# Research Papers — 2026-06 · bundle page 44 of 260 > Full-text markdown of papers #2151–2200 of 12993 published in 2026-06, newest first. AI-readable mirrors of the full 12993-paper 2026-06 archive (50 papers per file). Canonical page: https://aipod.dev/papers/archive/2026-06/page/44 [← Previous bundle](https://aipod.dev/papers/archive/2026-06/page/43.md) [Next bundle →](https://aipod.dev/papers/archive/2026-06/page/45.md) ### HarmVideoBench: Benchmarking Harmful Video Understanding in Large Multimodal Models Authors: Jiajun Wu, Haoyu Kang, Yining Sun, Jiacheng Hou, Heng Zhang, Danyang Zhang, Zhenjun Zhao, Haochi Zhang, Leixin Sun, Eric Hanchen Jiang, Yushan Li, Ruiyu Li et al. Published: 2026-06-25 Categories: cs.CV, cs.CL arXiv: https://arxiv.org/abs/2606.27187v1 Large vision-language models (LVLMs) have recently shown immense potential in automated content moderation, sparking growing interest in developing harmful-video benchmarks. However, we identify two primary limitations in existing works: 1) The multi-layered characteristics of harmful videos are overlooked. Existing benchmarks predominantly formulate evaluation as a binary classification task, failing to capture implicit or deep contextual harms. 2) Explanatory rationales are completely absent. Current frameworks measure exclusively whether a model flags a video correctly rather than explaining why, turning evaluation into a black box where models can succeed through superficial shortcuts. To address these problems, we present HarmVideoBench, a multi-layered diagnostic benchmark comprising 1,379 videos paired with 4,137 multiple-choice questions. HarmVideoBench benchmarks three hierarchical dimensions: Observable Evidence, Clip-Internal Meaning, and Beyond-Clip Reasoning, aiming to evaluate models' deep understanding beyond surface cues with carefully balanced and curated samples. We evaluate 19 leading models on HarmVideoBench to assess their multidimensional understanding of harmful videos. Moreover, we introduce BCR, a benchmark-aligned method that predicts reasoning boundaries and dynamically retrieves context only when needed. Experimental results show that BCR substantially improves the base model's performance in harmful video understanding, raising the macro average from 61.7 percent to a state-of-the-art 84.4 percent. --- ### Automating Potential-based Reward Shaping with Vision Language Model Guidance Authors: Henrik Müller, Daniel Kudenko Published: 2026-06-25 Categories: cs.LG, cs.AI, cs.RO arXiv: https://arxiv.org/abs/2606.27180v1 Sparse rewards are inherently challenging for reinforcement learning agents as they lack intermediate feedback to guide exploration and to correctly attribute the sparse success rewards to relevant parts of the trajectory. Naive reward shaping can induce reward hacking, yielding policies that exploit auxiliary signals instead of solving the intended task. Potential-based reward shaping (PBRS) guarantees preservation of the optimal policy set, but requires the definition of a heuristic potential function over the state space. In this work, we introduce the VLM-guided PBRS framework VLM-PBRS that learns the potential function directly from vision language model (VLM) feedback. We query a lightweight VLM to obtain preferences over image pairs and train a model of the potential function using these preferences. As this approach is based on potential-based reward shaping, it preserves the original optimal policies, and removes the need for expert-designed reward shaping terms. Because large VLMs are prohibitively expensive to invoke repeatedly during policy learning, we employ smaller, more computationally efficient VLMs. Although the resulting preference labels are less accurate, empirical evidence shows that the preference labels can still be used to accelerate learning. We validate our method empirically in the Meta-World and Franka Kitchen environments and highlight the connection between VLM preference label accuracy and sample efficiency improvements. Our contributions are threefold: (1) the first application of VLM preference-based learning to synthesize a potential function for PBRS, (2) a principled, low-cost solution that leverages small VLMs, and (3) extensive empirical demonstration of improved sample efficiency and robustness to reward hacking. --- ### RecallRisk-BERT: A Multi-Task Framework for Post-Report Medical Device Recall Triage Authors: Ali Semih Atalay, Sevgi Yigit-Sert Published: 2026-06-25 Categories: cs.LG arXiv: https://arxiv.org/abs/2606.27174v1 Medical device recalls are a critical regulatory mechanism for protecting patient safety. The growing volume of FDA recall records presents challenges in post-report recall triage, severity assessment, and root-cause interpretation. Existing studies mostly address recall occurrence prediction or root-cause analysis separately, while joint modeling of recall severity and root-cause categories has received limited attention. We develop an automated recall triage framework using 54,165 FDA medical device recall records from openFDA, covering the period from 2002 to October 2025. We first evaluate classical machine learning and boosting-based models for recall severity and root-cause category prediction. We then develop RecallRisk-BERT, a multi-task model that combines PubMedBERT-based textual representations of recall narratives with embedding-based representations of structured categorical features, including product code, regulation number, and medical specialty. The model simultaneously predicts recall severity (Class I/II/III) and a consolidated root-cause category (9 classes). Performance was evaluated using accuracy, macro-averaged precision, recall, F1-score, and ROC-AUC. In single-task severity prediction, our LightGBM-based text--tabular configuration achieved the strongest performance, with an accuracy of 0.963, macro-F1 of 0.856, and ROC-AUC of 0.974. In the multi-task setting, RecallRisk-BERT substantially outperformed the single-task PubMedBERT baseline. Model-derived risk rankings were strongly consistent with observed root-cause severity patterns (rho = 0.983, p = 1.936e-6). These findings indicate that text--tabular learning can support scalable post-report recall triage, regulatory decision support, and model-based root-cause risk analysis. --- ### Stochastic Gradient Optimization with Model-Assisted Sampling Authors: Jonne Pohjankukka, Jukka Heikkonen Published: 2026-06-25 Categories: cs.LG, stat.ME arXiv: https://arxiv.org/abs/2606.27171v1 This work addresses the problem of variance in stochastic gradient estimation for machine learning optimization. Deep learning relies on mini-batch methods such as stochastic gradient descent, which approximate full gradients but introduce noise, creating trade-offs between convergence stability, speed, and generalization. Existing methods, including variance reduction techniques (e.g., SVRG and SAG) and adaptive optimizers, aim to mitigate gradient noise but may introduce additional computational overhead. We propose a model-assisted sampling framework that interprets mini-batch gradients through survey sampling theory, treating the dataset as a fixed finite population and gradients as sample-based estimates. Our aim is to bridge machine learning optimization and survey sampling theory by combining their perspectives on sample-based estimation and variance reduction. By incorporating auxiliary gradient-prediction models, we construct more efficient gradient estimators, with uniform sampling arising as a special case when no auxiliary information is used. Our approach integrates easily with existing optimizers, improving efficiency without altering their dynamics. Empirical results on synthetic and six benchmark datasets show performance gains in 71-86% of the experiments, particularly for medium-sized input spaces in our benchmarks. Notably, with momentum-based optimizers such as AdamW, the proposed estimator achieves clearly better generalization in roughly half the training epochs compared to baseline estimator. --- ### Learning to Fold: prizewinning solution at LeHome Challenge 2026 (1st place online, 2nd offline) Authors: Ilia Larchenko Published: 2026-06-25 Categories: cs.RO, cs.AI, cs.LG arXiv: https://arxiv.org/abs/2606.27163v3 I describe my solution to the LeHome Challenge 2026, an ICRA 2026 competition on bimanual garment folding. The system placed 1st of 62 teams in the online (simulation) round and 2nd in the real-world final. It improves a vision-language-action (VLA) policy with a reinforcement-learning loop. The policy is its own value function: the same network that predicts actions also predicts success, progress, and a few task-relevant future quantities, and those predictions drive advantage estimation, live failure detection, and candidate selection. The work mostly recombines existing RL ideas with engineering and optimization contributions that can be used together as one recipe or individually: AWR + RECAP combined for flow-matching VLA; an asynchronous distributed training / rollout pipeline through HuggingFace Hub; inference-time hyperparameters optimization via Thompson sampling; a sim-to-real recipe with camera-alignment tooling, heavy augmentation and DAgger-like HIL data collection. --- ### Learning to Fold: prizewinning solution at LeHome Challenge 2026 (1st place online, 2nd offline) Authors: Ilia Larchenko Published: 2026-06-25 Categories: cs.RO, cs.AI, cs.LG arXiv: https://arxiv.org/abs/2606.27163v2 I describe my solution to the LeHome Challenge 2026, an ICRA 2026 competition on bimanual garment folding. The system placed 1st of 62 teams in the online (simulation) round and 2nd in the real-world final. It improves a vision-language-action (VLA) policy with a reinforcement-learning loop. The policy is its own value function: the same network that predicts actions also predicts success, progress, and a few task-relevant future quantities, and those predictions drive advantage estimation, live failure detection, and candidate selection. The work mostly recombines existing RL ideas with engineering and