In natural human-AI dialogue, users' corrections, requests, and self-descriptions carry reusable signals about how they want work to be done. These signals reveal individual process preferences, user-specific standards for evidence, depth, analytical focus, procedure, and delivery. We ask whether a preference revealed in one dialogue can shape an assistant's later answer when the task does not restate it. ProcessPrefBench uses three months of natural dialogues from two knowledge workers to build 66 source-target cases separately judged for source evidence and target applicability. Sources are preselected, so the benchmark tests application rather than retrieval. We test three models with target-only, full-history, and distilled-preference prompts. Target-specific 0-2 rubrics score preference realization, and a task gate flags severe failures. Compared with target-only prompting, full-history and imperative distilled-preference prompting raise preference realization by 0.204 and 0.244 points on the 0-2 scale, respectively, with no reliable advantage from distillation. Fabricated-evidence failures rise from 5.8% under target-only prompting to 10.1% under full-history and 10.4% under imperative prompting. ProcessPrefBench highlights a tension in learning from dialogue, where drawing on prior interactions can improve preference realization yet also invite fabricated evidence. We hope it helps build assistants that honor a user's process preferences without sacrificing factual reliability.
Time series forecasting is shifting from locally trained, per-series models to global models trained across many heterogeneous series, which requires learning what sets the series apart. When the raw data cannot be centralized, such a global model is trained with federated learning (FL), one client per data owner. State-of-the-art forecasters such as the Temporal Fusion Transformer (TFT) tell series apart through a static input that describes each series and does not change over time. In the special case of a federated setting where each client holds a single series, this static input is constant within each client and varies only across clients. We ask whether a federated TFT can learn from it. Per-batch FedAdam, which aggregates gradients after every batch, uses the static input almost as well as central training, whereas communication-efficient schemes with local epochs (FedAvg, FedAdam+Prox) extract far less from it and gain nothing significant on either application dataset. At the same local computation, an Adam-compatible SCAFFOLD keeps the static pathway responsive on wind power and gains significantly there. On building energy its error still drops, but the static pathway falls silent, so drift and communication frequency cannot be fully separated as the cause. Only per-batch FedAdam matches or beats the per-client local models across all datasets. We confirm the effect on building-energy meter data, a wind-power dataset, and a controlled synthetic dataset. The loss is not specific to the TFT: on the synthetic and the wind data two further architectures with different static-conditioning mechanisms show it as well.
Small LoRA adapters trained on Tajik news and books lowered Qwen3's Tajik text loss at every size we tried, from 0.6B to 8B, by 0.60 to 1.45 nats per token, but none of them improved Tajik reading comprehension where it can be measured. On Belebele the 1.7B and 8B adapters left accuracy unchanged and the 4B adapter lowered it by 3.7 points; at 0.6B the score rose by 4.2 points but sits near chance and coincides with a change in which answer letter the model prefers, so we do not read it as understanding. At 8B, continued pretraining on about 375 million tokens of raw Tajik text was followed by a Belebele gain of 9.9 points. After a further supervised fine-tuning stage, our main model, e3, is 11.4 points above Qwen3-8B on Belebele; its English MMLU changed by -1.2 points, and a loss larger than 3.3 points is excluded. The same heavy recipe applied to Qwen3-0.6B-Base brought its text loss close to that of the 8B model after books-only pretraining, while its multiple-choice answers stayed near chance and its English MMLU fell by 7.6 points. We report 15 trained models and their 5 starting models from June to October 2025, scored in 2026 on one harness with paired intervals, and we report what went wrong: lost training records, a thinking-mode dataset with almost no reasoning in it, later branches whose training data was never documented, and a 4-bit release that contains no transformer layers.
