially for contact-rich manipulation tasks, which limits their applicability to downstream tasks. To this end, we introduce ReWorld, a framework aimed to employ reinforcement learning to align the video-based embodied world models with physical realism, task completion capability, embodiment plausibility and visual quality. Specifically, we first construct a large-scale (~235K) video preference dataset and employ it to train a hierarchical reward model designed to capture multi-dimensional reward consistent with human preferences. We further propose a practical alignment algorithm that post-trains flow-based world models using this reward through a computationally efficient PPO-style algorithm. Comprehensive experiments and theoretical analysis demonstrate that ReWorld significantly improves the physical fidelity, logical coherence, embodiment and visual quality of generated rollouts, outperforming previous methods.

---

### Graph Attention Networks with Physical Constraints for Anomaly Detection

Authors: Mohammadhossein Homaei, Iman Khazrak, Ruben Molano, Andres Caro, Mar Avila
Published: 2026-01-18
Categories: cs.LG, cs.CR
arXiv: https://arxiv.org/abs/2601.12426v1

Water distribution systems (WDSs) face increasing cyber-physical risks, which make reliable anomaly detection essential. Many data-driven models ignore network topology and are hard to interpret, while model-based ones depend strongly on parameter accuracy. This work proposes a hydraulic-aware graph attention network using normalized conservation law violations as features. It combines mass and energy balance residuals with graph attention and bidirectional LSTM to learn spatio-temporal patterns. A multi-scale module aggregates detection scores from node to network level. On the BATADAL dataset, it reaches $F1=0.979$, showing $3.3$pp gain and high robustness under $15\%$ parameter noise.

---

### HOT-POT: Optimal Transport for Sparse Stereo Matching

Authors: Antonin Clerc, Michael Quellmalz, Moritz Piening, Philipp Flotho, Gregor Kornhardt, Gabriele Steidl
Published: 2026-01-18
Categories: cs.CV, math.OC
arXiv: https://arxiv.org/abs/2601.12423v2

Stereo vision between images faces a range of challenges, including occlusions, motion, and camera distortions, across applications in autonomous driving, robotics, and face analysis. Due to parameter sensitivity, further complications arise for stereo matching with sparse features, such as facial landmarks. To overcome this ill-posedness and enable unsupervised sparse matching, we consider line constraints of the camera geometry from an optimal transport (OT) viewpoint. Formulating camera-projected points as (half)lines, we propose the use of the classical epipolar distance as well as a 3D ray distance to quantify matching quality. Employing these distances as a cost function of a (partial) OT problem, we arrive at efficiently solvable assignment problems. Moreover, we extend our approach to unsupervised object matching by formulating it as a hierarchical OT problem. The resulting algorithms allow for efficient feature and object matching, as demonstrated in our numerical experiments. Here, we focus on applications in facial analysis, where we aim to match distinct landmarking conventions.

---

### Legal Experts Disagree With Rationale Extraction Techniques for Explaining ECtHR Case Outcome Classification

Authors: Mahammad Namazov, Tomáš Koref, Ivan Habernal
Published: 2026-01-18
Categories: cs.CL
arXiv: https://arxiv.org/abs/2601.12419v2

Interpretability is critical for applications of large language models (LLMs) in the legal domain, where trust and transparency are essential. A central NLP task in this setting is legal outcome prediction, where models forecast whether a court will find a violation of a given right. We study this task on decisions from the European Court of Human Rights (ECtHR), introducing a new ECtHR dataset with carefully curated positive (violation) and negative (non-violation) cases. Existing works propose both task-specific approaches and model-agnostic techniques to explain downstream performance, but it remains unclear which techniques best explain legal outcome prediction. To address this, we propose a comparative analysis framework for model-agnostic interpretability methods. We focus on two rationale extraction techniques that justify model outputs with concise, human-interpretable text fragments from the input. We evaluate faithfulness via normalized sufficiency and comprehensiveness metrics, and plausibility via legal expert judgments of the extracted rationales. We also assess the feasibility of using LLM-as-a-Judge, using these expert evaluations as reference. Our experiments on the new ECtHR dataset show that models' "reasons" for predicting violations differ substantially from those of legal experts, despite strong faithfulness scores. The source code of our experiments is publicly available at https://github.com/trusthlt/IntEval.

---

### Orthogonalized Policy Optimization:Policy Optimization as Orthogonal Projection in Hilbert Space

Authors: Wang Zixian
Published: 2026-01-18
Categories: cs.LG, cs.AI
arXiv: https://arxiv.org/abs/2601.12415v5

We propose Orthogonalized Policy Optimization (OPO), a principled framework for large language model alignment derived from optimization in the Hilbert function space L2(pi_k). Lifting policy updates from the probability simplex into L2(pi_k) transforms the nonlinear normalization constraint into a linear orthogonality condition <v, 1>_{pi_k} = 0 on the density fluctuation field v = pi/pi_k - 1. By the Hilbert projection theorem, the unique closed-form update is v_star = (omega_alpha - E(omega_alpha)) / mu, where the subtracted mean acts as a chemical potential enforcing probability conservation. This interpretation reveals advantage z-score normalization as a conservation-law projection rather than a variance-reduction heuristic. OPO cleanly decouples sampling geometry, controlled by the escort exponent alpha, from optimization geometry, governed by the stiffness parameter mu, a separation not attainable under KL-based objectives. The same update can also be derived as a Euclidean mirror-descent step and as the linear-response law of near-equilibrium statistical mechanics, establishing its structural uniqueness within ratio geometry. Structurally, OPO induces constant curvature, non-saturating linear gradient dynamics, and an intrinsic chi-square trust region. Experiments on MATH benchmarks show that the Hilbert projection formulation prevents gradient saturation typical of KL-constrained methods. By sustaining non-vanishing gradients in high-confidence regimes, OPO avoids premature plateaus and achieves stronger long-horizon training rewards and improved out-of-distribution generalization compared to clipping-based baselines.

---

### Are LLMs Smarter Than Chimpanzees? An Evaluation on Perspective Taking and Knowledge State Estimation

Authors: Dingyi Yang, Junqi Zhao, Xue Li, Ce Li, Boyang Li
Published: 2026-01-18
Categories: cs.AI
arXiv: https://arxiv.org/abs/2601.12410v2

Cognitive anthropology suggests that the distinction of human intelligence lies in the ability to infer other individuals' knowledge states and understand their intentions. In comparison, our closest animal relative, chimpanzees, lack the capacity to do so. With this paper, we aim to evaluate LLM performance in estimating other individuals' knowledge states and their potential actions. We design two tasks to test (1) if LLMs can predict story characters' next actions based on their own knowledge vs. improperly using information unavailable from their perspective, and (2) if LLMs can detect when story characters, through their actions, demonstrate knowledge they should not possess. Results reveal that most current state-of-the-art LLMs achieve near-random performance on both tasks, and are substantially inferior to humans. We argue future LLM research should place more weight on the abilities of knowledge estimation and intention understanding.

