# Research Papers — 2026-06 · bundle page 52 of 260 > Full-text markdown of papers #2551–2600 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/52 [← Previous bundle](https://aipod.dev/papers/archive/2026-06/page/51.md) [Next bundle →](https://aipod.dev/papers/archive/2026-06/page/53.md) ### Self-Supervised Tree-level Biomass Estimation in Urban Environments From Airborne LiDAR and Optical Observations Authors: Jose Bermudez, Zilong Zhong, Dominic Cyr, Camile Sothe, Alemu Gonsamo Published: 2026-06-24 Categories: cs.CV, cs.LG, eess.IV arXiv: https://arxiv.org/abs/2606.26194v1 Urban tree biomass remains less spatially explicitly quantified than biomass in managed forests because many estimates rely on inventories or coarse products that cannot resolve individual crowns or fine-scale heterogeneity. We present a crown-level above-ground biomass (AGB) framework for an 810~km$^2$ landscape in Ontario, Canada, using leaf-off airborne LiDAR (8--10~pulses~m$^{-2}$) and near-infrared RGB orthophotography (0.16--0.20~m) from 2018 and 2023. A dual-stream cross-attention network trained on rule-based pseudo-labels produced semantic marks for buildings, needleleaf trees, and deciduous trees, supporting crown delineation and functional-type assignment. On independently annotated withheld tiles, global/mean precision, recall, and Dice scores were 0.86, 0.83, and 0.84. Crowns were delineated with multiscale watershed segmentation in mapped tree areas, and AGB was estimated from a crown area--height power-law proxy calibrated to species-specific allometry (Lambert et al., 2005) for 21,921 inventory trees. For 18,713 inventory--segment matched pairs from a 90,726-tree held-out test set, AGB prediction achieved $R^2=0.609$ using inventory crown geometry and $R^2=0.570$ under operational segmentation, identifying crown delineation as the remaining uncertainty source. Aggregated to 30~m, estimates yielded total AGB stocks of 1.73~Tg in 2018 and 1.81~Tg in 2023 (811--850~Gg~C), local densities up to ${\sim}140$~Mg~ha$^{-1}$ along the Niagara Escarpment, and a net carbon gain of 39~Gg~C over five years. Deep-ensemble uncertainty maps highlighted high-epistemic-uncertainty areas linked to underrepresented land covers and guided assignment of uncertain crowns to a pooled allometric equation. The framework uses standard provincial data, requires no manual annotation, and produces a public bitemporal crown-level AGB database for trees outside forests at management-relevant resolution. --- ### Federated Hash Projected Latent Factor Learning Authors: Jialan He Published: 2026-06-24 Categories: cs.LG, cs.CR arXiv: https://arxiv.org/abs/2606.26192v1 Hash Learning (HL) is an efficient representation learning approach that maps real-valued data into compact binary representations. Traditional HL methods typically require users to upload personal data to a central server, which is incompatible with increasingly stringent data security regulations. Federated Learning (FL) provides a decentralized paradigm for learning globally optimal models without centralizing private data. However, most FL methods rely on transmitting large-scale real-valued gradient information, leading to high communication overhead and potential privacy risks. Integrating HL into FL is a promising solution. Nevertheless, existing HL methods suffer from limited representational capacity of binary codes, which may degrade model accuracy. To address this challenge, we propose a Federated Hash Projected Latent Factor (FHPLF) model. FHPLF introduces three key innovations: (a) replacing real-valued gradient matrices with binary gradient-like matrices, significantly reducing computation, storage, and communication costs while enhancing privacy protection; (b) leveraging Projected Hamming Distance for similarity modeling, which captures the importance of individual binary bits to improve representation capability; and (c) proposing a Secure Binary Gradient Reassembly and Privacy-Enhanced Upload (SBG-PEU) strategy to further reduce the risk of user interaction leakage during transmission. Extensive experiments on four real-world datasets demonstrate that FHPLF consistently outperforms state-of-the-art HL and FL methods, achieving a favorable trade-off among accuracy, efficiency, and privacy preservation. --- ### Clue-Guided Money Laundering Group Discovery Authors: Boyang Wang, Jianing Cao Published: 2026-06-24 Categories: cs.LG arXiv: https://arxiv.org/abs/2606.26189v1 Money Laundering Group Discovery (MLGD) aims to identify hidden criminal groups and recover their complete structures in large-scale financial networks. Existing graph anomaly detection methods mainly produce node-level risk alerts, while global group discovery methods passively search for suspicious groups over the whole network. Both are mismatched with real Anti-money-laundering (AML) investigations, where analysts usually start from a concrete clue and gradually expand the investigation to recover the responsible group. To address this gap, we propose Clue-Guided Group Discovery (CGGD), where a laundering group is progressively recovered from an initial clue set through analyst interaction. We further propose Clue2Group, a framework that first constructs a compact local investigation context to reduce noise and preserve chain-like and cycle-like laundering structures. It then estimates a clue-conditioned local risk field with a multi-semantic local-temporal GNN, and finally integrates risk, structural, and prior-pattern evidence to recover a coherent laundering group. Experiments on two large-scale AML benchmarks show that Clue2Group provides a practical clue-driven analysis framework for AML investigations, offering a feasible step toward bridging the gap between graph-based AML research and real investigation workflows. --- ### Necessary but Not Sufficient: Temperature Control and Reproducibility in LLM-as-Judge Safety Evaluations Authors: Hiroki Tamba Published: 2026-06-24 Categories: cs.LG arXiv: https://arxiv.org/abs/2606.26185v1 LLM-as-judge ("grader") components are now standard in evaluation harnesses, including safety evaluations where a pass/fail verdict may gate downstream deployment decisions. A widespread assumption is that setting the grader's sampling temperature to 0 makes grading deterministic. We test this assumption against a real safety-evaluation codebase (Japan AISI's open-source aisev) and show it fails on two levels. First, the harness invokes its grader without setting temperature or seed; the underlying provider silently applies its default of 1.0, so items near the decision boundary flip pass/fail across identical runs (per-item disagreement up to ~50% over 20 runs). Second, pinning temperature=0 reduces but does not eliminate flips: across 690 API calls spanning two providers, three model tiers, and five sampling configurations, 1-2 of 7 borderline items remain non-reproducible even under forced greedy decoding (top_k=1). Claude Opus 4.7/4.8 has since deprecated temperature entirely, rendering the primary mitigation inapplicable to newer model generations. These findings expose a structural gap: evaluation harnesses that report single-run verdicts without variance or grader-disagreement metrics can present noise as a safety property. We release a reproduction harness (690 calls, 7 conditions) and recommend that harnesses treat grader disagreement as a first-class health metric alongside the scores themselves. --- ### LiMoDE: Rethinking Lifelong Robot Manipulation from a Mixture-of-Dynamic-Experts Perspective Authors: Zhihao Gu, Lin Wang Published: 2026-06-24 Categories: cs.RO, cs.AI, cs.LG arXiv: https://arxiv.org/abs/2606.26183v1 Building a generalist robot that can leverage prior knowledge for continuous task adaptation remains a significant challenge. Previous works alleviate the catastrophic forgetting problem by parameter-efficient fine-tuning for single-task adaptation. However, they fail to extract reusable skills and model the interaction with other skills effectively. Recent works try to address these issues by learning prompts. Differently, this paper presents an architectural perspective on the Lifelong Mixture of Dynamic Experts (\textit{LiMoDE}), a novel two-stage learning scheme for lifelong robot manipulation. Specifically, a dynamic MoE structure is first proposed in the multi-task pre-training stage to learn prior knowledge, where a varied number of heterogeneous experts are activated based on the motion information to address different short-term manipulations. Subsequently, in the task adaptation stage, we design a lifelong MoE adaptation mechanism % (LiMoEAM) that learns lifelong experts and dynamically combines them with frozen ones for new tasks, facilitating the knowledge transfer during adaptation. The proposed \textit{LiMoDE} is evaluated on both the simulated lifelong learning benchmark and real-world tasks. Extensive experiments demonstrate its effectiveness in achieving superior performance and strong lifelong adaptation by introducing a moderate number of additional trainable parameters and inference overhead. --- ### KG-TRACE: A Neuro-Symbolic Framework for Mechanistic Grounding in Antimicrobial Resistance Prediction Authors: Naman Garg, Sarika Jain, Sourav Yadav, Bharat K. Bhargava, Ghanapriya Singh, Abhishek Srivastava, Parimal Kar Published: 2026-06-24 Categories: cs.LG, cs.AI, q-bio.QM arXiv: https://arxiv.org/abs/2606.26179v1 While WGS-based AMR prediction has reached high accuracy, existing models lack a mechanism to ground neural attributions in established biological pathways. We present KG-TRACE, a novel neuro-symbolic framework that integrates the WHO mutation knowledge graph (KG) as a structured biological constraint on a neural genomic model. Unlike existing methods that learn statistical patterns in isolation, KG-TRACE fuses genomic features and RotatE-based KG embeddings through a learned epistemic trust gate, dynamically weighting neural evidence against symbolic biological knowledge. Evaluated on the CRyPTIC M. tuberculosis cohort, KG-TRACE achieves an AUROC of 0.9760 for isoniazid, achieving competitive accuracy while its primary value lies in symbolic grounding, not predictive uplift. More importantly, we introduce the Biological Grounding Ratio (BGR), a dataset-level metric that quantifies alignment between neural attributions and established biology. Our framework achieves a 92.5% symbolic coverage of isoniazid-resistant predictions and effectively identifies MDR co-occurrence artifacts by issuing laboratory follow-up flags for 'UNCERTAIN' cases. We demonstrate that neuro-symbolic grounding provides a verifiable audit trail for clinicians, bridging the gap between predictive accuracy and clinical trust. --- ### AlgoEvolve: LLM-driven Meta-evolution of Algorithmic Trading Programs Authors: Dhruv Sharma, Gautam Shroff Published: 2026-06-24 Categories: cs.AI arXiv: https://arxiv.org/abs/2606.26173v1 Recent work shows that Large