Recent video world models allow users to explore generated environments through camera navigation. Continuous hand motion provides a fine-grained action channel for interacting with the people and objects encountered during exploration. We present Touch\'{e}World, a framework that integrates gesture-driven interaction into a camera-controlled video world model. Given an initial image and a scene-description prompt, the model generates video conditioned on temporally aligned camera and hand controls. We construct joint camera--hand training data from heterogeneous egocentric videos through annotation completion, consistency-based filtering, and temporal alignment. Building on a pretrained camera-controlled world model, we combine spatial hand-skeleton conditioning with numerical hand-motion features through parameter-efficient adaptation. Causal adaptation and few-step distillation support chunk-level online generation, allowing users to update their controls based on generated visual feedback. We further introduce Touch\'{e}Bench to evaluate hand-motion following, camera--hand coordination, online interaction, and object response. Together, these components establish a framework for studying continuous exploration and hand-driven interaction within generative worlds.
Attention is central to modern AI models for language, vision, and scientific modeling. However, direct evaluation compares every query with every key, requiring quadratic time in the context length, which limits its scalability to long contexts. Consequently, characterizing the complexity bounds of attention is of fundamental importance. For every fixed dimension $d$, we give an algorithm for $m$ queries, $n$ key-value pairs, and additive error $\varepsilon$, with expected time $ \widetilde{\mathrm{O}}\bigl((mn)^{k/(k+1)}+m+n\bigr),\qquad k=\lfloor d/2\rfloor. $ For fixed dimension and a fixed number of value channels, soft-O hides factors of the form $\operatorname{poly}_d\bigl(b,\log(m+n),\log B,\log(1/\varepsilon)\bigr),$ where $b$ bounds input encoding lengths and $B$ bounds coordinate and value magnitudes. In particular, the dependence on $B$ and $1/\varepsilon$ is polylogarithmic, which allows arbitrary positive temperature. When $m=n$ and $d\ge2$, the bound is $\widetilde{\mathrm{O}}(n^{2-2/(\lfloor d/2\rfloor+1)})$: near-linear in dimensions two and three, with exponent $4/3$ in dimensions four and five and $3/2$ in dimensions six and seven. It improves the previous $\widetilde{\mathrm{O}}(n^{2-1/d})$ upper bound and matches the best known batch maximum-inner-product bound in each fixed dimension. Additionally, we reduce integer batch maximum inner product computation to one attention call with the same dimension and input sizes $m,n$.
Given a collection of $N$ graphs, graph-level clustering (GLC) seeks to partition the $N$ graphs into semantically meaningful groups, which serves as a fundamental task in analyzing graph-structured data. While Graph Isomorphism Networks (GINs) have become the de facto encoders for GLC, they still fail to distinguish graphs due to their neglect of global topology. On top of that, existing deep clustering frameworks suffer from severe "clustering collapse" issues, particularly in imbalanced multi-class scenarios. To address these challenges, we propose S$^3$GLC, a novel framework that integrates global topological characteristics with local neighborhood features for balanced GLC. Theoretically, we leverage the Hoffman-Wielandt inequality to construct Spectral Signature as a rigorous lower bound for graph edit distance (GED), providing a computationally efficient proxy for global topological features. We additionally propose a joint manifold-preserving sparse regression module to fuse Spectral Signature with Spatial Representation, eliminating noise while preserving the complementary information. Furthermore, to mitigate clustering collapse, we introduce a balanced clustering objective based on optimal transport, harnessing the Sinkhorn-Knopp normalization to enforce equipartition constraints on cluster assignments. Extensive experiments on 10 real-world datasets demonstrate that S$^3$GLC significantly outperforms 15 state-of-the-art baselines, achieving remarkable improvements on benchmark datasets such as REDDIT-12K and OGBG-PPA.
Reservoir computing provides an efficient framework for time-series prediction because only the readout layer is trained while the recurrent reservoir remains fixed. In this study, we examine how the negative sign ratio of a signed Erdos-Renyi reservoir affects the internal dynamics and forecasting performance of an echo state network on the Mackey-Glass benchmark. To characterize the reservoir responses, we use two Shannon-entropy-based observables, namely node-state entropy and node-frequency entropy, as empirical proxies for the diversity of reservoir states and temporal activation patterns. We find that intermediate negative sign ratios are associated with richer internal dynamics and substantially lower prediction error than the sign-homogeneous extremes for the task considered here. These results suggest that link-sign composition can be a useful structural parameter in reservoir design for chaotic forecasting. The present findings should be interpreted within the scope of the current benchmark and reservoir setting, and broader validation will be required to establish their generality across other tasks, topologies, and controlled spectral conditions.