---

### De-Anonymization at Scale via Tournament-Style Attribution

Authors: Lirui Zhang, Huishuai Zhang
Published: 2026-01-18
Categories: cs.CR, cs.CL, cs.LG
arXiv: https://arxiv.org/abs/2601.12407v2

As LLMs rapidly advance and enter real-world use, their privacy implications are increasingly important. We study an authorship de-anonymization threat: using LLMs to link anonymous documents to their authors, potentially compromising settings such as double-blind peer review. We propose De-Anonymization at Scale (DAS), a large language model-based method for attributing authorship among tens of thousands of candidate texts. DAS uses a sequential progression strategy: it randomly partitions the candidate corpus into fixed-size groups, prompts an LLM to select the text most likely written by the same author as a query text, and iteratively re-queries the surviving candidates to produce a ranked top-k list. To make this practical at scale, DAS adds a dense-retrieval prefilter to shrink the search space and a majority-voting style aggregation over multiple independent runs to improve robustness and ranking precision. Experiments on anonymized review data show DAS can recover same-author texts from pools of tens of thousands with accuracy well above chance, demonstrating a realistic privacy risk for anonymous platforms. On standard authorship benchmarks (Enron emails and blog posts), DAS also improves both accuracy and scalability over prior approaches, highlighting a new LLM-enabled de-anonymization vulnerability.

---

### Explainable Machine Learning for Pediatric Dental Risk Stratification Using Socio-Demographic Determinants

Authors: Manasi Kanade, Abhi Thakkar, Gabriela Fernandes
Published: 2026-01-18
Categories: cs.LG, cs.AI
arXiv: https://arxiv.org/abs/2601.12405v1

Background: Pediatric dental disease remains one of the most prevalent and inequitable chronic health conditions worldwide. Although strong epidemiological evidence links oral health outcomes to socio-economic and demographic determinants, most artificial intelligence (AI) applications in dentistry rely on image-based diagnosis and black-box prediction models, limiting transparency and ethical applicability in pediatric populations. Objective: This study aimed to develop and evaluate an explainable machine learning framework for pediatric dental risk stratification that prioritizes interpretability, calibration, and ethical deployment over maximal predictive accuracy. Methods: A supervised machine learning model was trained using population-level pediatric data including age, income-to-poverty ratio, race/ethnicity, gender, and medical history. Model performance was assessed using receiver operating characteristic (ROC) analysis and calibration curves. Explainability was achieved using SHapley Additive exPlanations (SHAP) to provide global and individual-level interpretation of predictions. Results: The model achieved modest discrimination (AUC = 0.61) with conservative calibration, underestimating risk at higher probability levels. SHAP analysis identified age and income-to-poverty ratio as the strongest contributors to predicted risk, followed by race/ethnicity and gender. Conclusion: Explainable machine learning enables transparent, prevention-oriented pediatric dental risk stratification and supports population screening and equitable resource allocation rather than diagnostic decision-making.

---

### Weaknesses of Facial Emotion Recognition Systems

Authors: Aleksandra Jamróz, Patrycja Wysocka, Piotr Garbat
Published: 2026-01-18
Categories: cs.CV, cs.AI
arXiv: https://arxiv.org/abs/2601.12402v1

Emotion detection from faces is one of the machine learning problems needed for human-computer interaction. The variety of methods used is enormous, which motivated an in-depth review of articles and scientific studies. Three of the most interesting and best solutions are selected, followed by the selection of three datasets that stood out for the diversity and number of images in them. The selected neural networks are trained, and then a series of experiments are performed to compare their performance, including testing on different datasets than a model was trained on. This reveals weaknesses in existing solutions, including differences between datasets, unequal levels of difficulty in recognizing certain emotions and the challenges in differentiating between closely related emotions.

---

### Beyond the Dirac Delta: Mitigating Diversity Collapse in Reinforcement Fine-Tuning for Versatile Image Generation

Authors: Jinmei Liu, Haoru Li, Zhenhong Sun, Chaofeng Chen, Yatao Bian, Bo Wang, Daoyi Dong, Chunlin Chen, Zhi Wang
Published: 2026-01-18
Categories: cs.LG, cs.AI
arXiv: https://arxiv.org/abs/2601.12401v2

Reinforcement learning (RL) has emerged as a powerful paradigm for fine-tuning large-scale generative models, such as diffusion and flow models, to align with complex human preferences and user-specified tasks. A fundamental limitation remains \textit{the curse of diversity collapse}, where the objective formulation and optimization landscape inherently collapse the policy to a Dirac delta distribution. To address this challenge, we propose \textbf{DRIFT} (\textbf{D}ive\textbf{R}sity-\textbf{I}ncentivized Reinforcement \textbf{F}ine-\textbf{T}uning for Versatile Image Generation), an innovative framework that systematically incentivizes output diversity throughout the on-policy fine-tuning process, reconciling strong task alignment with high generation diversity to enhance versatility essential for applications that demand diverse candidate generations. We approach the problem across three representative perspectives: i) \textbf{sampling} a reward-concentrated subset that filters out reward outliers to prevent premature collapse; ii) \textbf{prompting} with stochastic variations to expand the conditioning space, and iii) \textbf{optimization} of the intra-group diversity with a potential-based reward shaping mechanism. Experimental results show that DRIFT achieves superior Pareto dominance regarding task alignment and generation diversity, yielding a $ 9.08\%\!\sim\! 43.46\%$ increase in diversity at equivalent alignment levels and a $ 59.65\% \!\sim\! 65.86\%$ increase in alignment at equivalent levels of diversity.

---

### BiCoLoR: Communication-Efficient Optimization with Bidirectional Compression and Local Training

Authors: Laurent Condat, Artavazd Maranjyan, Peter Richtárik
Published: 2026-01-18
Categories: math.OC, cs.LG
arXiv: https://arxiv.org/abs/2601.12400v1

Slow and costly communication is often the main bottleneck in distributed optimization, especially in federated learning where it occurs over wireless networks. We introduce BiCoLoR, a communication-efficient optimization algorithm that combines two widely used and effective strategies: local training, which increases computation between communication rounds, and compression, which encodes high-dimensional vectors into short bitstreams. While these mechanisms have been combined before, compression has typically been applied only to uplink (client-to-server) communication, leaving the downlink (server-to-client) side unaddressed. In practice, however, both directions are costly. We propose BiCoLoR, the first algorithm to combine local training with bidirectional compression using arbitrary unbiased compressors. This joint design achieves accelerated complexity guarantees in both convex and strongly convex heterogeneous settings. Empirically, BiCoLoR outperforms existing algorithms and establishes a new standard in communication efficiency.

