# Research Papers — 2026-08 · bundle page 67 of 240 > Full-text markdown of papers #3301–3350 of 11995 published in 2026-08, newest first. AI-readable mirrors of the full 11995-paper 2026-08 archive (50 papers per file). Canonical page: https://aipod.dev/papers/archive/2026-08/page/67 [← Previous bundle](https://aipod.dev/papers/archive/2026-08/page/66.md) [Next bundle →](https://aipod.dev/papers/archive/2026-08/page/68.md) ### Toward Sub-1 kB Identity-Preserving Face Compression: A Benchmark of Codecs, a Custom Learned Codec, and Studies of Resolution, Demographic Fairness, Recompression, and Adversarial Robustness Authors: Petr Hurtik, Jakub Sochor Published: 2026-08-24 Categories: cs.CV arXiv: https://arxiv.org/abs/2608.22866v1 Storing face images under a hard sub-kilobyte budget, as required for identity documents, smart-card biometrics and bandwidth-constrained verification, forces a codec to discard most of the signal while keeping what a face matcher actually reads: identity. Generic codecs optimize pixel fidelity, not the embedding distances that drive verification, so which codec, resolution and setting best preserve identity at 1024 bytes or less, and how that degrades at 512, is unclear. We benchmark ten general and face-specific codecs across resolutions, byte budgets, two datasets (controlled Color FERET, in-the-wild AI-Solutions-KK) and four anchor face matchers, with a fourteen-model ViT and CNN roster confirming the ranking is backbone-invariant. We then train a custom identity-preserving codec that hits the byte budget exactly via binary search over a frozen gain table, and run four studies: resolution, demographic fairness, recompression, and no-box adversarial robustness. Sub-kilobyte identity preservation is feasible, but which codec to deploy depends entirely on the budget. At 1024 bytes and the 112 px working resolution the problem is close to solved: modern codecs hold Color FERET equal-error rate under 0.35 percent on the ArcFace anchor. At 512 bytes the field re-sorts: AVIF, HEIF, JPEG XL and legacy JPEG collapse to 28 to 98 percent false-non-match rate at FMR 1e-4, while WebP, JPEG-AI and our byte-budgeted learned codecs stay out of that band, with 24.3 percent for WebP against 6.9 percent for our accurate variant in the wild. That re-sort, not the 1024-byte ranking, is the operational result: a codec chosen at 1 kB is not the codec to deploy at half that. --- ### JANUS: Online Jacobian-Aligned Infill for Black-Box Optimization Authors: Hongyuan Yu, Pufan Xu, Jiaojiao Yi, Yiding Tian, Mingrui Sun, Jiayuan Lu, Changyuan Wen Published: 2026-08-24 Categories: cs.NE arXiv: https://arxiv.org/abs/2608.22862v1 Population optimizers such as CMA-ES, DE, and multi-objective evolutionary algorithms drive search mainly through selection signals that are scalar or rank based: such a signal indicates that one candidate outperforms another, but not the local direction responsible for the improvement. JANUS (\emph{Jacobian-Aligned Newton-Unified Search}) is a plug-and-play infill module that extracts this missing local geometric signal without replacing the host optimizer. It estimates a local Jacobian from the recent evaluation trace; the same Jacobian yields both a damped Gauss--Newton exploitation candidate and a trace-preserving exploration metric, reserving a fraction of the host's per-generation candidate slots for geometry-guided infill rather than spending evaluations on top of the host's budget. Unlike MetaBBO methods, JANUS needs no offline training or task distribution, estimating this geometry on the fly from the current run alone, while the host keeps full control of selection, survival, covariance adaptation, and step-size control. Under same-protocol comparisons, JANUS improves the CMA-ES host on \textbf{11--15/16} BBOB functions across $d\in\{30,100,500\}$. It also attains the best mean error on \textbf{13 of the 16} functions at $d{=}500$ in the complete NN-BBO/MetaBBO baseline comparison, with no training cost, and yields a $936\times$ geometric-mean improvement over the host on a $d{=}1000$ BBOB subset. On structured and multi-objective tasks, JANUS gives the best mean cost on 1135-dimensional UAV path planning ($-12.8\%$ vs.\ the strongest baseline), and it improves SMS-EMOA/AGE-MOEA2 hosts on 12/38 multi-objective tasks with zero significant regressions. Code is available at https://github.com/hongyuanyu/JANUS. --- ### Following Motion for Sequential Modeling in Video Frame Interpolation Authors: Jaehyun Park, Nam Ik Cho Published: 2026-08-24 Categories: cs.CV arXiv: https://arxiv.org/abs/2608.22861v1 State Space Models (SSMs) have surfaced as a promising architecture in Video Frame Interpolation (VFI), as they can capture long-range dependencies with linear computational complexity. However, their predefined scanning order limits their effectiveness in modeling the dynamic motion trajectories inherent in VFI problems. To tackle this challenge, we propose Motion-Guided Mamba for Video Frame Interpolation (MGMVFI), an adaptation of the selective state space model tailored explicitly for VFI. MGMVFI introduces Motion-Guided Serialization (MGS), which leverages optical flow to define a motion-adaptive 1D input order for the SSM. This aligns the causal state updates with semantically related tokens, enabling motion-consistent feature propagation, particularly for large and dynamic motions. Additionally, to mitigate the unreliable feature representations caused by inaccurate optical flow estimates, we introduce contextual synthesis that utilizes the surrounding spatial context for robust inter-frame feature synthesis. These components are seamlessly integrated within our tailored Mamba architecture, which also employs a lightweight refinement block to enhance local detail reconstruction at a reduced computational cost. Extensive experiments on standard VFI benchmarks demonstrate that MGMVFI achievesstate-of-the-artperformance,particularly on complex and dynamic motions, thereby establishing a new direction for sequence modeling in video interpolation. --- ### WARP: Wasserstein-Aligned RAG for Population Opinions Authors: Aman Singh Thakur, Aditya Agrawal, Alwarappan Nakkiran, Alex Karlsson Published: 2026-08-24 Categories: cs.IR, cs.CL arXiv: https://arxiv.org/abs/2608.22859v2 RAG systems are increasingly used to summarize what large collections of documents say. A user asks "What do people think about X?" and receives an answer that reads as consensus. But standard top-k retrieval ranks documents by query similarity, not by how faithfully they represent the population, so minority views quietly disappear. Existing fixes fall short. Diversity re-rankers like MMR and DPP spread retrieved documents apart, but with no target distribution to aim for. Calibration methods based on KL or JS divergence do target one, yet treat opinion bins as unordered: confusing strong positive with strong negative costs no more than an adjacent-bin miss. We introduce WARP, a family of post-retrieval algorithms that calibrate retrieved evidence to the population's opinion distribution. WARP first recovers underrepresented opinions that cosine ranking may bury, then uses Wasserstein-1 distance to select documents whose sentiment-intensity distribution matches the population target, capturing the ordinal structure ignored by KL and JS divergence. We develop three variants for dense, sparse, and variable candidate pools, trading off calibration quality and speed. Across three review domains spanning 35K documents, 156 queries, and 26 entities, WARP's domain-matched variants reduce distributional error by at least 43% with sub-second latency. These gains carry through to generation: a five-judge LLM panel prefers WARP-generated answers in 86% of decided comparisons at k <= 5. --- ### Mapping the Concept Landscape: Structural Perception of Global Distributions for Transparent Data Pruning Authors: Dongyue Wu, Tao Ma Published: 2026-08-24 Categories: cs.LG, cs.CV arXiv: https://arxiv.org/abs/2608.22858v1 Existing data pruning methods predominantly rely on high-dimensional feature embeddings to measure sample importance. However, these compressed vectors often obscure fine-grained semantic interactions, leading to suboptimal coverage of rare semantic concepts in the pruned subsets. In this paper, we propose Mapping the Concept Landscape (MCL), a novel structural perception framework for transparent data pruning. Instead of abstract embeddings, we represent each image-caption pair as an explicit sample-level graph comprising entities, events, and attributes. By integrating these individual graphs into a comprehensive dataset-level graph, we characterize the global distribution of semantic concepts and quantify their rarity across the entire corpus. Based on this structured perception, we develop a greedy concept-coverage maximization algorithm that iteratively selects samples to maximize the marginal gain of high-value, under-represented concepts. Experimental results on various benchmarks demonstrate that our method not only achieves superior pruning efficiency compared to state-of-the-art methods but also provides a transparent and interpretable audit trail for the selection process. --- ### SAVER: Selective Auditing of Verbal Evidence for Error Recovery in VLM Change Reasoning Authors: Youdi Li Published: 2026-08-24 Categories: cs.CL arXiv: https://arxiv.org/abs/2608.22857v1 Vision-language models (VLMs) frequently fail at visual change reasoning, even when their vision encoders contain sufficient information. We observe that correct VLM outputs tend to contain explicit verbal evidence (object names, colors, spatial locations) that supports the claimed change, while incorrect outputs often lack such evidence. We propose SAVER (Selective Auditing of Verbal Evidence for Error Recovery), a lightweight, rule-based method that parses VLM responses for this evidence and triggers structured reprompting only when evidence is missing or inconsistent. Across three change detection benchmarks and four VLMs, SAVER significantly improves accuracy on tasks where errors stem from the model failing to articulate what it saw (expression failures), with gains up to +25.8% on CLEVR-Change. The evidence patterns can also be generated by an LLM in a single call, matching the hand-tuned gate on CLEVR-Change. Ablation experiments confirm that the evidence gate, not reprompting alone, drives the improvement. --- ### Same Agent, Different Answers: A Repeat-Aware Audit of Corpus-Induced Answer Churn in Retrieval-Augmented QA Authors: Jingjie Ning, Xueqi Li Published: 2026-08-24 Categories: cs.IR, cs.CL arXiv: https://arxiv.org/abs/2608.22856v1 A retrieval-augmented QA system can return different answers after an index expansion even when its requested model identifier, prompt, retrieval policy, evidence depth, rendering, and exposed generation controls are held fixed. Aggregate accuracy may hide these changes