Text-to-SVG systems are commonly evaluated by applying raster-image metrics to rasterised outputs, yet whether these scores capture fine-grained prompt faithfulness remains unclear. We introduce SVGFaith, a two-layer meta-evaluation framework in which the more faithful candidate of each pair is determined by construction or by human judgment. SVGFaith-Diagnostic contains 1,842 controlled contrastive pairs across eight capability axes, split evenly between two directions: two renderings compared against one prompt, and two descriptions compared against one rendering. SVGFaith-Arena contains 1,327 human-annotated comparisons of real generator outputs. On SVGFaith-Diagnostic, CLIP and three other image--text embedding metrics reach at least 0.789 pairwise accuracy on object identity but at most 0.672 on spatial relations and attribute binding, where chance is 0.5. We further introduce SVGFaith-Score, an interpretable pointwise metric. A domain-tuned decomposer turns each prompt into a dependency graph of atomic visual requirements, a visual question answering model scores whether each requirement is satisfied, and a requirement's score is discounted when its prerequisites score low. In addition to the overall score, the per-requirement scores show which requirements are judged unsatisfied and which inherit discounts from their prerequisites. SVGFaith-Score significantly outperforms all nine baselines, including a Qwen3.5-122B-A10B judge, on Arena and in both Diagnostic directions. Code and data: https://github.com/Yukinonooo/SVGFaith.
Large vision-language models (LVLMs) have recently been extended to handle multiple input images, but this extension dramatically increases the visual token budget, making visual-token pruning not merely an efficiency option but a practical necessity. Existing visual-token pruning methods, primarily designed for the single-image setting, rely on attention-derived saliency signals to estimate per-token importance. However, this attention-centric design transfers poorly to the multi-image regime: it causes substantial performance degradation on multi-image benchmarks, suffers from strong positional bias, and loses discriminative attention signals as the number of input images increases. To test whether attention-selected tokens are truly important for multi-image reasoning, we conduct two complementary causal probes Leave-One-In and Leave-One-Deletion, which measure token importance by restoring individual tokens to a zeroed visual sequence or removing them from the full sequence. Attention-based pruners provide only marginal gains over random selection, showing that attention is a weak proxy for visual-token importance in multi-image reasoning. Motivated by this gap, we propose Relation and Information Aware Pruning (RI-Prune), a training-free, attention-free pre-decoder selector for multi-image visual token pruning. RI-Prune scores each visual token using a z-normalized combination of three signals computed from projector outputs: query relevance, cross-image correspondence with compact per-image prototypes, and intra image representativeness. By integrating these signals, RI-Prune performs balanced per-image Top-K selection to preserve information across input images. Across multiple multi-image benchmarks and three LVLM backbones, RI-Prune achieves a stronger trade-off between token reduction and task performance than prior pruning methods, particularly under aggressive token budgets.
Search-augmented recommenders can use the latest search episode and earlier searches. We introduce FocalSwap to test how these inputs contribute to the next recommendation. For each recommendation, it changes the latest search slot or all search inputs while holding the target, candidates, pre-search history, and scorer fixed. It compares neutral masks with matched observed-search replacements, recording rank changes, RR gains relative to each control, and preference for the true focal search. A planted signal and GenSAR's current-query task confirm that the audit detects known input dependence. We evaluate behavioral and lexical evidence on KuaiSAR and Qilin, and recommendation checkpoints trained using the released SESRec and UniSAR implementations. Across three independently trained UniSAR checkpoints, the estimated all-source benefit depends on the control. On the same 66,139 rows, replacing all search inputs with a neutral ID yields a positive signed contribution, whereas replacing them with category/profile-matched observed search sequences gives smaller estimates, with an interval below zero for one checkpoint. The paired RR gaps range from $+2.770 \times 10^{-3}$ to $+6.604 \times 10^{-3}$, and every adjusted interval is above zero. Neutral masks change 59.7-61.7% of target ranks, while matched replacements change 44.1-44.8%. The control contrast persists across five alternative donor assignments per checkpoint. Under the primary neutral-mask and category-matched controls, positive latest-slot contribution is not established in SESRec or UniSAR. These results show why masking evidence should be compared with observed-search replacements when attributing a benefit to broader search state or the latest episode.
