# Research Papers — 2026-01 · bundle page 118 of 189 > Full-text markdown of papers #5851–5900 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/118 [← Previous bundle](https://aipod.dev/papers/archive/2026-01/page/117.md) [Next bundle →](https://aipod.dev/papers/archive/2026-01/page/119.md) ### How Many Human Judgments Are Enough? Feasibility Limits of Human Preference Evaluation Authors: Wilson Y. Lee Published: 2026-01-14 Categories: cs.CL, cs.LG arXiv: https://arxiv.org/abs/2601.09084v2 Human preference evaluations are widely used to compare generative models, yet it remains unclear how many judgments are required to reliably detect small improvements. We show that when preference signal is diffuse across prompts (i.e., all prompt types are similarly informative), proportional allocation is minimax-optimal: no allocation strategy substantially improves detectability. Empirical analysis of large-scale human preference datasets shows that most comparisons fall into this diffuse regime, exhibiting small preference margins that require far more judgments than typically collected, even in well-sampled comparisons. These limits persist across evaluation protocols and modalities, including chat, image generation, and code generation with execution feedback. In contrast, curated benchmarks that reduce prompt induced variability systematically induce larger margins and improve detectability through a $1.5\times$ reduction in prompt-level variance. Our results show that inconclusive or negative human evaluation outcomes frequently reflect underpowered evaluation rather than model equivalence, underscoring the need to account explicitly for effect size, budget, and protocol design. --- ### SRT: Accelerating Reinforcement Learning via Speculative Rollout with Tree-Structured Cache Authors: Chi-Chih Chang, Siqi Zhu, Zhichen Zeng, Haibin Lin, Jiaxuan You, Mohamed S. Abdelfattah, Ziheng Jiang, Xuehai Qian Published: 2026-01-14 Categories: cs.LG arXiv: https://arxiv.org/abs/2601.09083v1 We present Speculative Rollout with Tree-Structured Cache (SRT), a simple, model-free approach to accelerate on-policy reinforcement learning (RL) for language models without sacrificing distributional correctness. SRT exploits the empirical similarity of rollouts for the same prompt across training steps by storing previously generated continuations in a per-prompt tree-structured cache. During generation, the current policy uses this tree as the draft model for performing speculative decoding. To keep the cache fresh and improve draft model quality, SRT updates trees online from ongoing rollouts and proactively performs run-ahead generation during idle GPU bubbles. Integrated into standard RL pipelines (\textit{e.g.}, PPO, GRPO and DAPO) and multi-turn settings, SRT consistently reduces generation and step latency and lowers per-token inference cost, achieving up to 2.08x wall-clock time speedup during rollout. --- ### Exploring Reliable Spatiotemporal Dependencies for Efficient Visual Tracking Authors: Junze Shi, Yang Yu, Jian Shi, Haibo Luo Published: 2026-01-14 Categories: cs.CV arXiv: https://arxiv.org/abs/2601.09078v1 Recent advances in transformer-based lightweight object tracking have established new standards across benchmarks, leveraging the global receptive field and powerful feature extraction capabilities of attention mechanisms. Despite these achievements, existing methods universally employ sparse sampling during training--utilizing only one template and one search image per sequence--which fails to comprehensively explore spatiotemporal information in videos. This limitation constrains performance and cause the gap between lightweight and high-performance trackers. To bridge this divide while maintaining real-time efficiency, we propose STDTrack, a framework that pioneers the integration of reliable spatiotemporal dependencies into lightweight trackers. Our approach implements dense video sampling to maximize spatiotemporal information utilization. We introduce a temporally propagating spatiotemporal token to guide per-frame feature extraction. To ensure comprehensive target state representation, we disign the Multi-frame Information Fusion Module (MFIFM), which augments current dependencies using historical context. The MFIFM operates on features stored in our constructed Spatiotemporal Token Maintainer (STM), where a quality-based update mechanism ensures information reliability. Considering the scale variation among tracking targets, we develop a multi-scale prediction head to dynamically adapt to objects of different sizes. Extensive experiments demonstrate state-of-the-art results across six benchmarks. Notably, on GOT-10k, STDTrack rivals certain high-performance non-real-time trackers (e.g., MixFormer) while operating at 192 FPS(GPU) and 41 FPS(CPU). --- ### Lean Clients, Full Accuracy: Hybrid Zeroth- and First-Order Split Federated Learning Authors: Zhoubin Kou, Zihan Chen, Jing Yang, Cong Shen Published: 2026-01-14 Categories: cs.LG, cs.DC, cs.IT, cs.NI, eess.SP arXiv: https://arxiv.org/abs/2601.09076v1 Split Federated Learning (SFL) enables collaborative training between resource-constrained edge devices and a compute-rich server. Communication overhead is a central issue in SFL and can be mitigated with auxiliary networks. Yet, the fundamental client-side computation challenge remains, as back-propagation requires substantial memory and computation costs, severely limiting the scale of models that edge devices can support. To enable more resource-efficient client computation and reduce the client-server communication, we propose HERON-SFL, a novel hybrid optimization framework that integrates zeroth-order (ZO) optimization for local client training while retaining first-order (FO) optimization on the server. With the assistance of auxiliary networks, ZO updates enable clients to approximate local gradients using perturbed forward-only evaluations per step, eliminating memory-intensive activation caching and avoiding explicit gradient computation in the traditional training process. Leveraging the low effective rank assumption, we theoretically prove that HERON-SFL's convergence rate is independent of model dimensionality, addressing a key scalability concern common to ZO algorithms. Empirically, on ResNet training and language model (LM) fine-tuning tasks, HERON-SFL matches benchmark accuracy while reducing client peak memory by up to 64% and client-side compute cost by up to 33% per step, substantially expanding the range of models that can be trained or adapted on resource-limited devices. --- ### Human-AI Co-design for Clinical Prediction Models Authors: Jean Feng, Avni Kothari, Patrick Vossler, Andrew Bishara, Lucas Zier, Newton Addo, Aaron Kornblith, Yan Shuo Tan, Chandan Singh Published: 2026-01-14 Categories: cs.AI, cs.CL, stat.ME arXiv: https://arxiv.org/abs/2601.09072v1 Developing safe, effective, and practically useful clinical prediction models (CPMs) traditionally requires iterative collaboration between clinical experts, data scientists, and informaticists. This process refines the often small but critical details of the model building process, such as which features/patients to include and how clinical categories should be defined. However, this traditional collaboration process is extremely time- and resource-intensive, resulting in only a small fraction of CPMs reaching clinical practice. This challenge intensifies when teams attempt to incorporate unstructured clinical notes, which can contain an enormous number of concepts. To address this challenge, we introduce HACHI, an iterative human-in-the-loop framework that uses AI agents to accelerate the development of fully interpretable CPMs by enabling the exploration of concepts in clinical notes. HACHI alternates between (i) an AI agent rapidly exploring and evaluating candidate concepts in clinical notes and (ii) clinical and domain experts providing feedback to improve the CPM learning process. HACHI defines concepts as simple yes-no questions that are used in linear models, allowing the clinical AI team to transparently review, refine, and validate the CPM learned in each round. In two real-world prediction tasks (acute kidney injury and traumatic brain injury), HACHI outperforms existing approaches, surfaces new clinically relevant concepts not included in commonly-used CPMs, and improves model generalizability across clinical sites and time periods. Furthermore, HACHI reveals the critical role of the clinical AI team, such as directing the AI agent to explore concepts that it had not previously considered, adjusting the granularity of concepts it considers, changing the objective function to better align with the clinical objectives, and identifying issues of data bias and leakage. --- ### Resolving Predictive Multiplicity for the Rashomon Set Authors: Parian Haghighat, Hadis Anahideh, Cynthia Rudin Published: 2026-01-14 Categories: cs.LG arXiv: https://arxiv.org/abs/2601.09071v2 The existence of multiple, equally accurate models for a given predictive task leads to predictive multiplicity, where a Rashomon set of models achieve similar accuracy but diverge in their individual predictions. This inconsistency undermines trust in high-stakes applications where we want consistent predictions. We propose three approaches to reduce inconsistency among predictions for the members of the Rashomon set. The first approach is outlier correction. An outlier has a label that none of the good models are capable of predicting correctly. Outliers can cause the Rashomon set to have high variance predictions in a local area, so fixing them can lower variance. Our second approach is local patching. In a local region around a test point, models may disagree with each other because some of them are biased. We can detect and fix such biases using a validation set, which also reduces multiplicity. Our third approach is pairwise reconciliation, where we find pairs of models that disagree on a region around the test point. We