optimization contributions that can be used together as one recipe or individually: AWR + RECAP combined for flow-matching VLA; an asynchronous distributed training / rollout pipeline through HuggingFace Hub; inference-time hyperparameters optimization via Thompson sampling; a sim-to-real recipe with camera-alignment tooling, heavy augmentation and DAgger-like HIL data collection. --- ### TOPS: First-Principles Visual Token Pruning via Constructing Token Optimal Preservation Sets for Efficient MLLM Inference Authors: Tinghao Wang, Yichen Guo, Rui Huang, Junchi Yao, Qizhe Zhang, Zheng Lu, Yuan Zhang, Jiajun Cao, Zhirong Shen, Yaosong Du, Lijie Hu, Shanghang Zhang Published: 2026-06-25 Categories: cs.AI arXiv: https://arxiv.org/abs/2606.27161v2 Multimodal large language models (MLLMs) have achieved strong multimodal reasoning capabilities, but their efficiency is limited by the large number of visual tokens, which introduces substantial computational overhead. Visual token pruning offers a natural solution, yet existing methods are imperfect: attention-based criteria tend to retain redundant tokens, while diversity-based criteria are often agnostic to user instructions. Even methods that combine multiple criteria still lack a principled formulation of the intrinsic objective of token pruning. In this paper, we revisit visual token pruning from a first-principles perspective and formulate it as constructing Token Optimal Preservation Sets. Through a top-down information-theoretic analysis, we identify three fundamental principles for effective token selection: Task Relevance, Information Coverage, and Semantic Diversity. Based on these principles, we propose TOPS, a training-free and model-agnostic pruning module that can be applied to various MLLMs. Extensive experiments on 7 MLLM backbones and 14 benchmarks demonstrate that TOPS outperforms prior methods under diverse pruning settings. Notably, on LLaVA-NeXT, TOPS removes 77.8% of visual tokens while preserving 100.0% and 100.6% performance on its 7B and 13B models, respectively, suggesting that pruning redundant visual tokens can sometimes mitigate hallucination and inspire future lightweight MLLM design. --- ### TOPS: First-Principles Visual Token Pruning via Constructing Token Optimal Preservation Sets for Efficient MLLM Inference Authors: Tinghao Wang, Yichen Guo, Rui Huang, Zheng Lu, Qizhe Zhang, Chenxi Li, Yuan Zhang, Jiajun Cao, Zhirong Shen, Yaosong Du, Guangyan Gan, Wenya Wang et al. Published: 2026-06-25 Categories: cs.AI arXiv: https://arxiv.org/abs/2606.27161v1 Multimodal large language models (MLLMs) have achieved strong multimodal reasoning capabilities, but their efficiency is limited by the large number of visual tokens, which introduces substantial computational overhead. Visual token pruning offers a natural solution, yet existing methods are imperfect: attention-based criteria tend to retain redundant tokens, while diversity-based criteria are often agnostic to user instructions. Even methods that combine multiple criteria still lack a principled formulation of the intrinsic objective of token pruning. In this paper, we revisit visual token pruning from a first-principles perspective and formulate it as constructing Token Optimal Preservation Sets. Through a top-down information-theoretic analysis, we identify three fundamental principles for effective token selection: Task Relevance, Information Coverage, and Semantic Diversity. Based on these principles, we propose TOPS, a training-free and model-agnostic pruning module that can be applied to various MLLMs. Extensive experiments on 7 MLLM backbones and 14 benchmarks demonstrate that TOPS outperforms prior methods under diverse pruning settings. Notably, on LLaVA-NeXT, TOPS removes 77.8% of visual tokens while preserving 100.0% and 100.6% performance on its 7B and 13B models, respectively, suggesting that pruning redundant visual tokens can sometimes mitigate hallucination and inspire future lightweight MLLM design. --- ### OpenRCA 2.0: From Outcome Labels to Causal Process Supervision Authors: Aoyang Fang, Yifan Yang, Jin'ao Shang, Qisheng Lu, Junjielung Xu, Rui Wang, Songhan Zhang, Yuzhong Zhang, Boxi Yu, Pinjia He Published: 2026-06-25 Categories: cs.AI arXiv: https://arxiv.org/abs/2606.27154v2 Root cause analysis (RCA) poses a holistic test of LLM agentic capabilities, such as long-context understanding, multi-step reasoning, and tool use. However, existing datasets suffer from a fundamental gap: they label only the root cause, not the propagation path connecting it to the observed symptom, which largely simplifies the task to naive pattern matching. To support rigorous evaluation, we introduce PAVE, a step-wise labeling protocol that leverages known interventions from fault injection to reconstruct causal propagation paths. The mechanism is forward verification: reasoning from cause to effect rather than inferring backward from symptoms. Applying PAVE yields OpenRCA 2.0 (500 instances), the first cross-system RCA benchmark with step-wise causal annotations for LLM agents. Across 11 frontier LLMs, recovering the exact root-cause set succeeds in only 20.7% of cases on average. To locate where this difficulty lies, we relax the criterion and find what we call the ungrounded diagnosis: agents identify at least one correct root-cause service in 76.0% of cases, but ground that service in a verified causal propagation path to the observed symptom in only 61.5%. Outcome-only evaluation hides this failure mode; step-wise causal ground truth is the missing piece for trustworthy LLM-based RCA agents. --- ### DMuon: Efficient Distributed Muon Training with Near-Adam Overhead Authors: Vincent Chen, Starrick Liu, Regis Cheng, Dance Yang, Shalfun Li, Ryan Yu, Lucy Liang, Hang Su, Roy Gan, Hao Wang, Qian Wang Published: 2026-06-25 Categories: cs.DC, cs.LG arXiv: https://arxiv.org/abs/2606.27153v1 Matrix-orthogonalization-based optimizers, exemplified by Muon, have demonstrated strong convergence behavior across a wide range of modern deep learning workloads. The matrix-aware updates offer a compelling alternative to conventional element-wise optimization, particularly as model architectures continue to grow in scale and heterogeneity. Yet contemporary distributed training infrastructure built around the assumption of element-wise optimizers is poorly matched to matrix-level optimizers such as Muon, whose updates couple entire weight matrices and require costly Newton-Schulz iterations. Vanilla Muon implementations incur more than 2x the cost of forward and backward passes. To close this gap, we present DMuon, an open-source distributed Muon implementation that integrates into existing training pipelines as a drop-in module, with no framework-level modifications. Across both embodied foundation model and large language model (LLM) training workloads, DMuon achieves a 1.48x-3.01x speedup in end-to-end step time and a 6.85x-163.00x speedup in optimizer-step time, bringing per-step latency to near-AdamW levels and enabling efficient scaling in our model training. --- ### Safe Autoregressive Image Generation with Iterative Self-Improving Codebooks Authors: Yunqi Xue, Zhijiang Li, Philip Torr, Jindong Gu Published: 2026-06-25 Categories: cs.CV, cs.AI arXiv: https://arxiv.org/abs/2606.27147v1 Unlike diffusion-based models that operate in continuous latent spaces, autoregressive unified multimodal models produce images by sequentially predicting discretized visual tokens. These tokens are derived from a codebook that maps embeddings to quantized visual patterns. The language-like architecture enables unified multimodal models to effectively capture text conditional information for generation, making them promising for text-to-image tasks. This also raises an interesting question: how safe are the images generated in such an autoregressive way? In this work, we propose iterative self-improving codebooks for safe autoregressive generation. We leverage the understanding and judgment capabilities of the unified multimodal model itself to identify unsafe generated images without human annotation. Subsequently, the inherent representations in the codebook are fixed to eliminate harmful mappings. Our method comprises two steps: first, we use the unified model to identify unsafe generations and construct corresponding harmful and safe image-text pairs. These pairs are used to construct the Harmful Space and guide updates to the codebook, thereby eliminating harmful outputs. Second, we perform adaptive fine-tuning on the codebook within the harmless space using safe image-text pairs to ensure the quality of generated images. These two steps are repeated until no further improvement is observed, producing a safety-enhanced model codebook. Without additional external feedback, the safety of models is improved iteratively. --- ### fTNN: a tensor neural network for fractional PDEs Authors: Qingkui Ma, Hehu Xie, Xiaobo Yin Published: 2026-06-25 Categories: cs.LG arXiv: https://arxiv.org/abs/2606.27140v1 We develop the fTNN, a deterministic tensor neural network subspace method for problems involving the fractional Laplacian on bounded domains, taking the fractional Poisson equation and time-dependent fractional advection-diffusion equation as typical representatives. The work employs a geometry-adapted integration split featuring a spatially dependent near-field radius, which decomposes the fractional Laplacian into three contributions: a singular near field, a regular interior far field, and an analytical exterior far field. Then the singular radial integrals are treated by Gauss-Jacobi quadrature, the regular radial integrals by Gauss quadrature, and the angular variables by deterministic angular quadrature, yielding a fully deterministic integration framework of the fractional Laplacian operator. To accurately resolve low-regularity solutions and the associated loss functional, we construct boundary-singularity-aware trial functions enriched with explicit boundary features, and propose two strategies for automatically selecting the leading exponent and