when gains and losses cancel, while ordinary generation variability makes one-shot comparisons overstate update effects. We call the hidden phenomenon accuracy-blind answer churn and introduce the \emph{Snapshot Compatibility Audit}, which estimates excess answer churn by subtracting same-snapshot repeat disagreement from cross-snapshot disagreement. We instantiate it by expanding one frozen FineWeb prefix from one to seven shards. In a preregistered 400-question Natural Questions study, normalized-exact and blinded-semantic excess churn are 6.44 and 10.25 percentage points while exact-match accuracy changes by only $-1.50$ points. A post-hoc analysis finds repeat-stable semantic flips on 40/400 questions. A separately preregistered 200-question TriviaQA study yields smaller, directionally consistent excess churn while exact-match accuracy moves in the opposite direction. An outcome-blind post-hoc 100-question subset replication with a second DeepSeek generator and serving configuration finds 8.75 pp of semantic excess churn even as exact match rises by 3.00 percentage points. Answer-level compatibility can therefore fail without a conspicuous or consistently directed utility shift. Retrieval-augmented releases should audit compatibility alongside utility. --- ### Thinking at the Right Size: Amortized Distillation Across Post-Trained LLMs Authors: Yan Zhou, Sara Kangaslahti, Jonathan Geuter, Nihal V. Nayak, Marco Fumero, Francesco Locatello, David Alvarez-Melis Published: 2026-08-24 Categories: cs.LG arXiv: https://arxiv.org/abs/2608.22854v1 Practical deployment of large language models (LLMs) requires families of post-trained variants---instruction-tuned, reasoning-tuned, and chat-style models---each at multiple sizes to meet diverse latency and memory budgets. Producing each (variant, size) pair independently is prohibitive, so model families typically span only a handful of coarse-grained sizes per post-trained variant. Boomerang distillation (Kangaslahti et al., 2026) reduces this cost along the size axis for base models. Through model size interpolation, it constructs models of intermediate sizes from a single teacher-student pair without additional training. However, it still treats each post-trained variant as a separate object of optimization. We introduce ADAPT---Amortized Distillation Across Post-Trained LLMs---a framework for amortizing distillation across both axes of a model family: size and post-training variant, producing $L \times K$ models for $L$ interpolated sizes across $K$ post-trained variants with a single distillation run. ADAPT combines two components. First, a two-phase distillation procedure constructs post-trained students through pre-training alignment and supervised fine-tuning distillation, enabling smooth size--performance interpolation on generation and reasoning tasks. Second, weight-delta initialization approximates this construction across post-trained variants by transferring the distillation-induced weight change from the base model to students initialized from different post-trained variants. The resulting continuum of interpolated models also enables adaptive model-size selection at inference time, improving the compute--accuracy trade-off for long-form reasoning tasks. --- ### Your AI, On a Dial: Controlling Investment Bias in LLMs with a Single Neuron Authors: Sahong Park, Suhwan Park, Hoyoung Lee, Gakyung Kwon, Wonbin Ahn, Jaewon Choi, Alejandro Lopez-Lira, Yoon Kim, Chanyeol Choi, Hyeongwoo Kong, Yongjae Lee Published: 2026-08-24 Categories: cs.AI, cs.CL, q-fin.GN arXiv: https://arxiv.org/abs/2608.22852v1 Large language models (LLMs) are increasingly used in investment decision-making, yet prior work shows that they exhibit systematic, model-specific investment preferences. We study whether a model's overall investment stance can be calibrated to a specified direction and strength. We introduce an investment-bias dial, an inference-time intervention on a single neuron that continuously adjusts a model-level decision prior---its overall tendency toward buying or selling---without targeting specific firms or investment attributes. Using matched positive and negative evidence, we evaluate five open-weight LLMs and find that the dial produces monotonic changes in investment stance without modifying prompts or model parameters. At the response level, the dial shifts both investment decisions and the evidential emphasis of generated rationales under identical inputs. In an agentic retrieval setting, the dial also changes what information the model searches for, which evidence it selects, and which evidence is reflected in its final analysis. In a long-context evaluation, the dial maintains stable stance control as context length increases, whereas a matched system-prompt instruction progressively attenuates. We further show that changes in the dial propagate to security rankings and downstream portfolio composition in an exploratory backtest. Overall, our results show that an LLM's aggregate investment stance can be calibrated toward a specified target at inference time. --- ### RIBOSPAN: A Long-Context RNA Foundation Model for Versatile RNA Modeling Authors: Ziyuan Wang, Bohao Tang, Fei Zhang, Shuo Han, Pengfei Liu Published: 2026-08-24 Categories: cs.LG, q-bio.GN arXiv: https://arxiv.org/abs/2608.22849v2 Full-length RNAs, particularly messenger RNAs, often exceed the context lengths used to pretrain existing RNA foundation models, limiting complete-transcript modeling at single-nucleotide resolution. We present RIBOSPAN, a 1.61-billion-parameter bidirectional RNA foundation model natively pretrained with context lengths up to 10,240 nt. RIBOSPAN combines dense bidirectional self-attention, single-nucleotide tokenization, and attention-isolated sequence packing to enable high-resolution modeling of complete long RNAs. Native 10K pretraining preserves strong reconstruction at 10,240 tokens and, in a controlled long-context benchmark, maintains strong contextual responsiveness and context-specific representation separation while keeping perturbation-induced changes highly localized. Inference-time YaRN scaling recovers much of the contextual organization lost by direct short-context extrapolation, but induces substantially greater distal representation diffusion. Frozen RNA-type evaluations show that RIBOSPAN learns state-of-the-art RNA representations, with a particularly clear advantage on long RNAs. Across downstream biological benchmarks, RIBOSPAN emerges as the strongest encoder-only RNA foundation model, achieving state-of-the-art performance in both full-transcript biological property prediction and zero-shot mutation-fitness modeling. Building on the same backbone, we develop a multidimensionally conditioned discrete-diffusion framework for full-length mRNA generation and redesign, including synonymous-codon diffusion for protein-preserving CDS optimization. Together, RIBOSPAN establishes a powerful long-context foundation for transferable RNA representation learning, biological prediction, and full-transcript mRNA design. --- ### GSAR: Goal-State-Anchor Rewards for Mobile GUI Agents with Self-Evolving Data Synthesis Authors: Long Zhang, Yuhan Chen, Chaoran Zhang, Wanxia Cao, Kun Huang, Pengzhi Gao, Wei Liu, Jian Luan, Chenliang Li, Lixin Zou Published: 2026-08-24 Categories: cs.AI arXiv: https://arxiv.org/abs/2608.22847v1 Vision-Language Models (VLMs) based GUI agents stand to benefit significantly from online reinforcement learning (RL). However, their training is bottlenecked by two fundamental issues: current data synthesis methods for GUI Agents rely on specific environments and struggle to generate diverse data, while existing evaluators either suffer from limited scalability or provide inaccurate and unreliable reward signals. To overcome these challenges, we introduce GSAR (Goal-State-Anchor Reward), a RL reward framework that supports scalable task generation and delivers reliable reward signals for stable and efficient policy optimization. Our approach features self-evolving data synthesis, which produces multiple environments through task execution and generates diverse tasks and goal states. Complementing this, a state-anchor mechanism automatically annotates task-relevant UI elements in successful goal states as reference anchors. During RL training, these reference anchors provide accurate, scalable reward signals that substantially enhance efficiency. Extensive evaluations demonstrate that our framework achieves over 90% accuracy on offline trajectory verification and performs closest to rule-based methods. Furthermore, agents trained using our reward framework exhibit strong performance on both AndroidWorld and our constructed benchmark, establishing a scalable approach for GUI agent training. --- ### FinixDoc: Rethinking Financial Document Parsing Beyond Saturated Benchmarks Authors: Hang Wang, Jin Zhang, Guoliang Xu, Pengyue Lu, Yao Li, Zijiao Zhang, Tianyu Huang, Weiqi Xiong, Yulong Wang, Chuqiao Lu, Wenkang Huang, Kai Yang et al. Published: 2026-08-24 Categories: cs.AI arXiv: https://arxiv.org/abs/2608.22842v1 Financial document parsing requires accuracy, structural consistency, and verifiability that current benchmarks often fail to reflect. We present FinixDoc, an end-to-end agentic parsing system for real-world financial documents, with FinixDoc-VL, a 4B-scale vision-language model built on Qwen3-VL-4B, as its core parser. To characterize the gap between benchmark and deployment performance, we introduce a Document Parsing Capability Matrix organized along two practical axes: visual quality and document scale. Guided by this matrix, FinixDoc-VL is trained with a domain-adapted recipe combining homoglyph-aware contrastive learning and multi-stage reinforcement learning with composite domain-specific rewards. To better leverage our accumulated advantage in low-quality financial-document data and support large-scale, high-quality data production, we further build a human-in-the-loop Data Factory pipeline with confidence-aware expert review. For evaluation, we construct FinixDocBench, a financial-domain evaluation suite covering digital-native, camera-captured, ultra-large-page, and internal-workflow scenarios, with a compliance-reviewed subset released alongside this technical report. On its main subsets, FinixDoc-VL achieves the highest overall score (81.43) among evaluated baselines, outperforming the next-best open-source model by 5.13 points, with the largest gains on internal financial workflows (FinixInner: 84.08 vs. 78.73). --- ### Hierarchy-Aware Supervised Uncertainty Estimation for Black-box LLM Taxonomic Reasoning Authors: Shuting Xie, Nathaniel Lesperance, Graham W. Taylor Published: 2026-08-24 Categories: cs.LG, cs.AI arXiv: https://arxiv.org/abs/2608.22839v1 Large language models (LLMs) are increasingly used for scientific decision support, yet reliable confidence estimation remains difficult in