Streaming speech recognition is increasingly consumed by machines, such as translators and voice agents, that act on each token as it appears. A token that is later revised costs wasted work or a wrong action. In this work, we present Confucius4-R2T2 (short for Real Real-Time Transcription), a stable-prefix streaming ASR model that never modifies its earlier output while maintaining high accuracy and low latency. R2T2 is obtained by post-training Qwen3-ASR-1.7B on 1,500 hours of Chinese and English audio while freezing its pre-trained audio encoder and projector. The adaptation is based on the concept of Longest Stable Prefix (LSP), a new notion that we introduce in this work. We prove that LSP is monotone under the prefix order, consistent with the full-audio output, and maximal among labels that equal a whole earlier output. Learning from LSP labels converts a full-context model into a stable-prefix streaming one without any architectural change. Extensive evaluations show that R2T2 brings streaming recognition close to the base model's full-context quality while delivering stable text with low emission lag. Its gains extend beyond the post-training languages, demonstrating a practical path from full-context ASR to stable streaming. We open-source the model weights and code at https://huggingface.co/netease-youdao/Confucius4-R2T2 and https://github.com/netease-youdao/Confucius4-R2T2.
Universal image restoration aims to handle diverse degradations within a single model, but a reliable scaling strategy remains underexplored. Increasing backbone depth and width can lead to performance saturation, while degradation-aware specialization depends on accurate degradation estimation and is vulnerable to errors on out-of-distribution and real-world inputs. We propose UniRestorer++, a scalable framework based on multi-granularity degradation modeling. It progressively refines degradation groups and introduces group-specific experts, coupling capacity growth with specialization. Degradation-estimation-based routing selects a root-to-leaf path, while granularity-estimation-based routing chooses a level along that path to balance specialization and robustness. Spatially dense degradation representation maps further enable region-aware routing for spatially non-uniform corruptions. We develop CDD-50, a benchmark covering 50 degradation categories from single to penta-degradation. Experiments show approximately log-linear PSNR trends over the evaluated model-size range and stronger scaling gains on out-of-distribution and real-world data than vanilla MoE scaling. UniRestorer++ achieves state-of-the-art performance across benchmarks and demonstrates zero-shot generalization and few-shot transfer on real-world images. Code: https://github.com/mrluin/UniRestorer_plusplus
Underwater spatial reasoning depends on object boundaries and geometric cues that are obscured by scattering, color distortion, and low visibility. Existing underwater resources capture these cues through images, annotations, and depth maps, but provide fragmented supervision for vision-language models. We present a spatial data engine that connects visual appearance with object-level geometry through enhancement, 2D and geometry-assisted question generation, and expert filtering. It produces 4.27M QA pairs across five capability groups. We use this corpus to train USI-Brain and introduce an Underwater Prior-guided Rectifier (UPR) that applies gated corrections to visual features. USI-Bench evaluates the resulting capabilities with 6,525 held-out QA pairs. On its 15 task metrics, USI-Brain and USI-Brain + UPR improve over Qwen3-VL-8B-Instruct by 5.76 and 7.63 points on average, respectively, using signed differences on 100-point scales. Ablations show that geometry-assisted supervision improves depth and ordering by 4.8 and 6.1 points with UPR disabled, while UPR adds 18.0 points in scene description under matched training data. Both variants use raw images at evaluation. Across NautData, UWBench, and MarineEval, USI-Brain achieves the best or joint-best score on 28 of 32 metrics. Model weights and USI-Bench will be released at https://github.com/SKY-Zhu-91/USI-Brain.