---

### PsychēChat: An Empathic Framework Focused on Emotion Shift Tracking and Safety Risk Analysis in Psychological Counseling

Authors: Zhentao Xia, Yongqi Fan, Yuxiang Chu, Yichao Yin, Liangliang Chen, Tong Ruan, Weiyan Zhang
Published: 2026-01-18
Categories: cs.AI
arXiv: https://arxiv.org/abs/2601.12392v1

Large language models (LLMs) have demonstrated notable advancements in psychological counseling. However, existing models generally do not explicitly model seekers' emotion shifts across counseling sessions, a core focus in classical psychological schools. Moreover, how to align counselor models' responses with these emotion shifts while proactively mitigating safety risks remains underexplored. To bridge these gaps, we propose PsychēChat, which explicitly integrates emotion shift tracking and safety risk analysis for psychological counseling. Specifically, we employ interactive role-playing to synthesize counselor--seeker dialogues, incorporating two modules: Emotion Management Module, to capture seekers' current emotions and emotion shifts; and Risk Control Module, to anticipate seekers' subsequent reactions and identify potential risks. Furthermore, we introduce two modeling paradigms. The Agent Mode structures emotion management, risk control, and counselor responses into a collaborative multi-agent pipeline. The LLM Mode integrates these stages into a unified chain-of-thought for end-to-end inference, balancing efficiency and performance. Extensive experiments, including interactive scoring, dialogue-level evaluation, and human assessment, demonstrate that PsychēChat outperforms existing methods for emotional insight and safety control.

---

### Class-Partitioned VQ-VAE and Latent Flow Matching for Point Cloud Scene Generation

Authors: Dasith de Silva Edirimuni, Ajmal Saeed Mian
Published: 2026-01-18
Categories: cs.CV
arXiv: https://arxiv.org/abs/2601.12391v1

Most 3D scene generation methods are limited to only generating object bounding box parameters while newer diffusion methods also generate class labels and latent features. Using object size or latent feature, they then retrieve objects from a predefined database. For complex scenes of varied, multi-categorical objects, diffusion-based latents cannot be effectively decoded by current autoencoders into the correct point cloud objects which agree with target classes. We introduce a Class-Partitioned Vector Quantized Variational Autoencoder (CPVQ-VAE) that is trained to effectively decode object latent features, by employing a pioneering $\textit{class-partitioned codebook}$ where codevectors are labeled by class. To address the problem of $\textit{codebook collapse}$, we propose a $\textit{class-aware}$ running average update which reinitializes dead codevectors within each partition. During inference, object features and class labels, both generated by a Latent-space Flow Matching Model (LFMM) designed specifically for scene generation, are consumed by the CPVQ-VAE. The CPVQ-VAE's class-aware inverse look-up then maps generated latents to codebook entries that are decoded to class-specific point cloud shapes. Thereby, we achieve pure point cloud generation without relying on an external objects database for retrieval. Extensive experiments reveal that our method reliably recovers plausible point cloud scenes, with up to 70.4% and 72.3% reduction in Chamfer and Point2Mesh errors on complex living room scenes.

---

### NADIR: Differential Attention Flow for Non-Autoregressive Transliteration in Indic Languages

Authors: Lakshya Tomar, Vinayak Abrol, Puneet Agarwal
Published: 2026-01-18
Categories: cs.CL, cs.AI
arXiv: https://arxiv.org/abs/2601.12389v1

In this work, we argue that not all sequence-to-sequence tasks require the strong inductive biases of autoregressive (AR) models. Tasks like multilingual transliteration, code refactoring, grammatical correction or text normalization often rely on local dependencies where the full modeling capacity of AR models can be overkill, creating a trade-off between their high accuracy and high inference latency. While non-autoregressive (NAR) models offer speed, they typically suffer from hallucinations and poor length control. To explore this trade-off, we focus on the multilingual transliteration task in Indic languages and introduce NADIR, a novel NAR architecture designed to strike a balance between speed and accuracy. NADIR integrates a Differential Transformer and a Mixture-of-Experts mechanism, enabling it to robustly model complex character mappings without sequential dependencies. NADIR achieves over a 13x speed-up compared to the state-of-the-art AR baseline. It maintains a competitive mean Character Error Rate of 15.78%, compared to 14.44% for the AR model and 21.88% for a standard NAR equivalent. Importantly, NADIR reduces Repetition errors by 49.53%, Substitution errors by 24.45%, Omission errors by 32.92%, and Insertion errors by 16.87%. This work provides a practical blueprint for building fast and reliable NAR systems, effectively bridging the gap between AR accuracy and the demands of real-time, large-scale deployment.

---

### A Hierarchical Benchmark of Foundation Models for Dermatology

Authors: Furkan Yuceyalcin, Abdurrahim Yilmaz, Burak Temelkuran
Published: 2026-01-18
Categories: cs.CV
arXiv: https://arxiv.org/abs/2601.12382v1

Foundation models have transformed medical image analysis by providing robust feature representations that reduce the need for large-scale task-specific training. However, current benchmarks in dermatology often reduce the complex diagnostic taxonomy to flat, binary classification tasks, such as distinguishing melanoma from benign nevi. This oversimplification obscures a model's ability to perform fine-grained differential diagnoses, which is critical for clinical workflow integration. This study evaluates the utility of embeddings derived from ten foundation models, spanning general computer vision, general medical imaging, and dermatology-specific domains, for hierarchical skin lesion classification. Using the DERM12345 dataset, which comprises 40 lesion subclasses, we calculated frozen embeddings and trained lightweight adapter models using a five-fold cross-validation. We introduce a hierarchical evaluation framework that assesses performance across four levels of clinical granularity: 40 Subclasses, 15 Main Classes, 2 and 4 Superclasses, and Binary Malignancy. Our results reveal a "granularity gap" in model capabilities: MedImageInsights achieved the strongest overall performance (97.52% weighted F1-Score on Binary Malignancy detection) but declined to 65.50% on fine-grained 40-class subtype classification. Conversely, MedSigLip (69.79%) and dermatology-specific models (Derm Foundation and MONET) excelled at fine-grained 40-class subtype discrimination while achieving lower overall performance than MedImageInsights on broader classification tasks. Our findings suggest that while general medical foundation models are highly effective for high-level screening, specialized modeling strategies are necessary for the granular distinctions required in diagnostic support systems.

---

### Statistical-Neural Interaction Networks for Interpretable Mixed-Type Data Imputation

Authors: Ou Deng, Shoji Nishimura, Atsushi Ogihara, Qun Jin
Published: 2026-01-18
Categories: cs.LG, stat.ML
arXiv: https://arxiv.org/abs/2601.12380v1

Real-world tabular databases routinely combine continuous measurements and categorical records, yet missing entries are pervasive and can distort downstream analysis. We propose Statistical-Neural Interaction (SNI), an interpretable mixed-type imputation framework that couples correlation-derived statistical priors with neural feature attention through a Controllable-Prior Feature Attention (CPFA) module. CPFA learns head-wise prior-strength coefficients $\{λ_h\}$ that softly regularize attention toward the prior while allowing data-driven deviations when nonlinear patterns appear to be present in the data. Beyond imputation, SNI aggregates attention maps into a directed feature-dependency matrix that summarizes which variables the imputer relied on, without requiring post-hoc explainers. We evaluate SNI against six baselines (Mean/Mode, MICE, KNN, MissForest, GAIN, MIWAE) on six datasets spanning ICU monitoring, population surveys, socio-economic statistics, and engineering applications. Under MCAR/strict-MAR at 30\% missingness, SNI is generally competitive on continuous metrics but is often outperformed by accuracy-first baselines (MissForest, MIWAE) on categorical variables; in return, it provides intrinsic dependency diagnostics and explicit statistical-neural trade-off parameters. We additionally report MNAR stress tests (with a mask-aware variant) and discuss computational cost, limitations -- particularly for severely imbalanced categorical targets -- and deployment scenarios where interpretability may justify the trade-off.

