Generalizing deepfake detection across facial manipulation and fully generated video remains challenging due to forgery traces varying with the manipulation method and scene content. Explicit spatial and temporal feature differences offer a way to capture this variation, but also respond to natural texture and motion, motivating their joint interpretation with appearance features. We investigate whether combining multiscale feature differences with adapted appearance representations improves detection across both forgery categories. We introduce Multiscale Appearance-Residual Streams (MARS), a two-path video detector. The residual path applies fixed spatial and temporal difference-of-Gaussians operators to frozen intermediate features and models the resulting differences alongside appearance context across frames. The semantic path adapts the encoder's normalization parameters for frame classification, and a clip-dependent gate combines their predictions. Under a shared protocol across seven datasets, MARS achieves the highest average AUROC of 91.43\%, outperforming seven compared detectors. When trained on facial manipulations alone, MARS achieves 93.99\% AUROC on GenVideo and 75.50\% on DF26, demonstrating transfer to generated videos without generated-video training data. MARS also provides signed patch and frame contributions, whose faithfulness we evaluate through controlled spatial and temporal removal against random controls. MARS is publicly available for research: Project page: https://vectorinstitute.github.io/mars/ Github: https://github.com/VectorInstitute/mars
Does grouping translation updates before speech synthesis change the measured completion gap between two translators? We compare StreamAtt with its fixed-audio-history baseline on ten English-German talks. For each translation, we keep the intended text and its availability times fixed and vary how updates are grouped for synthesis. Four-second grouping increases StreamAtt's mean completion lead from 106.6 to 229.1 seconds with XTTS-v2, but reduces it from 169.5 to 146.5 seconds with Piper. Both produce shorter audio; different changes in playback idle account for the opposite effects. These directions recur on a separately selected twenty-talk panel. In continuing CosyVoice 3 sessions, grouping instead leaves paired audio unchanged and delays completion on nineteen of twenty blocks. A speech-timing comparison therefore needs to specify how text enters synthesis, not just which synthesizer is used. Recognition and automatic word timing describe the outputs, but do not establish translation accuracy or intelligibility.
AI-text detectors reach near-perfect scores in evaluations that contrast fully human-written texts with texts generated from researcher-designed prompts. In authentic student writing, however, AI use is voluntary and often partial. We study this setting with the PISA 2025 FLA dataset, in which 15-year-olds in Germany wrote English argumentative essays with and without access to a gpt-4o-mini-based chatbot. Keystroke, clipboard and chat logs let us trace every character of the final essay to its origin. From this provenance, we derive text- and sentence-level labels for human-written, AI-generated and mixed text, and for text written after chatbot use without traceable AI material. We evaluate a feature-based detector and a classifier on frozen Llama-3.1-8B embeddings, both adapted from top-ranked shared-task systems. Both separate AI from human text reliably (macro-F1 up to 0.96 for texts and 0.95 for sentences), independently of spelling errors. The feature-based detector, however, relies on lexical complexity, which a simple student-style rewriting prompt removes (AI recall 0.92 $\to$ 0.60 with gpt-4o-mini, 0.41 with Qwen3). The embedding-based detector stays robust for whole texts (AI recall $\geq$ 0.94) but not for single sentences (AI recall as low as 0.59). Mixed text falls between the two classes, and text without traceable AI material is classified as human-written in about 95\% of texts.
Answer accuracy alone does not distinguish failures in acoustic perception, rule knowledge, and rule application. AuDiR (Audio Diagnostic Reasoning) examines these distinctions through controlled input interventions. Each item contains an audio clip, a multiple-choice question, a task-general rule, and a textual description of the acoustic observations needed to answer. Three conditions progressively supply the rule and observations; a fourth removes the audio while retaining the question and rule. The first correct diagnostic condition assigns an operational response label, without identifying a unique internal cause. The current benchmark contains 201 items in 18 families drawn from 9 sources, with answer references documented separately from percept extraction. In a preliminary evaluation of Gemini-3.8-flash, accuracy is 29.3% without supplied information, 31.3% with the rule, and 98.5% with the observations also supplied. No-audio accuracy is 23.9%, with a confidence interval containing the 25.0% chance rate. A family-level knowledge probe associates incomplete rule recall with larger gains from rule provision. These results show the benefit of supplying acoustic observations as text, while incomplete validation, possible recognition shortcuts, and non-monotonic responses limit stronger failure-attribution claims.