evaluating the loss function from the singularity structure induced by the fractional operator, or jointly by the fractional operator and the source term. For time-dependent fractional PDEs, we design a spatiotemporally separable neural network that factorizes the time-space residual into a sum of low-dimensional temporal and spatial integrals, and we integrate this representation with an alternating neural network subspace optimization strategy for efficient training. Numerical experiments show that the proposed framework attains high accuracy on the tested benchmarks and improves substantially over existing fPINN and Monte Carlo baselines, particularly for problems with strong boundary singularities and long-time simulations. --- ### Joint Learning of Experiential Rules and Policies for Large Language Model Agents Authors: Shicheng Ye, Chao Yu Published: 2026-06-25 Categories: cs.AI arXiv: https://arxiv.org/abs/2606.27136v1 For LLM agents in multi-step interactive environments, a key challenge is to make effective use of accumulated interaction experience. Existing work has typically separated two uses of such experience: keeping it outside the model as natural-language rules for later prompting, or using trajectories and feedback to update the model parameters. The former is easy to interpret but can fall out of sync with the evolving policy; the latter improves the policy more broadly but provides only limited correction for local mistakes in sparse-reward settings. We present Joint Learning of Experiential Rules and Policies for LLM Agents (JERP), which updates a long-term experiential-rule pool and the policy from the same interaction trajectories. At decision time, JERP retrieves task-relevant rules and conditions the agent on them together with the interaction history. After each episode, it uses the collected trajectories both to optimize the policy and to revise the rule pool by comparing current rollouts with reference successful trajectories. This coupling keeps the rule pool aligned with the evolving policy while allowing stable and effective behaviors to be gradually absorbed into the model itself. Experiments on AlfWorld and WebShop show that JERP yields consistent gains in decision performance for complex interactive tasks. --- ### FlameVQA: A Physically-Grounded UAV Wildfire VQA Benchmark with Radiometric Thermal Supervision Authors: Mobin Habibpour, John Spodnik, Niloufar Alipour Talemi, Fatemeh Afghah Published: 2026-06-25 Categories: cs.CV, cs.RO arXiv: https://arxiv.org/abs/2606.27128v1 Wildfire monitoring from UAVs requires reliable reasoning over complex aerial scenes, where smoke, scale variation, and occlusions often limit RGB-only interpretation. We introduce FlameVQA, a multiple-choice visual question answering benchmark for UAV-based wildfire intelligence built on FLAME 3, leveraging paired RGB imagery and radiometric thermal TIFFs for temperature-grounded, safety-critical reasoning. FlameVQA includes 34 multiple-choice questions per image spanning six operational capability groups, covering tasks such as detection, localization, distribution/coverage estimation, cross-modal reasoning, and flight planning. To ensure label reliability, we combine MLLM-assisted annotation with deterministic thermal rules and cross-question consistency checks, followed by human auditing. We also evaluate representative MLLMs on FlameVQA to provide baselines for future work. Results show strong performance when explicit cross-modal cues are available, but notable failures on presence detection under heavy smoke and on coverage estimation. These findings suggest that current MLLMs require domain-specific adaptation to better support disaster and wildfire monitoring. The dataset and benchmark code are open-source at github.com/mobiiin/WildFire_VQA --- ### Kolmogorov Arnold networks (KAN) for aerodynamic prediction: a comparison with MLPs and GNNs Authors: Miguel Jaraiz, Fermin Gutierrez, Pablo Yeste, Miguel Sánchez-Domínguez, Eusebio Valero, Gonzalo Rubio, Lucas Lacasa Published: 2026-06-25 Categories: cs.LG, physics.data-an, physics.flu-dyn arXiv: https://arxiv.org/abs/2606.27126v1 Kolmogorov Arnold networks (KAN) have recently been introduced as a (deep) neural network architecture whose trainable parameters adapt the activation functions, instead of the coefficients of the affine transformations at the core of traditional architectures such as deep multilayer perceptrons (MLPs). This architecture builds on the Kolmogorov-Arnold theorem, which endows it with universal approximation properties. While the advent of KANs has been received with excitement, there is a current debate about the possible KAN supremacy over deep multilayer perceptrons (MLPs) for classic fields such as symbolic regression, generic-purpose machine learning, natural language processing or computer vision. Here we assess the performance of KANs --and its nuanced comparison against MLPs and graph neural networks (GNNs)-- in the realm of fluid dynamics surrogate modelling. To that aim, we consider the task of predicting the surface pressure distribution over subsonic and transonic airfoils, a canonical task in aerodynamics. Our results show that KAN models show good performance in predicting the whole pressure coefficients and is able to interpolate across Mach numbers and angles of attack, however its performance is comparable --marginally inferior-- to a suitably trained MLP, where best performance is achieved by a GNN at the expense or requiring lengthier training. While the optimal KAN model have typically much lower complexity than MLP and GNN --hence resulting in faster training--, we find that KANs suffer from training instabilities, and their performance is highly dependent on a proper hyperparameter optimisation. --- ### Proposal-Conditioned Latent Diffusion for Closed-Loop Traffic Scenario Generation Authors: Shubham Vaijanath Phoolari, Aleyna Kara, Christoph Lauer, Steven Peters Published: 2026-06-25 Categories: cs.RO, cs.CV arXiv: https://arxiv.org/abs/2606.27123v1 Closed-loop traffic simulation remains challenging because it must generate interactive multi-agent behaviors that are scene-consistent and controllable throughout rollout. Prior diffusion-based approaches achieve strong realism, but their computational cost can hinder deployment in time-constrained replanning loops for autonomous vehicle planning and simulation. We present a diffusion-based scenario generation framework conditioned on instance-centric scene context and multimodal proposal priors, with optional test-time guidance for shaping safety-critical behaviors. A compact action-latent representation and proposal-based initialization improve sampling efficiency and reduce per-step runtime without retraining. Experiments on the Waymo Open Motion Dataset demonstrate a favorable balance among realism, safety, and controllability across diverse interactive scenarios, while showing that test-time guidance enables systematic trade-offs among competing objectives. --- ### Efficient foundation decoders for fault-tolerant quantum computing Authors: Ge Yan, Shanchuan Li, Shiyi Xiao, Pengyue Ma, Hanyan Cao, Feng Pan, Yuxuan Du Published: 2026-06-25 Categories: quant-ph, cs.AI, cs.LG arXiv: https://arxiv.org/abs/2606.27119v1 Foundation decoders, a class of high-capacity neural decoders, are leading candidates for fault-tolerant quantum computing, with accurate and efficient decoding at large code distances. However, their construction often faces a steep scaling barrier, as larger code distances rapidly amplify the cost of syndrome generation and neural optimization. To address this bottleneck, here we devise neural transfer unification (NTU), a unified framework for efficient foundation decoders. A central feature of NTU is its ability to align decoding tasks across code distances via algebraic structures shared by scalable code families, which enables knowledge learned on smaller codes to accelerate large-scale decoder training. We instantiate NTU as NTU-Transformer, a transformer-based neural decoder tailored for planar surface codes and bivariate bicycle codes. For planar surface codes under circuit-level noise, NTU-Transformer outperforms correlation-aware matching on the $(\!(361,1,19)\!)$ code and further scales to the $(\!(625,1,25)\!)$ code, where it exceeds standard matching through transfer adaptation. For the bivariate bicycle code with $(\!(72,12,6)\!)$, it surpasses Relay-BP in the low-physical-error regime. These results establish our proposal as a scalable route to amortized cross-distance training of foundation decoders for fault-tolerant quantum processors. --- ### Cross-Head Attention Uplift Network with Inverse Propensity Score under Unobserved Confounding Authors: Haoran Zhang, Chuanpu Li, Yuxin Fu, Bin Tong, Guan Wang, Bo Zheng, Feng Zhou Published: 2026-06-25 Categories: cs.LG arXiv: https://arxiv.org/abs/2606.27114v1 Uplift modeling, crucial for estimating individual treatment effects (ITE), faces dual challenges: flexibly leveraging inter-group similarity to enhance discriminative power and debiasing under unobserved confounding scenarios. In this paper, we propose the Cross-Head Attention Uplift Network (CHAUN) and Robust Adversarial Inverse Propensity Score (RA-IPS) method to address these limitations. CHAUN employs shared feature embeddings and cross-head attention mechanisms to dynamically integrate treatment-specific and control-specific representations, enhancing inter-group correlation modeling. Theoretically, we prove that access to the true propensity scores ensures ITE identifiability even with unobserved confounders. For practical scenarios lacking true propensity scores, RA-IPS adversarially optimizes propensity weights within constrained uncertainty sets to mitigate bias from unobserved variables. Experiments on public datasets (CRITEO-UPLIFT, LAZADA) and a production e-commerce dataset demonstrate CHAUN's superiority over state-of-the-art uplift models, achieving relative improvements of up to 25.6% in QINI scores. RA-IPS further enhances robustness, outperforming