Tabular synthesis is critical for privacy-preserving sharing and augmentation, yet diffusion models rely on implicit mechanisms to capture inter-column relationships. We introduce Geometry-Aware Tabular Diffusion (GATD), which augments tabular diffusion denoisers with pairwise angles and lengths computed from column value differences and used as inputs and auxiliary targets. Our MLP instantiation achieves state-of-the-art benchmark performance while using 3.5x fewer parameters on average (up to 25x for classification tasks): on ten datasets, it wins 8/10 Shape, 7/10 Trend, and 9/10 downstream utility (F1/RMSE), reducing Shape and Trend error by 27% and 20%. Default loss weights transfer to GNN and Transformer denoisers, improving Shape on 27/30 and Trend on 25/30 architecture-dataset cells. A matched ablation shows supervision (not extra inputs or capacity) drives the gain. This shows explicit relational supervision is a portable inductive bias for tabular diffusion.

Large Language Models (LLMs) are increasingly deployed as continuously evolving services, where frequent base-model updates may invalidate previously deployed task-specific Low-Rank Adaptation (LoRA) adapters. For service providers managing numerous downstream model services, retraining each LoRA adapter from scratch for every updated base model is computationally prohibitive and delays service rollout. Meanwhile, the simpler alternative, i.e., naively applying the original LoRA adapter to the updated base model, often leads to degraded service quality due to adapter-backbone incompatibility. To address this problem, we propose ReLoRA, a knowledge-reusing re-adaptation framework that efficiently restores service-ready LoRA adapters for evolving LLM services while preserving or improving task performance. Specifically, ReLoRA comprises two key optimization steps: 1) Adaptive LoRA initialization leverages Bayesian optimization to construct a compatibility-aware starting point by fusing information from both the previously deployed task adapter and the base model's evolution; 2) Fine-tuning with scheduled regularization first rapidly steers the adapter to a high-quality region via strong regularization, followed by relaxed regularization for task-specific refinement. This design enables rapid service-quality recovery with reduced re-adaptation overhead. Extensive experiments demonstrate that ReLoRA reduces time-to-readiness by up to 8.9$\times$ and improves accuracy by up to 4.6\% compared to baselines.

Coronary artery stenosis is a common cardiovascular disease, with severe, untreated cases posing significant risks of heart attack. Although coronary (X-ray) angiograms remain the standard for stenosis diagnosis, they are invasive, time- and resource-intensive, and therefore only performed on patients with a high probability of disease based on symptoms and prior clinical tests. However, a subset of patients, especially those without symptoms, may remain undiagnosed. Detecting indications of stenosis from ECGs, which are fast, cheap, non-invasive, and thus routinely acquired even in asymptomatic patients, would support early diagnosis. However, as no reliable stenosis-specific signal has been identified in ECGs, they can not currently be used for stenosis risk stratification. To address this, we introduce StenCE, a pretraining framework, allowing stratification of patients based on features derived directly from ECGs. Evaluations across varying stenosis severity thresholds and additional ECG disease classification tasks demonstrate consistent performance improvements across different ECG encoders, outperforming previous work. The obtained models successfully detect signals for stenosis diagnosis in ECGs and are the first to achieve high performance in severe stenosis classification. The source code is available at https://github.com/NikolaCenic/ecg-stenosis-cls.

ESG and climate risk data remain fragmented across heterogeneous Scope 1, Scope 2, and Scope 3 reporting environments, while conventional validation pipelines lack provenance aware auditability, hidden drift detection, and reproducibility oriented governance. This paper proposes a deterministic climate risk intelligence framework integrating single source of truth orchestration, temporal anomaly detection, imbalance aware ensemble learning, and explainability oriented governance for auditable ESG validation. To support open reproducibility, we construct and release a synthetic ESG validation benchmark calibrated against publicly reported characteristics of the GHG Protocol, PCAF, and ISSB standards. The methodology incorporates temporal drift analysis, SMOTE based rare event optimization, ensemble learning, provenance aware orchestration, and TreeSHAP based interpretability for governance inspection and audit reconstruction. We evaluate the framework against statistical classifiers, anomaly detection methods, temporal forecasting baselines, and a threshold based system using classification metrics (recall, F1, ROC AUC), calibration metrics (ECE, Brier score), and a governance oriented audit trace completeness metric measuring the fraction of flagged anomalies for which a deterministic source to escalation provenance chain can be reconstructed. Results are reported as mean and standard deviation across stratified five fold cross validation with paired significance testing. The framework reframes ESG reporting toward deterministic climate risk governance infrastructure supporting reproducibility, explainability, and operational auditability.