Generating daily activity schedules from individual attributes (e.g., age and income) is valuable for activity-based travel modeling and human behavior simulation. The task is challenging because exact activity structures follow a highly imbalanced, long-tailed distribution, while the same structure can admit multiple duration allocations under a fixed full-day budget. Direct generation must therefore learn many exact activity structures from highly uneven observations while modeling their possible temporal realizations. However, many exact structures share recurring activity organization. Based on this observation, we propose Act-HiMotif, a hierarchical motif-based generative model that introduces learned, reusable activity-structure motifs as an intermediate representation between individual profiles and specific activity structures. Act-HiMotif first distinguishes home-only days from active days, then samples a motif from a profile-conditioned prior to guide activity-structure generation, and finally applies a budget-constrained duration allocator to generate activity durations that sum exactly to the full-day horizon. This hierarchy models home-only routing, activity-structure generation, and duration allocation at their corresponding levels, while motifs allow related structures to share statistical support without losing their distinct sequences. Experiments on large-scale travel survey data show that Act-HiMotif outperforms strong baselines in both in-distribution and spatial generalization settings, demonstrating the effectiveness of hierarchical generation and reusable activity-structure motifs for producing realistic profile-conditioned daily activity schedules.
Anti-money laundering (AML) investigations are organized around accounts, while the suspicious evidence for an account may reside in only a few of its transactions. Yet many account-level detectors summarize transactions before learning, and message-passing models may aggregate once more over neighboring accounts. We name this discrepancy the aggregation mismatch and study it under an account-window benchmark label defined by the presence of at least one flagged transaction. We recast account-level AML detection as multiple instance learning, in which an account-window is a bag and its transactions are instances. We propose TRACK, which encodes every transaction with a shared encoder and selectively pools the encodings into an account decision. To emulate coarse supervision, we withhold the transaction labels of five AML datasets from TRACK throughout training. Under a time-ordered protocol, TRACK achieves the highest average precision among the tabular and graph baselines trained from account labels on every dataset. Despite using only account-window supervision, it remains competitive with edge classifiers trained directly from transaction labels. Its instance scores also rank flagged transactions well above chance and approach or exceed directly supervised ranking performance across datasets.
There is a fundamental limit to the best prediction error of any model for a given data distribution. In classification, the Bayes error quantifies this limit and serves as both a benchmark for trained models and a criterion to detect overfitting. The state-of-the-art Bayes error estimation techniques for binary classification use only soft labels (positive class probabilities) and require no input features or auxiliary training. We extend such notions to constrained classification problems, such as fairness. Fair classification is an important case of constrained or multi-objective classification. It requires a model not only to be accurate but also to yield fair outcomes (equal or near-equal true positive rates) across different demographic groups (e.g., race and gender). This leads to an inherent fairness-accuracy tradeoff or a Pareto frontier that is used to compare fair classifiers against each other. The focus of our work is to estimate the fundamental limit for this fairness-accuracy tradeoff. We derive an explicit expression for the fair Bayes error rate (i.e., best achievable error rate subject to given fairness constraints), and construct consistent, instance-free estimators for it using only soft labels and group information. We provably bound the bias and the sample complexity of our estimators. Empirically, our method produces a robust, low-variance estimate of the optimal fairness-accuracy tradeoff curve, even when the soft labels are noisy and derived from a classification model on hard labels.
When a language model chooses a number between 1 and 30, every integer in the range is valid, yet the model can favor particular answers. We use numerical choice to study how preference is processed across layers and tokens. Perturbation probing and gradient attribution nominate neurons for activation interventions. In a selected Qwen3-4B case, a pre-answer edit changes 17 to 15 while both runs first emit 1. After the edit ends, selective cache and attention-output replacements trace part of its continuing influence on the second digit. Replacing one attention head's output or the other heads' outputs jointly produces similar immediate readout contributions; both interventions restore 17 but have different final log-odds effects on 7 versus 5. Holding later multilayer perceptron (MLP) outputs at their values from the edited run without either replacement reduces this effect gap from 0.119 to 0.0025 nat. Later computation therefore changes how similar local contributions affect the continuation. Retrospective estimates across eighteen contexts capture much of the probability movement but miss the selected answer switch. Prompt tests also separate influence from answer changes: one held-out prompt retains influence without switching, and only one of twelve fresh wordings qualifies for the reader/MLP test; there, joint MLP clamping again reduces the gap between the two reader effects while the answer remains 17. These results use numerical choice as a controlled case for studying persistent state and conditional downstream processing; they do not establish the origin or a complete circuit of natural numerical preference.