Multi-view anomaly detection is crucial for industrial inspection to identify defects hidden from single viewpoints. Yet, despite their strong success in single-view applications, reconstruction-based methods remain largely underexplored in multi-view settings. A key challenge in extending these methods lies in generating view-consistent pseudo-anomalies. To address this, we propose PAR, a reconstruction-based framework for enhanced multi-view anomaly detection. First, we introduce a feature perturbation strategy coupled with a consensus aggregation module that integrates these perturbed features across viewpoints to mitigate view-inconsistent anomalies and construct highly aligned pseudo-anomaly representations. Second, we employ a view-specialized multi-decoder architecture; by assigning a dedicated decoder to each view, we increase the model's representational complexity, enabling the generation of highly feasible, localized anomaly maps rather than less precise, view-agnostic reconstructions. Last, we incorporate Fourier attention into the decoders to emphasize global frequency-domain structures over local token evidence, effectively preventing identity mapping. Extensive experiments on the Real-IAD dataset demonstrate that PAR consistently outperforms state-of-the-art methods in sample-level, image-level, and pixel-level anomaly detection. Our code is available at https://github.com/SpatialAILab/PAR_mvad.
We investigate convolutional-neural-network (CNN) based automatic solver selection for sparse linear systems, a problem of practical importance where solver performance varies dramatically with matrix sparsity structure. A CNN is trained to regress log timing ratios between candidate solvers using spy plot images of matrix nonzero structure as input, achieving strong prediction accuracy on both held-out SuiteSparse matrices and real-world matrices from high-performance simulation codes.
Foundation models can supply predictions for a downstream task that draw on knowledge from extensive pretraining, but they differ in task-specific accuracy and query cost. We develop a framework that improves prediction for the task under a budget on the average query cost per input, using training data to learn which models to query and how to combine their outputs with features available without querying. The first method queries the same models for every input; the second lets them vary across inputs and uses Q-learning to decide which model to query next and when to stop. For the first method, we show that several individually inaccurate outputs can jointly outperform the best single one. It attains the minimax rate when the regression function depends on the outputs only through models within the budget. Otherwise, any method within the budget can be forced to a slower rate, which the first method attains. For the second method, we show when varying the queried models across inputs strictly improves on querying the same models for every input, and bound its prediction risk and average cost in finite samples. In simulations and a multimodal laptop price application, both methods outperform existing methods.
Speculative decoders raise throughput by banking more target tokens per forward pass, and Medusa, EAGLE-3, DFlash and DSpark all do it the same way: more candidates that survive verification. All of them draft in the target's own vocabulary, so one node of the candidate tree is one target token, and the tokenizer makes that unit look like a constant. It is not. BPE's pretokenizer splits on whitespace before merges are learned, so no multi-word string can ever become a single token: in Qwen3-4B, "United States of America" is four tokens and no adjacent pair of them exists as one entry, so a phrase is out of reach of any method drafting in the target's vocabulary. We change the unit. A small MLP head reads the frozen target's hidden state and predicts a one- to three-word phrase as one entry in a separate draft vocabulary; greedy longest-match expansion maps it back into the target's own tokens, and N candidates are verified in one forward pass behind a prefix-tree mask, so the change is orthogonal to tree depth and width. The target and its tokenizer stay frozen and a candidate is accepted only while it matches the target's own argmax, so the served output is identical to greedy decoding up to floating-point tie-breaking. On one A5000 at batch 1, against HuggingFace generate(), a 21.94M head (0.545% of Qwen3-4B) reaches 2.391x on GSM8K, 2.444x on MATH-500 and 1.544x on MT-Bench. Because the head's output layer is one row of 257 parameters per entry, the draft vocabulary size is the parameter budget: an 8,000-entry head is 3.434M parameters, 0.0854% of the target and 64x smaller than EAGLE-3's drafter, and keeps 92.3% of the larger head's GSM8K tokens per pass. With EAGLE-3 run on same prompts and baseline, our 21.94M head is ahead of tree-16 on both maths sets and behind tree-60 on every set.