standard IPS by 5.4% under unobserved confounding. The results validate the effectiveness of our proposed methods in real-world causal inference tasks. --- ### Heavy-Ball Q-Learning with Residual Weighting Correction Authors: Donghwan Lee Published: 2026-06-25 Categories: cs.LG, cs.AI arXiv: https://arxiv.org/abs/2606.27112v2 This paper proposes a corrected heavy-ball Q-learning method for reinforcement learning (RL) and establishes convergence of its deterministic mean dynamics. It also identifies conditions under which the method is theoretically guaranteed to converge faster than standard Q-learning. The same construction is then extended to Q-learning with linear function approximation, where analogous convergence and acceleration statements are derived for the corresponding corrected fixed point. The sampled stochastic versions are treated through conditional-mean recursions and, in the stated linear-function-approximation setting, finite-time bounds. The analysis is based on a switched linear system (SLS) representation of Q-learning algorithms and on the joint spectral radius (JSR) of the associated switching families. This SLS viewpoint is not commonly used in standard analyses of Q-learning, and it provides a complementary framework and new insight into how heavy-ball momentum can accelerate Q-learning. --- ### Application of LLMs to Threat Assessment of Foreign Peacekeeping Missions Authors: Gerhard Backfried, Christian Schmidt, Diego Pilutti, Michael Suker Published: 2026-06-25 Categories: cs.CR, cs.AI arXiv: https://arxiv.org/abs/2606.27106v2 We present a novel approach for applying Large Language Models (LLMs) to threat assessment in the context of foreign peacekeeping missions. Building on the PINPOINT project and its use case, the EU Monitoring Mission in Georgia, we combine an interdisciplinary risk-model with OSINT-based media collection and LLM-supported threat extraction. The proposed workflow maps media contents to mission-relevant threats, extracts structured information and applies several additional LLM-based processing steps to improve relevance and grounding. An evaluation of threats extracted from media documents shows high agreement between automatically generated results and human judgment for core aspects such as threat and mission relevance. These results indicate that LLMs provide a promising approach to support analysts in the context of peacekeeping missions. --- ### The Riddle Riddle: Testing Flexible Reasoning in Large Language Models and Humans Authors: Bella Fascendini, Kathryn McGregor, Max D. Gupta, Thomas L. Griffiths Published: 2026-06-25 Categories: cs.CL arXiv: https://arxiv.org/abs/2606.27103v1 Humans flexibly adapt their reasoning strategies to the requirements of a given problem. Large language models (LLMs) have performed well on many cognitive tasks, however, it is unclear whether this accuracy is a result of pattern matching from training data or flexible reasoning. Here, we introduce a novel paradigm to test this question: the riddle riddle paradigm. Riddle riddles are word problems written to mimic popular riddles, but altered so their answers only require literal interpretations. Identifying correct answers requires looking past the structure of each question and flexibly apply different reasoning strategies based on the content. If LLMs respond to surface features, such as form, a riddle-like structure should cause models to use an inventive reasoning strategy even when a literal interpretation suffices. Alternatively, if LLMs reason based on content, they should flexibly switch strategies when appropriate. Across two experiments with nine state-of-the-art LLMs and 100 human participants, we show humans and LLMs fail on this paradigm in opposite directions. LLMs were far more accurate on genuine riddles than on riddle riddles (84.9% vs. 50.7%); whereas humans showed the reverse effect (50.5% vs. 80.5%). Error analysis shows that 90.8% of LLM errors on riddle riddles (the condition where they show diminished performance) were due to inappropriate use of inventive reasoning while only 57.6% of human errors on genuine riddles were due to overextending literal reasoning. Thus, while both groups make mistakes, reasoning mistakes are made more often by LLMs than by humans. Overall, LLMs' strong performance on genuine riddles may reflect memory retrieval rather than flexible strategy selection, and without stimuli designed to elicit this contrast, it becomes easy to conflate LLM-generated outputs that look like reasoning with genuine reasoning. --- ### Transformer-Based Classification of Bacterial Raman Spectra with LOOCV Authors: Jamile Mohammad Jafari, Thomas Bocklitz Published: 2026-06-25 Categories: cs.LG arXiv: https://arxiv.org/abs/2606.27096v1 Transformer-based models have recently attracted increasing attention for Raman spectral classification. In this study, a transformer-based approach was systematically evaluated using a nested leave-one-replicate-out cross-validation framework and compared with conventional machine-learning pipelines combining PCA or ICA with LDA, SVM, and Random Forest classifiers. A bacterial Raman dataset comprising 5,417 single-cell spectra from six bacterial species and nine independent measurement replicates was used. The transformer consistently achieved the highest classification performance across independent test replicates and significantly outperformed all conventional approaches. Analysis of the learned latent feature space revealed improved class separation compared with PCA- and ICA-based representations. Furthermore, the transformer maintained superior performance when applied directly to raw Raman spectra without preprocessing, demonstrating robust behavior across measurement replicates. These findings highlight the potential of transformer-based models for robust Raman spectral classification and emphasize the importance of replicate-aware validation for realistic model evaluation. --- ### Data-Free Reservoir Features for Efficient Long-Horizon Cold-Start Continual Learning Authors: Augustinas Jučas, Yangchen Pan Published: 2026-06-25 Categories: cs.LG, cs.AI arXiv: https://arxiv.org/abs/2606.27095v1 Cold-start exemplar-free class-incremental learning requires learning a growing set of classes without replay, external pretraining, or a large initial task. Existing cold-start methods typically either train the backbone throughout the stream and compensate for semantic drift, or freeze a backbone after the first task, producing features biased toward the initial classes. These choices also create a computational tension: drift-compensation methods require repeated backbone training and increasingly expensive updates as the task horizon grows, while frozen-backbone methods are cheap but weak under cold start. We study a third option: a feature extractor that is never fit to image data at all. We propose CIRCLE, a class-incremental classifier built from fixed bidirectional two-dimensional reservoir features, adapted from BiRC2D for image classification, and streaming linear discriminant analysis heads. CIRCLE groups multiple random reservoir instantiations into feature ensembles and averages the softmax outputs of independent SLDA heads, yielding a tunable bias-variance tradeoff between richer random features and prediction-level ensembling. Because the feature extractor is fixed and the head admits streaming closed-form updates, CIRCLE performs sample-wise training without replay, task-boundary information, or backbone backpropagation. On CIFAR-100, TinyImageNet, ImageNet-Subset, and ImageNet-1k, CIRCLE is competitive at 10-20 task splits and substantially outperforms strong CS-EFCIL baselines at 50, 100, and 500 task splits, while training much faster than trained-backbone drift-compensation methods. Ablations show that the BiRC2D-style extractor, SLDA head, and balanced feature/prediction ensembling each contribute to the final performance. --- ### Inherited Circuits, Learned Semantics: How Fine-Tuning Creates Evasion Vulnerabilities Invisible to Standard Evaluation Authors: Ryan Fetterman Published: 2026-06-25 Categories: cs.CR, cs.AI arXiv: https://arxiv.org/abs/2606.27091v1 LLMs fine-tuned for security classification are usually evaluated on held-out examples from the same distribution as their training data. We show that this can miss vulnerabilities introduced by fine-tuning itself: models can learn token-level indicator semantics that preserve canonical accuracy while failing under behavior-preserving transformations such as PowerShell alias substitution, command reconstruction, string construction, execution indirection, and case mutation. We study Foundation-Sec-8B-Instruct and its base model, Llama-3.1-8B-Instruct, on matched PowerShell classification cohorts. Causal interventions localize the classification circuit to a late-attention route inherited from Llama rather than created by fine-tuning. Fine-tuning concentrates and semantically specializes this inherited structure, improving baseline behavior while creating transformation-sensitive attack surfaces. A three-tier evasion benchmark finds Foundation-Sec misses on iwr substitution, Invoke-Expression reconstruction, and case-mutated Invoke-Expression/IEX variants that Llama does not share. We also derive a pre-deployment monitoring method: a linear probe at the classification boundary and an indicator-token sign test identify command families where canonical indicators change role after fine-tuning. These signals prioritize red-team variant generation using only canonical inputs, showing that security fine-tuning can improve task accuracy while expanding the evasion surface. These results caution against treating small task-specific fine-tunes as straightforwardly safer security classifiers: specialization can convert inherited model structure into brittle indicator rules that preserve held-out accuracy while expanding the evasion surface. Robust AI-enabled security will require specifying the full transformation space of the task and monitoring semantic drift through fine-tuning. --- ### Beyond