modify predictions that disagree, making them less biased. These three approaches can be used together or separately, and they each have distinct advantages. The reconciled predictions can then be distilled into a single interpretable model for real-world deployment. In experiments across multiple datasets, our methods reduce disagreement metrics while maintaining competitive accuracy. --- ### From Symbolic to Natural-Language Relations: Rethinking Knowledge Graph Construction in the Era of Large Language Models Authors: Kanyao Han, Yushang Lai Published: 2026-01-14 Categories: cs.CL, cs.AI arXiv: https://arxiv.org/abs/2601.09069v1 Knowledge graphs (KGs) have commonly been constructed using predefined symbolic relation schemas, typically implemented as categorical relation labels. This design has notable shortcomings: real-world relations are often contextual, nuanced, and sometimes uncertain, and compressing it into discrete relation labels abstracts away critical semantic detail. Nevertheless, symbolic-relation KGs remain widely used because they have been operationally effective and broadly compatible with pre-LLM downstream models and algorithms, in which KG knowledge could be retrieved or encoded into quantified features and embeddings at scale. The emergence of LLMs has reshaped how knowledge is created and consumed. LLMs support scalable synthesis of domain facts directly in concise natural language, and prompting-based inference favors context-rich free-form text over quantified representations. This position paper argues that these changes call for rethinking the representation of relations themselves rather than merely using LLMs to populate conventional schemas more efficiently. We therefore advocate moving from symbolic to natural-language relation descriptions, and we propose hybrid design principles that preserve a minimal structural backbone while enabling more flexible and context-sensitive relational representations. --- ### Mi:dm 2.0 Korea-centric Bilingual Language Models Authors: Donghoon Shin, Sejung Lee, Soonmin Bae, Hwijung Ryu, Changwon Ok, Hoyoun Jung, Hyesung Ji, Jeehyun Lim, Jehoon Lee, Ji-Eun Han, Jisoo Baik, Mihyeon Kim et al. Published: 2026-01-14 Categories: cs.CL, cs.AI arXiv: https://arxiv.org/abs/2601.09066v1 We introduce Mi:dm 2.0, a bilingual large language model (LLM) specifically engineered to advance Korea-centric AI. This model goes beyond Korean text processing by integrating the values, reasoning patterns, and commonsense knowledge inherent to Korean society, enabling nuanced understanding of cultural contexts, emotional subtleties, and real-world scenarios to generate reliable and culturally appropriate responses. To address limitations of existing LLMs, often caused by insufficient or low-quality Korean data and lack of cultural alignment, Mi:dm 2.0 emphasizes robust data quality through a comprehensive pipeline that includes proprietary data cleansing, high-quality synthetic data generation, strategic data mixing with curriculum learning, and a custom Korean-optimized tokenizer to improve efficiency and coverage. To realize this vision, we offer two complementary configurations: Mi:dm 2.0 Base (11.5B parameters), built with a depth-up scaling strategy for general-purpose use, and Mi:dm 2.0 Mini (2.3B parameters), optimized for resource-constrained environments and specialized tasks. Mi:dm 2.0 achieves state-of-the-art performance on Korean-specific benchmarks, with top-tier zero-shot results on KMMLU and strong internal evaluation results across language, humanities, and social science tasks. The Mi:dm 2.0 lineup is released under the MIT license to support extensive research and commercial use. By offering accessible and high-performance Korea-centric LLMs, KT aims to accelerate AI adoption across Korean industries, public services, and education, strengthen the Korean AI developer community, and lay the groundwork for the broader vision of K-intelligence. Our models are available at https://huggingface.co/K-intelligence. For technical inquiries, please contact midm-llm@kt.com. --- ### Beyond Consensus: Perspectivist Modeling and Evaluation of Annotator Disagreement in NLP Authors: Yinuo Xu, David Jurgens Published: 2026-01-14 Categories: cs.CL arXiv: https://arxiv.org/abs/2601.09065v2 Annotator disagreement is widespread in NLP, particularly for subjective and ambiguous tasks such as toxicity detection and stance analysis. While early approaches treated disagreement as noise to be removed, recent work increasingly models it as a meaningful signal reflecting variation in interpretation and perspective. This survey provides a unified view of disagreement-aware NLP methods. We first present a domain-agnostic taxonomy of the sources of disagreement spanning data, task, and annotator factors. We then synthesize modeling approaches using a common framework defined by prediction targets and pooling structure, highlighting a shift from consensus learning toward explicitly modeling disagreement, and toward capturing structured relationships among annotators. We review evaluation metrics for both predictive performance and annotator behavior, and noting that most fairness evaluations remain descriptive rather than normative. We conclude by identifying open challenges and future directions, including integrating multiple sources of variation, developing disagreement-aware interpretability frameworks, and grappling with the practical tradeoffs of perspectivist modeling. --- ### Efficient Multilingual Dialogue Processing via Translation Pipelines and Distilled Language Models Authors: Santiago Martínez Novoa, Nicolás Rozo Fajardo, Diego Alejandro González Vargas, Nicolás Bedoya Figueroa Published: 2026-01-14 Categories: cs.CL arXiv: https://arxiv.org/abs/2601.09059v1 This paper presents team Kl33n3x's multilingual dialogue summarization and question answering system developed for the NLPAI4Health 2025 shared task. The approach employs a three-stage pipeline: forward translation from Indic languages to English, multitask text generation using a 2.55B parameter distilled language model, and reverse translation back to source languages. By leveraging knowledge distillation techniques, this work demonstrates that compact models can achieve highly competitive performance across nine languages. The system achieved strong win rates across the competition's tasks, with particularly robust performance on Marathi (86.7% QnA), Tamil (86.7% QnA), and Hindi (80.0% QnA), demonstrating the effectiveness of translation-based approaches for low-resource language processing without task-specific fine-tuning. --- ### StegoStylo: Squelching Stylometric Scrutiny through Steganographic Stitching Authors: Robert Dilworth Published: 2026-01-14 Categories: cs.CR, cs.CL, cs.IR arXiv: https://arxiv.org/abs/2601.09056v5 Stylometry -- the identification of an author through analysis of a text's style (i.e., authorship attribution) -- serves many constructive purposes: it supports copyright and plagiarism investigations, aids detection of harmful content, offers exploratory cues for certain medical conditions (e.g., early signs of dementia or depression), provides historical context for literary works, and helps uncover misinformation and disinformation. In contrast, when stylometry is employed as a tool for authorship verification -- confirming whether a text truly originates from a claimed author -- it can also be weaponized for malicious purposes. Techniques such as de-anonymization, re-identification, tracking, profiling, and downstream effects like censorship illustrate the privacy threats that stylometric analysis can enable. Building on these concerns, this paper further explores how adversarial stylometry combined with steganography can counteract stylometric analysis. We first present enhancements to our adversarial attack, $\textit{TraceTarnish}$, providing stronger evidence of its capacity to confound stylometric systems and reduce their attribution and verification accuracy. Next, we examine how steganographic embedding can be fine-tuned to mask an author's stylistic fingerprint, quantifying the level of authorship obfuscation achievable as a function of the proportion of words altered with zero-width Unicode characters. Based on our findings, steganographic coverage of 33% or higher seemingly ensures authorship obfuscation. Finally, we reflect on the ways stylometry can be used to undermine privacy and argue for the necessity of defensive tools like $\textit{TraceTarnish}$. --- ### Deep Incomplete Multi-View Clustering via Hierarchical Imputation and Alignment Authors: Yiming Du, Ziyu Wang, Jian Li, Rui Ning, Lusi Li Published: 2026-01-14 Categories: cs.LG arXiv: https://arxiv.org/abs/2601.09051v1 Incomplete multi-view clustering (IMVC) aims to discover shared cluster structures from multi-view data with partial observations. The core challenges lie in accurately imputing missing views without introducing bias, while maintaining semantic consistency across views and compactness within clusters. To address these challenges, we propose DIMVC-HIA, a novel deep IMVC framework that integrates hierarchical imputation and alignment with four key components: (1) view-specific autoencoders for latent feature extraction, coupled with a view-shared clustering predictor to produce soft cluster assignments; (2) a hierarchical imputation module that first estimates missing cluster assignments based on cross-view contrastive similarity, and then reconstructs missing features using intra-view, intra-cluster statistics; (3) an energy-based semantic alignment module, which promotes intra-cluster compactness by minimizing energy variance around low-energy cluster anchors; and (4) a contrastive assignment alignment module, which enhances cross-view consistency and encourages confident, well-separated cluster predictions. Experiments on benchmarks demonstrate that our framework achieves superior performance under varying levels of missingness. --- ### SITA: Learning