Camouflaged object detection has improved substantially, but most standard benchmarks evaluate models only on clean images. This is not realistic because real cameras often capture blur, sensor noise, weather effects, and compression artifacts. We present COD10K-C, a corruption robustness benchmark based on COD10K. It includes 8 corruption types and 5 severity levels, giving 40 conditions and 81,040 evaluation pairs in total. We evaluate three popular camouflaged object detection models, SINet-v2, PFNet, and ZoomNet, as well as a lightweight model called RobustCODLite. All models show clear performance drops on corrupted images. Motion blur and Gaussian blur cause the largest drops, with SINet-v2 losing 18.5 Dice points under motion blur. Brightness and fog are less harmful. RobustCODLite uses corruption augmentation, a frequency-prior branch, and an uncertainty-consistency loss. It retains 92.3% of its clean Dice score under corruption, compared with 87.7% for SINet-v2, 84.8% for ZoomNet, and 84.1% for PFNet. On the hardest corruptions, RobustCODLite matches or outperforms models that perform better on clean data. We will release the COD10K-C GitHub repository to support future research in robust camouflaged object detection.

In computational pathology, Whole Slide Images (WSIs) survival analysis is crucial for patient prognosis assessment, but it faces multiple technical challenges. Although the Transformer captures long-range dependencies through its self-attention mechanism, its $O(N^2)$ time complexity causes a severe computational bottleneck in large-scale WSIs graph structures. The Mamba model breaks through the Transformer's computational bottleneck with linear complexity. But, owing to Mamba's high sensitivity to the order of input data, traditional node sorting methods in Graph Mamba, such as those based on node degree or subgraph size, fail to adequately account for the topological connectivity of graph data. This inadequacy consequently restricts the performance of Mamba's sequential modeling. Moreover, its unidirectional architecture cannot leverage the bidirectional spatial structure of images. To address these challenges, this paper proposes a novel Graph Mamba survival analysis framework based on topology-aware ordering (TopoMamSurv) to adapt to the sequential sensitivity of Mamba. Our visualization experiments further confirmed that the nodes extracted through the topology-aware ordering (TAO) strategy indeed exhibit higher similarity. Furthermore, we designed a bidirectional Mamba module and integrated a Graph Convolutional Network (GCN) to achieve bidirectional spatial context modeling of images, forming a hierarchical feature learning architecture for "local aggregation - global capture." This framework effectively reconciles the contradiction between long-range dependency modeling, computational efficiency, and spatial structure utilization in WSIs analysis through its systematic design of TAO, bidirectional semantic modeling, and hierarchical feature fusion. This framework has been validated for its comprehensive performance advantage on five TCGA datasets.

Within-dataset class-split evaluation is widely used as a proxy for fully unconditional out-of-distribution anomaly detection. We show that this protocol can become ill-posed when the held-out anomaly class overlaps the normal mixture in representation space. In this regime, anomaly scores may collapse toward chance or even invert, and the preferred score direction can depend on the unknown anomaly class. We introduce a simple training-free diagnostic, neighborhood class leakage, and show that it predicts score-direction instability across Fashion-MNIST, CIFAR-10, and Imagenette, in both pixel and VAE latent spaces. Our results suggest that class-split AD benchmarks should be treated as geometry-dependent stress tests rather than unconditional evidence of anomaly-detection ability.