black-box settings. We study uncertainty estimation for hierarchical taxonomic reasoning generated by a black-box LLM in a long-tailed biodiversity monitoring pipeline. Using proxy features extracted by an open-source tool LLM, we train lightweight supervised estimators with hierarchy-aware supervision to predict rank-wise correctness. Across three tool LLMs, the supervised estimators consistently outperform a token-likelihood baseline for micro discrimination and selective prediction under a single global rejection threshold, improving micro AUROC from 0.57 to 0.75--0.80. The best results are achieved by a rank-specific multi-head design (H3), suggesting that accounting for hierarchical output structure is important when a unified abstention rule is required. Our code is publicly available at https://github.com/uoguelph-mlrg/hierarchy-aware-llm-uq --- ### Mirror descent algorithms with logarithmic barriers Authors: Alberto De Marchi, Yura Malitsky, Adrien B. Taylor Published: 2026-08-24 Categories: math.OC, cs.LG, math.NA arXiv: https://arxiv.org/abs/2608.22834v1 This work derives convergence guarantees for mirror descent and proximal mirror descent algorithms when a logarithmic barrier is used as a distance-generating function. Standard approaches cannot be applied when the solution lies on the boundary, where the Bregman divergence blows up. We show that, in a specific setting, both methods enjoy an $O(\log k / k)$ rate, which is also tight. In addition, our contributions include: (i) a new technique for handling the blow-up; (ii) a resolution of a gap in the theory of relative smoothness; and (iii) a comparison of the proposed approach with interior-point methods. --- ### Minimal Local Simulation Foundations for LLM- and VLM-Driven Agents in 2D and 3D Environments Authors: Ryuki Hyodo Published: 2026-08-24 Categories: cs.MA, cs.AI arXiv: https://arxiv.org/abs/2608.22833v1 Large language models (LLMs) and vision-language models (VLMs) are expanding the range of behaviors that can be represented in agent-based simulations, but many contemporary platforms are difficult to study, modify, or run on ordinary computers. We present two intentionally minimal simulation foundations for education and rapid prototyping. SD-AgentFoundry-2D provides a two-dimensional multi-agent environment in which locally hosted LLM agents move, communicate, respond to place occupancy, and encounter spatially localized fire events. SD-AgentFoundry-3D provides a three-dimensional digital-twin environment in which a locally hosted VLM receives first-person images and produces natural-language movement instructions. Both codebases are designed to run locally on macOS, Windows, and Linux and are deliberately left open to modification rather than developed as finished applications. Together, they offer accessible starting points for learning about generative social simulation and for building domain-specific extensions. --- ### Let the Bullets Fly: Multimodal Fake News Detection with Temporal-Aligned Generative Danmaku Authors: Xiansheng Luo, Chaowei Zhang, Zewei Zhang, Yi Zhu, Jipeng Qiang Published: 2026-08-24 Categories: cs.AI arXiv: https://arxiv.org/abs/2608.22832v1 The social interactions among crowds via \textit{Danmaku} (a.k.a., bullet comments) on modern multimedia platforms can facilitate both viewpoint conflicts and consensus, providing fine-grained discriminative social signals that can benefit fake news detection. However, the inherent accumulation latency of \textit{Danmaku} in real-world scenarios violates the real-time necessity of fake news detection, making the studies of \textit{Danmaku}-related fake news detection underexplored. To break this violation, we simulate this temporal-aware user interactive process by proposing a novel temporal \textbf{Gen}erative \textbf{da}nmaku framework, called \textbf{Genda}, which consists of: (1) a \textit{Danmaku} Trigger for predicting the timing and intensity of user reactions; and (2) a \textit{Danmaku} Generator for synthesizing corresponding semantic and emotional expressions, thereby mutually constructing a temporally aligned and human-like pseudo \textit{Danmaku} streams. To make the generated \textit{Danmaku} useful for identifying fake news videos, we further design a \textit{Danmaku}-guided Temporal Multimodal fake news detection model - \textbf{DM-FEND}, which enables fine-grained multimodal interactions among video, audio, text, and \textit{Danmaku}, enhancing dynamic modalities alignment and semantic noise inhibition. The experimental results demonstrate that \emph{DM-FEND} consistently outperforms state-of-the-art baselines across both Chinese (FakeSV) and English (FakeTT) benchmarks. Further ablations validate the crucial role of temporal \textit{Danmaku} modeling in enhancing robustness and discriminative capability. Finally, this study offers a bright and robust solution for multimodal fake news detection in modern social interactive fashions by bridging the temporal inconsistency between news and user behaviors. --- ### Beyond the Harness: End-to-End Optimization of Context Artifacts for Enterprise Text-to-SQL Authors: Kate Gwimm, Carson Eisenach Published: 2026-08-24 Categories: cs.AI, cs.LG arXiv: https://arxiv.org/abs/2608.22830v1 Deploying LLMs for enterprise Text-to-SQL is bottlenecked less by the model than by what context reaches it: business logic spans thousands of tables, and no model can ingest a full catalog at once. We argue that the most effective place to intervene is therefore the \emph{knowledge-base context} the model consumes, and that this context should be \emph{constructed} from historical usage rather than tuned for as a fixed input. Using a query-DAG decomposition--the same family of intermediates that enterprise benchmarks like BEAVER annotate, here recovered from production SQL--we compare the value of oracle query graphs versus retrieved knowledge-base context. In this ablation, retrieved knowledge-base context provides the largest marginal improvement when added to the full oracle graph. Building on this, we optimize a distillation procedure that turns historical query profiles into reusable SQL reference cards. On a benchmark of 5176 production queries from a major online retailer, optimizing these context artifacts yields larger gains (${\sim}12$--$25\%$ AST similarity) than optimizing the retrieval harness (${\sim}3$--$12\%$). On the public BEAVER benchmark, which lacks the production-usage signals available in our internal setting, the picture is more mixed: table cards alone perform about the same as raw historical SQL. The best optimized variant retrieves both cards and raw SQL, scoring $9.00\%$ versus $6.33\%$ (p-value $0.12$) for the comparable baseline on a held-out $N{=}300$ subset, using retrieved context and harness changes but no agentic loop. --- ### VeCAS: Vessel-Focused Contrast-Free Angiogram Synthesis for Vascular Interventions Authors: De-Xing Huang, Chen-Yu Wang, Hao Liang, Xiao-Hu Zhou, Mei-Jiang Gui, Tian-Yu Xiang, Qin-Yi Zhang, Chen Wang, Xiao-Liang Xie, Shi-Qi Liu, Ming-Yuan Liu, Zhen-Chang Wang et al. Published: 2026-08-24 Categories: cs.CV arXiv: https://arxiv.org/abs/2608.22828v1 X-ray angiography relies on iodinated contrast agents to visualize vascular structures during image-guided interventions. However, contrast administration carries risks of adverse events, motivating the development of contrast-free alternatives. Generating X-ray angiograms directly from non-contrast X-ray images offers a potential solution, but existing approaches remain limited by (i) insufficient control over vascular localization and (ii) inefficient modeling of redundant background content. To address these challenges, we propose VeCAS, a two-stage vessel-focused contrast-free angiogram synthesis framework that separates vascular structure localization from angiographic appearance synthesis. In Stage I, a discriminative model localizes vascular structures in non-contrast X-ray images, while cross-modality latent distillation transfers vessel-sensitive knowledge from X-ray angiograms during training. In Stage II, a vessel-focused inpainting model synthesizes angiographic appearance within the localized vascular regions while preserving the non-vascular background. Experiments on an in-house lower-limb vascular intervention dataset show that VeCAS outperforms the comparison methods in terms of vascular structural fidelity and image quality. Visual Turing tests and physician assessments indicate the perceptual realism of the synthesized angiograms. In addition, robotic guidewire navigation experiments in vascular phantoms show that VeCAS guidance reduces the time to target by 41.4% and the number of operation steps by 40.7% compared with non-contrast guidance. Together, these results suggest the potential of VeCAS to serve as ``meta contrast agent'' for vascular interventions. --- ### DIME: Query-Efficient Framework for Membership Inference on Diffusion Models Authors: Tue Do, Daniel Alabi Published: 2026-08-24 Categories: cs.LG, cs.CR arXiv: https://arxiv.org/abs/2608.22824v1 Membership inference attacks expose whether individual records were used to train a model, yet existing attacks on diffusion models are largely heuristic and can require substantial query budgets. We introduce DIME (Denoiser Ideal Membership Error), a theoretically grounded and query-efficient framework for membership inference on diffusion models. Our starting point is an exact characterization of the optimal diffusion denoiser for a finite training set, which reveals that membership leakage is governed by the denoiser's implicit reconstruction error. This error decomposes into two complementary signals: a bias term, capturing reconstruction accuracy, and a previously unexplored local crowding term, capturing the geometry of nearby training examples. Both admit efficient estimators using only model queries, yielding a practical attack with as few as two queries. Across CIFAR-10/100, STL10-U, CelebA, and ImageNet, DIME consistently outperforms prior attacks at comparable or substantially lower query cost, improving TPR at 1% FPR by up to $3\times$; remarkably, its two-query variant can outperform existing 30-query baselines. Finally, we suggest, discuss, and evaluate specific defenses to counteract such powerful membership tests. --- ### SiZeUp: Fast 3D Proxy from Aerial Images via Depth Ordinal Loss Authors: Wenjun Zhou, Yunshan Li, Qiaoyu Zhu, Weidan Xiong, Hao Zhang, Daniel Cohen-Or, Hui Huang Published: 2026-08-24 Categories: cs.CV arXiv: https://arxiv.org/abs/2608.22821v2 We present SiZeUp, a fast and scalable approach for constructing large-scale 3D urban proxy models directly from calibrated oblique aerial imagery. Our method adopts a height-from-footprint representation, reducing 3D building abstraction to a low-dimensional optimization problem in which building footprints are extruded by a single height parameter. To enable