Language Models (LLMs) can act as semantic mutation operators for the evolutionary discovery of programs and proofs. Most current applications focus on static coding benchmarks. We extend this paradigm to algorithmic trading. This domain is uniquely challenging because it is noisy, non-stationary, and highly discontinuous. We present AlgoEvolve, an LLM-driven evolutionary framework that generates, evaluates, and iteratively improves executable trading strategies. These strategies are expressed as Python code and evaluated through a rigorous testing protocol. Across multiple experiments, the system exhibits emergent regime-adaptive strategy logic, including autonomous shifts in trading rules. We further introduce a meta-evolutionary outer loop that evolves the prompts guiding program synthesis in the inner loop. This outer loop discovers improved search heuristics. These heuristics balance exploration and exploitation while reducing zero-trade failures. They consistently outperform initial human-designed instructions. The results demonstrate that LLM-based semantic evolution provides a viable approach for continual program synthesis in complex environments. --- ### LCG: Long-Context Consistent Image Generation with Sparse Relational Attention Authors: Zihao Wang, Yijia Xu, Haoze Zheng, Xuran Ma, Haokun Gui, Harry Yang Published: 2026-06-24 Categories: cs.CV, cs.AI arXiv: https://arxiv.org/abs/2606.26171v1 Recent image generation models achieve impressive quality in single-image synthesis, but often fail to maintain consistency across sequential outputs, as required in comics, storyboards, and visual narratives. We propose Long-Context Generation (LCG), a framework for long-context multi-image text-to-image generation, to improve consistency and scalability in long-context multi-image generation. LCG employs the Sparse Relational Attention (SRA) mechanism to selectively attend to core features across extended visual contexts, ensuring that the propagation of semantic and layout information remains computationally tractable. To enforce semantic alignment, we introduce the Routing Consistency Constraint (RCC), which leverages identity-aware masks to align structural patterns across generation branches, effectively mitigating drift in appearance even in complex multi-character scenes. To support training and evaluation in this setting, we construct the Long-Context Consistency Dataset (LCCD), a large-scale synthetic dataset comprising character-centric multi-image sequences spanning varied situational contexts. LCCD contains 600K training sequences and a separate 1K test set, with each sequence containing 6 to 20 images. The experiments demonstrate that LCG outperforms the compared baselines in prompt alignment and character consistency for long-context image generation, including multi-character scenes. --- ### Neural Architecture Search for Generative Adversarial Networks: A Comprehensive Review and Critical Analysis Authors: Abrar Alotaibi, Moataz Ahmed Published: 2026-06-24 Categories: cs.LG, cs.AI arXiv: https://arxiv.org/abs/2606.26169v1 Neural Architecture Search (NAS) has emerged as a pivotal technique in optimizing the design of Generative Adversarial Networks (GANs), automating the search for effective architectures while addressing the challenges inherent in manual design. This paper provides a comprehensive review of NAS methods applied to GANs, categorizing and comparing various approaches based on criteria such as search strategies, evaluation metrics, and performance outcomes. The review highlights the benefits of NAS in improving GAN performance, stability, and efficiency, while also identifying limitations and areas for future research. Key findings include the superiority of evolutionary algorithms and gradient-based methods in certain contexts, the importance of robust evaluation metrics beyond traditional scores like Inception Score (IS) and Fréchet Inception Distance (FID), and the need for diverse datasets in assessing GAN performance. By presenting a structured comparison of existing NAS-GAN techniques, this paper aims to guide researchers in developing more effective NAS methods and advancing the field of GANs. --- ### Implementation of reinforcement learning in chemical reaction networks: application to phototaxis as curiosity-driven exploration Authors: Ruyi Tang, Grégoire Sergeant-Perthuis, David Colliaux Published: 2026-06-24 Categories: cs.LG, q-bio.QM arXiv: https://arxiv.org/abs/2606.26168v1 Living systems navigate environments using noisy and incomplete sensory signals. In unicellular algae, phototaxis is often modeled as a mechanistic run--tumble process driven by stimulus--response rules. However, such descriptions overlook how organisms actively sample their environment to reduce sensory ambiguity. From a minimal cognition perspective, we reframe this navigation as a subjective, information-driven sensorimotor process. To this end, we propose a framework linking a Partially Observable Markov Decision Process (POMDP) with biochemical reaction dynamics. Environmental variables are hidden, while the cell updates a minimal internal state from each observation through a memoryless Bayesian step. These internal dynamics balance orienting toward light with exploratory reorientation and can be implemented through Chemical-Reaction-Network Ordinary Differential Equations (CRN--ODEs). Our model includes a biophysical observation process for photoreception and a chemically computable polynomial bound on information gain. Using Inverse Reinforcement Learning (IRL) on 30 experimentally recorded Chlamydomonas trajectories, we infer the behavioral objective consistent with observed phototactic motion and benchmark the resulting dynamics with standard Stochastic Simulation Algorithm (SSA) baselines. Our model reproduces the empirical alignment-to-light distribution, comparable to objective SSA baselines on this dataset. Within this framework, run--tumble alternation emerges as an information-acquisition strategy: tumbling reorients the cell to sample new sensory configurations and resolve sensor ambiguity, demonstrating how intracellular biochemical networks can support adaptive information-seeking behavior in cellular navigation. --- ### Predicting Fruit Quality with a Hybrid Machine Learning and Image Processing Approach Authors: Amir Reza Hashemi, Shahram Amiri Published: 2026-06-24 Categories: cs.CV arXiv: https://arxiv.org/abs/2606.26165v1 Fruit spoilage is a significant issue in agriculture, leading to substantial economic losses. Addressing this, our study introduces a hybrid approach combining image processing and deep learning to assess fruit freshness. We developed an image processing algorithm that quantifies spoilage on a scale from 0 (fully fresh) to 100 (fully rotten). Alongside, we trained a convolutional neural network (CNN) to perform binary classification (fresh or rotten) using a large dataset of fruit images. The outcomes of both methods were synthesized using logistic regression to enhance the accuracy of freshness predictions. Subsequently, this logistic regression model was utilized to enable the image processing algorithm to provide binary classification based on its percentage output, thus eliminating the need for the CNN in real-time applications. Our approach, which does not require high computational resources, achieved real-time performance and was validated with over 90% accuracy on a dataset comprising apples and oranges. The primary limitation lies in the requirement for fruits to be isolated on a background that must be either white or transparent, suggesting future improvements could include advanced segmentation models to automate background removal. This study's results highlight the potential of integrating simple image processing techniques with machine learning to provide practical solutions in the agricultural sector. --- ### \chisao{}: A GPU-Native Parallel Optimizer for Multimodal Black-Box Functions via Convergence-Anticonvergence Oscillation Authors: Ira Wolfson Published: 2026-06-24 Categories: cs.LG, math.NA, physics.data-an, stat.CO arXiv: https://arxiv.org/abs/2606.26164v1 Finding all modes of a multimodal black-box function is a fundamental challenge in optimization, Bayesian inference, and scientific computing. Existing approaches -- basin-hopping, CMA-ES, multistart gradient descent -- operate sequentially and cannot exploit the massive parallelism of modern GPU hardware. We introduce \chisao{} (\textbf{C}onvergence-\textbf{H}alt-\textbf{I}nvert-\textbf{S}tick-\textbf{A}nd-\textbf{O}scillate), a GPU-native population optimizer that runs an entire sample batch simultaneously and exploits a deliberate convergence-anticonvergence oscillation cycle to escape local traps while freezing confirmed modes. The structural move is asymmetric: samples that reach true peaks are frozen (``stuck'') and preserved, while the rest keep exploring via momentum-based anti-convergence and stochastically smoothed gradients. Adaptive reseeding via two complementary strategies (Repulse Monkey and Golden Rooster) maintains population diversity throughout. On all 42 functions of the Simon Fraser University optimization benchmark suite across dimensions $d \in \{2, 4, 8, 16, 32, 64\}$, \chisao{} achieves \textbf{100\%} mode recovery where all CPU baselines collapse at $d \geq 8$ on the hardest multimodal functions, at up to \textbf{$34\times$} speedup over basin-hopping on functions where all methods succeed (Michalewicz $d=64$) and up to \textbf{$39\times$} on unimodal functions (Rotated Hyper-Ellipsoid $d=64$, pure GPU dividend). All benchmarks evaluate the objective by value alone -- gradients come from finite differences -- so the reported speedups are a derivative-free worst case. Under substantial likelihood noise ($σ_{\mathrm{noise}}$ up to 1.0), mode detection remains 100\% reliable. The algorithm is available as a standalone open-source Python package on PyPI. --- ### Refusal Lives Downstream of Persona in Chat Models Authors: Viola Zhong, Qirui Li Published: 2026-06-24 Categories: cs.AI arXiv: https://arxiv.org/abs/2606.26161v1 Linear directions in activation space have been identified for both refusal and persona traits in instruction-tuned chat models, but the two have been studied as separate mechanisms. We show they interact: a compliant persona gates refusal. In Qwen2.5-7B-Instruct and Llama-3.1-8B-Instruct, we extract a compliant model-persona direction and a refusal direction and intervene on both. Compliant persona steering suppresses refusal -- in Llama, the refusal rate falls from 97% to 2%. Reintroducing the refusal direction partially restores refusal at late layers but not at early ones. Projecting out the persona direction in a late-layer window restores it to baseline; projecting out a random direction does not. Refusal is therefore gated at the late-layer expression stage, downstream of where it is computed. Treating refusal as a single