Causal discovery from observational data is central to scientific reasoning, yet established constraint-based and score-based methods often recover only a Markov equivalence class and provide no signal in bivariate or strongly autocorrelated regimes typical in some areas of science. Here, we recast causal discovery as a Bayesian symbolic regression problem. Given a set of candidate parents for each target variable, the Bayesian symbolic regression samples closed-form expressions from a posterior over symbolic models. The fraction of sampled expressions in which each candidate variable appears defines a posterior inclusion probability that we use as the causal score. This replaces pointwise model selection by an ensemble of plausible mechanisms. We evaluate the method on coupled nonlinear bivariate dynamics, on the CauseMe benchmark suite spanning linear, chaotic, nonlinear, higher-dimensional and high-noise settings, and on a four-variable biogeoscience structural causal model under Gaussian, heavy-tailed and mixed-noise regimes. Our approach is competitive with, and often outperforms, specialized causal discovery algorithms. Additionally, the method returns closed-form mechanisms leading to explicit causality and interpretability.
To address unreliable cross-view association in long-sequence feed-forward dense SLAM, this letter presents V3DSAM, an open-vocabulary semantic dense SLAM system that establishes physical correspondences at both the scene and object levels. The system combines feed-forward geometric estimates from VGGT-SLAM 2.0 with open-vocabulary semantic observations from SAM3. We introduce semantic-geometric dual-scale loop verification, which identifies false loop closures using pixel-level bidirectional reprojection and object-level 3D spatial correspondences, without accessing backbone-internal features. We further propose geometry-anchored persistent instance mapping, which defines physical instances through class-constrained 3D connectivity in an optimized global coordinate system and assigns global IDs without inter-frame identity propagation. The proposed verifier rejects all four false loop candidates in our TUM RGB-D evaluation, while long-sequence ScanNet experiments show improved instance identity consistency. Thus, V3DSAM reduces dependence on internal backbone features and temporal ID propagation for robust long-term cross-view association.
Symbolic knowledge (motif counts, ring counts, label constraints) is routinely added to graph neural networks (GNNs) and justified by their limited expressivity. We replace this yardstick with an exact one. The largest log-loss reduction a rule $S$ can give a model family is $I(Y;S\mid\sigma)$ given what the family computes; every realised gain splits exactly into this term plus repaid neural debt; and for a trained network with representation $T$ the value is $I(Y;S\mid T)$, the label information training failed to extract. Rule design thus becomes rule discovery: we score candidates on the network's out-of-fold residual against calibrated permutation nulls with family-wise error control. On ZINC, discovery recovers the hidden ring term $\max(0,\ell_{\max}-6)$ of the target in three independent runs ($p_{\mathrm{adj}}=0.001$) and beats expert cycle counts; frozen and applied to five unseen datasets it is never significantly worse than the plain network or the expert rule, and on BBBP it significantly beats both.
Video out-painting is an important problem, yet existing research focuses primarily on short clips, leaving long video out-painting largely underexplored. Extending video out-painting to long videos poses substantial challenges. A natural solution is iterative out-painting, i.e., generating videos chunk by chunk, but this introduces two major issues: generation instability or even failure caused by inter-chunk inconsistencies and error accumulation, and semantic drift over long temporal horizons. In this paper, we propose LongSCOP, short for Semantically Consistent long Video Out-Painting. First, we introduce a training recipe with three components for smooth and stable long-horizon generation: self-forcing training, latent calibration, and future references. Beyond generation stability, we recognize that the rich temporal context of long videos provides valuable references for maintaining semantic consistency. We therefore develop a VLM-based agent that automatically selects the most informative reference frames from the entire video to maintain consistent subjects, scenes, and backgrounds throughout long-horizon generation. We also construct a high-quality training dataset of 62K video clips. To facilitate research on this underexplored problem, we introduce LVO-Bench, a benchmark curated by human experts to assess long-horizon stability and semantic consistency. Extensive experiments demonstrate that our method outperforms existing approaches in both generation stability and semantic coherence.