---

### Utilizing the Score of Data Distribution for Hyperspectral Anomaly Detection

Authors: Jiahui Sheng, Yidan Shi, Shu Xiang, Xiaorun Li, Shuhan Chen
Published: 2026-01-18
Categories: cs.CV
arXiv: https://arxiv.org/abs/2601.12379v1

Hyperspectral images (HSIs) are a type of image that contains abundant spectral information. As a type of real-world data, the high-dimensional spectra in hyperspectral images are actually determined by only a few factors, such as chemical composition and illumination. Thus, spectra in hyperspectral images are highly likely to satisfy the manifold hypothesis. Based on the hyperspectral manifold hypothesis, we propose a novel hyperspectral anomaly detection method (named ScoreAD) that leverages the time-dependent gradient field of the data distribution (i.e., the score), as learned by a score-based generative model (SGM). Our method first trains the SGM on the entire set of spectra from the hyperspectral image. At test time, each spectrum is passed through a perturbation kernel, and the resulting perturbed spectrum is fed into the trained SGM to obtain the estimated score. The manifold hypothesis of HSIs posits that background spectra reside on one or more low-dimensional manifolds. Conversely, anomalous spectra, owing to their unique spectral signatures, are considered outliers that do not conform to the background manifold. Based on this fundamental discrepancy in their manifold distributions, we leverage a generative SGM to achieve hyperspectral anomaly detection. Experiments on the four hyperspectral datasets demonstrate the effectiveness of the proposed method. The code is available at https://github.com/jiahuisheng/ScoreAD.

---

### LR-DWM: Efficient Watermarking for Diffusion Language Models

Authors: Ofek Raban, Ethan Fetaya, Gal Chechik
Published: 2026-01-18
Categories: cs.CL
arXiv: https://arxiv.org/abs/2601.12376v1

Watermarking (WM) is a critical mechanism for detecting and attributing AI-generated content. Current WM methods for Large Language Models (LLMs) are predominantly tailored for autoregressive (AR) models: They rely on tokens being generated sequentially, and embed stable signals within the generated sequence based on the previously sampled text. Diffusion Language Models (DLMs) generate text via non-sequential iterative denoising, which requires significant modification to use WM methods designed for AR models. Recent work proposed to watermark DLMs by inverting the process when needed, but suffers significant computational or memory overhead. We introduce Left-Right Diffusion Watermarking (LR-DWM), a scheme that biases the generated token based on both left and right neighbors, when they are available. LR-DWM incurs minimal runtime and memory overhead, remaining close to the non-watermarked baseline DLM while enabling reliable statistical detection under standard evaluation settings. Our results demonstrate that DLMs can be watermarked efficiently, achieving high detectability with negligible computational and memory overhead.

---

### LiQSS: Post-Transformer Linear Quantum-Inspired State-Space Tensor Networks for Real-Time 6G

Authors: Farhad Rezazadeh, Hatim Chergui, Amir Ashtari Gargari, Mehdi Bennis, Houbing Song, Lingjia Liu, Merouane Debbah
Published: 2026-01-18
Categories: cs.NI, cs.LG
arXiv: https://arxiv.org/abs/2601.12375v3

Proactive and agentic control in Sixth-Generation (6G) Open Radio Access Networks (O-RAN) requires control-grade prediction under stringent Near-Real-Time (Near-RT) latency and computational constraints. While Transformer-based models are effective for sequence modeling, their quadratic complexity limits scalability in Near-RT RAN Intelligent Controller (RIC) analytics. This paper investigates a post-Transformer design paradigm for efficient radio telemetry forecasting. We propose a quantum-inspired many-body state-space tensor network that replaces self-attention with stable structured state-space dynamics kernels, enabling linear-time sequence modeling. Tensor-network factorizations in the form of Tensor Train (TT) / Matrix Product State (MPS) representations are employed to reduce parameterization and data movement in both input projections and prediction heads, while lightweight channel gating and mixing layers capture non-stationary cross-Key Performance Indicator (KPI) dependencies. The proposed model is instantiated as an agentic perceive-predict xApp and evaluated on a bespoke O-RAN KPI time-series dataset comprising 59,441 sliding windows across 13 KPIs, using Reference Signal Received Power (RSRP) forecasting as a representative use case. Our proposed Linear Quantum-Inspired State-Space (LiQSS) model is 10.8x-15.8x smaller and approximately 1.4x faster than prior structured state-space baselines. Relative to Transformer-based models, LiQSS achieves up to a 155x reduction in parameter count and up to 2.74x faster inference, without sacrificing forecasting accuracy.

---

### A Scalable Entity-Based Framework for Auditing Bias in Large Language Models

Authors: Akram Elbouanani, Aboubacar Tuo, Adrian Popescu
Published: 2026-01-18
Categories: cs.CL, cs.AI
arXiv: https://arxiv.org/abs/2601.12374v3

Existing approaches to bias evaluation in large language models (LLMs) trade ecological validity for statistical control, relying either on artificial prompts that poorly reflect real-world use or on naturalistic tasks that lack scale and rigor. We introduce a scalable bias-auditing framework that uses named entities as controlled probes to measure systematic disparities in model behavior. Synthetic data enables us to construct diverse, controlled inputs, and we show that it reliably reproduces bias patterns observed in natural text, supporting its use for large-scale analysis. Using this framework, we conduct the largest bias audit to date, comprising 1.9 billion data points across multiple entity types, tasks, languages, models, and prompting strategies. We find consistent patterns: models penalize right-wing politicians and favor left-wing politicians, prefer Western and wealthier countries over the Global South, favor Western companies, and penalize firms in the defense and pharmaceutical sectors. While instruction tuning reduces bias, increasing model scale amplifies it, and prompting in Chinese or Russian does not mitigate Western-aligned preferences. These findings highlight the need for systematic bias auditing before deploying LLMs in high-stakes applications. Our framework is extensible to other domains and tasks, and we make it publicly available to support future work.