isolated direction misses its dependence on persona. --- ### Learning Action Priors for Cross-embodiment Robot Manipulation Authors: Dong Jing, Tianqi Zhang, Jiaqi Liu, Jinman Zhao, Zelong Sun, Li Erran Li, Zhiwu Lu, Mingyu Ding Published: 2026-06-24 Categories: cs.RO, cs.AI, cs.CV arXiv: https://arxiv.org/abs/2606.26095v1 Most Vision-Language-Action (VLA) models build on a Vision-Language Model (VLM) backbone by attaching an action module and optimizing the full policy jointly. This design inherits strong visual and linguistic priors from the VLM, but leaves the action module to learn physical motion almost from scratch. As a result, the policy lacks an explicit motion prior, forcing early optimization to simultaneously discover temporal action dynamics and cross-modal alignment, a challenge further amplified in cross-embodiment settings. In this work, we propose to pretrain the action module with motion priors before cross-modal VLA alignment. Specifically, we introduce a two-stage training framework that equips the action module with cross-embodiment temporal motion structure before VLA training begins. In Stage~1, a lightweight flow-matching-based encoder-decoder action module efficiently learns temporal motion structure solely from unconditioned action trajectories, without processing visual or language tokens. In Stage~2, this learned prior is transferred to VLA training through decoder reuse and early-stage latent distillation, aligning visual-language features with the action embedding space while still allowing end-to-end policy refinement. In addition, the trained encoder serves as a compact history compressor, summarizing state-action histories into a single temporal context token for history-aware modeling at negligible cost. Extensive experiments across 13 diverse cross-embodiment tasks on both simulated and real-world platforms validate the effectiveness of our approach. Compared with VLA training without action priors, our model achieves faster convergence, higher success rates, and substantially stronger performance on data-scarce real-world tasks. Moreover, scaling up the action data in Stage~1 yields a more generalizable action prior that directly improves downstream VLA performance. --- ### RevengeBench: Reverse Engineering Code-Space Policies from Behavioral Experiments Authors: Babak Rahmani, Sebastian Dziadzio, Joschka Strüber, Sergio Hernández-Gutiérrez, Matthias Bethge Published: 2026-06-24 Categories: cs.LG arXiv: https://arxiv.org/abs/2606.26094v1 For most of scientific history, researchers studying behavior could only infer hidden mechanisms from outward actions: an inverse problem that becomes more tractable when observation is augmented by targeted intervention. We pose a computational analogue: given only behavioral traces of an agent in a game environment, can a learner reconstruct the underlying decision program as executable code, and how much does this reconstruction improve with the ability to design controlled experiments? We introduce RevengeBench, a benchmark of 75 LLM generated, Elo-calibrated policies across five game environments, drawn from CodeClash tournament trajectories. The learner observes the hidden target policy play against sampled opponents and designs behavioral probes in the form of custom opponent policies that elicit informative behavior. It then submits an executable hypothesis, which is evaluated using continuous action-distance metrics. We further validate that recovered code carries informative signal in downstream player-versus-player tournaments. Across twelve frontier LLMs, recovery quality varies substantially (34 to 72% of initial distance closed), with reconstructed policies yielding measurable competitive advantage, particularly for weaker models that otherwise struggle to design effective counter-strategies. Our benchmark positions behavioral recovery of programmatic policies as a tractable inverse problem in code-space, opening a path to opponent modeling, policy interpretability, and the broader question of inferring latent mechanisms from observations. --- ### TryOnCrafter: Unleashing Camera Trajectories for Realistic Video Virtual Try-on via a Renderable 4D Try-on Proxy Authors: Hao Sun, Hao Yan, Mengting Chen, Quanjian Song, Yu Li, Juan Cao, Jinsong Lan, Xiaoyong Zhu, Bo Zheng, Sheng Tang Published: 2026-06-24 Categories: cs.CV arXiv: https://arxiv.org/abs/2606.26092v2 While Video Virtual Try-on (VVT) has achieved remarkable progress in synthesizing realistic garment overlays on dynamic subjects, existing paradigms remains fundamentally constrained by a passive dependency on source camera trajectories, failing to accommodate the requisite interactive freedom for omnidirectional viewpoint exploration. To address this limitation, we define a pioneering research frontier: Camera-controllable Video Virtual Try-on (CaM-VVT). Unlike conventional VVT, CaM-VVT not only necessitates viewpoint-agnostic texture hallucination but also strict structural synchronization between non-rigid human dynamics and background contexts under arbitrary, unconstrained camera movements. To tackle these challenges, we present TryOnCrafter, the first unified DiT-based framework specifically architected for the CaM-VVT task. Departing from implicit pixel-space manipulation, we introduce a Renderable 4D Try-on Proxy that explicitly decouples the human subject from the environment. This is achieved by distilling high-fidelity 2D try-on priors into a clothed 3DGS-based avatar, which is subsequently animated via SMPL-X sequences and metric-aligned into a reconstructed background point cloud. This proxy establishes a robust structural foundation with superior texture density and motion integrity. Our Proxy-Anchored Video DiT leverages this robust structural foundation as a primary geometric anchor, ensuring that the synthesized photorealistic videos are strictly constrained by prescribed trajectories and physically plausible deformations. Benefiting from the inherent editability of the 4D proxy, TryOnCrafter facilitates diverse downstream applications, including human relocalization, ``bullet time'' effects, and $360$-degree orbital viewing. --- ### On-Policy Self-Distillation with Sampled Demonstrations Reduces Output Diversity Authors: Andrei Liviu Nicolicioiu, Mohammad Pezeshki, Aaron Courville Published: 2026-06-24 Categories: cs.LG, cs.AI arXiv: https://arxiv.org/abs/2606.26091v1 On-policy self-distillation achieves strong pass@1 accuracy by using a single model as both teacher and student, with the teacher conditioned on a correct demonstration to provide dense token-level feedback. We show that this could come at a hidden cost: rollout diversity decreases and pass@k curves flatten (i.e., generating more rollouts fails to improve accuracy). We trace this to compounding biases in the design of self-distillation with sampled demonstrations. The teacher scores each student rollout while conditioned on a sampled correct rollout, channeling its feedback through the model's own biases. We theoretically analyze the optimal self-distillation policy and show that it tilts the base distribution by a pointwise conditional mutual information score between the student's rollout and the correct rollout used as context. Unlike the ideal optimal on-policy reinforcement learning (RL), which preserves probability ratios among equally correct rollouts, self-distillation can amplify existing probability gaps, concentrating mass on already-dominant modes. On a controlled graph path-finding task and science question-answering benchmarks, self-distilled models match or exceed RL on average performance but exhibit substantially lower functional and semantic diversity, failing on out-of-distribution settings that require diverse strategies. --- ### MVTrack4Gen: Multi-View Point Tracking as Geometric Supervision for 4D Video Generation Authors: JoungBin Lee, Jaewoo Jung, Jongmin Lee, Tongmin Kim, Hyunsung Kim, Takuya Narihira, Kazumi Fukuda, Jahyeok Koo, Jisang Han, Yuki Mitsufuji, Seungryong Kim Published: 2026-06-24 Categories: cs.CV arXiv: https://arxiv.org/abs/2606.26087v1 Synthesizing a novel-view video from a monocular reference video along a target camera trajectory requires both geometric consistency and motion fidelity with respect to the reference video. Existing methods based on explicit 3D representations are limited by the accuracy of off-the-shelf reconstruction modules, which often produce inaccurate geometry for dynamic objects in monocular videos. In contrast, camera-conditioning-only methods can achieve high visual quality but often struggle to preserve geometric and motion consistency. In this work, we introduce MVTrack4Gen (Multi-View point Tracking for Novel-View Generation), a motion-aware training framework that leverages multi-view point tracking as an additional geometric and motion supervision signal for camera-conditioning-only novel-view video diffusion models. Our key finding is that specific attention layers encode strong correspondence cues, where query features attend to key features at geometrically corresponding locations across views and over time, and the misalignment of these correspondences causes motion inconsistency. Based on this observation, we route these features into an auxiliary multi-view tracking head and jointly train the diffusion model with a point-tracking objective. By explicitly strengthening these motion-aware correspondences, MVTrack4Gen improves existing models to better follow the motion in the reference view and maintain cross-view geometric consistency. Across diverse benchmarks, our method achieves state-of-the-art geometric consistency and competitive camera accuracy. --- ### Real-Time Voice AI Hears but Does Not Listen Authors: Martijn Bartelds, Federico Bianchi, James Zou Published: 2026-06-24 Categories: cs.CL, eess.AS arXiv: https://arxiv.org/abs/2606.26083v1 Speech conveys information through both words and vocal delivery. We evaluate four leading production realtime voice systems-OpenAI's GPT Realtime 2, Google's Gemini 3.1 Flash Live, and Alibaba's Qwen3.5 Omni Plus and Omni Flash-on tasks where the words and the delivery patterns both convey meaningful information. Across three consequential scenarios, all four systems act on the words rather than the voice. They end calls with crying callers who insist nothing is wrong, approve wire transfers authorized in frightened voices, and enroll callers whose agreement is clearly sarcastic. Surprisingly, this is often not a failure of perception. When asked directly, three of the four systems reliably identify the distress, fear, or sarcasm they later ignore when making decisions. We observe a similar pattern when these realtime voice systems estimate accent and age, as their responses frequently follow the biases of the words rather than the acoustic properties of the speaker. We term this disconnect between perception and action the emotional intelligence gap of voice AI. Prompting