Industrial visual inspection must both decide whether a product is defective and localize the defect, yet pixel-level masks are costly to collect at scale. Most anomaly-segmentation methods learn only from defect-free images and score deviations from normality. A true defect and an unusual-but-normal region, however, can both deviate substantially and receive similarly high scores. We propose Contrastive Dual Gaussian Processes (CDGP), a weakly supervised framework that models normal and anomaly inducing-variable predictive distributions over dense tokens. Its posterior-dominance statistic standardizes their predictive-mean difference by the joint predictive uncertainty, providing both spatial evidence and image-level confidence. This evidence complements hierarchical normal-reconstruction residuals for fine localization. All calibration uses training data only, without human pixel annotations or test-time fitting. Across MVTec AD~2, KSDD2, and VisA, CDGP ranks first among the evaluated methods on all MVTec AD~2 localization metrics and is first-place or competitive on KSDD2 and VisA. Factorized and matched linear-head controls delimit the contribution and scope of the linear-kernel Gaussian process (GP) formulation.
Cyber threat intelligence (CTI) is increasingly consumed not by human analysts but by LLM agents that compose multi-step investigations at query time. The harness side of this shift has matured rapidly (planning loops, tool protocols, context management), but the corpus side has not: threat reports and vulnerability databases are still packaged for retrieval-augmented generation, as opaque chunks behind an embedding index. We argue that this substrate, not model capability, is the bottleneck on agentic CTI investigation, and present CTIFoundry, an agent-native corpus scaffold. At build time, CTIFoundry materializes the latent structure of a CTI corpus: a deterministic ontology graph over four authoritative knowledge bases (CVE, CWE, CAPEC, ATT&CK) whose official cross-references become typed, traversable edges; a span-grounded report layer whose canonical, alias-resolved cross-vendor entities index provenance-carrying chunks; and hybrid dense+lexical retrieval surfaces. At query time this structure is exposed through seven typed tools and three procedural skills mounted on a stock open-source agent harness. On the public CTIConnect benchmark, swapping only the action surface lifts the identically-harnessed agent by +0.19 to +0.28 overall F1 across a four-model, two-provider panel: a small model on CTIFoundry surpasses a flagship on the flat substrate, and the gain is not bought with search effort, since on both Claude models the scaffolded agent is more accurate at roughly half the tool calls. An ablation attributes it: typed structure carries the larger share, procedural skills convert structure into discipline, and the two compose super-additively, because skills bind only to structure that exists.
We design randomized approximation schemes for the partition function of antiferromagnetic Ising models with uniform external field on random regular bipartite graphs. Our algorithm generalizes the approach of Kocurek, Oveis Gharan and Tjowasi (arXiv, 2026) for hard-core models on the same random graph model beyond the uniqueness threshold. We show that, as long as $λ$ is upper bounded by a constant and $λ(1 - β) \lesssim Δ^{-1/2}$, an efficient randomized algorithm approximates the partition function with high probability. The algorithm first truncates configurations that are large on either side of the bipartition and then samples from Gibbs distributions conditioned on fixed sizes on one or both sides. To choose an optimal truncation bound, we establish concentration properties of the Gibbs distribution on random regular bipartite graphs. Then we apply high-dimensional expansion and prove trickle-down theorems to obtain fast samplers for the conditioned distributions.
Electric power networks are increasingly exposed to weather-sensitive failure mechanisms that require asset-level, spatially explicit risk modelling for effective intervention planning. This study contributes a modular, robust, and explainable probability-of-failure (PoF) modelling framework for utility asset management. The central contribution is an asset-level architecture that can be scaled to new environmental data sources and additional PoF types without reworking the underlying pipeline. This is particularly relevant for industry settings, where risk models must remain operationally maintainable while adapting to changing data availability, asset-management priorities, and climate-driven hazard conditions. We demonstrate the framework for vegetation-related and lightning-related failure modes using a harmonised geospatial machine-learning pipeline. The implementation integrates multi-source predictors, including topography (SRTM), vegetation condition (MODIS Normalised Difference Vegetation Index - NDVI), lightning climatology (LIS VHRMC), OpenStreetMap-derived proximity features, and utility operational records. The resulting architecture is computationally efficient, operationally extensible, and suitable for utility-scale deployment. It provides actionable asset-level risk stratification for inspection prioritisation, vegetation management, asset hardening, and resilience planning, supporting earlier intervention and more climate-resilient network operations.