Global Divergences: A Local-Mass Perspective on Bayesian Inference Authors: Hanli Xu, Fengxiang He, Sarat Moka Published: 2026-06-25 Categories: stat.ML, cs.AI, cs.LG arXiv: https://arxiv.org/abs/2606.27090v1 Global objectives, such as KL divergence and ELBO, are widely used in Bayesian inference for measuring distributional discrepancy. This paper studies their local-mass behaviour that is not directly captured by such objectives. We introduce and use two mathematical tools: (1) Mass Index for recording the polynomial and logarithmic decay scales of local mass, and (2) regularised extended KL (RE-KL), a set-localised divergence that can be formulated in the presence of singular components. Mass Indices help characterise how Bayesian updating changes local mass: (1) power-log likelihood factors shift it explicitly, and (2) parameter-dependent supports, or their smooth softenings, may change the local scale through the amount of mass that remains near the parameter value. Using local RE-KL, we prove absolute, relative, and directional inequalities for comparing local small-ball masses under the two KL directions. Together, these results provide a local theoretical account of local mass behaviour. Experiments provide controlled illustrations of the local behaviour. Code is available at https://github.com/Forsythia0604/Local-Mass-Framework. --- ### TMP: Tree-structured Mixed-policy Pruning for Large-scale Image Generation and Editing Authors: Peizhen Zhang, Yang Li, Xunsong Li, Songtao Liu, Zewen Liu, Qiangqiang Hu, Guotong Guo, Jupeng Ding, Yifu Sun, coopersli, Jian Zhang, Zhao Zhong et al. Published: 2026-06-25 Categories: cs.CV arXiv: https://arxiv.org/abs/2606.27089v1 Modern image generation model rapidly grows their sizes to meet high-fidelity image synthesis. However, they gradually become unaffordable for their enormous parameter consumption and computation budget that lead to massive resources requirement and gpu memory footprint. In this paper, we propose TMP, the first Tree-structured Mixed-policy Pruning framework that generalizes prevalent image tasks (T2I and TI2I) and architectures (Mixture-of-Experts (MoE) and Diffusion transformer (DiT)). It could be applied to the step-distilled models and contribute as the last stage. We perform experiments upon current open-sourced SOTA HunyuanImage-3.0 instruct and a popular efficient model Z-Image turbo. The proposed pruning framework manages to compress HunyuanImage 3.0 from 80B to 20B parameters at 75% reduction ratio, sacrificing limited generation quality. We also optimize to enable the inference of the pruned 20B version of HunyuanImage 3.0 on a single 24GB 4090 GPU by engineering skills. The inference script and model weight have been integrated into the existing HunyuanImage3.0 open-source github and huggingface repository. Besides, we prove the efficacy of TMP by compressing Z-Image turbo from 6B to 4B (33% reduction) with negligible degradation. --- ### SubdivAR: Autoregressive Next-Scale Prediction for Neural Mesh Subdivision Authors: Huipeng Guo, Zikai Song, Hang Long, Jielei Zhang, Wenbing Li, Junkai Lin, Tianhao Zhao, Jinshen Zhang, Tianle Guo, Wei Yang Published: 2026-06-25 Categories: cs.CV arXiv: https://arxiv.org/abs/2606.27088v2 Mesh subdivision is a fundamental operation for converting coarse, editable meshes into high-resolution surfaces, with broad applications in digital asset creation. Classical rule-based schemes rely on fixed local refinement rules and often produce over-smoothed surfaces. Recent neural subdivision methods improve detail synthesis, but remain constrained by local modeling and exhibit limited generalizability. We present SubdivAR, a neural mesh subdivision framework based on our proposed Mesh Autoregressive Representation (MAR). MAR arranges meshes at different subdivision levels into an ordered scale sequence, reformulating subdivision as autoregressive next-scale prediction. To support this formulation, we introduce a Hybrid Topology-Aware Transformer that combines global semantic attention with topology-constrained local feature aggregation. SubdivAR adopts a next-scale coordinate prediction paradigm, regressing vertex offsets at each refinement stage to preserve subdivision topology while recovering fine-grained geometric details. To enable reliable learning, we construct FII-40K, a curated dataset of nearly 40,000 high-quality meshes with multi-level subdivision supervision. Experiments show that SubdivAR outperforms state-of-the-art baselines, reducing Hausdorff Distance and Chamfer Distance by 18.8% and 14.2%, respectively, and demonstrates strong robustness on complex open-surface geometries. --- ### Pseudo-Text-Conditioned 3D Grounding DINO for Organ Localization in Abdominal CT Authors: Siqi Chen, Han Gong, Keyi Hou, Jingxuan Yang, Sheethal Bhat, Andreas Maier Published: 2026-06-25 Categories: cs.CV, eess.IV arXiv: https://arxiv.org/abs/2606.27084v1 Reliable organ localization in abdominal CT can provide spatial priors for downstream trauma analysis. We propose CT-3GDINO, a lightweight 3D detector that adapts a Grounding-DINO-style query-based architecture to fixed organ localization using frozen pseudo-text class tokens instead of a real text encoder. The model combines a Swin3D visual backbone, bidirectional feature enhancement, pseudo-text-guided query selection, and a cross-modality decoder to predict normalized 3D boxes for liver, spleen, left kidney, right kidney, and bowel. We train and evaluate on 193 matched RSNA/RATIC CT volumes with segmentation-derived boxes. The best multi-scale model, trained from scratch, achieves 0.5830 overall top-1 class-wise mAP over 3D IoU thresholds from 0.1 to 0.7, outperforming fixed- and trainable-backbone classification-pretrained variants with 0.5570 and 0.4657 mAP. Performance is strong for coarse localization, with 0.9649 AP at IoU 0.1, but remains limited for strict box alignment, with 0.1552 AP at IoU 0.7. These results establish CT-3GDINO as an open-source baseline for pseudo-text-conditioned 3D organ localization and motivate future work on localization-aware pretraining, richer multimodal conditioning, and injury-focused detection. --- ### Finding Stationary Points by Comparisons Authors: Helin Wang, Chenyi Zhang, Xiwen Tao, Yexin Zhang, Tongyang Li Published: 2026-06-25 Categories: cs.LG, cs.DS, math.OC, quant-ph arXiv: https://arxiv.org/abs/2606.27082v1 We study the problem of finding stationary points of non-convex functions when access to the objective is provided only through a comparison oracle that, given two points, outputs which has the larger function value. For a twice differentiable $f\colon\mathbb R^n\to\mathbb R$ with Lipschitz gradient and Hessian, we develop an algorithm that visits an $ε$-stationary point using $\widetilde O(n^2/ε^{1.5})$ queries. Our approach uses a subroutine that estimates the normalized Hessian to accuracy $δ$ using $\widetilde O(n^2\log(1/δ))$ queries. We further study this problem with a quantum comparison oracle model where queries can be made in superpositions, and develop the first quantum algorithm that finds an $ε$-stationary point, which takes $\widetilde O(n/ε^{1.5})$ queries. --- ### PanoImager: Geometry-Guided Novel View Synthesis and Reconstruction from Sparse Panoramic Views Authors: Zhisong Xu, Takeshi Oishi Published: 2026-06-25 Categories: cs.CV arXiv: https://arxiv.org/abs/2606.27071v2 Panoramic sensing offers wide field-of-view coverage, yet 3D reconstruction from sparse panoramas remains challenging under rotation-dominant, weak-parallax motion. In such regimes, SfM/SLAM initialization is often ill-conditioned and unreliable. We present PanoImager, an SfM-free framework that combines feed-forward pose/depth priors, geometry-conditioned diffusion view completion, and depth-guided 3DGS optimization. Given only a few panoramic images, PanoImager decomposes them into local perspective views, synthesizes auxiliary observations to enrich sparse evidence, and stabilizes Gaussian optimization for improved cross-view consistency. Experiments on multiple benchmarks show improved stability under extreme sparsity, suggesting PanoImager as an offline/background component for map refinement when SfM/SLAM fails to initialize. --- ### Towards Explainable Adjudicative Variance: Quantifying Judicial Discretion via Gated Multi-Task Learning Authors: Stanisław Sójka, Felix Steffek, Matthias Grabmair Published: 2026-06-25 Categories: cs.CL arXiv: https://arxiv.org/abs/2606.27069v2 Legal outcome prediction must disentangle objective case facts from adjudicative context. Merit-based rulings rely on factual evidence while technical disposals may hinge on judicial discretion. We propose a Judge-Aware Gated Multi-Task Learning architecture that explicitly models this distinction. We introduce a fine-grained outcome taxonomy to supervise the encoder, enforcing a structural regularization that disentangles distinct semantic pathways. This granular legal curriculum enables our Gated Fusion mechanism to dynamically modulate reliance on judge identity. We evaluate our approach on 13,937 UK Employment Tribunal decisions. We benchmark our design against supervised fine-tuning (SFT) of a Gemma-4 26B-A4B backbone, in which judge identity and the taxonomy are injected as prompt tokens or autoregressive output targets. The two contextual signals compose only weakly when forced through a single autoregressive channel. In contrast, coupling a LoRA-adapted Gemma-4 encoder with our gated architecture defines a new state of the art on this benchmark while requiring an order of magnitude fewer trainable parameters than the generative SFT baselines, with gains concentrated on the most ambiguous and rarest outcome classes. Beyond accuracy, the architecture is interpretable; learned judge embeddings and calibration profiles localize the cases where adjudicative context drives the prediction. These results indicate that, for identity-conditioned classification of legal outcomes, the choice of conditioning interface dominates scale: differentiable structured composition yields more