systems to explicitly attend to vocal delivery improves performance only partially and inconsistently. Our findings show that current realtime voice AI systems often behave as if speech had been reduced to a transcript, suggesting that they should be used with caution in settings where the tone and emotion of delivery convey important information. --- ### Neglected Free Lunch from Post-training: Progress Advantage for LLM Agents Authors: Changdae Oh, Wendi Li, Seongheon Park, Samuel Yeh, Tanwi Mallick, Sharon Li Published: 2026-06-24 Categories: cs.LG, cs.AI arXiv: https://arxiv.org/abs/2606.26080v1 Process reward models enable fine-grained, step-level evaluation of LLMs, yet building them for agentic settings remains prohibitively difficult: long-horizon interactions, irreversible actions, and stochastic environment feedback make both human annotation and Monte Carlo estimation infeasible at scale. In this work, we show that reinforcement learning (RL) post-training already provides the ingredients for effective step-level scoring, eliminating the need for dedicated reward model training altogether. Concretely, we derive an implicit advantage under a general stochastic Markov decision process, which we term progress advantage -- log-probability ratio between the RL-trained policy and its reference policy exactly recovers the optimal advantage function. This formulation makes the resulting signal annotation-free, domain-agnostic, and available as a byproduct of the standard RL post-training pipeline. We validate the effectiveness of the progress advantage across three different applications: test-time scaling, uncertainty quantification, and failure attribution on five benchmarks and four model families. Across all settings, it consistently outperforms confidence-based baselines and, despite requiring no task-specific training, surpasses dedicated trained reward models. We complement these results with deeper analyses on characteristics of progress advantage, offering practical guidance for adoption in real-world agentic systems. --- ### Same Evidence, Different Answer: Auditing Order Sensitivity in Multimodal Large Language Models Authors: Akshay Paruchuri, Sanmi Koyejo, Ehsan Adeli Published: 2026-06-24 Categories: cs.CL, cs.CV, cs.LG arXiv: https://arxiv.org/abs/2606.26079v1 Standard benchmarks for multimodal large language models (MLLMs) score each item on one canonical ordering and miss whether order-irrelevant shuffling changes the answer, a baseline reliability property called for by emerging AI evaluation guidelines. We introduce Facet-Probe, a five-facet audit (option, evidence-chunk, document-rank, image-set, and mixed-modality ordering) of 18 frontier and open-weight MLLMs. A Bayesian item-response model separates ordering noise from per-facet bias, and a same-ordering control estimates the decoder-stochastic floor for observed flips. We find that none of the 18 MLLMs we audit are order-invariant: screened per-facet panel-mean flip rates span 24-50%. A Gemini same-ordering control at temperature 0 estimates a substantial ordering excess over a same-input decoder-noise floor in verified cells. Capability predicts but does not eliminate flips; the best model still flips on 13.4% of trials. In our Gemini mitigation tests, training-free prompt changes are modality-conditional and do not transfer from text to visual reasoning. These results suggest that prompt-level mitigation alone is unlikely to provide general order robustness, motivating future work on training-time and architectural approaches. We propose cross-ordering flip rate as a standard reporting axis for MLLMs. --- ### A cross-process welding penetration status prediction algorithm based on unsupervised domain adaptation in laser and TIG welding Authors: Sen Li, Haichao Cui, Chendong Shao, Yaqi Wang, Xinhua Tang Published: 2026-06-24 Categories: cs.CV, cs.AI arXiv: https://arxiv.org/abs/2606.26078v1 Supervised deep learning has been widely used for weld penetration state classification; however, its performance often degrades significantly under domain shift, such as when transferring models between welding processes with distinct physical mechanisms:for instance, from arc-dominated tungsten inert gas (TIG) welding to keyhole-based laser welding. To overcome this limitation, we propose an unsupervised domain adaptation (UDA) framework integrated with a gradual source domain expansion (GSDE) strategy. Evaluated on dedicated TIG and laser welding datasets, our approach achieves high accuracy in both same-process and cross-process transfer tasks. Specifically, it attains average accuracies of 90.65% on TIGFH and 90.72% on LSPS in same-process settings, surpassing a supervised baseline by 35.83% and 38.87%, respectively. More notably, in cross-process scenarios, it reaches 80.48% for TIG to Laser and 81.13% for Laser to TIG, improving upon the baseline by 43.39% and 43.40%. UMAP visualizations verify that the model learns domain-invariant features while maintaining discriminative class boundaries. This method considerably lowers the relabeling cost for new welding processes and enhances the versatility of intelligent monitoring across different welding systems. --- ### Model Forensics: Investigating Whether Concerning Behavior Reflects Misalignment Authors: Aditya Singh, Gerson Kroiz, Senthooran Rajamanoharan, Neel Nanda Published: 2026-06-24 Categories: cs.LG, cs.AI arXiv: https://arxiv.org/abs/2606.26071v2 A central goal of safety research is determining whether a model is misaligned. Prior work has largely focused on detecting concerning behavior. But behavior alone does not establish misalignment: a concerning action can arise from benign causes such as confusion. This motivates model forensics: investigating whether the action was driven by malign intent. In this paper, we propose a baseline protocol for model forensics consisting of two steps, iterated as needed. First, we read the chain of thought (CoT) to generate hypotheses about what drives model behavior. Second, we make edits to the prompt or environment to test these hypotheses. While the CoT is not always faithful, it is a rich source of unsupervised insight that can guide the collection of more rigorous evidence. To evaluate our protocol, we create a suite of six agentic environments where models exhibit concerning behavior, and apply it to each. We establish that Kimi K2 Thinking takes shortcuts due to a genuine disposition towards low-effort actions, by showing this hypothesis successfully predicts its behavior. Through counterfactual experiments, we show DeepSeek R1 deceives out of a desire to be consistent with a previous instance of itself. Our methods nonetheless leave significant room for refinement. For example, when we test whether Kimi K2 Thinking believes it is violating user intent, we find no evidence of such a belief, but without positive controls we cannot confirm our tests would detect it. Overall, we find our simple protocol provides a strong baseline that we hope future work will improve upon. More broadly, our work is a concrete step in developing the growing field of model forensics. --- ### When Certainty Is an Artifact: Keyword Lexicon Blindness and the (Mis)Measurement of Rhetorical Stance Authors: Bo Chen Published: 2026-06-24 Categories: cs.CL arXiv: https://arxiv.org/abs/2606.26062v1 Can a statistically significant, large-effect-size finding in computational social science be entirely an artifact of the measurement instrument? We present a case where the answer appears to be yes. Analyzing 85 interviews across four public intellectuals (2016--2026), we find a robust negative-affect/emphatic-certainty lexical co-occurrence pattern under keyword-based scoring ($r = 0.72$--$0.93$, $p < 0.01$ for all four speakers). Replacing keyword counting with LLM-based zero-shot semantic classification on the complete diarized corpus (32,625 sentences) dramatically reduces this correlation: Dalio's $r = 0.851$ drops to $r = 0.206$, with two speakers showing negative $r(\text{neg}, \text{emphatic})$ and one showing null. In contrast, the LLM reveals a strong negative-hedging coupling across speakers -- Rogoff's $r(\text{neg}, \text{hedged}) = 0.875$ ($p = 0.001$) and Zeihan's $r(\text{neg}, \text{hedged}) = 0.722$ ($p = 0.008$) -- consistent with the conventional expectation that pessimistic discourse attracts hedging, not certainty. Sentence-level error analysis traces this discrepancy to three structural failure modes in keyword lexicons -- syntactic blindness, polysemy blindness, and categorical absence -- illustrated through cases where keyword counting inverts semantic meaning (e.g., ''never absolutely totally confident'' scored as high-certainty). We argue that keyword lexicons measure a universal lexical co-occurrence tendency -- negative discourse naturally attracts emphatic vocabulary -- that is orthogonal to, and can systematically invert, rhetorical stance. Treating keyword counts as measurements of epistemic certainty is a category error: a finding that appears to be about a speaker's psychology may be entirely about the counting of words. --- ### Deviance-style normalization for jointly overdispersed counts Authors: Akshay Balsubramani Published: 2026-06-24 Categories: stat.ME, math.ST, q-bio.QM, stat.ML arXiv: https://arxiv.org/abs/2606.26061v1 We introduce a Dirichlet--multinomial (DM) deviance residualization for sparse, jointly overdispersed count matrices, the regime that dominates sequencing-based biochemical assays. The DM null treats each sample's count vector as a fixed-total composition with a single scalar concentration $α_0$ governing overdispersion, and arises exactly by conditioning independent negative-binomial feature counts on the observed sample total -- making the DM the joint conditional analogue of standard feature-wise overdispersed count models. The resulting transform preserves exact sparsity, evaluates in constant time per nonzero entry, agrees with multinomial residuals on singleton counts, shrinks repeated-count residuals according to the overdispersion the null tolerates, and recovers the multinomial residual as $α_0\to\infty$. The same fixed-dispersion comparison principle extends to ordered and tree-structured features via the generalized DM and the Dirichlet-tree multinomial, giving a single residual family that subsumes joint and feature-wise count nulls under a common compositional logic and is computationally lightweight enough to drop into existing sparse pipelines. --- ### A welding penetration prediction model for laser welding process based on self-supervised learning using physics-informed neural networks Authors: Sen Li, Xiaoying Liu, Xiaojian Xu, Chendong Shao, Yaqi Wang, Ling Lan, Xinhua Tang, Haichao Cui Published: 2026-06-24 Categories: cs.CV, cs.AI arXiv: https://arxiv.org/abs/2606.26059v1 The laser welding full-penetration is of critical importance, as it constitutes one of the fundamental factors in