Speaker-Invariant and Tone-Aware Speech Representations for Low-Resource Tonal Languages Authors: Tianyi Xu, Xuan Ouyang, Binwei Yao, Shoua Xiong, Sara Misurelli, Maichou Lor, Junjie Hu Published: 2026-01-14 Categories: cs.CL arXiv: https://arxiv.org/abs/2601.09050v2 Tonal low-resource languages are widely spoken but remain underserved by modern speech technologies. A central challenge is learning speech representations that are robust to nuisance variation, such as speaker gender, while preserving lexical tone, which carries word meaning. We propose SITA, a lightweight adaptation recipe for pretrained wav2vec-style self-supervised speech encoders. Rather than designing a new backbone or objective, SITA combines existing objectives in a staged optimization framework to reduce tone collapse while preserving ASR capability. Stage 1 improves speaker invariance without erasing tonal contrasts by combining a cross-gender contrastive loss with a tone-repulsive loss that separates same-word, different-tone realizations. Stage 2 restores recognition-oriented linguistic information through CTC fine-tuning and knowledge distillation on upper encoder layers. We evaluate SITA primarily on Hmong, a tonal language with limited digital resources and a small speaker pool. Against multilingual, speaker-adversarial, label-aware, and semi-supervised baselines, SITA achieves the best trade-off between cross-gender lexical retrieval and tone separation, while maintaining ASR accuracy close to an ASR-adapted XLS-R teacher. Results on Mandarin show consistent gains, suggesting that SITA is a general plug-in recipe for tonal speech representation learning. --- ### Is Grokking Worthwhile? Functional Analysis and Transferability of Generalization Circuits in Transformers Authors: Kaiyu He, Zhang Mian, Peilin Wu, Xinya Du, Zhiyu Chen Published: 2026-01-14 Categories: cs.CL, cs.AI arXiv: https://arxiv.org/abs/2601.09049v1 While Large Language Models (LLMs) excel at factual retrieval, they often struggle with the "curse of two-hop reasoning" in compositional tasks. Recent research suggests that parameter-sharing transformers can bridge this gap by forming a "Generalization Circuit" during a prolonged "grokking" phase. A fundamental question arises: Is a grokked model superior to its non-grokked counterparts on downstream tasks? Furthermore, is the extensive computational cost of waiting for the grokking phase worthwhile? In this work, we conduct a mechanistic study to evaluate the Generalization Circuit's role in knowledge assimilation and transfer. We demonstrate that: (i) The inference paths established by non-grokked and grokked models for in-distribution compositional queries are identical. This suggests that the "Generalization Circuit" does not represent the sudden acquisition of a new reasoning paradigm. Instead, we argue that grokking is the process of integrating memorized atomic facts into an naturally established reasoning path. (ii) Achieving high accuracy on unseen cases after prolonged training and the formation of a certain reasoning path are not bound; they can occur independently under specific data regimes. (iii) Even a mature circuit exhibits limited transferability when integrating new knowledge, suggesting that "grokked" Transformers do not achieve a full mastery of compositional logic. --- ### POWDR: Pathology-preserving Outpainting with Wavelet Diffusion for 3D MRI Authors: Fei Tan, Ashok Vardhan Addala, Bruno Astuto Arouche Nunes, Xucheng Zhu, Ravi Soni Published: 2026-01-14 Categories: eess.IV, cs.CV arXiv: https://arxiv.org/abs/2601.09044v1 Medical imaging datasets often suffer from class imbalance and limited availability of pathology-rich cases, which constrains the performance of machine learning models for segmentation, classification, and vision-language tasks. To address this challenge, we propose POWDR, a pathology-preserving outpainting framework for 3D MRI based on a conditioned wavelet diffusion model. Unlike conventional augmentation or unconditional synthesis, POWDR retains real pathological regions while generating anatomically plausible surrounding tissue, enabling diversity without fabricating lesions. Our approach leverages wavelet-domain conditioning to enhance high-frequency detail and mitigate blurring common in latent diffusion models. We introduce a random connected mask training strategy to overcome conditioning-induced collapse and improve diversity outside the lesion. POWDR is evaluated on brain MRI using BraTS datasets and extended to knee MRI to demonstrate tissue-agnostic applicability. Quantitative metrics (FID, SSIM, LPIPS) confirm image realism, while diversity analysis shows significant improvement with random-mask training (cosine similarity reduced from 0.9947 to 0.9580; KL divergence increased from 0.00026 to 0.01494). Clinically relevant assessments reveal gains in tumor segmentation performance using nnU-Net, with Dice scores improving from 0.6992 to 0.7137 when adding 50 synthetic cases. Tissue volume analysis indicates no significant differences for CSF and GM compared to real images. These findings highlight POWDR as a practical solution for addressing data scarcity and class imbalance in medical imaging. The method is extensible to multiple anatomies and offers a controllable framework for generating diverse, pathology-preserving synthetic data to support robust model development. --- ### Horseshoe Mixtures-of-Experts (HS-MoE) Authors: Nick Polson, Vadim Sokolov Published: 2026-01-14 Categories: stat.ML, cs.LG arXiv: https://arxiv.org/abs/2601.09043v1 Horseshoe mixtures-of-experts (HS-MoE) models provide a Bayesian framework for sparse expert selection in mixture-of-experts architectures. We combine the horseshoe prior's adaptive global-local shrinkage with input-dependent gating, yielding data-adaptive sparsity in expert usage. Our primary methodological contribution is a particle learning algorithm for sequential inference, in which the filter is propagated forward in time while tracking only sufficient statistics. We also discuss how HS-MoE relates to modern mixture-of-experts layers in large language models, which are deployed under extreme sparsity constraints (e.g., activating a small number of experts per token out of a large pool). --- ### SCaLE: Switching Cost aware Learning and Exploration Authors: Neelkamal Bhuyan, Debankur Mukherjee, Adam Wierman Published: 2026-01-14 Categories: cs.LG, cs.DS, math.OC, math.PR, stat.ML arXiv: https://arxiv.org/abs/2601.09042v1 This work addresses the fundamental problem of unbounded metric movement costs in bandit online convex optimization, by considering high-dimensional dynamic quadratic hitting costs and $\ell_2$-norm switching costs in a noisy bandit feedback model. For a general class of stochastic environments, we provide the first algorithm SCaLE that provably achieves a distribution-agnostic sub-linear dynamic regret, without the knowledge of hitting cost structure. En-route, we present a novel spectral regret analysis that separately quantifies eigenvalue-error driven regret and eigenbasis-perturbation driven regret. Extensive numerical experiments, against online-learning baselines, corroborate our claims, and highlight statistical consistency of our algorithm. --- ### Can LLMs interpret figurative language as humans do?: surface-level vs representational similarity Authors: Samhita Bollepally, Aurora Sloman-Moll, Takashi Yamauchi Published: 2026-01-14 Categories: cs.CL, cs.AI arXiv: https://arxiv.org/abs/2601.09041v1 Large language models generate judgments that resemble those of humans. Yet the extent to which these models align with human judgments in interpreting figurative and socially grounded language remains uncertain. To investigate this, human participants and four instruction-tuned LLMs of different sizes (GPT-4, Gemma-2-9B, Llama-3.2, and Mistral-7B) rated 240 dialogue-based sentences representing six linguistic traits: conventionality, sarcasm, funny, emotional, idiomacy, and slang. Each of the 240 sentences was paired with 40 interpretive questions, and both humans and LLMs rated these sentences on a 10-point Likert scale. Results indicated that humans and LLMs aligned at the surface level with humans, but diverged significantly at the representational level, especially in interpreting figurative sentences involving idioms and Gen Z slang. GPT-4 most closely approximates human representational patterns, while all models struggle with context-dependent and socio-pragmatic expressions like sarcasm, slang, and idiomacy. --- ### Depth-Wise Representation Development Under Blockwise Self-Supervised Learning for Video Vision Transformers Authors: Jonas Römer, Timo Dickscheid Published: 2026-01-14 Categories: cs.CV arXiv: https://arxiv.org/abs/2601.09040v1 End-to-end backpropagation couples all layers through a global error signal, enabling coordinated learning but requiring long-range credit assignment. Motivated by recent progress in blockwise self-supervised learning (BWSSL), we ask whether masked video transformers can be trained without end-to-end backpropagation. Applying BWSSL to masked video modeling remains relatively underexplored and must handle spatiotemporal context and long-range temporal structure. More broadly, analyses that compare BWSSL and end-to-end training in terms of learning dynamics and depth-wise representation development remain sparse. We apply blockwise learning to a masked autoencoding video vision transformer by partitioning the encoder into blocks, each of which is optimized with a local masked reconstruction loss. Across model sizes and partition granularities, training converges and yields representations close to matched end-to-end baselines under linear-probe and retrieval proxies. In order to compare intermediate representations, we analyze depth-wise decodability, inter-block similarity, and patch-level diagnostics. Blockwise training exposes higher-level structure earlier, while later blocks saturate and operate in a more geometry-preserving