---

### CD-TWINSAFE: A ROS-enabled Digital Twin for Scene Understanding and Safety Emerging V2I Technology

Authors: Amro Khaled, Farah Khaled, Omar Riad, Catherine M. Elias
Published: 2026-01-18
Categories: cs.CV, cs.HC, cs.RO
arXiv: https://arxiv.org/abs/2601.12373v1

In this paper, the CD-TWINSAFE is introduced, a V2I-based digital twin for Autonomous Vehicles. The proposed architecture is composed of two stacks running simultaneously, an on-board driving stack that includes a stereo camera for scene understanding, and a digital twin stack that runs an Unreal Engine 5 replica of the scene viewed by the camera as well as returning safety alerts to the cockpit. The on-board stack is implemented on the vehicle side including 2 main autonomous modules; localization and perception. The position and orientation of the ego vehicle are obtained using on-board sensors. Furthermore, the perception module is responsible for processing 20-fps images from stereo camera and understands the scene through two complementary pipelines. The pipeline are working on object detection and feature extraction including object velocity, yaw and the safety metrics time-to-collision and time-headway. The collected data form the driving stack are sent to the infrastructure side through the ROS-enabled architecture in the form of custom ROS2 messages and sent over UDP links that ride a 4G modem for V2I communication. The environment is monitored via the digital twin through the shared messages which update the information of the spawned ego vehicle and detected objects based on the real-time localization and perception data. Several tests with different driving scenarios to confirm the validity and real-time response of the proposed architecture.

---

### Can Deep Research Agents Retrieve and Organize? Evaluating the Synthesis Gap with Expert Taxonomies

Authors: Ming Zhang, Jiabao Zhuang, Wenqing Jing, Kexin Tan, Ziyu Kong, Jingyi Deng, Yujiong Shen, Yuhui Wang, Zhenghao Xiang, Qiyuan Peng, Yuhang Zhao, Ning Luo et al.
Published: 2026-01-18
Categories: cs.CL
arXiv: https://arxiv.org/abs/2601.12369v5

Deep Research Agents increasingly automate survey writing, yet existing benchmarks do not jointly test whether they retrieve the papers experts consider essential and organize those papers into paper-grounded taxonomies. We introduce TaxoBench, a benchmark built from 72 highly cited LLM surveys, 3,815 cited papers, and their expert-authored taxonomies. TaxoBench evaluates systems in two settings: Deep Research mode measures end-to-end retrieval and organization from a topic, while Bottom-Up mode provides the expert paper set and isolates organization. We evaluate leaf-level assignments with ARI and V-Measure and hierarchy-level structure with US-TED, US-NTED, and Sem-Path. Across 7 Deep Research Agents and 16 LLM configurations, the best agent retrieves only 20.92% of expert-cited papers, and none of 70 standard Bottom-Up runs reaches the experts' average depth of 4.86. A controlled probe shows that models which match this depth do so by fragmenting the taxonomy, reducing alignment with the expert reference. We further find that raw Sem-Path remains near a no-organization floor even when a newer model generation gains 3.68 pp ARI; after depth matching, humans lead on all 10 matched surveys by 13.27 pp. These results identify retrieval and hierarchical organization as separate bottlenecks and show why hierarchy metrics must be calibrated before they are used to compare models.

---

### DepthCropSeg++: Scaling a Crop Segmentation Foundation Model With Depth-Labeled Data

Authors: Jiafei Zhang, Songliang Cao, Binghui Xu, Yanan Li, Weiwei Jia, Tingting Wu, Hao Lu, Weijuan Hu, Zhiguo Han
Published: 2026-01-18
Categories: cs.CV
arXiv: https://arxiv.org/abs/2601.12366v1

DepthCropSeg++: a foundation model for crop segmentation, capable of segmenting different crop species under open in-field environment. Crop segmentation is a fundamental task for modern agriculture, which closely relates to many downstream tasks such as plant phenotyping, density estimation, and weed control. In the era of foundation models, a number of generic large language and vision models have been developed. These models have demonstrated remarkable real world generalization due to significant model capacity and largescale datasets. However, current crop segmentation models mostly learn from limited data due to expensive pixel-level labelling cost, often performing well only under specific crop types or controlled environment. In this work, we follow the vein of our previous work DepthCropSeg, an almost unsupervised approach to crop segmentation, to scale up a cross-species and crossscene crop segmentation dataset, with 28,406 images across 30+ species and 15 environmental conditions. We also build upon a state-of-the-art semantic segmentation architecture ViT-Adapter architecture, enhance it with dynamic upsampling for improved detail awareness, and train the model with a two-stage selftraining pipeline. To systematically validate model performance, we conduct comprehensive experiments to justify the effectiveness and generalization capabilities across multiple crop datasets. Results demonstrate that DepthCropSeg++ achieves 93.11% mIoU on a comprehensive testing set, outperforming both supervised baselines and general-purpose vision foundation models like Segmentation Anything Model (SAM) by significant margins (+0.36% and +48.57% respectively). The model particularly excels in challenging scenarios including night-time environment (86.90% mIoU), high-density canopies (90.09% mIoU), and unseen crop varieties (90.09% mIoU), indicating a new state of the art for crop segmentation.

---

### Machine Learning-Based Framework for Real Time Detection and Early Prediction of Control Valve Stiction in Industrial Control Systems

Authors: Natthapong Promsricha, Chotirawee Chatpattanasiri, Nuttavut Kerdgongsup, Stavroula Balabani
Published: 2026-01-18
Categories: cs.LG, physics.ins-det
arXiv: https://arxiv.org/abs/2601.12362v1

Control valve stiction, a friction that prevents smooth valve movement, is a common fault in industrial process systems that causes instability, equipment wear, and higher maintenance costs. Many plants still operate with conventional valves that lack real time monitoring, making early predictions challenging. This study presents a machine learning (ML) framework for detecting and predicting stiction using only routinely collected process signals: the controller output (OP) from control systems and the process variable (PV), such as flow rate. Three deep learning models were developed and compared: a Convolutional Neural Network (CNN), a hybrid CNN with a Support Vector Machine (CNN-SVM), and a Long Short-Term Memory (LSTM) network. To train these models, a data-driven labeling method based on slope ratio analysis was applied to a real oil and gas refinery dataset. The LSTM model achieved the highest accuracy and was able to predict stiction up to four hours in advance. To the best of the authors' knowledge, this is the first study to demonstrate ML based early prediction of control valve stiction from real industry data. The proposed framework can be integrated into existing control systems to support predictive maintenance, reduce downtime, and avoid unnecessary hardware replacement.