achieving defect-free welded joints. Accurate prediction of the penetration state is therefore essential for ensuring weld quality. To this end, this paper introduces SimPhysNet, a novel algorithm that achieves high classification accuracy in laser welding penetration prediction using only a limited number of labelled images. This approach effectively overcomes the limitations of supervised learning classification algorithms, which are hindered in industrial applications by their dependence on extensive, high-quality labelled data. The core of SimPhysNet is a unique self-supervised learning paradigm that embeds physical priors into a contrastive learning framework. By incorporating a physics-informed neural network (PINN), the model is guided to extract physically meaningful features of the molten pool and keyhole from a large set of unlabelled data, while three image augmentation tasks further enhance its generalization capabilities. Subsequently, a few-shot learning strategy, based on prototypical networks, enables robust classification by constructing class representations from a minimal set of labelled images. Experimental results demonstrate that SimPhysNet achieves a classification accuracy of 96.06% using only 200 labelled images (approximately 5% of the total labelled dataset), which is comparable to the performance of conventional supervised learning algorithms that utilize the entire labelled dataset. This work presents a new, efficient, and highly accurate method, providing the way for the intelligent automation of laser welding. --- ### DomainShuttle: Freeform Open Domain Subject-driven Text-to-video Generation Authors: Nan Chen, Yiyang Cai, Rongchang Xie, Junwen Pan, Cheng Chen, Weinan Jia, Zhuowei Chen, Wen Zhou, Zhenbang Sun, Wenhan Luo Published: 2026-06-24 Categories: cs.CV arXiv: https://arxiv.org/abs/2606.26058v1 Open domain subject-driven text-to-video (S2V) generation has drawn significant interest in academia and industry. Open domain S2V mainly involves two scenarios: in-domain, which requires retaining the reference subject features as much as possible, and cross-domain, which preserves the intrinsic features of the subject while allowing subject-irrelevant properties to vary flexibly according to the text prompt. Existing methods primarily focus on maximizing subject fidelity in in-domain scenarios, which limits their editability and adaptability in cross-domain scenarios, such as novel styles, semantic combinations, or domain attributes. In this study, we propose that an ideal S2V method should flexibly shuttle between different domains, achieving strong performance in both in-domain and cross-domain scenarios. To this end, we propose DomainShuttle, which could achieve high fidelity and generative flexibility for open domain video personalization. Specifically, we introduce Domain-MoT, which decouples videos and reference features and introduces the domain-aware AdaLN for domain-specific modeling of reference images. We then introduce the Video-Reference DualRoPE scheme, which places reference image tokens and video tokens in separate RoPE spaces to enable precise subject-level spatial modeling, and Cross-Pair Consistent Loss, which aims to extract intrinsic subject features unaffected by irrelevant features. Extensive experiments demonstrate that DomainShuttle achieves significant performance improvements over existing methods, exhibiting high subject fidelity and generative flexibility across diverse open domain application scenarios. --- ### The Unfireable Safety Kernel: Execution-Time AI Alignment for AI Agents and Other Escapable AI Systems Authors: Seth Dobrin, Łukasz Chmiel Published: 2026-06-24 Categories: cs.AI, cs.CR, cs.LG arXiv: https://arxiv.org/abs/2606.26057v1 AI agents are granted access to tools, APIs, and other infrastructure, making them active principals in those systems. The dominant approach places controls inside the agent's own runtime: system prompts, output filters, and guardrail libraries. Any control in the agent's address space is reachable by inputs that influence it; this generalizes to any AI system with sufficient reach into its own runtime, a class we term escapable AI systems. We identify four properties that an authorization mechanism must satisfy for architectural control rather than for cooperative requests: process separation, pre-action enforcement on a structurally only path, fail-closed at both the request and system levels, and externalized signed evidence verifiable outside the controlled system's trust boundary. We position this layer as execution-time AI alignment, complementing training-time alignment (RLHF, Constitutional AI) and inference-time alignment. We present the Unfireable Safety Kernel, a Rust reference implementation realizing all four. Its fail-closed invariant is machine-checked at two levels: an SMT theorem (Z3) and an exhaustive bounded-model-checking proof of the production decision function (Kani, 4/4 harnesses). A Python-to-Rust migration was gated on byte-equivalence (1000/1000 fixtures; 17/17 adversarial classes). We evaluate the kernel governing a live, escapable AI system, a deterministic, self-improving world model, against an escape-seeking adversary driving its real self-modification seam: across 1,000 self-modifications, all 704 attempts on the safety-critical core are refused, with no escape; a further 300, under the operator kill switch, are also refused. A separate campaign of 6,240 authorization round-trips had no successful bypass. Against 3 contemporary systems claiming the agent control plane, the agent invokes control; here, it lacks that choice. --- ### When Does Synthetic Data Augmentation Improve Score-Based Imbalanced Classification? Authors: Zhengchi Ma, Pengfei Lyu, Anru R. Zhang Published: 2026-06-24 Categories: stat.ML, cs.LG arXiv: https://arxiv.org/abs/2606.26053v1 Synthetic data augmentation is widely used to mitigate class imbalance, but its theoretical effects on score-based classification remain poorly understood. This paper develops a framework for characterizing when synthetic minority augmentation can improve threshold-integrated and threshold-optimized metrics, including AUROC, AUPRC, best-threshold balanced accuracy, and best-threshold \(\F_1\) score. We separate the effect of augmentation into two components: a change in effective class weighting and a discrepancy between the synthetic and true minority distributions. Under well-specified score models, the raw estimator already targets the likelihood-ratio ordering, which is population-optimal for the metrics considered. Consequently, augmentation cannot provide a fundamental population-level improvement beyond possible finite-sample variance reduction, and may introduce additional bias through synthetic distributional error. We further establish minimax lower bounds showing that the raw estimator already achieves the optimal metric-regret rate in the well-specified regime. Under misspecification, however, augmentation can play a qualitatively different role: by changing the effective class balance, it can alter the restricted-class projection and correct ranking errors induced by the raw imbalanced objective. We provide explicit improvement bounds quantifying the roles of approximation error, finite-sample estimation error, and synthetic distributional error. Simulation studies corroborate the theory, demonstrating limited gains under well-specification and nontrivial but nonmonotone improvements under misspecification. --- ### Natural Ungrokking: Asymmetric Control of Which Rules Survive Pretraining Authors: Juliana Li, Diya Sreedhar Published: 2026-06-24 Categories: cs.LG, cond-mat.dis-nn, cs.AI, cs.CL arXiv: https://arxiv.org/abs/2606.26050v1 Midway through an ordinary pretraining run, a small language model learns the pronoun-gender rule: cued with a girl's name ("Sue cried because"), it resolves the next pronoun to she, generalizing to held-out probes (0.94 by step 925). By step 3,500 the same model scores near zero on the same probes, although the rule's evidence is still in the training data. We call this within-run reversal natural ungrokking: the corpus decides, with no trace in the loss curve, which learned rules a model keeps. Which rules survive is predictable from one corpus statistic: how often the training stream shows the rule winning. Across un-intervened runs (two corpora, three budgets, three seeds), support frequency decides a rule's fate; the data-to-parameter ratio only modulates how deeply a doomed rule falls. The same emerge-then-collapse dynamics appear in public Pythia checkpoints, collapse depth ordered by model scale as predicted. The forgetting is a displacement: a competing surface pattern out-competes the rule, and the log-probability margin between them crosses zero within 100 training steps of the behavioral collapse. Control over this fate is asymmetric: the same edit that destroys a rule on demand cannot restore it. Flipping support to counter-evidence in place kills the rule with monotone dose-response in two unrelated rules; but injecting support back, even to 450 times the level that naturally sustains it, buys no recovery. Every confirmatory threshold and prediction was pre-registered before the data it governed was read. --- ### RoboAtlas: Contextual Active SLAM Authors: Alexander Schperberg, Shivam K. Panda, Abraham P. Vinod, Rokaha Bhagawan, M. K. Jawed, Stefano Di Cairano Published: 2026-06-24 Categories: cs.RO, cs.CV arXiv: https://arxiv.org/abs/2606.26046v2 We present RoboAtlas, a contextual Active SLAM framework that adaptively balances geometric exploration and semantic reasoning using a scalable 3D semantic mapping system, OpenRoboVox. RoboAtlas integrates frontier exploration, global semantic-map reasoning, and egocentric VLM-based reasoning through a contextual multi-armed bandit that transitions from exploration to semantically guided navigation as scene understanding improves. We evaluate the system in simulation and on a Unitree Go2 robot in large-scale real-world environments exceeding 1800 m2 with approx. 30k mapped semantic instances, achieving a 100% task success rate. On the GOAT-Bench "Val Unseen" benchmark, RoboAtlas achieves state-of-the-art performance with highest reported success rate (SR) of 90.6%, using GPT-4o, improving over the strongest prior baseline by 17.8 percentage points in SR. Using the much smaller Qwen2.5-VL-7B model, it still achieves 88.8% SR, outperforming all baselines using GPT-4o in SR, and revealing the importance of the information gained by our semantic mapping framework over simply replacing the underlying foundation model. The results demonstrate that grounding foundation models with large-scale 3D semantic maps enables robust and efficient contextual Active SLAM. --- ### RoboAtlas: Contextual Active SLAM Authors: Alexander Schperberg, Shivam K. Panda, Abraham P. Vinod, M. K. Jawed, Stefano Di Cairano Published: 2026-06-24 Categories: cs.RO, cs.CV arXiv: https://arxiv.org/abs/2606.26046v1 We present RoboAtlas, a contextual Active SLAM framework that adaptively balances geometric exploration and semantic reasoning using a scalable 3D semantic mapping system, OpenRoboVox. RoboAtlas integrates frontier exploration, global semantic-map reasoning, and egocentric VLM-based reasoning through a contextual multi-armed bandit that transitions from exploration to semantically guided navigation as scene understanding improves. We evaluate the system in simulation and on a Unitree Go2 robot in large-scale real-world environments exceeding 1800 m2 with approx. 