regime. It can also induce token-level shifts consistent with stronger early mixing that pooled metrics can miss. These findings point to late-block saturation and interface formation as contributors to the remaining gap. --- ### SpectraQuery: A Hybrid Retrieval-Augmented Conversational Assistant for Battery Science Authors: Sreya Vangara, Jagjit Nanda, Yan-Kai Tzeng, Eric Darve Published: 2026-01-14 Categories: cs.CL, cs.IR arXiv: https://arxiv.org/abs/2601.09036v1 Scientific reasoning increasingly requires linking structured experimental data with the unstructured literature that explains it, yet most large language model (LLM) assistants cannot reason jointly across these modalities. We introduce SpectraQuery, a hybrid natural-language query framework that integrates a relational Raman spectroscopy database with a vector-indexed scientific literature corpus using a Structured and Unstructured Query Language (SUQL)-inspired design. By combining semantic parsing with retrieval-augmented generation, SpectraQuery translates open-ended questions into coordinated SQL and literature retrieval operations, producing cited answers that unify numerical evidence with mechanistic explanation. Across SQL correctness, answer groundedness, retrieval effectiveness, and expert evaluation, SpectraQuery demonstrates strong performance: approximately 80 percent of generated SQL queries are fully correct, synthesized answers reach 93-97 percent groundedness with 10-15 retrieved passages, and battery scientists rate responses highly across accuracy, relevance, grounding, and clarity (4.1-4.6/5). These results show that hybrid retrieval architectures can meaningfully support scientific workflows by bridging data and discourse for high-volume experimental datasets. --- ### A Decompilation-Driven Framework for Malware Detection with Large Language Models Authors: Aniesh Chawla, Udbhav Prasad Published: 2026-01-14 Categories: cs.CR, cs.AI arXiv: https://arxiv.org/abs/2601.09035v1 The parallel evolution of Large Language Models (LLMs) with advanced code-understanding capabilities and the increasing sophistication of malware presents a new frontier for cybersecurity research. This paper evaluates the efficacy of state-of-the-art LLMs in classifying executable code as either benign or malicious. We introduce an automated pipeline that first decompiles Windows executable into a C code using Ghidra disassembler and then leverages LLMs to perform the classification. Our evaluation reveals that while standard LLMs show promise, they are not yet robust enough to replace traditional anti-virus software. We demonstrate that a fine-tuned model, trained on curated malware and benign datasets, significantly outperforms its vanilla counterpart. However, the performance of even this specialized model degrades notably when encountering newer malware. This finding demonstrates the critical need for continuous fine-tuning with emerging threats to maintain model effectiveness against the changing coding patterns and behaviors of malicious software. --- ### Comparison of Outlier Detection Algorithms on String Data Authors: Philip Maus Published: 2026-01-13 Categories: cs.LG arXiv: https://arxiv.org/abs/2603.11049v1 Outlier detection is a well-researched and crucial problem in machine learning. However, there is little research on string data outlier detection, as most literature focuses on outlier detection of numerical data. A robust string data outlier detection algorithm could assist with data cleaning or anomaly detection in system log files. In this thesis, we compare two string outlier detection algorithms. Firstly, we introduce a variant of the well-known local outlier factor algorithm, which we tailor to detect outliers on string data using the Levenshtein measure to calculate the density of the dataset. We present a differently weighted Levenshtein measure, which considers hierarchical character classes and can be used to tune the algorithm to a specific string dataset. Secondly, we introduce a new kind of outlier detection algorithm based on the hierarchical left regular expression learner, which infers a regular expression for the expected data. Using various datasets and parameters, we experimentally show that both algorithms can conceptually find outliers in string data. We show that the regular expression-based algorithm is especially good at finding outliers if the expected values have a distinct structure that is sufficiently different from the structure of the outliers. In contrast, the local outlier factor algorithms are best at finding outliers if their edit distance to the expected data is sufficiently distinct from the edit distance between the expected data. --- ### Do Open-Vocabulary Detectors Transfer to Aerial Imagery? A Comparative Evaluation Authors: Christos Tsourveloudis Published: 2026-01-13 Categories: cs.CV, cs.LG, cs.RO arXiv: https://arxiv.org/abs/2601.22164v1 Open-vocabulary object detection (OVD) enables zero-shot recognition of novel categories through vision-language models, achieving strong performance on natural images. However, transferability to aerial imagery remains unexplored. We present the first systematic benchmark evaluating five state-of-the-art OVD models on the LAE-80C aerial dataset (3,592 images, 80 categories) under strict zero-shot conditions. Our experimental protocol isolates semantic confusion from visual localization through Global, Oracle, and Single-Category inference modes. Results reveal severe domain transfer failure: the best model (OWLv2) achieves only 27.6% F1-score with 69% false positive rate. Critically, reducing vocabulary size from 80 to 3.2 classes yields 15x improvement, demonstrating that semantic confusion is the primary bottleneck. Prompt engineering strategies such as domain-specific prefixing and synonym expansion, fail to provide meaningful performance gains. Performance varies dramatically across datasets (F1: 0.53 on DIOR, 0.12 on FAIR1M), exposing brittleness to imaging conditions. These findings establish baseline expectations and highlight the need for domain-adaptive approaches in aerial OVD. --- ### Mem2ActBench: A Benchmark for Evaluating Long-Term Memory Utilization in Task-Oriented Autonomous Agents Authors: Yiting Shen, Kun Li, Wei Zhou, Songlin Hu Published: 2026-01-13 Categories: cs.CL, cs.AI arXiv: https://arxiv.org/abs/2601.19935v1 Large Language Model (LLM)-based agents are increasingly deployed for complex, tool-based tasks where long-term memory is critical to driving actions. Existing benchmarks, however, primarily test a angent's ability to passively retrieve isolated facts in response to explicit questions. They fail to evaluate the more crucial capability of actively applying memory to execute tasks. To address this gap, we introduce \textsc{Mem2ActBench}, a benchmark for evaluating whether agents can proactively leverage long-term memory to execute tool-based actions by selecting appropriate tools and grounding their parameters. The benchmark simulates persistent assistant usage, where users mention the same topic across long, interrupted interactions and expect previously established preferences and task states to be implicitly applied. We build the dataset with an automated pipeline that merges heterogeneous sources (ToolACE, BFCL, Oasst1), resolves conflicts via consistency modeling, and synthesizes 2,029 sessions with 12 user--assistant--tool turns on average. From these memory chains, a reverse-generation method produces 400 tool-use tasks, with human evaluation confirming 91.3\% are strongly memory-dependent. Experiments on seven memory frameworks show that current systems remain inadequate at actively utilizing memory for parameter grounding, highlighting the need for more effective approaches to evaluate and improve memory application in task execution. --- ### Agentic Digital Twins: A Taxonomy of Capabilities for Understanding Possible Futures Authors: Christopher Burr, Mark Enzer, Jason Shepherd, David Wagg Published: 2026-01-13 Categories: cs.CY, cs.AI, cs.ET arXiv: https://arxiv.org/abs/2601.18799v1 As digital twins (DTs) evolve to become more agentic through the integration of artificial intelligence (AI), they acquire capabilities that extend beyond dynamic representation of their target systems. This paper presents a taxonomy of agentic DTs organised around three fundamental dimensions: the locus of agency (external, internal, distributed), the tightness of coupling (loose, tight, constitutive), and model evolution (static, adaptive, reconstructive). From the resulting 27-configuration space, we identify nine illustrative configurations grouped into three clusters: "The Present" (existing tools and emerging steering systems), "The Threshold" (where emergent properties appear and coupling becomes constitutive), and "The Frontier" (where systems gain reconstructive capabilities). Our analysis explores how agentic DTs exercise performative power--not merely representing physical systems but actively participating in constituting them. Using traffic navigation systems as examples, we show how even passive tools can exhibit emergent performativity, while advanced configurations risk performative lock-in. Drawing on performative prediction theory, we trace a progression from passive tools through active steering to ontological reconstruction, examining how constitutive coupling enables systems to create self-validating realities. Understanding these configurations is essential for navigating the transformation from DTs as mirror worlds to DTs as architects of new ontologies. --- ### Investigating Self-regulated Learning Sequences within a Generative AI-based Intelligent Tutoring System Authors: Jie Gao, Shasha Li, Jianhua Zhang, Shan Li, Tingting Wang Published: 2026-01-13 Categories: cs.CY, cs.AI arXiv: https://arxiv.org/abs/2601.17000v1 There has been a growing trend in employing generative artificial intelligence (GenAI) techniques to support learning. Moreover, scholars have reached a consensus on the