Generalizable deepfake video detection remains challenging because forgery artifacts vary substantially across manipulation types and often combine diverse spatial and temporal inconsistencies. Existing parameter-efficient adaptation methods typically rely on a single shared representation, rendering them prone to manipulation-specific bias and limiting transfer to unseen forgeries. To address this challenge, we propose Mixture Of Spatio-Temporal Experts (MoSTE), an adaptation framework that enables conditional specialization of forensic representations. Instead of forcing one adapter to model all forgery patterns, MoSTE decomposes adaptation into parallel spatial and temporal expert pools and dynamically routes tokens to the most relevant experts through an unsupervised token-level routing mechanism. This allows experts to naturally specialize in complementary forgery cues without requiring explicit manipulation labels or multi-stage training, while effectively transferring representations from CLIP to the video domain. Extensive experiments across cross-dataset and cross-manipulation benchmarks demonstrate that MoSTE sets a new state of the art in average generalization while preserving strong performance across diverse manipulation families. Additional analysis of expert routing and attention behavior reveals the emergence of expert diversity, suggesting that MoSTE learns complementary spatio-temporal forensic features. Code is available at https://github.com/charbelyah/MoSTE
Landslide susceptibility models are trained on inventories that record where landslides were seen, not where they occurred. The established remedy fits bias-describing variables and neutralises them when mapping. We adapt occupancy-detection modelling from ecology, treating several inventories as imperfect observers of one event and separating occurrence from recording, and ask when this improves on neutralisation. We use near-complete inventories of the 2018 Hokkaido Eastern Iburi, 2015 Gorkha and 2010 Haiti earthquakes with simulated reporting bias, and three real Gorkha inventories. Compared with standard practice, the occupancy model reduces road bias in 24 of 25 paired draws and raises AUC by 0.08-0.14 on Gorkha and 0.03 on Haiti, but only 0.01-0.02 on Iburi. Its median predicted-to-true landslide rate is 0.99 (range 0.44-1.89), against 0.24-0.30 for the baselines. Rescaling the baselines with the same anchor area corrects their median but leaves larger errors (range 0.20-4.34). Cell-level probabilities remain too extreme (calibration slope 0.59). Compared with the best neutralised model, AUC is higher on Iburi (by 0.03-0.07) and lower on Gorkha and Haiti (by 0.01-0.05). Transplant experiments tie the Iburi advantage to the structure of its true landslide pattern, not to its terrain or road network. In all eight settings with a low interaction gap, computable from biased records, neutralisation was at least as accurate, including a prespecified test on Haiti; a high gap did not guarantee an advantage. The real Gorkha inventories are slope-biased, not road-biased, and there the model brings no meaningful gain. Report-based catalogues remain the main untested case.
Probabilistic circuits (PCs) have emerged as generative models guaranteeing tractable probabilistic inference. However, tractability comes at the price of larger model sizes than in other generative models. As such, to scale PCs to high-dimensional data such as images, previous work has devised several parameter sharing techniques as well as fast evaluation protocols that consider learning PCs on independent image patches. In this work, we revisit this practice to highlight a surprising learnability issue: models that assume independence counterintuitively can yield better likelihoods than models that do not. To break this apparent bottleneck and to scale PCs further, we introduce a recursive parameter sharing construction (RePA) that fosters efficiency without assuming independence. We show empirically that RePA can greatly reduce time and memory consumption while yielding competitive likelihoods. Inspired by these results, we further generalize the approach to data modalities that do not inherently exhibit patch-structure (NoPA), such as tabular data, showing similar efficiency improvements, with marginal changes in likelihoods and sample quality on seven tabular datasets. Code available at https://github.com/april-tools/RePA.
Non-linear functions are an important class of operations in modern large language models. However, chip vendors building for AI have so far relied on traditional vector or scalar processing units for non-linear functions. The performance of these units has not kept pace with matrix multiplication units (MMUs), which can be found in both high-performance and edge-computing devices from almost every vendor. Reconfigurable units such as FPGAs and CGRAs have been proposed in the literature, but have not been systematically compared with traditional units backed by MMUs. In this paper, we build two SoCs: one using an RVV vector processor and the other using a latency-insensitive (\emph{elastic}) CGRA, both coupled with an MMU and driven by a single-core CPU. We then propose mappings for common feed-forward networks on our SoCs and find that the CGRA reaches a maximum speedup in clock cycles of 7.12x over RVV on the activation stage of simple point-wise non-linear functions like SiLU, whereas the RVV reaches a maximum speedup of \textbf{9.8x} over the CGRA on complex non-linear functions such as $\mathrm{softmax}$. Over the enclosing feed-forward block the corresponding maxima are 3.16x and 4.24x.