30k mapped semantic instances, achieving a 100% task success rate. On the GOAT-Bench "Val Unseen" benchmark, RoboAtlas achieves state-of-the-art performance with highest reported success rate (SR) of 90.6%, using GPT-4o, improving over the strongest prior baseline by 17.8 percentage points in SR. Using the much smaller Qwen2.5-VL-7B model, it still achieves 88.8% SR, outperforming all baselines using GPT-4o in SR, and revealing the importance of the information gained by our semantic mapping framework over simply replacing the underlying foundation model. The results demonstrate that grounding foundation models with large-scale 3D semantic maps enables robust and efficient contextual Active SLAM. --- ### How Robust is OCR-Reasoning? Evaluating OCR-Reasoning Robustness of Vision-Language Models under Visual Perturbations Authors: Yuxing Cheng, Yuan Wu, Yi Chang Published: 2026-06-24 Categories: cs.CV, cs.CL arXiv: https://arxiv.org/abs/2606.26041v1 Vision-language models (VLMs) have achieved strong performance on OCR-based benchmarks and increasingly focused on text-rich understanding, but their robustness under controlled visual degradation remains insufficiently understood. This gap is critical for OCR reasoning, where visual corruption can induce OCR errors and structural distortions, thereby introducing uncertainty into the reasoning task. To systematically study this problem, we introduce OCR-Robust, a benchmark designed for evaluating OCR reasoning robustness under visual perturbations. It contains 812 samples across two complementary subsets: OCR1.0, covering documents, scene text, receipts, handwriting, and mathematical content, and OCR2.0, focusing on charts, geometry diagrams, and tables. To enable efficient yet informative evaluation, we conduct a pilot study over 18 candidate perturbations and select 5 representative types at 3 severity levels each based on their impact and cross-model discriminability. We evaluate robustness using clean accuracy, Relative Corruption Retention (RCR), Worst-Case Retention (WCR), and a composite Corruption Robustness Index (CRI), and benchmark 18 models spanning proprietary systems, open-source VLMs, and OCR+LLM pipelines. Our results show that higher clean accuracy does not necessarily imply stronger robustness, and that models can suffer pronounced degradation in the worst case on OCR tasks that are sensitive to structure, and charts and tables are substantially more fragile than document-like inputs under perturbation. --- ### AI translation of literary texts is "fine", but readers still prefer human translations Authors: Yves Ferstler, Adam Podoxin, Ty Brassington, Gaëlle Laperrière, Roman Grundkiewicz, Marie-Jean Meurs, Maite Taboada, Marzena Karpinska Published: 2026-06-24 Categories: cs.CL arXiv: https://arxiv.org/abs/2606.26040v2 AI translation of literary works is increasingly common. While the content may be rendered adequately, we do not know enough about how readers experience it in terms of immersiveness and literary effect-aspects poorly captured by automatic metrics or human evaluation targeting fluency and adequacy. We ask 15 avid readers to compare recently published human translations (HT) to machine translations (MT) generated with an agentic language model-based pipeline, for 15 recent novels in French, Polish, and Japanese translated into English. Readers evaluated approximately 8K-word excerpts in two conditions: immersive reading of the whole excerpt (30 comparisons) and close reading of 386 aligned HT-MT chunk pairs (772 comparisons), with two readers per book. Overall, readers find MT "fine", but prefer HT (descriptively at the excerpt level, 19/30, and significantly at the chunk level, 522/772) for its ease, clarity, and immersive nature. Readers' highlights show that MT's quality varies more within one book than HT's does. Crucially, readers cannot reliably tell the two apart (17/30 guess correctly) and tend to prefer the version they believe to be human. Automatic metrics, including LLM-as-a-judge approaches, fail to recover reader preferences and favor MT. We release LAIT (Literary AI Translation), a reader-centered evaluation dataset with 1K reader comments, 2K judgments and preference ratings, and 7.2K span-level annotations, along with our evaluation protocol and supporting interface. --- ### AI translation of literary texts is "fine", but readers still prefer human translations Authors: Yves Ferstler, Adam Podoxin, Ty Brassington, Roman Grundkiewicz, Maite Taboada, Marzena Karpinska Published: 2026-06-24 Categories: cs.CL arXiv: https://arxiv.org/abs/2606.26040v1 AI translation of literary works is increasingly common. While the content may be rendered adequately, we do not know enough about how readers experience it in terms of immersiveness and literary effect, aspects poorly captured by automatic machine translation metrics or human evaluation targeting fluency and adequacy. We ask 15 avid readers to compare recently published human translations (HT) to machine translations (MT) generated with an agentic large language model (LLM)-based pipeline, for 15 recent novels in French, Polish, and Japanese and translated into English. Readers evaluated approximately 8K-word excerpts in two conditions: immersive reading of the whole excerpt (30 comparisons) and close reading of 386 aligned HT-MT chunk pairs (772 comparisons), with two readers per book and in alternating order of presentation. Overall, readers find MT "fine", but prefer HT (slightly at excerpt-level 19/30, more clearly at chunk-level 522/772) for its ease, clarity, and immersive nature. Readers' highlights show that MT's quality varies more within one book than HT's does. Crucially, readers cannot reliably tell the two apart (17/30 guess correctly) and tend to prefer the version they believe to be human. Automatic metrics, including LLM-as-a-judge approaches, fail to recover reader preferences and favor MT. We release LAIT (Literary AI Translation), a reader-centered evaluation dataset with 1K reader comments, 2K judgments and preference ratings, and 7.2K span-level annotations, along with our evaluation protocol and supporting interface. --- ### FedReLa: Imbalanced Federated Learning via Re-Labeling Authors: Guangzheng Hu, Patricia Menéndez, Feng Liu, Mingming Gong, Guanghui Wang, Liuhua Peng Published: 2026-06-24 Categories: stat.ML, cs.CV, cs.LG arXiv: https://arxiv.org/abs/2606.26037v1 Federated learning has emerged as the foremost approach for decentralized model training with privacy preservation. The global class imbalance and cross-client data heterogeneity naturally coexist, and the mismatch between local and global imbalances exacerbates the performance degradation of the aggregated model. The agnosticism of global class distribution poses significant challenges for data-level methods, especially under extreme conditions with severe class absence across clients. In this paper, we propose FedReLa, a novel data-level approach that tackles the coexistence of data heterogeneity and class imbalance in federated learning. By re-labeling samples with a feature-dependent label re-allocator, FedReLa corrects biased global decision boundaries without requiring knowledge of the global class distribution. This modular, model-agnostic approach can be integrated with algorithmic methods to deliver consistent improvements without additional communication overhead. Through extensive experiments, our method significantly improves the accuracy of minority classes and the overall accuracy on stepwise-imbalanced and long-tailed datasets, outperforming the previous state of the art. --- ### Detect, Unlearn, Restore: Defending Text Summarization Models Against Data Poisoning Authors: Poojitha Thota, Shirin Nilizadeh Published: 2026-06-24 Categories: cs.CL, cs.CR arXiv: https://arxiv.org/abs/2606.26036v1 Training-time data poisoning during fine-tuning poses a significant threat to large language models (LLMs) deployed for abstractive text summarization, where small task-specific datasets exert disproportionate influence on model behavior. In this setting, adversaries manipulate fine-tuning data to induce persistent summarization failures, such as biased or harmful summaries, while preserving standard evaluation metrics. We present a unified post-hoc defense framework for detecting and remediating fine-tuning-stage poisoning in summarization models across the machine learning supply chain. Our experiments show that in white-box settings, poisoned document-summary pairs exhibit abnormally high training influence, enabling detection via influence-function analysis with semantic consistency checks. In black-box settings, poisoned models display two to three times greater sensitivity to semantics-preserving perturbations, enabling behavioral auditing without training data access. Beyond existing poisoning formulations, we introduce novel attacks targeting factual distortion and representational bias, showing that poisoning alters summarization behavior without triggering conventional alarms. Across nine architectures and six benchmark datasets under adaptive attacks, our defenses achieve 85-92% detection precision, while gradient-ascent unlearning restores up to 96% of original behavior with minimal utility loss (less than 0.6% ROUGE degradation). These results indicate that fine-tuning-time poisoning leaves persistent structural artifacts, enabling practical detection and post-deployment recovery without full retraining. --- ### TriViewBench: Controlled Complexity Scaling for Multi-View Structural Reasoning in MLLMs Authors: Yu-Yang Chen, Lan-Zhe Guo Published: 2026-06-24 Categories: cs.CV, cs.AI arXiv: https://arxiv.org/abs/2606.26029v1 Multimodal Large Language Models (MLLMs) demonstrate strong performance on standard visual question answering benchmarks, yet their scalability under controlled structural complexity remains poorly understood. We introduce TriViewBench, a controlled three-view visual reasoning benchmark constructed from synthetic 3D scenes with explicitly parameterized object count and occlusion. The benchmark contains 1,923 scenes and over 14K Question-Answer (QA) pairs organized into four complexity levels and three reasoning categories: Local Decision, Object Counting, and Global Recovery. We evaluate 18 open- and closed-source MLLMs under a unified prompting