critical role of self-regulated learning (SRL) in ensuring learning effectiveness within GenAI-assisted learning environments, making it essential to capture students' dynamic SRL patterns. In this study, we extracted students' interaction patterns with GenAI from trace data as they completed a problem-solving task within a GenAI-assisted intelligent tutoring system. Students' purpose of using GenAI was also analyzed from the perspective of information processing, i.e., information acquisition and information transformation. Using sequential and clustering analysis, this study classified participants into two groups based on their SRL sequences. These two groups differed in the frequency and temporal characteristics of GenAI use. In addition, most students used GenAI for information acquisition rather than information transformation, while the correlation between the purpose of using GenAI and learning performance was not statistically significant. Our findings inform both pedagogical design and the development of GenAI-assisted learning environments. --- ### Uncertainty Quantification for Named Entity Recognition via Full-Sequence and Subsequence Conformal Prediction Authors: Matthew Singer, Srijan Sengupta, Karl Pazdernik Published: 2026-01-13 Categories: cs.CL, cs.LG, stat.ML arXiv: https://arxiv.org/abs/2601.16999v1 Named Entity Recognition (NER) serves as a foundational component in many natural language processing (NLP) pipelines. However, current NER models typically output a single predicted label sequence without any accompanying measure of uncertainty, leaving downstream applications vulnerable to cascading errors. In this paper, we introduce a general framework for adapting sequence-labeling-based NER models to produce uncertainty-aware prediction sets. These prediction sets are collections of full-sentence labelings that are guaranteed to contain the correct labeling with a user-specified confidence level. This approach serves a role analogous to confidence intervals in classical statistics by providing formal guarantees about the reliability of model predictions. Our method builds on conformal prediction, which offers finite-sample coverage guarantees under minimal assumptions. We design efficient nonconformity scoring functions to construct efficient, well-calibrated prediction sets that support both unconditional and class-conditional coverage. This framework accounts for heterogeneity across sentence length, language, entity type, and number of entities within a sentence. Empirical experiments on four NER models across three benchmark datasets demonstrate the broad applicability, validity, and efficiency of the proposed methods. --- ### When Generative AI Meets Extended Reality: Enabling Scalable and Natural Interactions Authors: Mingyu Zhu, Jiangong Chen, Bin Li Published: 2026-01-13 Categories: cs.HC, cs.AI arXiv: https://arxiv.org/abs/2601.15308v1 Extended Reality (XR), including virtual, augmented, and mixed reality, provides immersive and interactive experiences across diverse applications, from VR-based education to AR-based assistance and MR-based training. However, widespread XR adoption remains limited due to two key challenges: 1) the high cost and complexity of authoring 3D content, especially for large-scale environments or complex interactions; and 2) the steep learning curve associated with non-intuitive interaction methods like handheld controllers or scripted gestures. Generative AI (GenAI) presents a promising solution by enabling intuitive, language-driven interaction and automating content generation. Leveraging vision-language models and diffusion-based generation, GenAI can interpret ambiguous instructions, understand physical scenes, and generate or manipulate 3D content, significantly lowering barriers to XR adoption. This paper explores the integration of XR and GenAI through three concrete use cases, showing how they address key obstacles in scalability and natural interaction, and identifying technical challenges that must be resolved to enable broader adoption. --- ### DeepSurvey-Bench: Evaluating Academic Value of Automatically Generated Scientific Surveys Authors: Guo-Biao Zhang, Xian-Ling Mao, Ding-Yuan Liu, Da-Yi Wu, Tian Lan, Huihui Li, Heyan Huang Published: 2026-01-13 Categories: cs.AI, cs.CL arXiv: https://arxiv.org/abs/2601.15307v2 The rapid development of automated survey generation technology has made it increasingly important to establish a comprehensive benchmark to evaluate the quality of generated surveys. Most existing benchmarks first construct ground-truth datasets by selecting human-written surveys based on limited selection criteria, such as citation counts and structural coherence, and evaluate generated surveys primarily based on conventional quality dimensions, including structural quality and reference relevance. However, these benchmarks have two key issues: (1) the datasets are insufficiently reliable because the selection criteria only identify highly cited or structurally coherent surveys without verifying their academic value; (2) the evaluation metrics mainly reflect the surface-level quality of generated surveys and are insufficient to assess their academic value. Together, these issues prevent existing benchmarks from effectively assessing the academic value of generated surveys. To address the above problems, we propose DeepSurvey-Bench, a comprehensive benchmark for evaluating the academic value of automatically generated surveys. Specifically, our proposed benchmark introduces a set of academic value evaluation criteria covering three dimensions: informational value, scholarly communication value, and research guidance value. We first construct a reliable dataset with academic value annotations based on these criteria, and then evaluate the academic value of generated surveys according to these criteria through a multi-LLM-as-a-judge approach. Extensive experiments demonstrate that DeepSurvey-Bench not only aligns closely with human assessments in evaluating the academic value of surveys, but also reveals underlying academic value beyond the reach of surface-level quality metrics, providing a foundation for fine-grained diagnosis and iterative improvement of generated surveys. --- ### Uncovering Latent Bias in LLM-Based Emergency Department Triage Through Proxy Variables Authors: Ethan Zhang Published: 2026-01-13 Categories: cs.AI arXiv: https://arxiv.org/abs/2601.15306v1 Recent advances in large language models (LLMs) have enabled their integration into clinical decision-making; however, hidden biases against patients across racial, social, economic, and clinical backgrounds persist. In this study, we investigate bias in LLM-based medical AI systems applied to emergency department (ED) triage. We employ 32 patient-level proxy variables, each represented by paired positive and negative qualifiers, and evaluate their effects using both public (MIMIC-IV-ED Demo, MIMIC-IV Demo) and restricted-access credentialed (MIMIC-IV-ED and MIMIC-IV) datasets as appropriate~\cite{mimiciv_ed_demo,mimiciv_ed,mimiciv}. Our results reveal discriminatory behavior mediated through proxy variables in ED triage scenarios, as well as a systematic tendency for LLMs to modify perceived patient severity when specific tokens appear in the input context, regardless of whether they are framed positively or negatively. These findings indicate that AI systems is still imperfectly trained on noisy, sometimes non-causal signals that do not reliably reflect true patient acuity. Consequently, more needs to be done to ensure the safe and responsible deployment of AI technologies in clinical settings. --- ### A Comparison of Polynomial-Based Tree Clustering Methods Authors: Pengyu Liu, Mariel Vázquez, Nataša Jonoska Published: 2026-01-13 Categories: cs.LG arXiv: https://arxiv.org/abs/2601.14285v1 Tree structures appear in many fields of the life sciences, including phylogenetics, developmental biology and nucleic acid structures. Trees can be used to represent RNA secondary structures, which directly relate to the function of non-coding RNAs. Recent developments in sequencing technology and artificial intelligence have yielded numerous biological data that can be represented with tree structures. This requires novel methods for tree structure data analytics. Tree polynomials provide a computationally efficient, interpretable and comprehensive way to encode tree structures as matrices, which are compatible with most data analytics tools. Machine learning methods based on the Canberra distance between tree polynomials have been introduced to analyze phylogenies and nucleic acid structures. In this paper, we compare the performance of different distances in tree clustering methods based on a tree distinguishing polynomial. We also implement two basic autoencoder models for clustering trees using the polynomial. We find that the distance based methods with entry-level normalized distances have the highest clustering accuracy among the compared methods. --- ### Beyond Affinity: A Benchmark of 1D, 2D, and 3D Methods Reveals Critical Trade-offs in Structure-Based Drug Design Authors: Kangyu Zheng, Kai Zhang, Jiale Tan, Xuehan Chen, Yingzhou Lu, Zaixi Zhang, Lichao Sun, Marinka Zitnik, Tianfan Fu, Zhiding Liang Published: 2026-01-13 Categories: cs.LG, cs.AI arXiv: https://arxiv.org/abs/2601.14283v1 Currently, the field of structure-based drug design is dominated by three main types of algorithms: search-based algorithms, deep generative models, and reinforcement learning. While existing works have typically focused on comparing models within a single algorithmic category, cross-algorithm comparisons remain scarce. In this paper, to fill the gap, we establish a benchmark to evaluate the performance of fifteen models across these different algorithmic foundations by assessing the pharmaceutical properties of the generated molecules and their docking affinities and poses with specified target proteins. We highlight the unique advantages of each algorithmic approach and offer recommendations for the design of future SBDD models. We emphasize that 1D/2D ligand-centric drug design methods can be used in SBDD by treating the docking function as a black-box oracle, which is typically neglected. Our evaluation reveals distinct patterns across model categories. 