Accurate and generalizable estimation of cognitive workload from electroencephalography (EEG) is critical for human-centered and safety-critical systems. Although EEG is widely used for workload assessment, the consistency of region-level EEG contributions across tasks, datasets, and subjects remains unclear. This paper presents a region-level evaluation framework for EEG-based workload prediction in which models are trained and evaluated using features extracted exclusively from electrodes belonging to anatomically defined scalp regions. We perform a large-scale analysis across four publicly available EEG workload datasets spanning diverse task demands, recording hardware, and electrode montages. Region importance is quantified using a model-agnostic, performance-based approach under both mixed-subject and subject-independent evaluation protocols, with results aggregated using a rank-based strategy to ensure robustness across experimental configurations. Across all datasets and subject-independent evaluations, frontal electrode groups outperform the full-scalp baseline by approximately 15-20% in relative rank position while using substantially fewer electrodes. Fronto-central regions exhibit the most stable predictive utility, whereas posterior and occipital regions contribute less consistently across experimental conditions. These findings indicate that workload-relevant EEG information is most consistently retained within frontal and fronto-central electrode groups, supporting the design of efficient and generalizable EEG-based workload monitoring systems.

The development of brain-computer interfaces (BCIs) based on electroencephalograms (EEGs) has advanced significantly mainly to machine learning. Although the majority of earlier research has been on increasing classification accuracy, relatively little focus has been placed on security and robustness. According to recent research, EEG-based BCIs are susceptible to adversarial attacks, which can cause misdiagnosis due to minute, well-crafted disturbances. Evaluating model robustness against such perturbations is therefore critical for ensuring reliable deployment. In this study, we propose a lightweight custom Convolutional Neural Network (CNN) architecture to investigate adversarial robustness in EEG-based BCIs. The suggested method is assessed using two EEG datasets and contrasted with three novel CNN models tailored to EEG, namely EEGNet, DeepConvNet, and SleepEEGNet, under gradient-based adversarial attack scenarios. According to experimental findings, the suggested model continuously performs better in classification under adversarial perturbations compared to baseline models, indicating improved robustness. These findings highlight the potential of lightweight architectures for enhancing the reliability of EEG-based BCI systems under adversarial conditions.

The curvature exponent $α$ in $h_k \propto σ_k^α$ -- governing how Hessian eigenvalues scale with gradient singular values -- varies systematically across layer types ($α\approx 2$ for convolutions, $\approx 1$ for transformer attention, $< 1$ for MLP up-projections). Why? We prove the Spectral Alignment Decomposition: $α= 2 + d\logΦ_k / d\logσ_k$, where $Φ_k$ measures alignment between Kronecker factor eigenbases and gradient singular directions. This reduces "why does $α$ vary?" to a geometric question we answer for LayerNorm, residual connections, and softmax heads. The decomposition implies a spectral transfer identity $s = αγ$ linking curvature exponent, effective gradient rank-decay $γ$, and Hessian decay exponent $s$. The identity is algebraic; its empirical content is that $α$ and $γ$, fit on independent data (HVPs vs. SVD), recover $s$ to ~2% median error across 93 layers, five architectures, and three datasets -- with no free parameters. A zeta-function bound on participation ratio shows curvature concentrates onto effectively one direction per layer. As a proof of concept, we derive the architecture-adaptive preconditioner $T(σ;α)$ and show that Spectral Newton -- implementing $T$ in the gradient singular basis -- outperforms AdamW on vision benchmarks where $α\approx 2$.

Dynamic pricing in short-term rental (STR) markets presents a distinctive challenge for online learning algorithms: pricing decisions carry significant financial risk, operators require explainability, and market feedback is sparse (one booking outcome per listed night). We introduce the Human-in-the-Loop Gated Bandit (HITL-GB) framework, in which a contextual bandit algorithm generates price recommendations but a human agent retains authority to accept, modify, or reject each recommendation before it is applied. We show that under this approval constraint, historical pricing data -- collected under a prior deterministic policy -- is structurally equivalent to on-policy warm-up data for initialising the bandit's posterior, bypassing the weeks-to-months cold-start period that renders pure online bandit learning impractical in sparse-feedback markets. We formalise the approval-gated reward signal, derive a regularised ridge-regression warm-up procedure from historical episodes, and validate the approach on real STR production data (anonymised urban market, 2 rooms, April 2022 -- April 2026, 1,461 nightly pricing episodes). Our warm-up procedure compresses effective cold-start from ~150 episodes to ~30 episodes when initialising agents from the Hierarchical Factored Thompson Sampling (HF-TS) family. We further argue that the structural equivalence result is domain-agnostic: any high-stakes domain where human approval is legally or operationally required -- including clinical drug dosing, credit origination, content moderation, and radiological diagnosis -- satisfies the same conditions and benefits from the same warm-up strategy. In regulated industries, mandatory human oversight is thus a statistical asset rather than a deployment constraint.