efficient and robust height estimation, we introduce an ordinal depth consistency loss that enforces agreement between the relative depth ordering of rendered proxies and depth priors predicted by a monocular depth model. This is realized through a differentiable renderer that maps parametric building proxies into multi-view depth images, allowing gradients to be propagated from depth supervision to building heights. Our ordinal formulation produces stable optimization in practice and avoids explicit feature matching or dense point cloud reconstruction. Rather than relying on metric depth, which can be unreliable under monocular scale ambiguity, our ordinal depth consistency loss operates on relative depths, providing a more reliable signal across views. Combined with an efficient dynamic view selection, our approach achieves a 23-52$\times$ speedup over state-of-the-art proxy reconstruction pipelines while maintaining comparable proxy-level coverage and volume consistency, making it well suited for large-scale urban modeling tasks. --- ### Fairness-Aware Mixture-of-Experts via Subgroup Reweighting and Gate Regularization Authors: Sunhee Hwang Published: 2026-08-24 Categories: cs.LG, cs.AI arXiv: https://arxiv.org/abs/2608.22820v1 Deep learning models often produce performance disparities across demographic groups, due to the training data imbalance with respect to sensitive attributes such as gender or age. To address this problem, existing work has explored fair representation learning, data re-sampling, and adversarial training, which can be broadly categorized into two main approaches. Single-stage methods typically learn a shared representation for fairness, but often struggle to handle heterogeneous subgroup distributions. Two-stage methods learn representations separately from the final prediction task, which can lead to misalignment between fairness objectives and downstream predictions. We identify routing-induced bias, a failure mode in which subgroup imbalance drives the gating network to route subgroups onto a few experts, and propose an end-to-end Mixture-of-Experts (MoE) framework that corrects it. Specifically, we apply subgroup reweighting to correct data imbalance, and introduce gate entropy regularization to prevent routing from collapsing onto subgroup attributes, keeping expert utilization both balanced and interpretable. Beyond improving fairness, the routing distribution offers an interpretable view of how subgroups are allocated across experts. Experimental results demonstrate that the proposed approach improves fairness while maintaining competitive predictive performance. --- ### Direct, Parallel, or Sequential? A Comparative Study of Training-Free Multi-Subject Image-to-Video Generation Authors: Yanliang Qi, Kexi Chen, Muchao Ye, Haomiao Ni Published: 2026-08-24 Categories: cs.CV arXiv: https://arxiv.org/abs/2608.22819v1 Text-conditioned image-to-video (I2V) generation has advanced rapidly, yet generating videos with multiple subjects remains challenging. A model must simultaneously preserve the appearance of each subject, assign distinct motions, and maintain coherent spatial and temporal interactions. This paper presents a systematic study of three representative paradigms for training-free multi-subject I2V generation: direct, parallel, and sequential generation. Direct generation applies a pretrained I2V model to the complete reference image and prompt, requiring all subjects and motions to be synthesized jointly. Parallel and sequential generation instead decompose the reference image and prompt into subject-specific visual and textual conditions. Parallel generation synthesizes each subject independently and subsequently composes the resulting videos, reducing the complexity of each generation step at the cost of weaker inter-subject context. Sequential generation first synthesizes a background video and then progressively introduces individual subjects. This preserves accumulated scene context but introduces sensitivity to subject ordering and error propagation. We empirically evaluate the three paradigms across diverse multi-subject scenes, comparing appearance preservation, motion fidelity, temporal consistency, and inter-subject coherence, while also characterizing their distinct failure modes. Our findings reveal the strengths and limitations of each paradigm and offer practical insights for designing controllable multi-subject video generation systems. --- ### Industrial-Instruction: An End-to-End Framework for Building Instruction-Tuning and Benchmark Datasets from Industrial Technical Reports Authors: Parsa Bakhtiari, Hassan Bashiri, Alireza Khalilipour, Masoud Nasiripour, Moharram Challenger Published: 2026-08-24 Categories: cs.CL arXiv: https://arxiv.org/abs/2608.22817v1 Industrial technical reports contain high-value knowledge for maintenance, troubleshooting, and product engineering, but their heterogeneous structure (dense prose, specifications, tables) makes them difficult to index and reason over with standard retrieval and QA pipelines, and no public instruction-tuning or benchmark datasets are built from such documents. We address this gap with Industrial-Instruction, contributing (i) two open QA datasets built from real industrial technical reports and (ii) the end-to-end pipeline that produces them. Using 906 public Panasonic documents (7,525 pages), we apply layout-aware extraction, build a semantic retrieval index, and synthesize multiple-choice QA grounded in retrieved evidence under five query-document relationships (irrelevant retrieval, single-/multi-document support, single-/multi-document answer). After filtering an initial 23.9k generated samples, each dataset provides approximately 13.6k QA pairs with source documents and a held-out benchmark split. Fine-tuning small open LLMs (under 10B parameters) improves Set-Match Accuracy from 28.5% to 42.0% and F1 from 46.6% to 63.5% on the Panasonic benchmark. We release two parallel versions built by the same pipeline: one generated with the open-weight Qwen3-30B-A3B-Instruct model and one with the closed, API-based Claude-Opus-4.6 model, enabling a direct comparison of open- versus frontier-model data generation. The Claude-Opus-4.6 dataset yields a cleaner raw corpus and larger fine-tuning gains, at roughly two orders of magnitude higher cost. MMLU evaluation shows models trained on the Claude-Opus-4.6 data retain essentially all general knowledge, versus a small but measurable forgetting effect for the Qwen-generated data. Together, these datasets and pipeline offer a practical, reproducible path toward scalable industrial benchmarks and training data from real-world documentation. --- ### SAGE: Stability-Aware Graph-Based Ensemble Feature Selection for Explainable Postpartum Depression Risk Prediction Authors: Md. Rokon Islam Emon, Syed Shariar Alam Shuvo, Shahriar Siddique Ayon, Abdullah Al Mamun, Ahnaf Atef Choudhury Published: 2026-08-24 Categories: cs.LG arXiv: https://arxiv.org/abs/2608.22809v1 Postpartum depression (PPD) poses a major burden on maternal and child health, especially in low- and middle-income countries where prevalence exceeds 19%. Despite advancements in machine learning for PPD prediction, current approaches are limited by opaque global explanations that lack clinical usefulness at the patient level, unstable feature selection, and poor generalization under class imbalance. We propose SAGE, a Stability-Aware Graph-Based Ensemble feature selection system that incorporates both local explainable AI and a genetically optimized artificial neural network (GA-ANN). Using a primary cohort of 766 postpartum women, SAGE combines information-theoretic relevance, PCA-based structure, and graph-based interactions with bootstrap stability weighting to identify robust and non-redundant predictors. The GA-ANN architecture, optimized using a genetic algorithm and enhanced with GAN based oversampling, achieved strong performance with 87.96% accuracy, 86.32% F1 score, and 0.88 AUC using only 16 features, outperforming baseline and other feature selection methods. Psychological and socioeconomic factors such as EPDS score, PHQ-9 score, feelings about motherhood, and abuse history are the main predictors, while demographic factors have less influence. The LIME-based explanations allow instance-based insight into selected features from the graph, enabling personalized risk assessment. The findings make SAGE a scalable, interpretable, and clinical tool for early identification of PPD in health-care limited resources. --- ### CatchBench: When Can an Agent Failure Be Caught? Authors: Yue Zhao, Mengyuan Li, Ruolin Li, Prince Zizhuang Wang, Shuli Jiang, Linsey Pang, Xiongye Xiao, Xiyang Hu Published: 2026-08-24 Categories: cs.LG, cs.MA, cs.PF arXiv: https://arxiv.org/abs/2608.22808v4 When can an agent failure be caught? A weak audit score alone cannot identify whether the record or the method is limiting. CatchBench therefore puts one auditor's question to three information states: the declared configuration before a run (PRE), a growing prefix of its trace (LIVE), and the finished trace (POST). Prior benchmarks fix one of these states or vary the telemetry; to our knowledge none scores all three under one task-method interface. Each state admits different questions, so seven task contracts carry their own labels and metrics rather than one leaderboard. Four are evidential; three are Gold-derived mechanism diagnostics. The release scores 72 entrants, from rule scanners and structural models to eleven LLM judges across nine model families (GPT, Claude, Gemini, Gemma, Llama, Qwen, DeepSeek, Mistral, Nova), over 1187 declared configurations and 1162 recorded runs. Every recorded comparison is published as a measured difference with its interval, uncorrected, and no board declares a winner it cannot show. The three sharpest results cut against our own data. One rule reads declaration order alone and reaches a perfect F1 on one of six configuration sources, so a score there measures how the corpus was built. Our admissibility bar then rejected one injected substrate and withheld evidential status from the other. A published structural gain also turns on which size reference it is measured against. A benchmark number is therefore not interpretable until the process behind its labels is published and tested for the shortcut it may leave. We report all three, and regenerate every ordering from released predictions with no model call. --- ### Change Detection in Probability Flow ODE: Online Testing in Diffusion Latent Spaces Authors: Artem Kraevskiy, Artem Prokhorov Published: 2026-08-24 Categories: cs.LG, stat.ML arXiv: https://arxiv.org/abs/2608.22807v2 A rapidly growing range of sequential data tasks, such as identifying trend reversals in financial markets, auto-segmenting video and audio recordings, detecting changes in movement direction from motion sensors cannot be fully addressed