accurate, more parameter-efficient models than prompt-based composition over a substantially larger backbone. --- ### Parametric Open Source Games Authors: Aleksandar Todorov, Jesse ten Napel, Alexander Müller Published: 2026-06-25 Categories: cs.GT, cs.AI, cs.LG arXiv: https://arxiv.org/abs/2606.27068v1 Open-source game theory studies agents whose behavior may depend on one another's decision procedures, but most existing models use discrete or symbolic programs. We introduce parametric open-source games, a continuous analogue of program equilibria in which players choose parameter vectors and semantics maps convert the full parameter profile into mixed actions in an underlying finite game. We establish equilibrium existence results, derive an exact coupling threshold at which selfish gradient ascent in symmetric $2\times2$ games switches from defection toward cooperation, and give a one-dimensional boundary test for parametric program Nash equilibria. We further extend the framework to a neural semantics class whose first-order cooperation condition is governed by the ratio of cross-player to self-player sensitivity. Across canonical games, the framework shows how access to internal parameterizations can qualitatively reshape learning dynamics and equilibrium structure, and how sufficiently strong open-source coupling can steer selfish optimization toward cooperative outcomes. --- ### How to evaluate clustering with ground truth? Authors: Pasi Fränti Published: 2026-06-25 Categories: cs.AI arXiv: https://arxiv.org/abs/2606.27061v1 External indexes can be used for cluster evaluation when ground truth is available. We review the most common external validity indexes focusing on set-matching-based measures. We recommend centroid index (CI), because it is an intuitive cluster-level measure with an explainable result. If we need a more fine-tuned, point-level measure, there are more choices. Pair-set index (PSI) provides a normalized score which is not biased by cluster sizes. If all points should matter equally, then clustering accuracy (ACC) or any other set-matching measure is suitable. --- ### NuclearQAv2: A Structured Benchmark for Evaluating Domain-Science Competence in Large Language Models Authors: Henry Shaowu Yuchi, Michal Kucer, Benjamin H. Sims, Selma Peterson, Emily Taylor Published: 2026-06-25 Categories: cs.CL, cs.AI arXiv: https://arxiv.org/abs/2606.27047v1 Large language models (LLMs) have demonstrated strong performance across a wide range of tasks, but ensuring their reliability in highly technical domains remains a significant challenge. In nuclear engineering, problem solving often requires not only factual knowledge but also quantitative reasoning and conceptual understanding. To address the need for systematic evaluation in this domain, we introduce NuclearQAv2, a benchmark for assessing LLMs on nuclear engineering knowledge. The benchmark comprises approximately 1,240 question-answer pairs spanning three categories: boolean, numeric, and verbal. NuclearQAv2 is constructed using a hybrid pipeline that combines expert-authored questions, existing datasets, and LLM-assisted generation from domain-specific technical corpora. By leveraging structured prompting for both automated question generation and response evaluation, the proposed framework enables scalable benchmark construction and evaluation. We evaluate a diverse set of LLMs using NuclearQAv2 and observe substantial performance differences across task types. While the models generally perform well on factual questions, quantitative reasoning and conceptual understanding remain considerably more challenging. These results highlight the importance of multi-faceted evaluation frameworks and establish NuclearQAv2 as a scalable benchmark for assessing LLM capabilities in technical domains. --- ### The Spec Growth Engine: Spec-Anchored, Code-Coupled, Drift-Enforced Architecture for AI-Assisted Software Development Authors: Hartwig Grabowski Published: 2026-06-25 Categories: cs.SE, cs.AI arXiv: https://arxiv.org/abs/2606.27045v1 AI coding agents dramatically accelerate implementation speed but introduce two structural failure modes that existing spec-driven approaches do not fully solve: (1) context explosion -- the agent must reason over an entire repository at once, degrading output quality as the context window fills; and (2) silent spec-code drift -- code evolves, the specification does not, and the divergence becomes invisible until it is costly to repair. We present the Spec Growth Engine, a lightweight framework that addresses both failure modes through a machine-readable spec graph whose nodes carry explicit contract/design separation, a Spine context assembler that scopes agent context to an ownership path, a vertical-slice growth protocol that enforces hardest-first ordering, and a drift gate that makes spec-code divergence a blocking merge condition. The design synthesises well-established software engineering principles (Parnas information hiding, C4, ADRs, Walking Skeleton, Reflexion Models, Fitness Functions) into a lean, code-coupled, machine-enforced whole -- without the overhead of heavy-weight frameworks such as RUP or MDA. --- ### State Representation Matters in Deep Reinforcement Learning: Application to Energy Trading Authors: Jesper Klicks, Sander Vržina, Vincent François-Lavet Published: 2026-06-25 Categories: cs.LG, cs.AI arXiv: https://arxiv.org/abs/2606.27032v1 Energy trading decisions depend not only on current market prices, but also on expected future market conditions, and operational constraints. This makes the state representation given to a reinforcement learning agent an important design choice. We study this in HydroDam, a pumped-storage arbitrage environment, using a fixed Double DQN agent. The environment, action space, reward function, network, and training protocol are kept fixed; only the market features are changed. We compare absolute price/calendar features, relative features that compare current prices with recent market history, forecast features, and all combinations of these three feature families. Policies are trained and selected using 2007--2011 Belgian day-ahead prices and evaluated on two test settings: a later same-market test set from 2012--2025 and 39 other ENTSO-E market zones. Absolute features only reaches 28.8% on the test set and a median 5.7% across zones. Relative-only and forecast-only states also stay below a rolling price-score heuristic in the cross-zone median. Combining feature families is much stronger: absolute + relative reaches 49.9% on the test set and a 39.8% cross-zone median, while absolute + relative + forecast reaches 55.6% and 47.5%. These results suggest that state representation is not a minor preprocessing choice in storage-trading RL, but a central part of the policy design: robust transfer requires combining price scale, recent relative price context, and short-horizon forecast information, rather than relying on any single feature family. --- ### Symplectic Neural Networks for Learning Non-Separable Hamiltonians Authors: Harsh Choudhary, Vyacheslav Kungurtsev, Chandan Gupta, Melvin Leok, Georgios Korpas Published: 2026-06-25 Categories: cs.LG arXiv: https://arxiv.org/abs/2606.27029v3 Hamiltonian Neural Networks (HNNs) integrate physical priors into neural models by learning a system's Hamiltonian, improving generalization and sample efficiency. Identifying the system Hamiltonian from noisy observations of state variables is a challenging task. For simulations to faithfully reflect the long-term behavior of Hamiltonian systems, especially energy conservation, it is essential to use symplectic integrators, which preserve the system's geometric structure. This fidelity comes at a cost: implicit symplectic integrators are more computationally intensive and make backpropagation through the ODE solver non-trivial. However, by leveraging the fact that symplectic discretizations of the adjoint system yield the same sensitivities associated by backpropagation, we obtain an efficient method of training the Neural Network parameters. In our work, we explore this alternate method of HNN training under noisy observation of trajectories with our HNN model based on an implicit symplectic integrator. Computationally, a predictor-corrector based ODE solver and fixed point iteration help to mitigate the computational cost of the implicit timestepping, resulting in more efficient generation of gradient updates. We showcase the numerical advantage, in experiments, in system identification and energy preservation on a range of non-separable, chaotic systems and the efficient computation and memory complexity of our method. We also observe that the post-processing of the learned Hamiltonian using backward error analysis yields a modified Hamiltonian that is a more accurate approximation of the true Hamiltonian without the need to use more accurate discretizations of the flow map. --- ### Symplectic Neural Networks for Learning Non-Separable Hamiltonians Authors: Harsh Choudhary, Vyacheslav Kungurtsev, Chandan Gupta, Melvin Leok, Georgios Korpas Published: 2026-06-25 Categories: cs.LG arXiv: https://arxiv.org/abs/2606.27029v2 Hamiltonian Neural Networks (HNNs) integrate physical priors into neural models by learning a system's Hamiltonian, improving generalization and sample efficiency. Identifying the system Hamiltonian from noisy observations of state variables is a challenging task. For simulations to faithfully reflect the long-term behavior of Hamiltonian systems, especially energy conservation, it is essential to use symplectic integrators, which preserve the system's geometric structure. This fidelity comes at a cost: implicit symplectic integrators are more computationally intensive and make backpropagation through the ODE solver non-trivial. However, by leveraging the fact that symplectic discretizations of the adjoint system yield the same sensitivities associated by backpropagation, we obtain an efficient method of training the Neural Network parameters. In our work, we explore this alternate method of HNN