While video streaming understanding has made significant strides, real-world applications, such as live sports broadcasting, autonomous driving, and multi-screen collaboration, inherently demand continuous, multi-stream interactions. However, existing benchmarks are confined to single-stream paradigms, leaving a critical gap in evaluating online, cross-stream reasoning. To bridge this, we introduce X-Stream, the first benchmark dedicated to multi-stream streaming understanding. Comprising 4,220 rigorously curated QA pairs across 932 videos, X-Stream evaluates 11 subtasks across multi-window, multi-view, and multi-device scenarios. Crucially, our dataset is constructed using a novel dual-verification pipeline that prevents over-reliance on a single stream. Furthermore, we pioneer the conceptualization of multi-modal large language models (MLLMs) as naive multiplexers, systematically evaluating their performance through the lens of Signal Multiplexing Theory. Our extensive online inference experiments reveal a stark reality: state-of-the-art MLLMs struggle significantly with concurrent streams, achieving only about 50% score and exhibiting poor proactive ability. Ultimately, X-Stream exposes the trade-off of current multiplexing schemes, providing both a practical evaluation protocol and empirical guidance for next-generation multi-stream agents.

Language agents spend substantial inference time solving individual tasks, yet the experience acquired in one episode is often underutilized in future episodes. Continual learning expects an agent to accumulate reusable experience across a stream of tasks, improve over time, and avoid interference from irrelevant experiences. Unfortunately, existing benchmarks struggle to evaluate continual learning in language agents rigorously. Most efforts focus on retrieval and reasoning over long-context conversations or documents, while recent lifelong-adaptation benchmarks often rely on naive task streams with limited analysis of cross-task relationships, making it difficult to understand what an agent learns and reuses over time. This paper presents an evaluation framework AgentCL for continual learning in agents, centered on controlled task streams and metrics for transfer gains. AgentCL constructs compositional streams where earlier sub-solutions, evidence, or workflows are intentionally reusable in later tasks, and contrasts them with naive streams where such reusability is not guaranteed. We use the benchmark to evaluate non-parametric memory designs for continual learning. To diagnose how memory design choices affect continual learning, we develop MemProbe, a probing method that stores interactions, insights, and skills, while filtering unreliable experiences during consolidation. Empirical analysis across coding, deep research, and language understanding/reasoning tasks shows that naive streams offer limited ability to distinguish memory designs, whereas controlled streams more clearly distinguish their plasticity. Meanwhile, naive and held-out settings often yield limited gains and can expose memory-induced degradation. These results highlight the need for stronger memory designs that balance plasticity and stable reuse.

Combining attention's global retrieval with the sequential importance signal of state space models (SSMs) is the open challenge of hybrid language modeling. Transformers see everywhere but cannot prioritize; SSMs know what matters but cannot revisit. Existing hybrids -- Jamba (block level) and Hymba (head level) -- place the two in separate compartments, so neither informs the other during the attention computation itself. We propose SISA (SSM-Informed Softmax Attention), which adds an SSM-derived importance term directly inside the attention score and realizes the full operation as a single SDPA call on augmented query/key vectors -- no recurrent state, no custom kernel. At 152M / 5B tokens, SISA reaches LAMBADA-greedy 17.3% (vs. Transformer 13.9 and Mamba-3 15.5) and attains NIAH 100% from step 1K, 7x faster than Transformer's retrieval convergence; at 369M, Mamba-3 leads LAMBADA while SISA preserves perfect NIAH and stock-SDPA execution. SISA thus defines a third design axis for SSM-attention hybrids -- score-level fusion -- beyond the block-level and head-level paradigms that have dominated the field.