without detection of distributional shifts in time-ordered data. We consider a sequential change-point detection problem where the conditional density switches at an unknown time, yet neither the pre- nor post-change distribution admits a closed-form. Classical likelihood-ratio statistics are inapplicable in this settings. A conditional diffusion model, trained on pre-change-point data with a frozen context encoder, defines a deterministic bijection via the probability flow ODE. Pre-change observations are mapped onto standard Gaussian latent variables. Post-change observations, processed through the same frozen map, deviate from this reference. We employ the Maximum Mean Discrepancy as the test statistic, derive closed-form expressions for its components under the Gaussian null, and establish its asymptotic distribution as a degenerate U-statistic. Afterwards we apply an online detection procedure of Shiryaev--Roberts to the resulting statistic with exact threshold calibration. The method detects arbitrary distributional shifts, including covariance rotations and higher-order structural breaks, without parametric assumptions on either regime. --- ### DIAG: Diagnostic Iterative Alignment and Generation for Data-Efficient Mathematical Preference Distillation Authors: Guhan Chen, Songtao Tian, Bohan Li, Hejin Wang, YeXin Xie, Zixiong Yu Published: 2026-08-24 Categories: cs.CL arXiv: https://arxiv.org/abs/2608.22806v1 Iterative preference optimization is essential for aligning Large Language Models on mathematical reasoning tasks, yet its efficiency is often throttled by signal scarcity: as the model improves, static problem sets become increasingly mismatched to the model's evolving competence, producing rollouts that are either too easy or too hard and therefore non-informative, which leads to a scarcity of valid preference pairs. We propose DIAG, a Diagnostic Iterative Alignment and Generation framework that adaptively reshapes the practice distribution to increase informative supervision and focus training near the student's current competence boundary. DIAG consists of two phases: (1) diagnosing valid preference-pair yield to calibrate the exploration-exploitation trade-off and allocate topic quotas via an Empirical Bayes shrinkage estimator, thereby prioritizing high-yield concepts; and (2) generating targeted practice, where a teacher synthesizes variants from the student's failure traces. We further provide a theoretical view interpreting DIAG as a teacher-mediated approximation to KL-regularized reweighting of the practice distribution toward the student's competence boundary, where valid preference-pair yield is maximized. Experiments show that DIAG boosts yield across iterations and delivers stronger reasoning performance under an iso-effective training budget, demonstrating that it can distill more informative preference supervision for mathematical reasoning. --- ### Contrastive Representation-Guided Genetic Minority Oversampling for Imbalanced Time-Series Classification Authors: Wenbin Pei, Yunrong Hao, Zhen Liu, Guan Wang, Bing Xue, Yiu-Ming Cheung, Qiang Zhang Published: 2026-08-24 Categories: cs.LG arXiv: https://arxiv.org/abs/2608.22804v1 Real-world time-series classification tasks often exhibit class imbalance, which can be extremely severe in some applications. To avoid training biased classifiers on imbalanced data, sampling is one of the most popular data pre-processing techniques because of its classifier-agnostic nature. However, due to the complex temporal dependencies in original time-series data and the scarcity of minority-class samples, existing sampling methods, including interpolation-based oversampling methods and deep learning-based generative models, usually suffer from limited generalization and poor diversity when generating new time-series samples. This paper proposes a Frequency-domain representation-guided Multi-tree Genetic Programming-based oversampling approach (FreMGP) to imbalanced time-series classification, where each individual represents a set of synthetic samples for the minority class. A frequency-domain class-discriminative representation module based on contrastive learning is also developed, guiding the evolutionary search toward high-quality synthetic time-series samples. Experiments on imbalanced time-series datasets demonstrate that FreMGP outperforms existing oversampling methods and consistently improves the performance of different classifiers, including both general machine learning and deep learning models. --- ### SDoH-Aware Narrative Anchoring Bias in Medical LLMs for Trustworthy Clinical Decision Support Authors: Ahnaf Atef Choudhury, Ramkrishna Saha Published: 2026-08-24 Categories: cs.CL, cs.AI arXiv: https://arxiv.org/abs/2608.22802v1 Medical large language models are often judged by how many clinical questions they answer correctly. That view is useful, but it misses a practical risk. A model may know the right answer and still change its response when the same case is written in a different patient voice. This paper evaluates that risk as SDoH aware narrative anchoring bias. We use NarrativeShield SDoH MedQA, a counterfactual medical question answering dataset in which each case appears in persona based narratives while the answer key remains fixed. The dataset is reshaped from wide format into case grouped persona rows. We evaluate three open source instruction tuned LLMs from the Qwen2.5 family: 1.5B, 3B, and 7B. The final experiment uses 300 clinical cases and produces 8,100 model responses across three prompting conditions. We report persona level accuracy, counterfactual consistency, correct consistency, and narrative sensitivity error. Qwen2.5 7B achieves the best accuracy at 56.33 percent and the best correct consistency at 40.33 percent. Paired McNemar exact tests show significant accuracy gains for 7B over 3B in all prompt settings. Even so, narrative sensitivity remains, with the lowest error still at 31.67 percent. These results suggest that trustworthy clinical decision support should be evaluated by both average correctness and stability across medically equivalent patient narratives. --- ### Triplet2Track: A Hierarchical System with Object-Centric Representations for Reliable Long-Horizon Manipulation Authors: Jianxiang Liu, Gaojing Zhang, Chuan Wen, Qipeng Liu, Yuxuan Zhao, Ning Guo, Wenzhao Lian Published: 2026-08-24 Categories: cs.RO, cs.AI arXiv: https://arxiv.org/abs/2608.22800v1 Ensuring reliability in uncertain environments remains difficult for long-horizon robotic manipulation. End-to-end VLA models are data-heavy and opaque, making diagnosis and verification difficult. Hierarchical pipelines are more interpretable, but their plans are often weakly grounded in observations, weakly aligned with low-level actions, and computed without online feedback, leading to open-loop behavior and hallucinations. To address these issues, we introduce the Triplet-to-Track System (TTS), a closed-loop long-horizon imitation learning system that uses human videos to reduce reliance on robot-collected data. TTS represents high-level subgoals as instance-grounded triplets, translates them into continuous track priors for execution, and monitors task progress from observations for online replanning. Across diverse real-world long-horizon tasks, TTS achieves a 74.8\% average success rate and supports object-level and compositional generalization. --- ### Reproducible Vision-Guided 6-DoF Robotic Manipulator with a Mixed Stepper-Driver Architecture and Browser-Native Control Authors: Lasan Perera, Deneth Priyadarshana, Dulana Pitiwaduge, Isitha Dinujaya, Mokshan Colambage Published: 2026-08-24 Categories: cs.RO, cs.CV, eess.SY arXiv: https://arxiv.org/abs/2608.22799v1 We present the NeuralNexus Arm, an open, low-cost 6-DOF robotic manipulator built by an undergraduate engineering team, together with the design decisions and debugging experience needed to reproduce it. The arm is driven by a single STM32H743 microcontroller on a custom printed circuit board (PCB) and combines two stepper-driver strategies on one controller: push-pull 3.3 V step/direction outputs for onboard TMC2209 drivers on the three wrist joints, and open-drain outputs for external CL57T and DM542 drivers on the three high-torque proximal joints. We describe the mechanical design, mixed-driver electronics, interrupt-driven firmware, a MATLAB/Simscape-based inverse-kinematics pipeline, a browser-native control interface using the Web Serial API, and a lightweight vision pipeline for object localisation and autonomous pick-and-place tasks. We also document non-obvious hardware and firmware failure modes encountered during the transition from a development board to the custom PCB as reproducibility guidance. All design files and firmware are released openly. The platform actuates all six axes under coordinated control at a 2 kHz update rate and executes both manual and pre-recorded motions from the browser interface. --- ### Performance of a domain-specific large language model in answering patient questions in psychiatry Authors: Alexander J. Hish, Arjun Nagendran, Scott N. Compton Published: 2026-08-24 Categories: cs.AI arXiv: https://arxiv.org/abs/2608.22797v1 Background This study was designed to evaluate whether a domain-specific large language model (LLM) trained exclusively on patient education resources can answer questions about psychiatric medications, in a manner superior to LLM chatbots. We developed an LLM ("MIND") fine-tuned for clinical fidelity, trained on patient education resources from authoritative medical organizations. Methods We compared the responses of MIND, ChatGPT, and OpenEvidence to patient questions about escitalopram, using two methods: (1) computer analysis according to a rubric measuring accuracy, clarity, completeness, nuance, safety, and referral appropriateness; (2) ratings from N=10 board-licensed psychiatrists on similar metrics. Results When rated by rubric, MIND was rated highest in all domains (p<0.001). When rated by psychiatrists, ChatGPT was rated accurate more often than MIND with a negligible effect size (p=0.021, r=0.073); MIND was rated complete more often than ChatGPT with a small effect size (p<0.001, r=0.160); and MIND and ChatGPT were rated safe with the same frequency (p=0.955, r=0.002). The majority of psychiatrists preferred the responses generated by ChatGPT (57.6%) compared to MIND (42.4%, p=0.003). Conclusions MIND was able to answer many questions about escitalopram in a manner deemed accurate, complete, and safe by psychiatrists the majority of the time. However, despite MIND's ability to provide more complete responses, psychiatrists preferred ChatGPT's responses. MIND represents a step towards building safe LLM systems to enhance patient education in psychiatry. --- ### VersaDB: A High-Performance AI Storage Database for Unifying Mutimodal Datasets Authors: Cong Wang, Zelin Liu, Yang Luo