3D structure-based models excel in binding affinities but show inconsistencies in chemical validity and pose quality. 1D models demonstrate reliable performance in standard molecular metrics but rarely achieve optimal binding affinities. 2D models offer balanced performance, maintaining high chemical validity while achieving moderate binding scores. Through detailed analysis across multiple protein targets, we identify key improvement areas for each model category, providing insights for researchers to combine strengths of different approaches while addressing their limitations. All the code that are used for benchmarking is available in https://github.com/zkysfls/2025-sbdd-benchmark --- ### Hallucination-Free Automatic Question & Answer Generation for Intuitive Learning Authors: Nicholas X. Wang, Aggelos K. Katsaggelos Published: 2026-01-13 Categories: cs.CL, cs.AI arXiv: https://arxiv.org/abs/2601.14280v1 Hallucinations in large language models (LLMs), defined as fluent yet incorrect or incoherent outputs, pose a significant challenge to the automatic generation of educational multiple-choice questions (MCQs). We identified four key hallucination types in MCQ generation: reasoning inconsistencies, insolvability, factual errors, and mathematical errors. To address this, we propose a hallucination-free multi-agent generation framework that breaks down MCQ generation into discrete, verifiable stages. Our framework utilizes both rule-based and LLM-based detection agents, as well as hallucination scoring metrics to optimize question quality. We redefined MCQ generation as an optimization task minimizing hallucination risk while maximizing validity, answerability, and cost-efficiency. We also introduce an agent-led refinement process that uses counterfactual reasoning and chain-of-thought (CoT) to iteratively improve hallucination in question generation. We evaluated a sample of AP- aligned STEM questions, where our system reduced hallucination rates by over 90% compared to baseline generation while preserving the educational value and style of questions. Our results demonstrate that structured multi-agent collaboration can mitigate hallucinations in educational content creation at scale, paving the way for more reliable LLM-powered learning tools. --- ### On the Limits of Learned Importance Scoring for KV Cache Compression Authors: Brady Steele Published: 2026-01-13 Categories: cs.LG, cs.AI arXiv: https://arxiv.org/abs/2601.14279v1 We investigate learned KV cache compression through Speculative Importance Prediction (SIP), a 1.7M parameter non-query-aware scorer that predicts token importance from KV representations alone. Despite architectural sophistication (multi-horizon lookahead, cross-attention), SIP does not outperform simple baselines, including random selection, across 5 seeds, 4 retention levels, and 3 tasks. Key findings: (1) position-based heuristics (keep first 4 + last N tokens) match or exceed learned approaches; (2) prefill attention provides equivalent signal to complex learned scorers; (3) marginal information in KV representations beyond position and prefill attention appears limited for importance prediction. We hypothesize that circular dependence between future queries and generation trajectories contributes to this difficulty. --- ### Handcrafted Feature-Assisted One-Class Learning for Artist Authentication in Historical Drawings Authors: Hassan Ugail, Jan Ritch-Frel, Irina Matuzava Published: 2026-01-13 Categories: cs.CV arXiv: https://arxiv.org/abs/2601.11627v1 Authentication and attribution of works on paper remain persistent challenges in cultural heritage, particularly when the available reference corpus is small and stylistic cues are primarily expressed through line and limited tonal variation. We present a verification-based computational framework for historical drawing authentication using one-class autoencoders trained on a compact set of interpretable handcrafted features. Ten artist-specific verifiers are trained using authenticated sketches from the Metropolitan Museum of Art open-access collection, the Ashmolean Collections Catalogue, the Morgan Library and Museum, the Royal Collection Trust (UK), the Victoria and Albert Museum Collections, and an online catalogue of the Casa Buonarroti collection and evaluated under a biometric-style protocol with genuine and impostor trials. Feature vectors comprise Fourier-domain energy, Shannon entropy, global contrast, GLCM-based homogeneity, and a box-counting estimate of fractal complexity. Across 900 verification decisions (90 genuine and 810 impostor trials), the pooled system achieves a True Acceptance Rate of 83.3% with a False Acceptance Rate of 9.5% at the chosen operating point. Performance varies substantially by artist, with near-zero false acceptance for some verifiers and elevated confusability for others. A pairwise attribution of false accepts indicates structured error pathways consistent with stylistic proximity and shared drawing conventions, whilst also motivating tighter control of digitisation artefacts and threshold calibration. The proposed methodology is designed to complement, rather than replace, connoisseurship by providing reproducible, quantitative evidence suitable for data-scarce settings common in historical sketch attribution. --- ### Synthetic Data for Veterinary EHR De-identification: Benefits, Limits, and Safety Trade-offs Under Fixed Compute Authors: David Brundage Published: 2026-01-13 Categories: cs.CR, cs.AI, cs.CL arXiv: https://arxiv.org/abs/2601.09756v1 Veterinary electronic health records (vEHRs) contain privacy-sensitive identifiers that limit secondary use. While PetEVAL provides a benchmark for veterinary de-identification, the domain remains low-resource. This study evaluates whether large language model (LLM)-generated synthetic narratives improve de-identification safety under distinct training regimes, emphasizing (i) synthetic augmentation and (ii) fixed-budget substitution. We conducted a controlled simulation using a PetEVAL-derived corpus (3,750 holdout/1,249 train). We generated 10,382 synthetic notes using a privacy-preserving "template-only" regime where identifiers were removed prior to LLM prompting. Three transformer backbones (PetBERT, VetBERT, Bio_ClinicalBERT) were trained under varying mixtures. Evaluation prioritized document-level leakage rate (the fraction of documents with at least one missed identifier) as the primary safety outcome. Results show that under fixed-sample substitution, replacing real notes with synthetic ones monotonically increased leakage, indicating synthetic data cannot safely replace real supervision. Under compute-matched training, moderate synthetic mixing matched real-only performance, but high synthetic dominance degraded utility. Conversely, epoch-scaled augmentation improved performance: PetBERT span-overlap F1 increased from 0.831 to 0.850 +/- 0.014, and leakage decreased from 6.32% to 4.02% +/- 0.19%. However, these gains largely reflect increased training exposure rather than intrinsic synthetic data quality. Corpus diagnostics revealed systematic synthetic-real mismatches in note length and label distribution that align with persistent leakage. We conclude that synthetic augmentation is effective for expanding exposure but is complementary, not substitutive, for safety-critical veterinary de-identification. --- ### Heterogeneous computing platform for real-time robotics Authors: Jakub Fil, Yulia Sandamirskaya, Hector Gonzalez, Loïc Azzalin, Stefan Glüge, Lukas Friedenstab, Friedrich Wolf, Tim Rosmeisl, Matthias Lohrmann, Mahmoud Akl, Khaleel Khan, Leonie Wolf et al. Published: 2026-01-13 Categories: cs.NE, cs.AI, cs.RO arXiv: https://arxiv.org/abs/2601.09755v1 After Industry 4.0 has embraced tight integration between machinery (OT), software (IT), and the Internet, creating a web of sensors, data, and algorithms in service of efficient and reliable production, a new concept of Society 5.0 is emerging, in which infrastructure of a city will be instrumented to increase reliability, efficiency, and safety. Robotics will play a pivotal role in enabling this vision that is pioneered by the NEOM initiative - a smart city, co-inhabited by humans and robots. In this paper we explore the computing platform that will be required to enable this vision. We show how we can combine neuromorphic computing hardware, exemplified by the Loihi2 processor used in conjunction with event-based cameras, for sensing and real-time perception and interaction with a local AI compute cluster (GPUs) for high-level language processing, cognition, and task planning. We demonstrate the use of this hybrid computing architecture in an interactive task, in which a humanoid robot plays a musical instrument with a human. Central to our design is the efficient and seamless integration of disparate components, ensuring that the synergy between software and hardware maximizes overall performance and responsiveness. Our proposed system architecture underscores the potential of heterogeneous computing architectures in advancing robotic autonomy and interactive intelligence, pointing toward a future where such integrated systems become the norm in complex, real-time applications. --- ### Critically Engaged Pragmatism: Scientific Norm and Social, Pragmatist Epistemology for AI Science Evaluation Tools Authors: Carole J. Lee Published: 2026-01-13 Categories: cs.CY, cs.AI arXiv: https://arxiv.org/abs/2601.09753v2 AI science evaluation tools aim to assess research credibility. As with traditional metrics such as impact factors, their edicts can be decontextualised and repurposed in problematic ways. To address this, I propose Critically-Engaged