Agentic reinforcement learning can induce tool abuse, where models overuse external tools even for queries solvable by internal reasoning. Existing approaches mitigate this issue with uniform tool-use penalties or hard limits, which reduce tool frequency but may also suppress useful tool-assisted exploration. We propose EAPO, an Efficient Agentic Policy Optimization framework that learns selective tool use. EAPO introduces tool-free trajectories into each rollout group, applies difficulty-aware reward shaping to penalize redundant tool calls mainly on easier queries, and uses confidence-aware token reweighting to improve policy learning. Across nine mathematical and knowledge-intensive reasoning benchmarks, EAPO consistently improves the accuracy efficiency trade-off on Qwen2.5-3B, Qwen2.5-7B, and Llama3.1-8B. Compared with GRPO, EAPO improves average performance by 10.45%, 7.27%, and 9.69%, while reducing average tool calls by 18.33%, 18.33%, and 24.59%, respectively. These results show that agents can learn when not to use tools without compromising tool-integrated reasoning.

Block diffusion speculative decoding accelerates LLM inference by predicting all tokens within a block simultaneously for the target model to verify in parallel. Predicting an entire block at once requires a sufficiently capable draft model and effective utilization of the target model's internal knowledge. However, the state-of-the-art method DFlash constrains all draft layers to share a single fused representation derived from only a few target layers, limiting per-layer expressiveness and hindering further scaling of draft capacity. In this paper, we present \modelname, which flares out the narrow conditioning bottleneck of DFlash through a lightweight layer-wise fusion mechanism: each draft layer attends to its own learnable combination of a broad set of target layers at negligible overhead, simultaneously injecting richer target knowledge and providing every draft layer with a distinct input. This enhanced per-layer expressiveness enables scaling the draft model to deeper architectures with consistent gains. We further scale training data from 800K to 2.4M samples to fully exploit the enlarged capacity. On six benchmarks spanning mathematical reasoning, code generation, and conversation, \modelname attains average wall-clock speedups of 5.52x on Qwen3-4B, 5.46x on Qwen3-8B, and 3.91x on GPT-OSS-20B, improving over DFlash by roughly 11\%, 8\%, and 5\% respectively. Our code is available at https://github.com/Tencent/AngelSlim.

Diffusion transformer (DiT) has been widely adopted in the generative diffusion field, advancing the denoising of query tokens through attention and Feed-Forward (\text{FFN}) layers. FFN actually acts as the key-value vocabulary for decoding visual contents where the value embeds the visual semantical knowledge. We present that focusing on critical query tokens corresponding to more complex details and encouraging the model to improve these tokens is essential for fine-grained visual generation. To this end, we propose FocusDiT, which applies a Masking scheme to focus on critical query tokens that are exclusively fed into FFN. The masked queries can retrieve visual tokens from the FFN vocabularies, and use them to decode their visual details. Extensive text-to-image experiments validate the effectiveness of token masking in enhancing generative performance.

Deep-research agents solve tasks through long trajectories of search, tool use, evidence inspection, and answer synthesis. Evaluation based on final answers shows whether an agent succeeds, but not which parts of the trajectory make the answer unreliable. We study span-level error localization for deep-research agents. We collect 2,790 real trajectories from two agent frameworks, three backbone models, and three benchmarks, convert raw logs into semantic spans, and annotate harmful error spans through LLM-assisted expert review. From these annotations, we build TELBench, a 1,000-instance benchmark for identifying error spans among normal exploration, failed searches, tentative hypotheses, and harmless noise. We further propose DRIFT, a claim-centric auditing framework that tracks agent claims, checks their support in trajectory evidence, and marks spans where unsupported or conflicting claims affect the answer path. Experiments across model families and auditing frameworks show that DRIFT improves span-level error localization and first-error accuracy by up to 30 percentage points. Our work provides a process-level view of reliability in deep-research agents.