Ran Zhang, Zhijian Guo, Hui Zhang, Fan Yu, Yanfei Cao, Naijie Gu, Jun Yu Published: 2026-08-24 Categories: cs.CV arXiv: https://arxiv.org/abs/2608.22795v1 The AI field has been rapidly developing, leading to the emergence of a large number of AI training datasets of various types. These datasets contain different modalities, including text, images, audio, etc., and may come in various data storage formats. With the advancement of AI hardware, AI computation units like GPUs, TPUs, and NPUs can greatly accelerate the training speed of AI models, which in turn increases the demand for faster data processing. When using existing AI processing frameworks to handle datasets with different modalities and storage formats, processing speeds may be suboptimal due to issues such as data layout and the way users handle the data. Therefore, using a unified database to store multiple data formats can better manage and optimize data access. In this paper, we introduce VersaDB, a database designed specifically for AI datasets with various modalities. We implemented a page-based storage system, separating structured and unstructured data. Additionally, we generated B+ tree-based index files to accelerate data access. VersaDB supports automatic sharding and maintains a hierarchical metadata management system, with corresponding metadata maintained at the page, shard, and global levels, forming the foundation for the efficient operation of the database. We also focused on ease of use by providing APIs for directly converting datasets into VersaDB, as well as APIs for converting popular AI data storage formats (e.g., CSV, TFRecord, .bin) into VersaDB.Our experiments show that using VersaDB can achieve up to 5.35x acceleration and maintain consistent performance across different parallelism levels. --- ### TRACE: A Self-Evolving Skill Bank for Consistent, Limit-Aware LLM Agents Authors: Wenhao Wu, Menghao Zhang, Xin Wang, Zhi Wang, Kun Shao, Jian Luan Published: 2026-08-24 Categories: cs.CL, cs.AI arXiv: https://arxiv.org/abs/2608.22793v2 Reliable deployment of LLM agents in user-facing products depends not on raw task-solving ability but on consistency and limit-awareness: behaving the same way across repeated trials, and recognizing when a request cannot, or cannot yet, be safely fulfilled. CAR-bench exposes this reliability gap in the domain of in-car assistants: an LLM-simulated user issues incomplete or ambiguous requests, requiring the agent to resolve uncertainty through multi-turn dialogue and tool use while strictly adhering to domain policies. Even frontier models show a substantial gap between what they can solve at least once (Pass@3) and what they solve consistently across trials (Pass^k). We bridge this gap with TRACE (TRAjectory-Contrastive Evolution), which iteratively improves a skill-based agent's behavioral knowledge without modifying model weights. This knowledge is organized as a Skill Bank of modular, retrievable skills, each encoding a self-contained set of tool-use rules and behavioral guidelines. TRACE evolves this bank through an agentic self-evolution loop: after each evaluation round, it groups trajectories by the skills invoked and refines each skill by contrasting successful and failed behaviors. The updated bank then guides subsequent rounds, while during deployment the Actor performs state-conditioned skill orchestration at every turn. On GPT-5.5, TRACE improves consistency (Pass^3) by 34.6 points, from 59.9% to 94.5%, while shrinking the gap between potential and reliable performance to just 4.0 points. On the official hidden set, TRACE achieved first place using GPT-5.6-Sol, attaining a Pass^3 score of 70%-a 40% relative improvement over the baseline. These results show that TRACE converts high model potential into stable, consistent performance gain. Project homepage: https://darwin-agent.github.io/Car-bench-TRACE. --- ### GuidedFlow: An Attention-Guided Framework for Anomaly Detection in Additive Manufacturing Authors: Sosmita Paul, Krishna Roy Published: 2026-08-24 Categories: cs.CV, cs.LG arXiv: https://arxiv.org/abs/2608.22789v1 Additive Manufacturing (AM) plays a vital role in the ongoing industrial revolution. However, quality control remains crucial and challenging due to printing defects or potential cyber-physical intrusions. Image or video-based anomaly detection is a key effort towards addressing these challenges. Various approaches have been explored in this domain, including reconstruction-based, embedding-based, and flow-based methods. Though normalizing flow-based methods address some of the core challenges of unforeseen defects and generalization while maintaining detection performance, existing approaches struggle with tiny/stringing defects common in 3D printing. In a small-data setting, this poses a limitation in generalization. To address these limitations, we propose \textbf{GuidedFlow}, a novel attention-guided normalizing flow model for anomaly detection and localization. GuidedFlow employs a pre-trained ResNet model, fine-tuned on the domain dataset. An attention-guided spatial and temporal flow framework models the dynamics across multiple scales and frames. A Spatio-Temporal Attention Network (SAN) enables the flow model to prioritize relevant contextual cues from input frames. We evaluate GuidedFlow on our AM3D-AD dataset, consisting of benign and anomalous real 3D printed object images and videos. We also conduct a comparative study using the MVTec-AD industrial image anomaly detection dataset. Experimental results demonstrate that GuidedFlow outperforms most of the state-of-the-art models with enhanced detection accuracy and AUROC. --- ### TailSieve: Partial-Rollout-Guided Tail Routing for LLM Rollouts Authors: Tianqi Xu, Lu Lv, Haoyang Huang, Wenjie Huang, Zhanming Shen, Yuhao Shen, Baolin Zhang, Xinyi Hu, Shuang Ge, Jun Dai, Tianyu Liu, Suorong Yang et al. Published: 2026-08-24 Categories: cs.AI, cs.LG arXiv: https://arxiv.org/abs/2608.22788v2 Large-scale rollouts have become a core component of modern LLM systems, spanning reinforcement learning (RL) post-training, on-policy distillation (OPD), and sampling-heavy evaluation pipelines. Unlike online serving, which is typically optimized for request-level latency and throughput, a small number of long-tail generations can dominate the end-to-end makespan of an entire rollout step. In practice, rollout requests are often routed uniformly across replicas, which can place extremely long generations inside high-concurrency decoding batches. To address this, we present TailSieve, a partial-rollout-guided framework that jointly controls tail routing and replica allocation for LLM rollouts. In an idealized setting with known completion lengths, we show that makespan-optimal routing in the long-tail regime combines tail isolation with load balancing, and that a simple top-k policy closely approximates this offline optimum. Leveraging the observation that long-tail prompts tend to remain long-tailed across policy updates, TailSieve uses partial rollouts as a training-free signal for identifying candidate tail groups. A hierarchical controller then jointly adapts the number of isolated groups and the replica split between the tail and bulk pools using collected response-work history and a measured concurrency-throughput model. TailSieve achieves up to 1.67x routing-only speedup over uniform group routing. The resulting low-concurrency tail pool further enables route-specialized speculative decoding with MTP or DFlash, achieving up to 2.59x speedup over uniform routing. Selected prompts are regenerated under the current policy, preserving on-policy generation and avoiding additional routing-induced length bias in steady state. --- ### ReCoG: Reciprocal Co-Evolution for Multimodal Graph Learning Authors: Rui Xue, Tianfu Wu Published: 2026-08-24 Categories: cs.LG arXiv: https://arxiv.org/abs/2608.22786v1 Multimodal graph learning requires jointly training over graph structure and heterogeneous node attributes, yet existing methods largely decouple these processes: prior multimodal graph neural networks (GNNs) focus on aligning modalities in a shared embedding space while operating on fixed or weakly adapted graph structures, and graph structure learning approaches infer topology from unimodal node representations without accounting for multimodal interactions. This separation fundamentally limits the ability of GNNs to capture semantically meaningful relationships in multimodal settings, where observed edges are often noisy, incomplete, or misaligned with underlying semantics. We propose ReCoG (Reciprocal Co-Evolution for Multimodal Graph Learning), a new learning paradigm that tightly couples graph structure learning and multimodal representation learning through end-to-end reciprocal interaction. Concretely, ReCoG integrates (i) a multimodal graph refiner that infers and corrects edges using cross-modal semantic evidence, and (ii) a coupled cross-modal message passing mechanism that performs joint intra- and inter-modality propagation over the refined graph. This unified design yields greater expressiveness than decoupled or two-stage formulations and allows dynamic interaction between topology and representation learning. Across diverse benchmarks for node classification and link prediction, ReCoG consistently outperforms strong multimodal graph structure learning baselines, including graph foundation models. Our results demonstrate that reciprocal co-evolution of structure and semantics is important for effective multimodal graph learning, challenging the prevailing separation between topology and representation learning. --- ### OmicSync: Reliability-Aware Spatial Multi-Omics Clustering with Evidence-Constrained LLM Reasoning Authors: Rabeya Tus Sadia, Qiang Ye, Qiang Cheng Published: 2026-08-24 Categories: cs.CV arXiv: https://arxiv.org/abs/2608.22785v2 Spatial multi-omics technologies jointly profile gene expression, surface proteins, and histology at each tissue spot, yet most spatial domain discovery methods provide only cluster assignments, without indicating assignment reliability, modality contributions, or why a domain decision should be trusted. We present OmicSync, a reliability-aware spatial multi-omics framework that couples unsupervised domain clustering with evidence-constrained LLM reasoning using model-derived per-spot signals, including assignment confidence, epistemic routing uncertainty, and modality-routing weights. These signals are converted into structured evidence dictionaries and used to generate standard, stepwise, counterfactual, contrastive, and uncertainty-focused explanations. OmicSync integrates a KAN-GCN backbone with spatial encoding, cross-modal fusion, uncertainty-aware routing, cell-type supervision, and missing-modality imputation. We further introduce OmicSync-R, which closes the reasoning-clustering loop by using automatically computed reasoning-quality