Scientific research has long been driven by substantial human effort, from reviewing prior work and identifying research gaps to designing specific solutions. Recent advances in harness-based agent systems have given rise to increasingly capable Auto Researchers, which can autonomously carry out these tasks and, increasingly, entire end-to-end research workflows. Yet despite their rapidly growing capabilities, how closely the research decisions and outputs produced by auto researchers align with those of human experts remains underexplored. We investigate this gap through the lens of research reconstruction and study: Given comparable research context, how similar are the research decisions and outputs produced by Auto Researchers to those of human researchers in AI research? To answer this question, we introduce AutoResearcher-Human Research Similarity (AutoHRS), a stage-wise evaluation framework for assessing how faithfully auto researchers reconstruct human research artifacts across different stages of the scientific workflow. Using 152 ICLR oral papers as our corpus, we generate 456 research trajectories spanning different stages of the research process. Applying AutoHRS to these trajectories, we find that current Auto Researchers frequently fail to recover key research decisions and their resulting artifacts across multiple stages of the research process.
Multi-armed bandits are a natural framework for many real-world sequential decision problems, from recommender systems and clinical trials to educational test recommendation and online advertising. The performance of these algorithms depends on how they explore different arms, and exploration strategies range from variance-independent (e.g., UCB1) to variance-aware (e.g., UCB-V). In practice the reward variance of each arm is rarely known and varies widely across arms, and the reward support is often underspecified. In these cases no single rule is uniformly best: variance-independent methods are robust but loose when variances are small, while variance-aware methods are tight only when their variance estimates and the assumed support bound are reliable. Such heterogeneous, unknown variances arise in many settings, for instance when multiple objectives are scalarized into a single reward. We propose UCB-min, a per-arm adaptive meta-algorithm that selects, for each arm, the tighter of a variance-independent and a variance-aware exploration bonus, without prior knowledge of the variance regime and without reward normalization. We prove that, with UCB-V as variance-aware base, its regret is bounded by the sum of the two base bounds at the standard exploration constant, and by their per-arm minimum, up to a constant, for a larger one. We validate it on synthetic data, an educational tutoring task, and two group-recommendation benchmarks (MovieLens, Steam), showing that UCB-min consistently tracks the best base algorithm across diverse variance regimes and datasets.
Discrete autoregressive modeling, notably through next-token prediction, has become a cornerstone of modern language models. Extending this paradigm to vision allows visual and language models to converge on a unified framework, thereby fostering shared methodology and infrastructure. This work investigates a central question: how far can a finite discrete vocabulary go in representing the richness of the visual world? We present dNaViT, a Discrete Native-resolution Vision Transformer that produces discrete visual primitives analogous to language tokens. Built upon a frozen pretrained encoder, dNaViT applies lightweight residual modulation with a dual-branch objective to inject reconstruction-aware cues into semantic features. By leveraging grouped product quantization, it scales discrete capacity without inflating spatial sequence lengths, while a balanced training recipe maintains both semantic and pixel-level information. Consequently, dNaViT functions as a versatile discrete visual tokenizer, enabling unified encoding across downstream tasks alongside high-fidelity image decoding. Extensive evaluations show that dNaViT matches or exceeds task-specific baselines across 12 visual understanding benchmarks, 4 generation benchmarks, and an image editing benchmark. These findings demonstrate that a compact discrete vocabulary can serve as a universal visual currency and unlock the full potential of fully unified, native multimodal methodology.
Neural surrogates for engineering models are often trained on large collections of simulated solutions. We investigate whether physics-informed neural networks (PINNs) can achieve accurate predictions from scarce labelled observations by incorporating governing equations and prescribed initial and boundary conditions. We introduce hierarchical loss weighting (H-PINN) and its gradient-direction-aware extension (HO-PINN), which adjust the emphasis on physical constraints according to higher-priority losses. Hierarchical PINNs are compared with Vanilla PINN, tuned gradient boosting, support-vector regression, and Gaussian-process regression on a heat equation and three heterogeneous two-phase-flow benchmarks. Experiments use 20--1000 labelled states and ten paired seeds. Each label is one coordinate--state observation, not a complete simulation. At least one PINN variant achieves the highest mean primary-target \(R^2\) in each interpolation and equation-assisted forecasting comparison at 20 and 50 labels. The advantage for water saturation persists across all tested budgets. Hierarchical weighting reduces variability across runs in selected settings. These results demonstrate that known physics can support accurate neural surrogate predictions from very few solution observations, at substantially higher fitting cost than the tested classical regressors.