Consumer-price measurement increasingly draws on alternative data sources -- scanner, web-scraped, and transaction/receipt data. A recurring obstacle is that product descriptions in such sources are short, noisy, and abbreviated, with no standard product code, so each item must first be mapped to a consumption classification (e.g., the UN COICOP scheme) before prices can be compared. This paper studies that mapping as a general, reproducible method. The pipeline is: (i) text normalization and tokenization of noisy item names; (ii) a prefix-tree (trie) rule-based pre-classifier driven by per-category key-phrases and stop-phrases; and (iii) a per-category binary confirmation model deciding whether an item belongs to a tentatively assigned category. For labels at scale we use a human-in-the-loop protocol in which annotators give a binary valid/reject judgment, aggregated by a dynamically updated reliability weight; the model joins the same rule, enabling continual fine-tuning. Our empirical finding is deflationary: in a controlled, leakage-free study (one category, real positives vs. hard negatives, five seeds), bag-of-words models essentially saturate the task (F1 about 0.99) -- a linear classifier matches a multilayer perceptron, explicit word-order (n-gram) features add nothing, and about 67 labeled examples already suffice. A Monte-Carlo study of the labeling protocol shows the reliability-weighted vote barely beats plain majority (its additive weights saturate) while Dawid-Skene recovers labels markedly better. We also discuss price-level quality control and design lessons for statistical offices considering transaction data. All figures are illustrative; no confidential data, code, or documentation is reproduced.

Metal surface defect detection is critical for maintaining product quality in industrial manufacturing. However, it faces significant challenges, including limited annotated data, difficulty in identifying subtle multi-scale defects, and poor generalization across diverse scenarios. To address these issues, this paper proposes a novel Contrastive Augmented Transformer (CAT) framework for robust defect detection. CAT employs a hierarchical Swin Transformer backbone and redesigns the feature pyramid network to effectively fuse low-level textures with high-level semantics, enabling precise modeling of subtle and multi-scale defect patterns. To enhance robustness under real-world noise conditions, we propose a domain-specific droplet augmentation algorithm. Furthermore, we incorporate a hard negative mining strategy into the contrastive loss to strengthen the model's discrimination ability in ambiguous defect regions. Experimental results on the KolektorSDD2 dataset demonstrate that CAT achieves a pixel-level AUROC of 99.54%, outperforming existing methods. In addition, CAT exhibits superior generalization and robustness on three unseen datasets, including KSDD1, MTD for tile defects, and MSDD for rail surface defects, demonstrating its potential for wide-scale industrial deployment.

The increasing application of Natural Language Processing (NLP) in healthcare demands language models specifically attuned to the complexities of clinical language. This work introduces KliniskVestBERT, a suite of three BERT-based encoder models pre-trained on a substantial corpus of real-world, de-identified Norwegian clinical texts from Helse Vest. We continue pretraining existing language models Nb-BERT-large, NorBERT3-large, and ModernBERT on our specialized clinical dataset. This dataset is based on a representative population of Helse Vest patients. The included document types are carefully curated to encompass a broad clinical spectrum in bokmål and nynorsk including discharge summaries, surgical reports, nursing notes etc. ensuring comprehensive representation of the linguistic landscape within Norwegian healthcare settings. Evaluation on three synthtetic Norwegian clinical benchmark datasets and two real-world problems demonstrates that each of our clinically specialized models consistently outperforms their baseline counterparts, highlighting the significant benefit of domain-specific pre-training for NLP tasks within the clinical domain. The project was a joint effort by all Helse Vest entities (Helse Bergen, Helse Fonna, Helse Førde and Helse Stavanger) with DIPS under the project lead of Helse Vest ICT.

Learning a good action embedding space is fundamental to scalable robot policy learning, yet existing methods treat action latents as task-specific intermediates rather than first-class representations. The resulting latents are unstructured, embodiment-specific, and weakly tied to motion semantics, limiting interpretability, controllability, and transferability across robots. We position the action embedding space itself as a first-class design target, with downstream policy quality emerging from representation quality. Exploiting motion's intrinsic periodicity, we factorize it into a phase manifold that captures cyclic structure via FFT-parametric coefficients, together with a pose branch that conditions the manifold on non-periodic configuration detail. Combined with motion-semantic distillation, this factorized structure yields a cross-embodiment motion manifold that is interpretable and embodiment-agnostic by design. Anchoring multiple humanoid robots to a shared human-pretrained manifold then produces a unified action embedding space across diverse platforms, achieving strong cross-embodiment retrieval and consistent gains on downstream robot tasks.