scores as REINFORCE rewards, allowing reasoning coherence to shape the latent structure without backpropagating through the language model. Across four 10x CytAssist FFPE spatial proteomics benchmarks, OmicSync achieves the best average rank on Human Tonsil (1.44), Glioblastoma (1.78), and Tonsil Add-on (1.22), and second-best on Human Breast Cancer (2.33). OmicSync-R further improves ARI on Human Breast Cancer from 45.73 to 46.72 and outperforms existing methods on six of nine clustering metrics. Together, OmicSync and OmicSync-R enable reliability-aware, spot-level auditable spatial domain discovery guided by evidence-constrained reasoning. --- ### Neural Operator based Multi-Field Reconstruction of Inner Solar Boundary State Authors: Vignesh Kumar Pandian Sathia, Reza Mansouri, Dustin J. Kempton, Pete Riley, Rafal A. Angryk Published: 2026-08-24 Categories: cs.LG, astro-ph.IM, astro-ph.SR, cs.CV arXiv: https://arxiv.org/abs/2608.22782v1 The Solar wind is a continuous flow of charged particles emanating from the solar surface and governed by complex, interacting magnetohydrodynamic processes. Accurate specification of inner-boundary conditions is essential for heliospheric modeling and solar-wind prediction. In many practical applications, only a subset of interacting multi-field variables is directly available, but for a comprehensive view of solar wind prediction and downstream magnetohydrodynamic simulations, a more complete boundary state is required. In this work, we study the problem of learning the multi-field multi-scale solar magnetohydrodynamic state at 30 solar radii ($R_\odot$) using operator learning. Specifically, given the radial velocity and radial magnetic field, we aim to reconstruct the non-radial velocity and magnetic field components, radial and non-radial current density, thermodynamic density, and pressure components. This mapping is highly nonlinear, spatially coupled, and multi-scale, making it a challenging task for data-driven scientific machine learning. To address this problem, we employ a Local Neural Operator (LocalNO) that learns mappings between input and output function spaces while retaining locality and resolution-awareness. Unlike conventional regression models and autoencoder models, neural operators are better suited for learning structured field-to-field transformations arising from physical systems. The resulting predictions along with inputs are intended to serve as boundary condition variables for future inner-heliospheric modeling pipelines. --- ### Can We Perform Online RL for Image Editing without Editing Rewards? Authors: Qichao Ma, Jikang Cheng, Ling Liang, Zhaofei Yu, Tiejun Huang, Renye Yan Published: 2026-08-24 Categories: cs.CV arXiv: https://arxiv.org/abs/2608.22780v1 Reinforcement learning (RL) enables direct preference optimization for image editing through editing-specific rewards, which remain less developed due to costly triplet supervision and complex task-dependent calibration. In contrast, text-to-image (T2I) generation benefits from a mature and diverse reward ecosystem spanning semantic alignment, aesthetics, realism, glyph shape, and other visual preferences. Extending this ecosystem to image editing would substantially broaden the range of visual preferences accessible to RL-based optimization, prompting the central question: \emph{Can We Perform Image Editing RL without Editing Rewards?} In this paper, we argue that the standard image editing dimensions have potential to be mapped to the T2I reward space: image quality can transfer directly, prompt following can be aligned through a description of the desired visual state, and reference consistency admits a coarse semantic conversion by encoding the source content to preserve. However, editing instructions specify relative changes, whereas T2I rewards require self-contained target descriptions; moreover, semantically valid captions from generic vision-language models may be incompatible with the frozen reward. Hence, we further introduce Lever-Edit, a two-stage framework that learns a reward-aligned captioner for counterfactual target descriptions, freezes it, and optimizes the editing policy solely with the transferred T2I reward. Experiments show competitive editing alignment and source preservation against editing-reward-based fine-tuning, while outperforming intuitive transfer baselines. --- ### LagrangeGS: Non-Conservative Lagrangian System on Dynamic 3D Gaussian Splatting Authors: Shogo Sato, Takuhiro Kaneko, Shoichiro Takeda, Tomoyasu Shimada, Riku Inoue, Kazuhiko Murasaki, Ryuichi Tanida Published: 2026-08-24 Categories: cs.CV arXiv: https://arxiv.org/abs/2608.22773v1 Dynamic 3D Gaussian Splatting (3DGS) achieves photorealistic reconstruction of time-varying scenes, and recent physics-aware extensions improve extrapolation by explicitly predicting velocity fields. However, these extensions merely fit vector fields to visual deformations without satisfying Lagrangian mechanics, leading to three major issues: (i) physically inconsistent trajectories, (ii) lack of time-reversibility, and (iii) geometric collapse during long-term extrapolation. In this paper, we propose LagrangeGS, which formulates dynamic 3DGS as a non-conservative Lagrangian system. While this Lagrangian formulation fundamentally solves (i), a direct application of general LNNs to dynamic 3DGS requires a large velocity-Hessian inversion for millions of Gaussian particles. To overcome this computational bottleneck, we approximate the velocity-Hessian as an identity matrix, decoupling particle dynamics for computational tractability. For (ii), we restrict the non-conservative forces to be explicitly time independent, enabling consistent backward integration. Finally, to address (iii), we introduce local rigid alignment that regularizes particle trajectories. Extensive evaluations on dynamic scene benchmarks demonstrate that LagrangeGS enables stable long-term extrapolation, consistent time reversal, and counterfactual physics-based editing without retraining. --- ### SPOC-SQL: Stage-wise Preference Optimization for Controllable Text-to-SQL Authors: Yingnan Chen, Chun Ding, Tianshi Xu, Xu Yang, Si Wu Published: 2026-08-24 Categories: cs.CL arXiv: https://arxiv.org/abs/2608.22772v1 Text-to-SQL aims to translate natural language questions into executable SQL queries over relational databases, requiring multi-stage structured reasoning over database schemas and query constraints. However, existing methods treat this task as single-step generation, where models optimize entire SQL sequences without targeted feedback at key decision points and lack support for interacting with and controlling the intermediate generation process. To address this issue, we propose SPOC-SQL, which decomposes Text-to-SQL into four sequential subtasks following standard SQL execution logic and designs stage-specific optimization strategies for the model to learn key decisions. Specifically, we propose the implementation of fine-grained preference optimisation at key decision points across SQL stages, with the objective of enhancing structured decision-making during query construction. Furthermore, a structured decomposition strategy is designed, facilitating stage-wise intervention and correction through explicit intermediate representations. This results in more controllable and reliable SQL generation. Experiments demonstrate that incorporating stage-wise human knowledge consistently improves performance, validating the effectiveness of stage perception controllable generation. --- ### DelistBench: Evaluating Search-Enabled LLMs for Auditable Corporate-Event Database Completion Authors: Xuan Yao, Shuping Li, Yang Dai, Yi Zhou, Ke-Wei Huang Published: 2026-08-24 Categories: cs.CL arXiv: https://arxiv.org/abs/2608.22770v2 Financial institutions need an independent way to detect missing, stale, and misclassified corporate-event records in vendor databases. We introduce Search-to-Record, a database-assurance task in which search-enabled large language models reconstruct institution-defined event records from public sources for a known security universe and historical cutoff, and DelistBench, a 1,200-record benchmark for security-level delisting announcements. We evaluate five models in paired closed-book and web-enabled conditions. Web access raises announcement-date accuracy within seven days by 34.0 to 48.0 percentage points and event-status accuracy by approximately 2.8 to 21.7 points; the best system achieves 81.5% overall joint accuracy within seven days. Economy web systems achieve 75.9-78.3% overall joint accuracy within seven days at 4.5-6.6% of the API cost of the most expensive web system. Risk-based triage identifies low-error subsets, although the highest-coverage operating point still sends 27.3% of the balanced test set to review. The evaluation identifies web retrieval as the main source of timing gains and shows that low-cost systems can approach the best system's accuracy. Together, Search-to-Record, DelistBench, and the evaluation provide concrete deployment guidance: calibrate triage to local event prevalence and market mix, preserve positive-event recall, and route positive and ambiguous cases to targeted review. --- ### The Retriever Should Remember: Experience-Amortized Reranking for Long-Term Agent Memory Authors: Qi Feng, Chris Ding, Jicong Fan Published: 2026-08-24 Categories: cs.AI arXiv: https://arxiv.org/abs/2608.22767v1 Long-term language-model agents accumulate memories across interactions, but their retrievers typically do not accumulate retrieval experience. Semantic retrieval is efficient, but embedding similarity does not always reflect whether a memory contains evidence relevant to the current query. Large language model (LLM) rerankers provide stronger query-conditioned relevance scores, yet stateless reranking repeatedly scores a large candidate pool and discards these scores after each query. We introduce EARM, an experience-amortized reranking framework that treats previously acquired LLM relevance scores as reusable retrieval experience. EARM stores sparse query--memory relevance scores in an online matrix, learns their shared structure through causal matrix completion, and combines a small set of newly observed scores with estimated scores to rerank the remaining candidates. The scoring budget decreases as experience accumulates, changing LLM reranking from a repeated per-query expense into a retrieval capability learned over an agent's lifetime. Experiments on long-term conversational memory show that mixed observed-and-estimated reranking improves answer accuracy over semantic retrieval by up to 6.62% and remains effective when only 17.5% of candidates receive direct LLM relevance scores, thereby substantially reducing the inference overhead of LLM reranking. These results motivate a broader view of agent memory: a long-lived agent should remember not only past content, but also