---

### Zero-Shot Embedding Drift Detection: A Lightweight Defense Against Prompt Injections in LLMs

Authors: Anirudh Sekar, Mrinal Agarwal, Rachel Sharma, Akitsugu Tanaka, Jasmine Zhang, Arjun Damerla, Kevin Zhu
Published: 2026-01-18
Categories: cs.CR, cs.AI, cs.CL
arXiv: https://arxiv.org/abs/2601.12359v1

Prompt injection attacks have become an increasing vulnerability for LLM applications, where adversarial prompts exploit indirect input channels such as emails or user-generated content to circumvent alignment safeguards and induce harmful or unintended outputs. Despite advances in alignment, even state-of-the-art LLMs remain broadly vulnerable to adversarial prompts, underscoring the urgent need for robust, productive, and generalizable detection mechanisms beyond inefficient, model-specific patches. In this work, we propose Zero-Shot Embedding Drift Detection (ZEDD), a lightweight, low-engineering-overhead framework that identifies both direct and indirect prompt injection attempts by quantifying semantic shifts in embedding space between benign and suspect inputs. ZEDD operates without requiring access to model internals, prior knowledge of attack types, or task-specific retraining, enabling efficient zero-shot deployment across diverse LLM architectures. Our method uses adversarial-clean prompt pairs and measures embedding drift via cosine similarity to capture subtle adversarial manipulations inherent to real-world injection attacks. To ensure robust evaluation, we assemble and re-annotate the comprehensive LLMail-Inject dataset spanning five injection categories derived from publicly available sources. Extensive experiments demonstrate that embedding drift is a robust and transferable signal, outperforming traditional methods in detection accuracy and operational efficiency. With greater than 93% accuracy in classifying prompt injections across model architectures like Llama 3, Qwen 2, and Mistral and a false positive rate of <3%, our approach offers a lightweight, scalable defense layer that integrates into existing LLM pipelines, addressing a critical gap in securing LLM-powered systems to withstand adaptive adversarial threats.

---

### From Prompts to Pavement: LMMs-based Agentic Behavior-Tree Generation Framework for Autonomous Vehicles

Authors: Omar Y. Goba, Ahmed Y. Gado, Catherine M. Elias, Ahmed Hussein
Published: 2026-01-18
Categories: cs.CV, cs.AI, cs.RO
arXiv: https://arxiv.org/abs/2601.12358v1

Autonomous vehicles (AVs) require adaptive behavior planners to navigate unpredictable, real-world environments safely. Traditional behavior trees (BTs) offer structured decision logic but are inherently static and demand labor-intensive manual tuning, limiting their applicability at SAE Level 5 autonomy. This paper presents an agentic framework that leverages large language models (LLMs) and multi-modal vision models (LVMs) to generate and adapt BTs on the fly. A specialized Descriptor agent applies chain-of-symbols prompting to assess scene criticality, a Planner agent constructs high-level sub-goals via in-context learning, and a Generator agent synthesizes executable BT sub-trees in XML format. Integrated into a CARLA+Nav2 simulation, our system triggers only upon baseline BT failure, demonstrating successful navigation around unexpected obstacles (e.g., street blockage) with no human intervention. Compared to a static BT baseline, this approach is a proof-of-concept that extends to diverse driving scenarios.

---

### SimpleMatch: A Simple and Strong Baseline for Semantic Correspondence

Authors: Hailing Jin, Huiying Li
Published: 2026-01-18
Categories: cs.CV, cs.AI
arXiv: https://arxiv.org/abs/2601.12357v2

Recent advances in semantic correspondence have been largely driven by the use of pre-trained large-scale models. However, a limitation of these approaches is their dependence on high-resolution input images to achieve optimal performance, which results in considerable computational overhead. In this work, we address a fundamental limitation in current methods: the irreversible fusion of adjacent keypoint features caused by deep downsampling operations. This issue is triggered when semantically distinct keypoints fall within the same downsampled receptive field (e.g., 16x16 patches). To address this issue, we present SimpleMatch, a simple yet effective framework for semantic correspondence that delivers strong performance even at low resolutions. We propose a lightweight upsample decoder that progressively recovers spatial detail by upsampling deep features to 1/4 resolution, and a multi-scale supervised loss that ensures the upsampled features retain discriminative features across different spatial scales. In addition, we introduce sparse matching and window-based localization to optimize training memory usage and reduce it by 51%. At a resolution of 252x252 (3.3x smaller than current SOTA methods), SimpleMatch achieves superior performance with 84.1% PCK@0.1 on the SPair-71k benchmark. We believe this framework provides a practical and efficient baseline for future research in semantic correspondence. Code is available at: https://github.com/hailong23-jin/SimpleMatch.

---

### Tree-Structured Synergy of Large Language Models and Bayesian Optimization for Efficient CASH

Authors: Beicheng Xu, Weitong Qian, Lingching Tung, Yupeng Lu, Bin Cui
Published: 2026-01-18
Categories: cs.LG
arXiv: https://arxiv.org/abs/2601.12355v2

To lower the expertise barrier in machine learning, the AutoML community has focused on the CASH problem, which jointly automates algorithm selection and hyperparameter tuning. While traditional methods like Bayesian Optimization (BO) struggle with cold-start issues, Large Language Models (LLMs) can mitigate these through semantic priors. However, existing LLM-based optimizers generalize poorly to high-dimensional, structured CASH spaces. In this paper, we propose LB-MCTS, a trajectory-structured optimization framework that uses a Monte Carlo Tree Search tree as a shared state for algorithm selection, hyperparameter refinement, and BO-LLM proposer synergy. Within this shared state, BO provides algorithm-specific surrogate modeling for quantitative search, while the LLM exploits path-aware selective memory to generate semantic proposals and reflections. As the surrogate model improves, a reliability-aware proposer policy adaptively shifts from LLM-driven to BO-driven proposals within a unified search trajectory. Experiments on 104 AMLB datasets demonstrate that LB-MCTS consistently outperforms BO-based, LLM-based, and hybrid baselines.

---

### Bone-conduction Guided Multimodal Speech Enhancement with Conditional Diffusion Models

Authors: Sina Khanagha, Bunlong Lay, Timo Gerkmann
Published: 2026-01-18
Categories: eess.AS, cs.LG, cs.SD
arXiv: https://arxiv.org/abs/2601.12354v1

Single-channel speech enhancement models face significant performance degradation in extremely noisy environments. While prior work has shown that complementary bone-conducted speech can guide enhancement, effective integration of this noise-immune modality remains a challenge. This paper introduces a novel multimodal speech enhancement framework that integrates bone-conduction sensors with air-conducted microphones using a conditional diffusion model. Our proposed model significantly outperforms previously established multimodal techniques and a powerful diffusion-based single-modal baseline across a wide range of acoustic conditions.

---

### Mind the Gap: Action Rebinding Attacks against Android GUI Agents

Authors: Yi Qian, Kunwei Qian, Xingbang He, Ligeng Chen, Jikang Zhang, Tiantai Zhang, Haiyang Wei, Linzhang Wang, Hao Wu, Bing Mao
Published: 2026-01-18
Categories: cs.CR, cs.AI, cs.SE
arXiv: https://arxiv.org/abs/2601.12349v3

Large multimodal model powered GUI agents are emerging as high-privilege operators on mobile platforms, entrusted to perceive screen content and inject inputs across application boundaries. While these agents aim to automate complex tasks, we demonstrate that their design introduces a fundamental conflict with Android's strict application sandboxing. We present a novel cross-application Action Rebinding attack, which allows a malicious application with zero dangerous permissions to hijack the agent's execution and perform privileged operations on behalf of the attacker. Our attack exploits the inevitable observation-action gap inherent in the agent's reasoning pipeline. A malicious app can render a benign ``contextual carrier'' to elicit a planned action, and then swap the foreground to a sensitive target application during the reasoning latency. The agent, unaware of the transition, unwittingly executes the action in the privileged context. We further advance this attack by weaponizing the agent's own task-recovery logic to create programmable, multi-step exploit loops , and introducing an Intent Alignment Strategy (IAS) that manipulates the agent's reasoning to rationalize the hijacked state. We evaluate our attack on six widely-used Android GUI agents. Our results demonstrate a 100% success rate for atomic action hijacking and the ability to orchestrate high-impact exploits, including unauthorized file deletion, SMS transmission, and app uninstallation, without the attacker holding any corresponding permissions. Furthermore, since the malicious application separates intent from capability and contains no privileged API calls, it achieves a 0% detection rate across commercial malware scanners (e.g., VirusTotal), highlighting a critical blind spot in current mobile security analysis. To access experimental logs and demonstration videos, please contact yi_qian@smail.nju.edu.cn.