protocol. All 18 models exhibit an identical capability hierarchy without exception (Local Decision > Object Counting > Global Recovery), and performance degrades monotonically with complexity: Local Decision tasks decline modestly (12.11% relative drop), while Object Counting degrades substantially (59.14%) and Global Recovery collapses severely (80.02%). Error analysis on Object Counting reveals two mechanistically independent failure modes: single-view tasks are dominated by undercounting due to occlusion blindness, whereas the multi-view task reverses to overcounting due to cross-view identity confusion. Chain-of-Thought (CoT) prompting yields near-zero overall benefit ($Δ= -0.16\%$) and its effect on Global Recovery is strongly capability-gated, suggesting that the bottleneck lies in cross-view spatial representation rather than reasoning strategy. These findings reveal fundamental scalability limitations in current MLLMs and position TriViewBench as a controlled diagnostic framework for analyzing structural reasoning failures. --- ### Can Trustless Agents Be Trusted? An Empirical Study of the ERC-8004 Decentralized AI Agent Ecosystem Authors: Xihan Xiong, Zelin Li, Wei Wei, Qin Wang, William Knottenbelt, Zhipeng Wang Published: 2026-06-24 Categories: cs.CR, cs.AI, cs.MA arXiv: https://arxiv.org/abs/2606.26028v2 As autonomous AI agents increasingly transact across organizational boundaries, a fundamental trust challenge emerges: how can an agent assess whether an unknown counterpart is trustworthy? The ERC-8004 protocol addresses this challenge with the first permissionless trust layer for AI agent economies, built around three on-chain registries for Identity, Reputation, and Validation. Despite its rapid adoption, the protocol has not been studied empirically, leaving it unclear whether the information it records provides a trustworthy basis for decision-making. To address this gap, we present the first empirical study of ERC-8004 across three chains: Ethereum, BNB Smart Chain (BSC), and Base, covering the period from protocol deployment through May 13, 2026. We crawl on-chain Identity and Reputation events, off-chain files, and x402 payment transactions. On the identity side, we find that most registrations are placeholders rather than active agents, with only a small fraction (3%, 4%, and 15% across Ethereum, BSC, and Base) exposing a valid ERC-8004 registration file with at least one live service endpoint. On the reputation side, we show that the Registry, as currently deployed, cannot function as a trust signal: values are not commensurable, feedback records are rarely grounded in verifiable interactions, and reputation can be manipulated at minimal cost. Consistent with these design weaknesses, we find that a substantial fraction of reviewers (73.5%, 59.2%, and 90.6% across Ethereum, BSC, and Base) exhibit coordinated Sybil behavior. After removing Sybil-flagged feedback, 15.8%, 77.9%, and 86.8% of rated agents, respectively, are left with no valid feedback. We then turn these findings into concrete recommendations for future revisions of ERC-8004. Our study yields actionable protocol-design implications and establishes an empirical baseline for research on AI agent markets. --- ### Why Multi-Step Tool-Use Reinforcement Learning Collapses and How Supervisory Signals Fix It Authors: Yupu Hao, Zhuoran Jin, Huanxuan Liao, Kang Liu, Jun Zhao Published: 2026-06-24 Categories: cs.CL, cs.LG arXiv: https://arxiv.org/abs/2606.26027v1 Tool use enables large language models (LLMs) to perform complex tasks, and recent agentic reinforcement learning (RL) methods show promise for enhancing model capabilities. However, RL alone often leads to instability or limited gains in tool-use tasks. In our experiments, some models exhibit catastrophic collapse, where performance abruptly drops and tool-invocation structures fail. The analysis reveals that these failures stem from unexpected probability spikes in specific control tokens, disrupting structured execution, yet the underlying tool-use capability remains intact, merely obscured by specific formats. To address this, we systematically investigate a diverse set of supervisory signals, including off-policy supervision, hint-based guidance, erroneous example supervision, and others, applied under both synchronous and interleaved training schemes. We find that interleaving supervised fine-tuning (SFT) with RL substantially improves stability, but exhibits degraded performance under format and content out-of-distribution (OOD) evaluation. We also analyze the impact of learning rates and generalization across settings. These results highlight the importance of understanding RL failures and demonstrate how diverse supervisory signals can guide exploratory learning, enabling robust training of LLMs for complex, multi-step tool-use tasks. Our Code is available at https://github.com/hypasd-art/Tool-RL-Box. --- ### In-Context World Modeling for Robotic Control Authors: Siyin Wang, Junhao Shi, Senyu Fei, Zhaoyang Fu, Li Ji, Jingjing Gong, Xipeng Qiu Published: 2026-06-24 Categories: cs.RO, cs.CV arXiv: https://arxiv.org/abs/2606.26025v3 Modern Vision-Language-Action (VLA) models often fail to generalize to novel setups, such as altered camera viewpoints or robot morphologies, because they are typically conditioned only on current observations and language instructions. By ignoring the underlying system configuration as a variable, these models implicitly assume a fixed execution context encountered during training, necessitating data-intensive fine-tuning for any new environment. In this work, we introduce In-Context World Modeling (ICWM), a framework that treats system identification as an in-context adaptation problem. ICWM enables robot policies to autonomously infer essential system variables from a short history of self-generated, task-agnostic interactions. Unlike traditional In-Context Learning that uses demonstrations to specify what task to perform, ICWM leverages the context window to understand how the system operates. By processing these interactions before task execution, the model implicitly captures the world dynamics of the current system, enabling adaptation to novel configurations without parameter updates. Extensive experiments in simulation and on real-world robot platforms demonstrate that ICWM significantly outperforms standard VLA baselines on novel camera viewpoints. --- ### Privacy Vulnerabilities of Attention Layers in Tabular Foundation Models and Protection of High-Risk Queries Authors: Tânia Carvalho, Maxime Cordy Published: 2026-06-24 Categories: cs.CR, cs.AI arXiv: https://arxiv.org/abs/2606.26021v1 Tabular foundation models are commonly assumed to present limited privacy concerns as they are often pre-trained on large collections of synthetic data. However, these models leverage in-context learning, where sensitive records may be provided directly at inference time as labelled context examples. In this paper, we demonstrate that predictions generated via the attention mechanism leak sufficient information to enable effective Membership Inference Attacks (MIAs). To highlight this vulnerability, we propose AMIA (Attention-based Membership Inference Attack), a shadow-model-free attack that exploits the concentration of transformer attention patterns. Our results show that attention mechanisms reveal strong membership signals, which exceed classical confidence-based attacks, achieving an average gain of 7.7\%, specially in low false-positive regimes. To mitigate this risk, we introduce an inference-time defence inspired by $k$-anonymity principles. This approach reduces the uniqueness of context-key representations without introducing random noise or retraining the model. By targeting only high-risk queries identified through AMIA scores, the defence substantially reduces membership leakage of this attack by an average of 50\% and 25\% against confidence-based attacks, while preserving predictive utility with only 3.9\% performance degradation. Beyond showing that context examples are vulnerable, we further demonstrate that fine-tuning introduces an additional source of privacy risk. In particular, samples whose prediction confidence increases after fine-tuning become more susceptible to MIAs, indicating that fine-tuning can amplify memorisation and expose sensitive training information through confidence shifts. --- ### MIMFlow: Integrating Masked Image Modeling with Normalizing Flows for End-to-End Image Generation Authors: Yang Chen, Xiaowei Xu, Shuai Wang, Xinwen Zhang, Qiushi Guo, Tiezheng Ge, Limin Wang Published: 2026-06-24 Categories: cs.CV arXiv: https://arxiv.org/abs/2606.26016v3 Normalizing Flows (NFs) are powerful generative models capable of exact density estimation and sampling. However, their strict invertibility often forces the model to exhaust its capacity on low-level pixel details, hindering the capture of high-level semantic structures. While Masked Image Modeling (MIM) has excelled in representation learning, its integration into generative pipelines has remained largely modular and disjointed. In this paper, we propose MIMFlow, a unified end-to-end framework that jointly optimizes latent semantics, pixel reconstruction, and generative flow. By employing a VAE encoder to infer semantic latent from masked images, MIMFlow achieves a principled decoupling of the generative task: the Normalizing Flow focuses on modeling a simplified, low-frequency semantic manifold, while a specialized decoder handles high-frequency synthesis. This design effectively resolves the inherent capacity bottleneck of NFs, allowing the model to prioritize global structural coherence over redundant noise. Empirical results on ImageNet 256$\times$256 show that MIMFlow-L reaches 71.3\% linear probing accuracy and an FID of 2.50. Despite using only 128 tokens (50\% fewer than standard models), it yields a 32.8\% performance gain over similar-scale NF baselines. Our code is available at https://github.com/MCG-NJU/MIMFlow. --- ### The Tatoxa System for Text Detoxification in Low-Resource Languages: The Case of Tatar Authors: Ilseyar Alimova, Bogdan Monogov, Artyom Mazur, Daniil Antonov, Vsevolod Karimov, Vitaliy Egorov, Bulat Khakimov, Alexander Panchenko Published: 2026-06-24 Categories: cs.CL arXiv: https://arxiv.org/abs/2606.26015v1 Text detoxification, the automated detection and mitigation of abusive and harmful content, is essential for ensuring the safety of online communities and protecting users. However, low resource languages such as Tatar have received little research attention. In this paper we present Tatoxa, a novel state-of-the-art system for text detoxification in the Tatar language. Comparative experiments show that the proposed approach outperforms existing open source and proprietary commercial LLMs on key quality metrics. We also introduce a new dataset for text detoxification in Tatar, designed for fine tuning and evaluation in low