how that content has proved useful for retrieval. --- ### Learning to Control Coupled-Dynamics Environments with Joint Markov Decision Processes Authors: Ege C. Kaya, Aliasghar Pourghani, Mahsa Ghasemi, Vijay Gupta, Abolfazl Hashemi Published: 2026-08-24 Categories: cs.LG, math.OC arXiv: https://arxiv.org/abs/2608.22765v1 Coupled-dynamics environments expose the one-step outcomes that would follow from several possible counterfactual actions under a common realization of exogenous randomness. The ordinary Markov decision process formalism allows one to reason about the marginal law of each action but discards dependence across these counterfactual outcomes. The Joint Markov decision process (JMDP) formalism preserves that dependence. Prior work established the formalism and solved the fixed-policy joint moment evaluation problem in JMDPs. This paper develops optimal-control methods. We define a nonparametric distributional Bellman optimality operator for JMDPs, and prove that when the induced marginal MDP has a unique optimal policy, its iterates converge in Wasserstein distance to the optimal joint return law. For the first two moments, we establish convergence under a weaker condition that permits several mean-optimal actions as long as their tie resolutions share a second-moment fixed point. We also derive sampled targets for neural approximation. --- ### LpWM: A Case for Sparse Representations in World Models Authors: Yilun Kuang, Yash Dagade, Quentin Le Lidec, Lucas Maes, Randall Balestriero, Yann LeCun Published: 2026-08-24 Categories: cs.LG arXiv: https://arxiv.org/abs/2608.22764v1 Joint-embedding predictive architectures (JEPAs) learn latent dynamics for planning and avoid representation collapse by matching features to maximum-entropy distributions such as isotropic Gaussians, yielding dense representations. However, it is unclear whether dense representations are the most favorable geometry for modeling dynamics. In this work, we ask whether a different geometry, sparse representations, can make action-conditioned latent dynamics easier to model, and what dynamical structure emerges from such representations. We first show that nonlinear Lipschitz dynamics can be approximated arbitrarily well by action-conditioned linear dynamics in a sufficiently high-dimensional one-hot latent space, with rollout error vanishing as the dimension grows. This motivates distributed sparse representations as a practical relaxation of one-hot sparsity. We introduce LpWorldModel (LpWM), a JEPA model regularized with Rectified Distribution Matching Regularization (RDMReg) to match encoder features to a Rectified Generalized Gaussian distribution, yielding non-negative sparse codes. Empirically, sparsity lowers the predictor complexity required for successful planning: on PushT, sparse LpWM outperforms dense LeWM by up to 57% in planning success at intermediate predictor capacities. This advantage also extends beyond Gaussian distribution matching, with LpWM outperforming dense VICReg representations across multiple predictor families. We further find that the learned sparse representations are mode-factored, with support encoding discrete dynamical regimes and feature magnitudes capturing continuous within-regime state. Together, these results suggest that sparse representations can reduce the predictor complexity required for control while revealing interpretable structure. --- ### Compositional Chain-of-Relations for Faithful Knowledge Graph Question Answering with Large Language Models Authors: Chenhui Liu, Jianpeng Zhou, Jiahai Wang Published: 2026-08-24 Categories: cs.AI arXiv: https://arxiv.org/abs/2608.22762v1 Knowledge graph question answering (KGQA) is a key task for evaluating KG-augmented Large Language Models (LLMs), and complex KGQA that requires multi-hop reasoning is especially challenging. Solving a complex query involves two coupled phases: candidate retrieval, which locates answer candidates over the KG, and constraint handling, which filters these candidates against the query constraints. Faithful reasoning requires grounding both phases in the KG. However, existing agent-based methods ground candidate retrieval through entity-centric exploration, while leaving constraint handling to the LLM's internal knowledge, which leads to two critical limitations. (1) Unreliable entity pruning: entity-centric exploration uses entities as search units and must prune them to a fixed-size subset at each hop. Because entity information in KGs is often incomplete and a fixed-size subset cannot retain all valid entities, such pruning inevitably drops valid entities and ultimately leads to wrong answers. (2) Ungrounded constraint handling: query constraints are resolved from the LLM's internal knowledge rather than the KG, leaving the final answers unverifiable and prone to hallucination. To address these limitations, this paper introduces a relation-centric exploration paradigm, which uses relations rather than entities as search units and thus avoids unreliable entity pruning. Built on this paradigm, this paper proposes Compositional Chain-of-Relations (CCoR), a simple and effective framework that grounds both phases in the KG with two relation chains: a main chain for candidate retrieval and a constraint chain that verifies query constraints through explicit KG exploration. Experiments on four KGQA benchmarks show that CCoR consistently improves accuracy, faithfulness, and efficiency over strong baselines, with more pronounced gains on complex queries. --- ### Don't Repeat Yourself: Stopping Verbatim Loops at Sampling Time Authors: Philipp Emanuel Weidmann, Allen Roush, Judah Goldfeder, Sanjay Basu, Ravid Shwartz-Ziv Published: 2026-08-24 Categories: cs.CL, cs.AI, cs.LG arXiv: https://arxiv.org/abs/2608.22761v1 Large Language Models generate text autoregressively, but open-ended generation is prone to verbatim looping, in which models repeat spans already present in context. Standard defenses such as repetition, presence, and frequency penalties and n-gram blocking act on token recurrence rather than the sequential structure of a loop, and often suppress looping only at strengths that also degrade formatting or fluency. We propose Don't Repeat Yourself (DRY), a sampling-time logit adjustment that penalizes a candidate token only when generating it would extend the current suffix into an exact continuation of a span seen earlier in the context. Sequence breakers protect chat templates and formatting tokens. Across models from 1.5B to 120B parameters, nine prompt families, and a 600-pair human study, DRY reduces suffix-extension rate by 47% while improving lexical diversity. An intervention-matched placebo produces no comparable reduction, identifying suffix matching as the operative mechanism. On AWQ-quantized 70B and 120B models, DRY reduces loop rate by roughly half while preserving MT-Bench, MMLU, and GSM8K performance, whereas standard alternatives lose measurable ground. DRY has been adopted by popular open-source LLM inference frameworks including llama.cpp, ExLlamaV2, and text-generation-webui, highlighting its practical impact on text generation. --- ### ByteAction: Byte-space Action Recognition Foundation Model Authors: Fangcheng Li, Zhen Yu, Kejun Wu, Qiong Liu, You Yang Published: 2026-08-24 Categories: cs.CV arXiv: https://arxiv.org/abs/2608.22760v1 Byte-space Action Recognition (BAR) aims to recognize human actions directly from compressed image bitstreams without any pixel decoding. By operating entirely in byte space, BAR is inherently independent of file integrity and pixel-level reconstruction, making it naturally applicable to privacy-sensitive scenarios and robust against bitstream corruption. In this paper, we propose ByteAction, a BAR foundation model that achieves accurate action recognition on corrupted image bitstreams. ByteAction follows a dual-view byte-level recognition framework. It constructs weakly and strongly corrupted bitstream views, which are augmented by Bitstream Pattern Augmentation (BPA) and encoded with a shared ByteFormer backbone. The model is optimized with both classification and corruption consistency objectives. Specifically, we propose Bitstream Pattern Augmentation (BPA), which reshapes one-dimensional byte sequences into two-dimensional byte matrix and applies region-level erasure to encourage the model to learn robust cross-region byte dependencies. We further propose a Corruption Consistency Training strategy that constrains the model to maintain stable predictions across different corruption severities through bidirectional KL divergence. Experiments on the image bitstream from Stanford40, PPMI, and PASCAL VOC 2012 Action demonstrate that ByteAction achieves state-of-the-art corruption robustness across all scenarios while maintaining competitive intact bitstream performance. --- ### XTC: Head-Aware Sampling by Excluding Top Choices Authors: Philipp Emanuel Weidmann, Allen Roush, Judah Goldfeder, Sanjay Basu, Ravid Shwartz-Ziv Published: 2026-08-24 Categories: cs.CL, cs.AI, cs.LG arXiv: https://arxiv.org/abs/2608.22758v1 Standard decoding rules for autoregressive language models promote diversity by rescaling the full next-token distribution or truncating its low-probability tail. These strategies overlook a common regime of open-ended generation in which several continuations are plausible but too much probability mass remains concentrated on the most generic choice. We introduce XTC (Exclude Top Choices), a lightweight head-aware decoding operator that targets this regime directly. XTC identifies tokens whose probabilities exceed an absolute plausibility threshold $τ$: when at least two qualify, it removes the dominant eligible choices with probability $ρ$ and retains only the weakest plausible alternative before renormalization. Across 60 experiments on Gemma 3 27B Q4, Gemma 3 12B Q6, and DeepSeek R1 14B Q6, with scaling validation on Llama 3.3 70B Q4, XTC improves the diversity-repetition Pareto frontier. On creative generation, Distinct-2 increases by 11--15% and repeat trigrams decrease by 27--47% across the four models. Combined with temperature scaling, gains reach 38% in Distinct-2 and 71% in repeat-trigram reduction over baseline. A blinded Amazon Mechanical Turk study with 150 Master raters yields a 62.3% creativity preference for XTC ($p<10^{-4}$) without reduced fluency, while a GPT-4o control judge reproduces the Anthropic-judge direction on every measure. On IFEval with Llama 3.3 70B Q4, XTC preserves prompt-level strict accuracy within 1.7 percentage points of baseline while recovering most of the diversity gain; a temperature setting matched on Distinct-2 reduces IFEval by 8.8 points. The effect is additive with temperature and repetition penalties, robust across quantization levels and model families, and consistent across twelve prompt genres. XTC has been adopted by llama.cpp, ExLlamaV2, and text-generation-webui. ---