Differentially private (DP) text synthesis promises to unlock sensitive corpora for model training, but it remains unclear whether DP synthetic data transmits genuinely new knowledge and capabilities present only in those corpora. This is because existing evaluations rely on tasks that are nearly solvable without training, so strong benchmark performance does not establish that DP synthesis can substitute original data access. Thus, we introduce ContinuousBench, a continuously and automatically-regenerated benchmark that measures capability gain from DP synthetic text. Each quarter, a new release pairs a never-before-seen training corpus with a derived QA set, constructed to be: (1) unsolvable sans-corpus; and (2) learnable under DP, as the tested knowledge is supported by hundreds of independent records. Researchers produce DP synthetic data from the training corpus and run our standardized training and evaluation harness on their synthetic data to measure gains. We instantiate two tracks: Geminon, a procedurally-generated dataset about fictional creatures; and News, a stream of newly crawled public news articles. Although standard benchmarks are nearly saturated, on ContinuousBench we find that non-private synthesis transfers substantial knowledge from the original corpus, while state-of-the-art DP synthesis methods generally fail to do so, even at $\varepsilon=100$.

Electroencephalography (EEG) is the cornerstone of non-invasive brain-computer interfaces (BCIs), yet conventional decoding relies on fragmented, task-specific architectures that severely limit cross-task scalability. While EEG foundation models pre-trained on massive corpora promise universal brain decoding, current post-training depends on task-isolated fine-tuning. This static paradigm restricts knowledge transfer across heterogeneous tasks, hinders model scalability, and incurs computational and storage overheads that scale linearly with task count. To overcome these bottlenecks, we formulate downstream adaptation as a cross-task continual learning problem and propose EvoBrain, a dynamic, task-aware continual learning framework for unified EEG decoding. EvoBrain addresses the plasticity-stability trade-off via two complementary components: (1) Neuro-Spectral Task Normalization (NSN) aligns incoming tasks with historical statistics while recalibrating spectral responses to handle distributional and neuro-spectral shifts; and (2) Response-Affinity Distillation (RAD), combined with time-dependent replay, preserves old-task response geometry and promotes selective knowledge transfer between spectrally compatible tasks, effectively mitigating forgetting. Extensive evaluations across six distinct BCI tasks demonstrate that EvoBrain consistently surpasses state-of-the-art methods across diverse foundation backbones, optimally balancing plasticity and stability. To our knowledge, this work pioneers cross-task continual learning in the EEG domain, advancing the realization of a unified, one-for-all brain decoding system.

As large language models (LLMs) are increasingly used for long-form generation, reliably evaluating long-form outputs has become a critical challenge. LLM-as-a-judge offers a scalable alternative to human evaluation, yet its reliability in long-form output evaluation remains underexamined: existing meta-evaluation benchmarks focus mainly on short-form outputs. Compared with short-form evaluation, long-form evaluation is not merely a matter of output length; it often requires judges to make more complex document-level assessments of overall organization, task-relevant coverage and depth, cross-section consistency, and scenario-specific quality criteria. In this work, we introduce LongJudgeBench, a comprehensive benchmark for evaluating LLM judges on long-form outputs across diverse real-world scenarios and judging protocols. We systematically evaluate a broad range of LLM judges, covering multiple base models and judging settings. Our results reveal a substantial reliability gap: current LLM judges remain unstable across scenarios, and rubrics or references are helpful but not always sufficient. We hope LongJudgeBench will support future research on more robust, context-aware, and human-aligned LLM-as-a-judge methods. Our code is available at https://github.com/cjj826/LongJudgeBench.

Positional encoding (PE) is widely viewed as necessary for transformers to process ordered sequences: without them, the next-token map appears permutation-invariant in its context tokens. This intuition underlies all prior universality results, which rely on positional information to prove that transformers with chain-of-thought can perform arbitrary computation, i.e., they are Turing complete. We revisit this belief in the regime most relevant to long-form reasoning, where generation proceeds through a finite sliding context window. Our opening perception is that the window mechanism itself (mildly) breaks the permutation symmetry. To distill and precisely capture the degree of this added expressiveness, we introduce an abstract autoregressive model, the HIST model, in which each update depends only on constant-size internal state and the token-count histogram within the current window. We prove that this HIST model is Turing complete by showing that the evolution of the window can reveal the token that has just left the window, which suffices to simulate Turing-complete Post machines. We then construct a sliding-window transformer over a constant-size token alphabet, without PE, and show that it can simulate the HIST model. Our result demonstrates that positional encodings are not indispensable for transformers to perform universal computation: The window sliding itself already breaks permutation symmetry and captures sufficient positional information.