Pragmatism as a scientific norm enjoining scientific communities to scrutinise the purposes and purpose-specific reliability of AI science evaluation tools. To foster Critically Engaged Pragmatism, creators of AI science evaluation tools should transparently and fully report design, training, and benchmarking details to facilitate assessments of purpose-specific reliability, liability to different types of error, and bias. What count as best practices for the transparent reporting of AI science evaluation tools should be updated as new forms of error, bias, and gamesmanship are discovered. Under this framework, AI science evaluation tools are not objective arbiters of scientific credibility. Rather, they are the object of critical discursive practices that ultimately ground the credibility of scientific communities. --- ### The Hierarchy of Agentic Capabilities: Evaluating Frontier Models on Realistic RL Environments Authors: Logan Ritchie, Sushant Mehta, Nick Heiner, Mason Yu, Edwin Chen Published: 2026-01-13 Categories: cs.AI arXiv: https://arxiv.org/abs/2601.09032v1 The advancement of large language model (LLM) based agents has shifted AI evaluation from single-turn response assessment to multi-step task completion in interactive environments. We present an empirical study evaluating frontier AI models on 150 workplace tasks within a realistic e-commerce RL environment from Surge. Our analysis reveals an empirically-derived \emph{hierarchy of agentic capabilities} that models must master for real-world deployment: (1) tool use, (2) planning and goal formation, (3) adaptability, (4) groundedness, and (5) common-sense reasoning. Even the best-performing models fail approximately 40\% of the tasks, with failures clustering predictably along this hierarchy. Weaker models struggle with fundamental tool use and planning, whereas stronger models primarily fail on tasks requiring contextual inference beyond explicit instructions. We introduce a task-centric design methodology for RL environments that emphasizes diversity and domain expert contributions, provide detailed failure analysis, and discuss implications for agent development. Our findings suggest that while current frontier models can demonstrate coherent multi-step behavior, substantial capability gaps remain before achieving human-level task completion in realistic workplace settings. --- ### Generalizable Geometric Prior and Recurrent Spiking Feature Learning for Humanoid Robot Manipulation Authors: Xuetao Li, Wenke Huang, Mang Ye, Jifeng Xuan, Bo Du, Sheng Liu, Miao Li Published: 2026-01-13 Categories: cs.RO, cs.AI arXiv: https://arxiv.org/abs/2601.09031v1 Humanoid robot manipulation is a crucial research area for executing diverse human-level tasks, involving high-level semantic reasoning and low-level action generation. However, precise scene understanding and sample-efficient learning from human demonstrations remain critical challenges, severely hindering the applicability and generalizability of existing frameworks. This paper presents a novel RGMP-S, Recurrent Geometric-prior Multimodal Policy with Spiking features, facilitating both high-level skill reasoning and data-efficient motion synthesis. To ground high-level reasoning in physical reality, we leverage lightweight 2D geometric inductive biases to enable precise 3D scene understanding within the vision-language model. Specifically, we construct a Long-horizon Geometric Prior Skill Selector that effectively aligns the semantic instructions with spatial constraints, ultimately achieving robust generalization in unseen environments. For the data efficiency issue in robotic action generation, we introduce a Recursive Adaptive Spiking Network. We parameterize robot-object interactions via recursive spiking for spatiotemporal consistency, fully distilling long-horizon dynamic features while mitigating the overfitting issue in sparse demonstration scenarios. Extensive experiments are conducted across the Maniskill simulation benchmark and three heterogeneous real-world robotic systems, encompassing a custom-developed humanoid, a desktop manipulator, and a commercial robotic platform. Empirical results substantiate the superiority of our method over state-of-the-art baselines and validate the efficacy of the proposed modules in diverse generalization scenarios. To facilitate reproducibility, the source code and video demonstrations are publicly available at https://github.com/xtli12/RGMP-S.git. --- ### Proactively Detecting Threats: A Novel Approach Using LLMs Authors: Aniesh Chawla, Udbhav Prasad Published: 2026-01-13 Categories: cs.CR, cs.AI arXiv: https://arxiv.org/abs/2601.09029v1 Enterprise security faces escalating threats from sophisticated malware, compounded by expanding digital operations. This paper presents the first systematic evaluation of large language models (LLMs) to proactively identify indicators of compromise (IOCs) from unstructured web-based threat intelligence sources, distinguishing it from reactive malware detection approaches. We developed an automated system that pulls IOCs from 15 web-based threat report sources to evaluate six LLM models (Gemini, Qwen, and Llama variants). Our evaluation of 479 webpages containing 2,658 IOCs (711 IPv4 addresses, 502 IPv6 addresses, 1,445 domains) reveals significant performance variations. Gemini 1.5 Pro achieved 0.958 precision and 0.788 specificity for malicious IOC identification, while demonstrating perfect recall (1.0) for actual threats. --- ### OpenDecoder: Open Large Language Model Decoding to Incorporate Document Quality in RAG Authors: Fengran Mo, Zhan Su, Yuchen Hui, Jinghan Zhang, Jia Ao Sun, Zheyuan Liu, Chao Zhang, Tetsuya Sakai, Jian-Yun Nie Published: 2026-01-13 Categories: cs.CL, cs.AI, cs.IR arXiv: https://arxiv.org/abs/2601.09028v2 The development of large language models (LLMs) has achieved superior performance in a range of downstream tasks, including LLM-based retrieval-augmented generation (RAG). The quality of generated content heavily relies on the usefulness of the retrieved information and the capacity of LLMs' internal information processing mechanism to incorporate it in answer generation. It is generally assumed that the retrieved information is relevant to the question. However, the retrieved information may have a variable degree of relevance and usefulness, depending on the question and the document collection. It is important to take into account the relevance of the retrieved information in answer generation. In this paper, we propose OpenDecoder, a new approach that leverages explicit evaluation of the retrieved information as quality indicator features for generation. We aim to build a RAG model that is more robust to varying levels of noisy context. Three types of explicit evaluation information are considered: relevance score, ranking score, and QPP (query performance prediction) score. The experimental results on five benchmark datasets demonstrate the effectiveness and better robustness of OpenDecoder by outperforming various baseline methods. Importantly, this paradigm is flexible to be integrated with the post-training of LLMs for any purposes and incorporated with any type of external indicators. --- ### Layer-Parallel Training for Transformers Authors: Shuai Jiang, Marc Salvadó-Benasco, Eric C. Cyr, Alena Kopaničáková, Rolf Krause, Jacob B. Schroder Published: 2026-01-13 Categories: cs.LG arXiv: https://arxiv.org/abs/2601.09026v2 We present a new training methodology for transformers using a multilevel, layer-parallel approach. Through a neural ODE formulation of transformers, our application of a multilevel parallel-in-time algorithm for the forward and backpropagation phases of training achieves parallel acceleration over the layer dimension. This dramatically enhances parallel scalability as the network depth increases, which is particularly useful for increasingly large foundational models. However, achieving this introduces errors that cause systematic bias in the gradients, which in turn reduces convergence when closer to the minima. We develop an algorithm to detect this critical transition and either switch to serial training or systematically increase the accuracy of layer-parallel training. Results, including BERT, GPT2, ViT, and machine translation architectures, demonstrate parallel-acceleration as well as accuracy commensurate with serial pre-training while fine-tuning is unaffected. --- ### Universal Latent Homeomorphic Manifolds: A Framework for Cross-Domain Representation Unification Authors: Tong Wu, Tayab Uddin Wara, Daniel Hernandez, Sidong Lei Published: 2026-01-13 Categories: eess.IV, cs.LG, eess.SP arXiv: https://arxiv.org/abs/2601.09025v2 We present the Universal Latent Homeomorphic Manifold (ULHM), a framework that unifies semantic representations (e.g., human descriptions, diagnostic labels) and observation-driven machine representations (e.g., pixel intensities, sensor readings) into a single latent structure. Despite originating from fundamentally different pathways, both modalities capture the same underlying reality. We establish \emph{homeomorphism}, a continuous bijection preserving topological structure, as the mathematical criterion for determining when latent manifolds induced by different semantic-observation pairs can be rigorously unified. This criterion provides theoretical guarantees for three critical applications: (1) semantic-guided sparse recovery from incomplete observations, (2) cross-domain transfer learning with verified structural compatibility, and (3) zero-shot compositional learning via valid transfer from semantic to observation space. Our framework learns continuous manifold-to-manifold transformations through conditional variational inference, avoiding brittle point-to-point mappings. We develop practical verification algorithms, including trust, continuity, and Wasserstein distance metrics, that empirically validate homeomorphic structure from finite samples. Experiments demonstrate: (1) sparse