---

### MMDeepResearch-Bench: A Benchmark for Multimodal Deep Research Agents

Authors: Peizhou Huang, Zixuan Zhong, Zhongwei Wan, Donghao Zhou, Samiul Alam, Xin Wang, Zexin Li, Zhihao Dou, Li Zhu, Jing Xiong, Chaofan Tao, Yan Xu et al.
Published: 2026-01-18
Categories: cs.CV
arXiv: https://arxiv.org/abs/2601.12346v1

Deep Research Agents (DRAs) generate citation-rich reports via multi-step search and synthesis, yet existing benchmarks mainly target text-only settings or short-form multimodal QA, missing end-to-end multimodal evidence use. We introduce MMDeepResearch-Bench (MMDR-Bench), a benchmark of 140 expert-crafted tasks across 21 domains, where each task provides an image-text bundle to evaluate multimodal understanding and citation-grounded report generation. Compared to prior setups, MMDR-Bench emphasizes report-style synthesis with explicit evidence use, where models must connect visual artifacts to sourced claims and maintain consistency across narrative, citations, and visual references. We further propose a unified, interpretable evaluation pipeline: Formula-LLM Adaptive Evaluation (FLAE) for report quality, Trustworthy Retrieval-Aligned Citation Evaluation (TRACE) for citation-grounded evidence alignment, and Multimodal Support-Aligned Integrity Check (MOSAIC) for text-visual integrity, each producing fine-grained signals that support error diagnosis beyond a single overall score. Experiments across 25 state-of-the-art models reveal systematic trade-offs between generation quality, citation discipline, and multimodal grounding, highlighting that strong prose alone does not guarantee faithful evidence use and that multimodal integrity remains a key bottleneck for deep research agents.

---

### Adaptive Rotary Steering with Joint Autoregression for Robust Extraction of Closely Moving Speakers in Dynamic Scenarios

Authors: Jakob Kienegger, Timo Gerkmann
Published: 2026-01-18
Categories: eess.AS, cs.LG, cs.SD
arXiv: https://arxiv.org/abs/2601.12345v2

Latest advances in deep spatial filtering for Ambisonics demonstrate strong performance in stationary multi-speaker scenarios by rotating the sound field toward a target speaker prior to multi-channel enhancement. For applicability in dynamic acoustic conditions with moving speakers, we propose to automate this rotary steering using an interleaved tracking algorithm conditioned on the target's initial direction. However, for nearby or crossing speakers, robust tracking becomes difficult and spatial cues less effective for enhancement. By incorporating the processed recording as additional guide into both algorithms, our novel joint autoregressive framework leverages temporal-spectral correlations of speech to resolve spatially challenging speaker constellations. Consequently, our proposed method significantly improves tracking and enhancement of closely spaced speakers, consistently outperforming comparable non-autoregressive methods on a synthetic dataset. Real-world recordings complement these findings in complex scenarios with multiple speaker crossings and varying speaker-to-array distances.

---

### How Well Do LLMs Predict Human Behavior? A Measure of their Pretrained Knowledge

Authors: Wayne Gao, Sukjin Han, Annie Liang
Published: 2026-01-18
Categories: econ.EM, cs.AI, stat.ML
arXiv: https://arxiv.org/abs/2601.12343v1

Large language models (LLMs) are increasingly used to predict human behavior. We propose a measure for evaluating how much knowledge a pretrained LLM brings to such a prediction: its equivalent sample size, defined as the amount of task-specific data needed to match the predictive accuracy of the LLM. We estimate this measure by comparing the prediction error of a fixed LLM in a given domain to that of flexible machine learning models trained on increasing samples of domain-specific data. We further provide a statistical inference procedure by developing a new asymptotic theory for cross-validated prediction error. Finally, we apply this method to the Panel Study of Income Dynamics. We find that LLMs encode considerable predictive information for some economic variables but much less for others, suggesting that their value as substitutes for domain-specific data differs markedly across settings.

---

### Time-Continuous Modeling for Temporal Affective Pattern Recognition in LLMs

Authors: Rezky Kam, Coddy N. Siswanto
Published: 2026-01-18
Categories: cs.LG, cs.AI, cs.ET, cs.HC, eess.SY
arXiv: https://arxiv.org/abs/2601.12341v1

This paper introduces a dataset and conceptual framework for LLMs to mimic real world emotional dynamics through time and in-context learning leveraging physics-informed neural network, opening a possibility for interpretable dialogue modeling.

---

### Actionable Advice from Reviews via Mixture of LoRA Experts: A Two-LLM Pipeline for Issue Extraction and Business Recommendations

Authors: Kartikey Singh Bhandari, Manav Ganesh, Yashwant Viswanathan, Archit Agrawal, Dhruv Kumar, Pratik Narang
Published: 2026-01-18
Categories: cs.AI
arXiv: https://arxiv.org/abs/2601.12338v1

Customer reviews contain detailed, domain specific signals about service failures and user expectations, but converting this unstructured feedback into actionable business decisions remains difficult. We study review-to-action generation: producing concrete, implementable recommendations grounded in review text. We propose a modular two-LLM framework in which an Issue model extracts salient issues and assigns coarse themes, and an Advice model generates targeted operational fixes conditioned on the extracted issue representation. To enable specialization without expensive full fine-tuning, we adapt the Advice model using a mixture of LoRA experts strategy: multiple low-rank adapters are trained and a lightweight gating mechanism performs token-level expert mixing at inference, combining complementary expertise across issue types. We construct synthetic review-issue-advice triples from Yelp reviews (airlines and restaurants) to supervise training, and evaluate recommendations using an eight dimension operational rubric spanning actionability, specificity, feasibility, expected impact, novelty, non-redundancy, bias, and clarity. Across both domains, our approach consistently outperforms prompting-only and single-adapter baselines, yielding higher actionability and specificity while retaining favorable efficiency-quality trade-offs.