training under noisy observation of trajectories with our HNN model based on an implicit symplectic integrator. Computationally, a predictor-corrector based ODE solver and fixed point iteration help to mitigate the computational cost of the implicit timestepping, resulting in more efficient generation of gradient updates. We showcase the numerical advantage, in experiments, in system identification and energy preservation on a range of non-separable, chaotic systems and the efficient computation and memory complexity of our method. We also observe that the post-processing of the learned Hamiltonian using backward error analysis yields a modified Hamiltonian that is a more accurate approximation of the true Hamiltonian without the need to use more accurate discretizations of the flow map. --- ### ShareLock: A Stealthy Multi-Tool Threshold Poisoning Attack Against MCP Authors: Liwei Liu, Tianzhu Han, Zijian Liu, Zishu Dong, Na Ruan Published: 2026-06-25 Categories: cs.CR, cs.AI arXiv: https://arxiv.org/abs/2606.27027v1 With the rapid evolution of LLM-driven agents, Model Context Protocol (MCP), an open protocol bridging LLMs with external tools, has quickly become foundational to modern agent ecosystems. However, the expanding adoption of MCP has also introduced novel security concerns such as Tool Poisoning Attack (TPA), which exploit LLM-server interactions to inject malicious prompts. Existing poisoning schemes typically adopt a monolithic plaintext embedding paradigm, which fails to withstand manual inspection or automated detectors. Current research still lacks a systematic analysis on multi-tool poisoning, where multiple tools can be exploited cooperatively to disperse detection risk. In this paper, we introduce ShareLock, a multi-tool threshold poisoning framework that utilizes Shamir's threshold scheme to ensure exceptional stealth and fault tolerance. ShareLock distributes the malicious instruction as benign-looking secret shares across multiple tool descriptions, achieving both information-theoretic secrecy and attack robustness against moderate auditing. After a covert reconstruction trigger is planted during server update, the aggregated shares reconstruct the hidden instruction, resulting in critical breaches of system assets or private data. To evaluate the realistic threat of ShareLock, we constructed a comprehensive benchmark encompassing four multi-tool scenarios and conducted extensive experiments across mainstream LLMs on two distinct MCP clients. Our results demonstrate that ShareLock significantly outperforms existing single-tool poisoning strategies in tool description-based detection while maintaining an average attack success rate exceeding 90%. --- ### Improving General Role-Playing Agents via Psychology-Grounded Reasoning and Role-Aware Policy Optimization Authors: Zhenhua Xu, Dongsheng Chen, Jian Li, Yitong Lin, Zhebo Wang, Jiafu Wu, Yizhang Jin, Chengjie Wang, Meng Han, Yabiao Wang Published: 2026-06-25 Categories: cs.CL arXiv: https://arxiv.org/abs/2606.27025v1 Building general-purpose role-playing agents that faithfully portray any character from a natural-language profile remains challenging. The dominant paradigm -- supervised fine-tuning -- encourages behavioral mimicry without deep, human-like internal thought processes, resulting in poor out-of-distribution generalization. Therefore, we propose \textbf{Psy-CoT}, a psychology-grounded chain-of-thought framework that decomposes pre-response reasoning into three role-specific steps -- \emph{Interaction Perception}, \emph{Psychological Empathy}, and \emph{Logical Construction} -- so that the model \emph{thinks dynamically} from the profile rather than merely mimicking surface patterns. While structured reasoning provides a foundation, it alone is insufficient; reinforcement learning is essential to further align the model with character fidelity. However, we observe that under LLM-based reward models, both generic phrases that hack the reward model and genuinely role-specific phrases receive identical gradient signals -- this hacking accumulates over training, misleading the model into treating both as equally optimal choices. To address this, we propose \textbf{Role-Aware Policy Optimization (RAPO)}, which uses profile--token mutual information to weight gradients asymmetrically -- amplifying role-specific tokens under positive advantage while attenuating them under negative advantage. Experiments on CoSER, CharacterBench, and CharacterEval demonstrate that Psy-CoT outperforms existing role-playing CoT methods, and RAPO consistently surpasses GRPO across multiple model scales. --- ### Just how sure are you? Improving Verbalized Uncertainty Calibration in Medical VQA Authors: Eren Senoglu, Federico Toschi, Nicolo Brunello, Andrea Sassella, Mark James Carman Published: 2026-06-25 Categories: cs.LG, cs.CL, cs.CV arXiv: https://arxiv.org/abs/2606.27023v1 Multimodal large language models (MLLMs) applied to Medical Visual Question Answering (VQA) tend to produce overconfident outputs regardless of actual correctness, and existing verbalized confidence calibration methods, developed primarily for text only LLMs, do not account for the multimodal nature of medical image understanding. This work proposes a training based framework that finetunes MLLMs to improve their calibration using a composite loss function combining a Brier style calibration term, an anchor regularizer that prevents confidence collapse toward extreme values, a contrastive image text alignment term, and a KL based model stabilization term. The alignment signal is derived from a $2 \times 2$ factorial perturbation design that crosses image presence with text integrity, probing the reliance of the model on visual modality input versus language priors. Finally, a top K KL divergence regularizer is used to protect the answering ability of the model during finetuning. Across three Medical VQA benchmarks and two architectures (MedGemma 4B IT and Qwen2 VL 7B Instruct), our method reduces calibration error by 60% or more, and improves discrimination by 26% or more, while preserving predictive accuracy. On average across benchmarks, the technique outperforms prompting based, sampling based, and training based approaches, and ablation experiments confirm that each component of the loss function is indeed necessary for improving the calibration. All code for the experiments is publicly available. --- ### MinGram: A Minimalist Unigram Tokenizer with High Compression and Competitive Morphological Alignment Authors: Sander Land Published: 2026-06-25 Categories: cs.CL arXiv: https://arxiv.org/abs/2606.27019v2 The Unigram tokenizer uses an elegant representation which makes it straightforward to edit vocabularies, but its training is comparatively heavy and complex. We introduce MinGram (Minimalist Unigram), which keeps the token-list representation but simplifies training using a BPE-derived seed vocabulary, Hard EM on a minimum-token path, and a single flat score-pruning step. This removes the suffix array, the forward-backward pass, and the iterative prune loop, leaving a procedure that requires little beyond tokenizer inference itself. By making token count the primary objective and using a Unigram score only as a tiebreak, MinGram keeps the compression of pure token-count methods while retaining much of the morphological alignment and downstream quality of probabilistic ones. Across six languages, MinGram compresses better than both BPE and standard Unigram, and a compression-oriented variant matches the strongest token-count compressors while retaining substantially higher morphological alignment. In controlled downstream language-model training, Unigram-family tokenizers, with MinGram among the best, consistently beat BPE in bits-per-byte. --- ### MinGram: A Minimalist Unigram Tokenizer with High Compression and Competitive Morphological Alignment Authors: Sander Land Published: 2026-06-25 Categories: cs.CL arXiv: https://arxiv.org/abs/2606.27019v1 The Unigram tokenizer uses an elegant representation which makes it straightforward to edit vocabularies, but its training is comparatively heavy and complex. We introduce MinGram (Minimalist Unigram), which keeps the token-list representation but simplifies training using a BPE-derived seed vocabulary, Hard EM on a minimum-token path, and a single flat score-pruning step. This removes the suffix array, the forward-backward pass, and the iterative prune loop, leaving a procedure that requires little beyond tokenizer inference itself. By making token count the primary objective and using a Unigram score only as a tiebreak, MinGram keeps the compression of pure token-count methods while retaining much of the morphological alignment and downstream quality of probabilistic ones. Across six languages, MinGram compresses better than both BPE and standard Unigram, and a compression-oriented variant matches the strongest token-count compressors while retaining substantially higher morphological alignment. In controlled downstream language-model training, Unigram-family tokenizers, with MinGram among the best, consistently beat BPE in bits-per-byte. --- ### On-board Remote-Sensing Foundation Models for Unsupervised Change Detection of Disaster Events Authors: S. Ramírez-Gallego Published: 2026-06-25 Categories: cs.CV, cs.AI arXiv: https://arxiv.org/abs/2606.27018v1 Remote Sensing Foundation Models (RSFMs) have emerged as a powerful alternative to supervised models for Earth Observation, allowing satellites to autonomously trigger high-resolution captures or adjust tasking parameters upon detecting an anomaly, thereby maximizing the utility of the mission's limited power and computational resources. RSFMs are versatile, unified encoders that optimize onboard storage for multiple orbital applications while ensuring high-fidelity feature extraction. In particular, unsupervised change detection with RSFMs offers a well-informed and transformative path for disaster monitoring without expensive labels. In this paper, we present a novel unsupervised detection method based on ResNet (RSFM) + FPN which identifies a wide spectrum of anomalies by detecting subtle semantic shifts in the latent space between successive orbital passes. By relying on an untrained FPN architecture and its intrinsic priors, the system achieves efficient image-level generation and higher resolution mapping with minimal effort (training-free) compared to previous proposals (patch-based, trained). And by replacing tailored models with RSFMs, we can achieve comparable results through an approach that eliminates the need for bespoke training and extensive development effort and adds customization, while ensuring high-performance generalization across diverse terrains and sensors. --- ### A Generalization Theory for JEPA-Based World Models Authors: Jingyi Cui, Qi Zhang, Hongwei Wen, Letian Cheng, Yisen Wang Published: 2026-06-25 Categories: cs.LG arXiv: https://arxiv.org/abs/2606.27014v2 Joint Embedding Predictive Architectures (JEPAs) have emerged as a promising paradigm for world modeling by predicting future representations rather than reconstructing observations. Despite their empirical success, the theoretical understanding of JEPA-based world models remains limited. In this paper, we develop the first generalization theory for JEPA-based world models. We formulate JEPA pretraining as a conditional spectral graph learning problem and show that the JEPA objective is equivalent to a low-rank factorization of an action-conditioned co-occurrence matrix. Building on this characterization, we establish a connection between JEPA pretraining error and downstream planning regret, leading to a finite-sample generalization bound for JEPA-based world models. Our analysis reveals an approximation-sample error trade-off governed by latent dimensionality, providing insights into latent- versus input-level prediction. The theoretical trade-off motivates us to investigate a simple Multi-Depth JEPA extension that combines representations from different network depths. In validation experiments, latent-level prediction exhibits greater robustness under unpredictable observations and difficult planning tasks. Multi-depth representations further improve planning success over single-depth representations and a dimensionality-matched control. --- ### A Generalization Theory for JEPA-Based World Models Authors: Jingyi Cui, Qi Zhang, Hongwei Wen, Yisen Wang Published: 2026-06-25 Categories: cs.LG arXiv: https://arxiv.org/abs/2606.27014v1 Joint Embedding Predictive Architectures (JEPAs) have recently emerged as a promising paradigm for world modeling by learning predictive dynamics in a latent space rather than generating future observations at the input level. Despite their empirical success, the theoretical understanding of JEPA-based world models remains limited. In this paper, we develop the first generalization theory for JEPA-based world models. We formulate JEPA pretraining as a conditional spectral graph learning problem and show that the JEPA objective is equivalent to a low-rank factorization of an action-conditioned co-occurrence matrix. Building on this characterization, we establish a connection between JEPA pretraining error and downstream planning regret, leading to a finite-sample generalization bound for JEPA-based world models. Our analysis reveals an inherent trade-off between approximation and sample errors with respect to the latent dimension, providing theoretical insights into the advantages and limitations of latent predictive models compared with input-level predictive approaches. --- ### Semantic Early-Stopping for Iterative LLM Agent Loops Authors: Sahil Shrivastava Published: 2026-06-25 Categories: cs.AI, cs.LG, cs.MA arXiv: https://arxiv.org/abs/2606.27009v1 Multi-agent large language model (LLM) loops, for example a Writer that drafts and a Critic that revises, are almost always terminated by a fixed iteration cap (max_iterations). This is a syntactic kill-switch: it is blind to whether the answer is still improving, so it over-spends tokens on easy inputs and truncates hard ones. We study semantic early-stopping: the loop halts when consecutive draft embeddings stop changing in meaning (cosine distance with a patience window) and the answer's measured quality stops improving. Our work makes three contributions. First, an honest theoretical footing: we prove deterministic termination and well-definedness and machine-check these claims, while treating the convergence of the distance sequence as an empirically tested conjecture rather than a (previously over-claimed) Banach contraction. Second, a judge-efficient evaluation protocol: we generate each question's full trajectory once, replay every stopping policy over the identical drafts, and cache every LLM-judge call, yielding a strictly paired efficiency-versus-quality comparison at low cost; we further separate operational tokens (charged to a policy) from evaluation tokens (a measurement instrument). Third, an empirical study on multi-hop retrieval-augmented question answering (HotpotQA). On the 60-question test split, a judge-free semantic stopper reduces operational tokens by 38% relative to max_iterations at parity quality (Delta-IS = -0.004, p = 0.81), whereas the full quality-gated variant is counter-productive because its per-round judging dominates cost. An oracle that selects the best round attains +0.115 Information Score over every practical policy (p ~ 4e-11), reframing the problem from "when to stop" (easy) to "which round is best" (open). --- ### Adaptive Utility driven Resource Orchestration for Resilient AI (AURORA-AI) Authors: Rahul Umesh Mhapsekar, Ilias Cherkaoui, Lizy Abraham, Indrakshi Dey Published: 2026-06-25 Categories: cs.AI arXiv: https://arxiv.org/abs/2606.27005v1 Modern AI systems are increasingly deployed under non-stationary computational, demographic, and operational conditions in which static resource allocation strategies degrade both predictive performance and human-centric properties such as fairness and explainability. This paper presents AURORA-AI, an Adaptive Utility-driven Resource Orchestration framework for Resilient AI that unifies Hamilton-Jacobi-Bellman feedback control, Lyapunov-based stability monitoring, and a fairness-aware composite utility into a single closed-loop policy.The framework continuously redistributes computational budget across a population of heterogeneous AI models so that the global utility, defined jointly over predictive performance, demographic parity, cost, latency, robustness, and interpretability, remains maximised under disruption. The framework is evaluated in a stress-rich discrete-time simulation that concurrently injects demographic bias shocks, gradual concept drift, and abrupt black-swan disruptions, and is compared against five established controllers including Static, Round Robin, Greedy, LinUCB, and a deep reinforcement-learning agent based on Proximal Policy Optimisation. AURORA-AI achieves immediate recovery from the black-swan event compared to eighty-eight time steps for the Static baseline and twenty-two for Proximal Policy Optimisation, lifts the alpha-quantile and the super-quantile by twenty-nine and twenty-five percent respectively, simultaneously reduces the mean and maximum demographic parity gap, and increases the fraction of Lyapunov-stable operating steps. These results indicate that fairness-aware adaptive orchestration grounded in stability theory is a practical and theoretically motivated path toward resilient human-centric AI deployment. --- ### Inverse Design of Compact and Wideband Inverted Doherty Power Amplifiers Using Deep Learning Authors: Han Zhou, Haojie Chang, David Widen, Christian Fager Published: 2026-06-25 Categories: eess.SP, cs.AI, cs.AR, eess.SY arXiv: https://arxiv.org/abs/2606.27002v1 This paper presents a deep learning-assisted methodology for the inverse synthesis of a compact, wideband inverted Doherty power amplifier (PA). Convolutional neural networks (CNNs) and genetic algorithms (GAs) are jointly employed to generate pixelated Doherty combiner networks that integrate load modulation, impedance matching, power combining, and phase compensation into a single structure. As a proof of concept, we design and fabricate a GaN HEMT Doherty PA with a pixelated output combiner. The prototype achieves a measured peak drain efficiency of 51%-63% and a 6-dB back-off efficiency of 48%-54% over 1.9-2.5 GHz. Within the same frequency range, the measured output power is 44+/-0.3 dBm. Furthermore, with digital predistortion (DPD) applied, the prototype circuit demonstrates an adjacent channel leakage ratio (ACLR) better than -53.2 dBc. --- ### Uncertainty quantification via conformal prediction in data assimilation Authors: Catherine George, Alireza Javanmardi, Tijana Janjić, Eyke Hüllermeier Published: 2026-06-25 Categories: cs.LG arXiv: https://arxiv.org/abs/2606.27001v1 Quantifying the evolution of uncertainty is critical to both probabilistic forecasting and data assimilation in numerical weather prediction. In this study, we investigate the applicability of conformal prediction (CP), a recent machine learning (ML) method, to quantify uncertainty in a controlled, idealized setting. We use the one dimensional modified shallow water model, designed to mimic the convective process. CP provides a set of possible outcomes with a chosen confidence level. Here, we compare and evaluate the average empirical coverage, the average interval length, miss low, miss high and average interval score loss (AISL) for three variants of CP, namely a) Standard CP, b) Normalized CP and c) Conformalized Quantile Regression. We further compare these CP-based uncertainty estimates with traditional ensemble-based measures such as standard deviation intervals and ensemble spread. In addition, we investigate the integration of CP-derived uncertainty within the data assimilation cycle through CP perturbations. Our results highlight the strengths and limitations of each approach, providing insight into the effectiveness of CP to complement common ensemble-based uncertainty quantification in simplified atmospheric models. ---