image recovery from 5% of CelebA pixels and MNIST digit reconstruction at multiple sparsity levels, (2) cross-domain classifier transfer achieving 86.73% accuracy from MNIST to Fashion-MNIST without retraining, and (3) zero-shot classification on unseen classes achieving 78.76% on CIFAR-10. Critically, the homeomorphism criterion determines when different semantic-observation pairs share compatible latent structure, enabling principled unification into universal representations and providing a mathematical foundation for decomposing general foundation models into domain-specific components. --- ### An Inexact Weighted Proximal Trust-Region Method Authors: Leandro Farias Maia, Robert Baraldi, Drew P. Kouri Published: 2026-01-13 Categories: math.OC, cs.LG arXiv: https://arxiv.org/abs/2601.09024v1 In (R. J. Baraldi and D. P. Kouri, Math. Program., 201:1 (2023), pp. 559-598), the authors introduced a trust-region method for minimizing the sum of a smooth nonconvex and a nonsmooth convex function, the latter of which has an analytical proximity operator. While many functions satisfy this criterion, e.g., the $\ell_1$-norm defined on $\ell_2$, many others are precluded by either the topology or the nature of the nonsmooth term. Using the $δ$-Fréchet subdifferential, we extend the definition of the inexact proximity operator and enable its use within the aforementioned trust-region algorithm. Moreover, we augment the analysis for the standard trust-region convergence theory to handle proximity operator inexactness with weighted inner products. We first introduce an algorithm to generate a point in the inexact proximity operator and then apply the algorithm within the trust-region method to solve an optimal control problem constrained by Burgers' equation. --- ### Tail-Sensitive KL and Rényi Convergence of Unadjusted Hamiltonian Monte Carlo via One-Shot Couplings Authors: Nawaf Bou-Rabee, Siddharth Mitra, Andre Wibisono Published: 2026-01-13 Categories: stat.ML, cs.LG, math.PR, math.ST, stat.CO arXiv: https://arxiv.org/abs/2601.09019v1 Hamiltonian Monte Carlo (HMC) algorithms are among the most widely used sampling methods in high dimensional settings, yet their convergence properties are poorly understood in divergences that quantify relative density mismatch, such as Kullback-Leibler (KL) and Rényi divergences. These divergences naturally govern acceptance probabilities and warm-start requirements for Metropolis-adjusted Markov chains. In this work, we develop a framework for upgrading Wasserstein convergence guarantees for unadjusted Hamiltonian Monte Carlo (uHMC) to guarantees in tail-sensitive KL and Rényi divergences. Our approach is based on one-shot couplings, which we use to establish a regularization property of the uHMC transition kernel. This regularization allows Wasserstein-2 mixing-time and asymptotic bias bounds to be lifted to KL divergence, and analogous Orlicz-Wasserstein bounds to be lifted to Rényi divergence, paralleling earlier work of Bou-Rabee and Eberle (2023) that upgrade Wasserstein-1 bounds to total variation distance via kernel smoothing. As a consequence, our results provide quantitative control of relative density mismatch, clarify the role of discretization bias in strong divergences, and yield principled guarantees relevant both for unadjusted sampling and for generating warm starts for Metropolis-adjusted Markov chains. --- ### Meta-learning to Address Data Shift in Time Series Classification Authors: Samuel Myren, Nidhi Parikh, Natalie Klein Published: 2026-01-13 Categories: cs.LG, cs.AI arXiv: https://arxiv.org/abs/2601.09018v1 Across engineering and scientific domains, traditional deep learning (TDL) models perform well when training and test data share the same distribution. However, the dynamic nature of real-world data, broadly termed \textit{data shift}, renders TDL models prone to rapid performance degradation, requiring costly relabeling and inefficient retraining. Meta-learning, which enables models to adapt quickly to new data with few examples, offers a promising alternative for mitigating these challenges. Here, we systematically compare TDL with fine-tuning and optimization-based meta-learning algorithms to assess their ability to address data shift in time-series classification. We introduce a controlled, task-oriented seismic benchmark (SeisTask) and show that meta-learning typically achieves faster and more stable adaptation with reduced overfitting in data-scarce regimes and smaller model architectures. As data availability and model capacity increase, its advantages diminish, with TDL with fine-tuning performing comparably. Finally, we examine how task diversity influences meta-learning and find that alignment between training and test distributions, rather than diversity alone, drives performance gains. Overall, this work provides a systematic evaluation of when and why meta-learning outperforms TDL under data shift and contributes SeisTask as a benchmark for advancing adaptive learning research in time-series domains. --- ### Multicultural Spyfall: Assessing LLMs through Dynamic Multilingual Social Deduction Game Authors: Haryo Akbarianto Wibowo, Alaa Elsetohy, Qinrong Cui, Alham Fikri Aji Published: 2026-01-13 Categories: cs.CL arXiv: https://arxiv.org/abs/2601.09017v1 The rapid advancement of Large Language Models (LLMs) has necessitated more robust evaluation methods that go beyond static benchmarks, which are increasingly prone to data saturation and leakage. In this paper, we propose a dynamic benchmarking framework for evaluating multilingual and multicultural capabilities through the social deduction game Spyfall. In our setup, models must engage in strategic dialogue to either identify a secret agent or avoid detection, utilizing culturally relevant locations or local foods. Our results show that our game-based rankings align closely with the Chatbot Arena. However, we find a significant performance gap in non-English contexts: models are generally less proficient when handling locally specific entities and often struggle with rule-following or strategic integrity in non-English languages. We demonstrate that this game-based approach provides a scalable, leakage-resistant, and culturally nuanced alternative to traditional NLP benchmarks. The game history can be accessed here https://huggingface.co/datasets/haryoaw/cultural-spyfall. --- ### TranslateGemma Technical Report Authors: Mara Finkelstein, Isaac Caswell, Tobias Domhan, Jan-Thorsten Peter, Juraj Juraska, Parker Riley, Daniel Deutsch, Geza Kovacs, Cole Dilanni, Colin Cherry, Eleftheria Briakou, Elizabeth Nielsen et al. Published: 2026-01-13 Categories: cs.CL, cs.AI arXiv: https://arxiv.org/abs/2601.09012v3 We present TranslateGemma, a suite of open machine translation models based on the Gemma 3 foundation models. To enhance the inherent multilingual capabilities of Gemma 3 for the translation task, we employ a two-stage fine-tuning process. First, supervised fine-tuning is performed using a rich mixture of high-quality large-scale synthetic parallel data generated via state-of-the-art models and human-translated parallel data. This is followed by a reinforcement learning phase, where we optimize translation quality using an ensemble of reward models, including MetricX-QE and AutoMQM, targeting translation quality. We demonstrate the effectiveness of TranslateGemma with human evaluation on the WMT25 test set across 10 language pairs and with automatic evaluation on the WMT24++ benchmark across 55 language pairs. Automatic metrics show consistent and substantial gains over the baseline Gemma 3 models across all sizes. Notably, smaller TranslateGemma models often achieve performance comparable to larger baseline models, offering improved efficiency. We also show that TranslateGemma models retain strong multimodal capabilities, with enhanced performance on the Vistra image translation benchmark. The release of the open TranslateGemma models aims to provide the research community with powerful and adaptable tools for machine translation. --- ### Block Decomposable Methods for Large-Scale Optimization Problems Authors: Leandro Farias Maia Published: 2026-01-13 Categories: math.OC, cs.LG arXiv: https://arxiv.org/abs/2601.09010v1 This dissertation explores block decomposable methods for large-scale optimization problems. It focuses on alternating direction method of multipliers (ADMM) schemes and block coordinate descent (BCD) methods. Specifically, it introduces a new proximal ADMM algorithm and proposes two BCD methods. The first part of the research presents a new proximal ADMM algorithm. This method is adaptive to all problem parameters and solves the proximal augmented Lagrangian (AL) subproblem inexactly. This adaptiveness facilitates the highly efficient application of the algorithm to a broad swath of practical problems. The inexact solution of the proximal AL subproblem overcomes many key challenges in the practical applications of ADMM. The resultant algorithm obtains an approximate solution of an optimization problem in a number of iterations that matches the state-of-the-art complexity for the class of proximal ADMM schemes. The second part of the research focuses on an inexact proximal mapping for the class of block proximal gradient methods. Key properties of this operator is established, facilitating the derivation of convergence rates for the proposed algorithm. Under two error decreases conditions, the algorithm matches the convergence rate of its exactly computed counterpart. Numerical results demonstrate the superior performance of the algorithm under a dynamic error regime over a fixed one. The dissertation concludes by providing convergence guarantees for the randomized BCD method applied to a broad class of functions, known as Hölder smooth functions. Convergence rates are derived for non-convex, convex, and strongly convex functions. These convergence rates match those furnished in the existing literature for the Lipschtiz smooth setting. ---