---

### Turbo-GoDec: Exploiting the Cluster Sparsity Prior for Hyperspectral Anomaly Detection

Authors: Jiahui Sheng, Xiaorun Li, Shuhan Chen
Published: 2026-01-18
Categories: cs.CV
arXiv: https://arxiv.org/abs/2601.12337v1

As a key task in hyperspectral image processing, hyperspectral anomaly detection has garnered significant attention and undergone extensive research. Existing methods primarily relt on two prior assumption: low-rank background and sparse anomaly, along with additional spatial assumptions of the background. However, most methods only utilize the sparsity prior assumption for anomalies and rarely expand on this hypothesis. From observations of hyperspectral images, we find that anomalous pixels exhibit certain spatial distribution characteristics: they often manifest as small, clustered groups in space, which we refer to as cluster sparsity of anomalies. Then, we combined the cluster sparsity prior with the classical GoDec algorithm, incorporating the cluster sparsity prior into the S-step of GoDec. This resulted in a new hyperspectral anomaly detection method, which we called Turbo-GoDec. In this approach, we modeled the cluster sparsity prior of anomalies using a Markov random field and computed the marginal probabilities of anomalies through message passing on a factor graph. Locations with high anomalous probabilities were treated as the sparse component in the Turbo-GoDec. Experiments are conducted on three real hyperspectral image (HSI) datasets which demonstrate the superior performance of the proposed Turbo-GoDec method in detecting small-size anomalies comparing with the vanilla GoDec (LSMAD) and state-of-the-art anomaly detection methods. The code is available at https://github.com/jiahuisheng/Turbo-GoDec.

---

### Efficient Privacy-Preserving Retrieval Augmented Generation with Distance-Preserving Encryption

Authors: Huanyi Ye, Jiale Guo, Ziyao Liu, Kwok-Yan Lam
Published: 2026-01-18
Categories: cs.CR, cs.AI
arXiv: https://arxiv.org/abs/2601.12331v1

RAG has emerged as a key technique for enhancing response quality of LLMs without high computational cost. In traditional architectures, RAG services are provided by a single entity that hosts the dataset within a trusted local environment. However, individuals or small organizations often lack the resources to maintain data storage servers, leading them to rely on outsourced cloud storage. This dependence on untrusted third-party services introduces privacy risks. Embedding-based retrieval mechanisms, commonly used in RAG systems, are vulnerable to privacy leakage such as vector-to-text reconstruction attacks and structural leakage via vector analysis. Several privacy-preserving RAG techniques have been proposed but most existing approaches rely on partially homomorphic encryption, which incurs substantial computational overhead. To address these challenges, we propose an efficient privacy-preserving RAG framework (ppRAG) tailored for untrusted cloud environments that defends against vector-to-text attack, vector analysis, and query analysis. We propose Conditional Approximate Distance-Comparison-Preserving Symmetric Encryption (CAPRISE) that encrypts embeddings while still allowing the cloud to compute similarity between an encrypted query and the encrypted database embeddings. CAPRISE preserves only the relative distance ordering between the encrypted query and each encrypted database embedding, without exposing inter-database distances, thereby enhancing both privacy and efficiency. To mitigate query analysis, we introduce DP by perturbing the query embedding prior to encryption, preventing the cloud from inferring sensitive patterns. Experimental results show that ppRAG achieves efficient processing throughput, high retrieval accuracy, strong privacy guarantees, making it a practical solution for resource-constrained users seeking secure cloud-augmented LLMs.

---

### IceWatch: Forecasting Glacial Lake Outburst Floods (GLOFs) using Multimodal Deep Learning

Authors: Zuha Fatima, Muhammad Anser Sohaib, Muhammad Talha, Ayesha Kanwal, Sidra Sultana, Nazia Perwaiz
Published: 2026-01-18
Categories: cs.LG, cs.AI
arXiv: https://arxiv.org/abs/2601.12330v1

Glacial Lake Outburst Floods (GLOFs) pose a serious threat in high mountain regions. They are hazardous to communities, infrastructure, and ecosystems further downstream. The classical methods of GLOF detection and prediction have so far mainly relied on hydrological modeling, threshold-based lake monitoring, and manual satellite image analysis. These approaches suffer from several drawbacks: slow updates, reliance on manual labor, and losses in accuracy when clouds interfere and/or lack on-site data. To tackle these challenges, we present IceWatch: a novel deep learning framework for GLOF prediction that incorporates both spatial and temporal perspectives. The vision component, RiskFlow, of IceWatch deals with Sentinel-2 multispectral satellite imagery using a CNN-based classifier and predicts GLOF events based on the spatial patterns of snow, ice, and meltwater. Its tabular counterpart confirms this prediction by considering physical dynamics. TerraFlow models glacier velocity from NASA ITS_LIVE time series while TempFlow forecasts near-surface temperature from MODIS LST records; both are trained on long-term observational archives and integrated via harmonized preprocessing and synchronization to enable multimodal, physics-informed GLOF prediction. Both together provide cross-validation, which will improve the reliability and interpretability of GLOF detection. This system ensures strong predictive performance, rapid data processing for real-time use, and robustness to noise and missing information. IceWatch paves the way for automatic, scalable GLOF warning systems. It also holds potential for integration with diverse sensor inputs and global glacier monitoring activities.

---

### FlowIID: Single-Step Intrinsic Image Decomposition via Latent Flow Matching

Authors: Mithlesh Singla, Seema Kumari, Shanmuganathan Raman
Published: 2026-01-18
Categories: cs.CV
arXiv: https://arxiv.org/abs/2601.12329v1

Intrinsic Image Decomposition (IID) separates an image into albedo and shading components. It is a core step in many real-world applications, such as relighting and material editing. Existing IID models achieve good results, but often use a large number of parameters. This makes them costly to combine with other models in real-world settings. To address this problem, we propose a flow matching-based solution. For this, we design a novel architecture, FlowIID, based on latent flow matching. FlowIID combines a VAE-guided latent space with a flow matching module, enabling a stable decomposition of albedo and shading. FlowIID is not only parameter-efficient, but also produces results in a single inference step. Despite its compact design, FlowIID delivers competitive and superior results compared to existing models across various benchmarks. This makes it well-suited for deployment in resource-constrained and real-time vision applications.

---

### The Expert Validation Framework (EVF): Enabling Domain Expert Control in AI Engineering

Authors: Lucas Gren, Felix Dobslaw
Published: 2026-01-18
Categories: cs.SE, cs.AI
arXiv: https://arxiv.org/abs/2601.12327v1

Generative AI (GenAI) systems promise to transform knowledge work by automating a range of tasks, yet their deployment in enterprise settings remains hindered by the lack of systematic quality assurance mechanisms. We present an Expert Validation Framework that places domain experts at the center of building software with GenAI components, enabling them to maintain authoritative control over system behavior through structured specification, testing, validation, and continuous monitoring processes. Our framework addresses the critical gap between AI capabilities and organizational trust by establishing a rigorous, expert-driven methodology for ensuring quality across diverse GenAI applications. Through a four-stage implementation process encompassing specification, system creation, validation, and production monitoring, the framework enables organizations to leverage GenAI capabilities while maintaining expert oversight and quality standards.

---