Charts are a primary medium for conveying quantitative and relational information, yet systematically evaluating chart parsing models remains difficult. Existing benchmarks focus on narrow chart types and leave diagrammatic structures such as flowcharts and mind maps largely unaddressed, while models produce outputs in incompatible formats, and datasets rarely include the printed or hand-drawn images encountered in practice. To address these issues, we introduce ChartArena, a comprehensive bilingual benchmark covering eight chart families spanning both numeric charts and diagrammatic structures, each evaluated across three visual scenarios: digital renderings, printed photos, and hand-drawn photos. The dataset is built via a human-agent collaborative annotation pipeline with multi-stage human verification to ensure annotation reliability. To enable fair cross-model comparison, we further design a format-agnostic evaluation protocol that maps heterogeneous outputs into two canonical semantic spaces, a normalized triple view and a directed graph view, and scores them with structure-aware metrics. Through extensive evaluation of 26 leading MLLMs, we observe three consistent findings: (i) frontier proprietary models such as Gemini 3.1 Pro lead overall, yet the strongest open-source systems are rapidly closing the gap; (ii) document parsing models handle numeric charts reasonably but fall sharply behind on diagrammatic structures; and (iii) expert chart parsers remain limited to narrow chart families. Across all models, radar charts and hand-drawn scenarios stay especially challenging. These findings show that ChartArena exposes clear capability gaps and provides a unified foundation for future progress. ChartArena is publicly available at https://github.com/pspdada/ChartArena.

We ask: when do Bayesian model averaging (BMA) weights over decision trees carry sufficient epistemic information to justify committed exploitation of the averaging distribution? We answer this question in closed form for Bayesian decision trees (BDTs) with Dirichlet-Multinomial leaf models and a Catalan-exponential tree-size prior (Schetinin&Jakaite, 2025), establishing a complete non-asymptotic theory of rational commitment thresholds.

Transitive inference is the challenge of inferring that A < C from knowing only adjacent relations (A < B, B < C). It is solved by humans and animals not through logical chaining but via an analogue mental number line, whose signature is the symbolic distance effect: distant comparisons are easier than nearby ones. We ask whether Transformers acquire the same primitive, training small models exclusively on adjacent comparisons from a hidden total order and evaluating generalization to unseen distant pairs. We find that out-of-distribution generalization emerges alongside a striking geometric reorganization: entity embeddings collapse onto a one-dimensional manifold whose principal axis recovers the hidden rank order with near-perfect fidelity, and this structure is sensitive to optimization in ways that produce grokking-like transient dynamics. Critically, even when accuracy is at ceiling, decision confidence and geometric separation both scale monotonically with rank distance, directly mirroring the symbolic distance effect observed across decades of behavioural experiments on humans, primates, and rodents. We further show the same rank-aligned geometry in a pretrained large language model, where it tracks the topology of each ordinal relation: linear for sizes and digits, cyclic for months. These results ground a 50-year-old behavioural regularity in the geometry of learned representations, offering a mechanistic account of transitive inference that bridges cognitive science and modern neural networks.

Navigation and manipulation are fundamental capabilities of embodied intelligence, enabling robots to interpret natural language commands and interact physically with their surroundings. However, current Vision-Language-Action (VLA) models remain constrained by task-specific architectures, specializing in either navigation or manipulation, which hinders the development of general-purpose robotic agents. To bridge this gap, we introduce OneVLA, a unified architecture that integrates these distinct tasks into a single, cohesive framework. Specifically, we design a unified action head capable of generating both navigation and manipulation actions without requiring task-specific variants. Furthermore, we propose a multi stage progressive training strategy-incorporating curated data construction and Chain-of-Thought (CoT) fine-tuning that facilitates strong positive transfer and mutual reinforcement between the two domains. Extensive experiments in both simulated and real-world environments demonstrate that OneVLA achieves state-of-the-art performance, significantly outperforming both specialized single-task and existing cross-task models. By unifying these core capabilities, OneVLA paves the way for truly general-purpose robotic systems. The model and source code will be publicly released.