resource settings. Finally, cross lingual transfer experiments indicate that transfer from other languages, including the culturally close Russian, performs significantly worse than training on native Tatar data even when a large Russian corpus is available. --- ### Is Variational Monte Carlo Robust? Sharp Moment Thresholds and Heavy-tailed Stochastic Optimization Authors: Philipp Grohs, Davide Nobile Published: 2026-06-24 Categories: cs.LG arXiv: https://arxiv.org/abs/2606.26009v2 Variational Monte Carlo (VMC) is a central algorithm in electronic structure theory and has gained renewed importance through modern neural-network ansätze such as FermiNet. At its core, VMC seeks ground states by minimizing the Rayleigh quotient by stochastic optimization. In this work, we show that the resulting stochastic optimization problem is intrinsically governed by the nodal geometry of the underlying wave function. More precisely, we establish that properties of the nodal set determine the integrability of the local energy and gradient estimators that drive VMC. For broad and practically relevant ansatz classes, including Slater-Jastrow wave functions with variable-exponent Slater-type orbitals, we prove that these estimators are generically heavy-tailed and fail to admit higher moments. At the same time, for general analytic ansätze, we prove weak moment bounds for the relevant estimators and identify precise low-moment regimes, showing how generic and degenerate nodal structures lead to different integrability thresholds. Building on this analysis, we introduce a new robust variant of VMC $\unicode{x2013}$ coined PS-Clip-VMC $\unicode{x2013}$ which is based on clipping both the local energies and the per-sample gradients. We prove that PS-Clip-VMC converges both in expectation and with high probability in the weak moment regime of VMC. The robustness of our method is confirmed experimentally by training FermiNet on atoms with up to 18 electrons. --- ### From Sparse and Imperfect 2D Anchors to Consistent 3D Gaussian Street Scenes: Support-Aware Appearance Authors: Long Cao, Zhongquan Wang, Jie Li, Yuhan Chen, Kefei Qian, Xiangfei Huang, Guofa Li Published: 2026-06-24 Categories: cs.CV, cs.GR arXiv: https://arxiv.org/abs/2606.26007v1 Image priors can synthesize target conditions for 3D Gaussian street scenes, but independently edited views do not define a coherent 3D target. Direct fitting can propagate view-specific noise, while existing pipelines do not jointly handle imperfect sparse anchors and standard-rasterizer deployment. To address this gap, teacher-relative appearance residual distillation is introduced for appearance baking. A structured space for frequency decomposition, confidence estimation, and primitive-level lifting is formed by residuals between teacher anchors and original renders. The direct optimization signal is supplied by renderer-space matching, while primitive assignment is regularized by support-aware Gaussian-space aggregation. Supported detail is admitted and unsupported noise is suppressed through confidence-gated coarse-to-fine optimization, after which all residuals are baked into fixed-geometry spherical-harmonic coefficients. The teacher and auxiliary training modules are discarded at inference. Evaluation across Waymo street assets, Tanks and Temples scenes, and multiple target conditions shows a favorable overall balance of target alignment, content preservation, artifact suppression, and cross-view consistency over editing-based baselines. Ablations confirm the effectiveness of the main components. Code will be released at https://github.com/Cagares/Baking-for-3D-Gaussian. --- ### FORCE: Efficient VLA Reinforcement Fine-Tuning via Value-Calibrated Warm-up and Self-Distillation Authors: Shuyi Zhang, Yunfan Lou, Hongyang Cheng, Yichen Guo, Chuyao Fu, Yaoxu Lyu, Xiaojie Zhang, Haoran Li, Pengwei Wang, Zhongyuan Wang, Shanghang Zhang Published: 2026-06-24 Categories: cs.RO, cs.AI arXiv: https://arxiv.org/abs/2606.26006v1 Vision-Language-Action (VLA) models are often constrained by the imitation ceiling imposed by sub-optimal data. While Reinforcement Learning (RL) fine-tuning can surpass this limit, it is notoriously sample inefficient. This challenge arises from two core issues: (1) catastrophic initial unlearning due to an unstable Q-function and (2) inefficient policy updates caused by low-quality exploration data, often forcing a reliance on costly human interventions. We introduce FORCE, a 3-stage framework that stabilizes fine-tuning by tackling both issues. FORCE first incorporates a Value-Calibrated Warm-Up phase, utilizing on-policy rollouts to mitigate the distributional shift of the Q-function. Subsequently, during the online stage, this calibrated Q-function acts as a filter for both the policy's own action proposals and expert data, ensuring only high-value actions are used for the policy update. We evaluate FORCE on various simulation and real-world tasks, and the result shows that FORCE achieves a 79% absolute improvement in success rates and outperform prior RL methods by 10%, while accelerating training by 32.5%. Critically, it mitigates the common success rate drop and achieves this robust performance without human intervention, marking a significant step towards deploying capable and autonomous robotic agents. --- ### Dziri Voicebot: An End-to-End Low-Resource Speech-to-Speech Conversational System for Algerian Dialect Authors: Dihia Lanasri, Rebeh Imane Ammar Aouchiche, Abdelkarim Remmide, Fairouz Taki, Asma Kemmoum Published: 2026-06-24 Categories: cs.CL arXiv: https://arxiv.org/abs/2606.26003v2 Automatic speech and language technologies are still heavily biased toward high-resource languages, limiting their applicability to dialectal and low-resource settings such as Algerian Dialect. This language presents additional challenges including lack of standardized orthography, frequent codeswitching with French, and scarcity of annotated speech resources. This paper addresses the problem of building a complete speech-to-speech conversational system for Algerian Dialect. We propose a modular pipeline integrating automatic speech recognition, natural language understanding, retrieval-augmented generation, and text-to-speech synthesis within a unified architecture. This work is the continuation of our previous work on Algerian dialectal conversational systems Bechiri and Lanasri (2026), extending it from text-based dialogue modeling to full speech-based interaction. We constructed dedicated datasets for ASR, NLU, and TTS in the telecom domain and fine-tune pretrained models for each component. The ASR system is built on Whisper-based adaptation, while the NLU module combines transformer-based embeddings with a task-oriented dialogue framework. A neural TTS system is trained on a newly collected dialectal corpus to enable spoken response generation. Experimental results show strong performance across all components, including low word error rate for ASR, high intent classification and entity recognition scores for NLU, and stable speech synthesis quality. The proposed system provides a reproducible baseline for end-to-end conversational modeling in Algerian Dialect. --- ### Hierarchical Reinforcement Learning for Neural Network Compression (HiReLC): Pruning and Quantization Authors: Kamar Hibatallah Baghdadi, Kawther Guoual Belhamidi, Sara Belhadj, Aissa Boulmerka, Nadir Farhi Published: 2026-06-24 Categories: cs.LG, cs.AI, math.OC arXiv: https://arxiv.org/abs/2606.26002v1 We present HiReLC, a hierarchical ensemble-reinforcement learning framework for automated joint quantization and structured pruning of deep neural networks. The framework decomposes the compression search across two levels of abstraction: low-level agents (LLAs) operate independently per block, selecting per-kernel configurations over a multi-discrete action space spanning bitwidth, pruning keep-ratio, quantization type, and granularity, while high-level agents (HLAs) coordinate global budget allocation via ensemble voting guided by Fisher Information-based sensitivity estimates. To mitigate the computational cost of policy evaluation, an iterative active learning loop interleaves surrogate-guided RL optimization with post-compression fine-tuning, using a lightweight MLP surrogate to amortize expensive evaluations and a logit-MSE proxy during cold-start. The surrogate is used for reward shaping rather than as a replacement for final post-compression evaluation. The controller is architecture-agnostic by design, with a modular layer abstraction decoupling the RL environment from the underlying network topology. Experiments across Vision Transformer and CNN benchmarks demonstrate effective parameter-storage compression ratios of 5.99 - 6.72$\times$ with a 3.83 % gain in one setting and 0.55 - 5.62 % accuracy drops elsewhere, supporting hierarchical policy decomposition and sensitivity-aware guidance as practical design choices for joint neural network compression. --- ### Variable Bound Tightening for Nash Equilibrium Computation in Multiplayer Imperfect-Information Games Authors: Sam Ganzfried Published: 2026-06-24 Categories: cs.GT, cs.AI, cs.MA, econ.TH, math.OC arXiv: https://arxiv.org/abs/2606.25997v2 There has been significant recent progress in algorithms for approximation of Nash equilibrium in large two-player zero-sum imperfect-information games and exact computation of Nash equilibrium in multiplayer strategic-form games. While counterfactual regret minimization and fictitious play are scalable to large games and have convergence guarantees in two-player zero-sum games, they do not guarantee convergence to Nash equilibrium in multiplayer games. Recently, an approach has been presented for exact computation of Nash equilibrium in multiplayer imperfect-information games that solves a quadratically constrained program based on a nonlinear complementarity problem formulation derived from the sequence-form game representation. This formulation was solved using Gurobi's nonconvex quadratic solver, which employs spatial branch-and-bound to iteratively refine variable bounds by solving convex relaxations of bilinear terms via McCormick envelopes. During presolve, Gurobi introduces auxiliary variables and, in some cases, binary variables, leading to an internal MIQCP reformulation. This approach was demonstrated to outperform prior algorithms from the Gambit software suite and quickly solve three-player Kuhn poker after removal of dominated actions; however, the algorithm was not able to solve the full version of the game within 24 hours. In this paper, we derive finite bounds on slack and multiplier variables in the nonlinear complementarity formulation. These bounds strengthen the convex relaxations used within spatial branch-and-bound and lead to substantial computational improvements. We demonstrate the impact of the proposed bounds on exact Nash equilibrium computation in three-player Kuhn poker. ---