Generative models trained on the output of earlier models lose knowledge across generations, rare knowledge first. The remedies that work all bring in information from outside the loop, and the theory of model collapse treats that supply as given: a fixed archive, drawn from a fixed distribution. We argue that the supply is the problem, and that cultural evolution has the theory of it. We read model collapse as Rogers' paradox: a population that comes to rely on copying ends up no better off than one that never copied, because copies go stale unless someone keeps learning for themselves. We test the identification in three steps. First, Rogers' model, fitted only to fully synthetic training, predicts to within a few percentage points how much rare knowledge language models, from word trigrams to a pretrained transformer, retain under other mixtures of fresh data. Second, when the world itself changes, a lineage that copies its own past needs more fresh data to keep up the faster the world changes, as Rogers' model says it must and as analyses that assume a fixed target cannot; a lineage that restarts every generation from fixed pretrained weights is largely insulated, because those weights are knowledge that is never copied. Third, we close the loop: when the people who write a model's training data can ask the model instead of finding things out, the supply of fresh data becomes endogenous. At equilibrium, the model is exactly as reliable as it is costly for people to find out for themselves, the population is no better off than without the model, and every increase in the share of model-written text must be paid for by more people learning for themselves. Users who check what they are told and relearn when it is wrong break the paradox. Filters that select plausible text are conformist and accelerate collapse; only checks that carry new information act as critical learning.
Attracting features toward learned in-distribution (ID) regions is an intuitive strategy for out-of-distribution (OOD) detection, but does such attraction improve detection? We investigate this question using class-conditioned Hamiltonian dynamics as a controlled testbed and identify an attraction-ranking mismatch: improving individual class-affinity scores need not improve the relative ordering of ID and OOD samples. Our analysis separates misalignment between the driving field and the detection score from ranking failure that persists even when they are aligned. An exact AUROC decomposition and a local boundary analysis identify relative score changes at the ID-OOD ranking boundary as the source of first-order AUROC change under regularity conditions. Experiments demonstrate that a score-aligned rule can improve performance on a held-out source validation split yet consistently degrade it after transfer, even without a single winning-class change. A separate perturbation study finds that near-boundary relative score changes track the direction of small AUROC changes in the tested settings. Finally, centroid approximation and shared scales connect the zero-step score to a static centroid family that includes the existing CTM detector. Together, these results explain why class attraction is insufficient to justify feature refinement and motivate evaluating its effect on the ranking boundary.
A 3D Gaussian Splatting (3DGS) reconstruction inherits arbitrary units from structure-from-motion, so it cannot be measured or used to place AR content until someone supplies a scale. Both standard recipes are fragile on real captures: on five indoor scenes with tape-measured ceilings, a quantile floor-to-ceiling heuristic errs by 27% and metric-depth alignment by 35-71% across three networks. We present METRIC-GS, a selective scale estimator for real indoor photographs: it reads the scale off the reconstruction's own semantic content and, just as importantly, refuses when the evidence is thin. A frozen 2D teacher labels Gaussians by visibility-aware voting; instance-level size priors for architectural and furniture classes (doors, tables, the floor-to-ceiling gap) are fused by a robust maximum-likelihood step; and an inlier-mass test turns thin evidence into an explicit abstention that doubles as a router to a heavyweight metric model. On 47 ScanNet++ scenes with laser ground truth, with a configuration frozen before public evaluation, the estimator alone cuts the heuristic's mean error from 21.1% to 16.6% (9.3% on the 28 scenes it acts on) at ~11 s per scene. MASt3R remains the more accurate standalone estimator (9.9%), but routing only our refusals to it matches running it everywhere with 60% fewer invocations; this selective hybrid is the configuration we advocate. Visibility-aware voting raises the scenes the estimator can act on from 16 to 28 at unchanged deployed accuracy. Applied unchanged to 13 phone and drone captures of our own, on which MASt3R fails its own registration gate 12 times, the system improves all five captures it acts on and degrades none. We also report that a current feed-forward metric model scores 2.3% on a benchmark inside its training mixture and 28-35% on captures it has never seen: a published metric-scale number is a best case, not a deployment estimate.