Dense text embeddings are widely used in data mining, retrieval, and downstream machine learning systems due to their compact and semantically rich representations, but recent embedding inversion attacks have shown that they can expose substantial information about the original text, leading to serious privacy leakage risks. A common defense is to release perturbed embeddings by adding Gaussian noise, which is simple yet effective against standard inversion attacks and does not significantly degrade embedding utility for downstream tasks. However, it remains unclear whether such noise-protected embeddings are sufficiently safe against adaptive attackers that explicitly account for the perturbation process. In this paper, we study text embedding inversion in a noise-protected setting, where the attacker can observe only noisy embeddings and has no access to clean embedding targets. We first analyze why existing generative inversion methods fail under this setting and identify a "Double Noise Trap", which fundamentally prevents standard generative inversion models from achieving high-quality reconstruction. To address this challenge, we propose DAEI, a denoising-aware embedding inversion pipeline that combines a residual denoising autoencoder with generative text inversion where the denoiser is trained in an unsupervised manner using Stein's unbiased risk estimate to enable denoising from noisy observations alone. Extensive experiments show that DAEI achieves approximately 154\% relative improvement in BLEU over the existing generative inversion baseline, while also improving token-level F1 and ROUGE-L by 32--60\%. The promising inversion performance of DAEI challenges the prevailing assumption that simple Gaussian perturbation is sufficient to prevent sensitive information leakage from embedding representations.
Using reinforcement learning to post-train joint video-audio generation models requires a reward signal. Existing methods construct this reward by combining metrics for individual quality dimensions, including audio quality, visual fidelity, and synchronization. However, these metrics evaluate perceptual dimensions separately and fail to capture the overall semantic and temporal coherence among the text prompt, video, and audio that shapes human preferences. Optimizing models against these metrics encourages reward hacking, generating video-audio content that achieves high scores on these metrics yet appears incoherent or unfaithful to human viewers. To address this problem, we first construct a large-scale human-preference dataset VAPref-10K for joint video-audio generation, comprising 9K prompts and 10.3K fine-grained paired comparisons from open-source generation models. We also introduce the VA-Judger-Bench benchmark with both in-domain and out-of-domain model comparisons to evaluate whether reward models truly align with human preferences. We further propose VA-Judger, a chain-of-thought omni-reward model for joint video-audio generation. In particular, VA-Judger first learns from pairs with clear quality gaps to establish structured output and coarse preference discrimination, then distills reliable preference explanations for harder near-quality comparisons via rejection sampling verified against human annotations, and finally performs dimension-wise reinforcement learning that decomposes human feedback into individual quality dimensions for denser reward signals than a single binary preference label. Experiments show that VA-Judger outperforms metric baselines in predicting human preferences on both in-domain and out-of-domain evaluations. Using its human-aligned rewards for post-training audio-video generation model also yields significant improvements in generation quality.
Platform-scale recommender systems often span multiple business streams such as organic recommendation, advertising, and merchant services, where user behaviors form a continuous cross-stream trajectory. Maintaining separate ranking systems fragments user representations and increases engineering cost. We propose \textbf{OneModel}, a unified framework for multi-stream final ranking. OneModel maps heterogeneous behaviors into shared event sequences, learns long-context user representations with an action-oriented backbone, and introduces \emph{Scenario-aware Information Modulation} to balance cross-stream transfer and stream-specific specialization. For production deployment, OneModel further adopts stratified user representation, multi-objective training, and optimized online serving with feature decomposition, user feature prefetching, shared user-tower computation, and graph-level inference optimization. We deploy OneModel in production at \emph{Xiaohongshu}, where it delivers consistent offline gains over strong baselines and scales favorably with context length and model capacity. Online A/B tests improve Time Spent by \textbf{+0.33\%} and Engagement by \textbf{+1.25\%} in Explore Feed, lift advertising value by \textbf{+3.43\%} and CTR by \textbf{+8.18\%} in Feed Advertising, and raise DGMV by \textbf{+1.1867\%} and GPM by \textbf{+2.1585\%} in Merchant Recommendation, validating unified multi-stream ranking as an effective production foundation.
Public ledgers increasingly authorize state transitions using prior transactions, finalized state, timing, and ordering rather than only a public key, message, and portable signature. We introduce ledger authenticators and $\LAEUF$, an unforgeability experiment for reactive authorization protocols whose public judgment algorithm reads a finalized transcript. The model separates authentication safety from ledger liveness and captures canonical transition freshness, adaptive corruption, exposure before inclusion, censorship, and adversarial ordering. We identify two conditional resource boundaries. An authenticator satisfying our single event conditions yields a contextual one-time signature. Within our rebindable reveal class, safety requires computational post-disclosure non-admissibility. When precursor admission uses only public computation and ledger scheduling, this condition is enforced by closing the evidence eligible to use a disclosed credential. If newly constructed evidence remains admissible after disclosure, censoring the honest reveal gives a forgery. We then define a joint ledger and quantum random oracle execution model in which quantum state persists across classical finalization cuts and oracle evaluations made through the ledger are charged. For a closed finalized target set of size at most $K$, we prove the bound $3β_{\mathsf{cut}}^2+3c_{\mathsf{co}}KQ^2/2^λ+6\ell/2^λ$, where $β_{\mathsf{cut}}$ accounts for fresh openings already present at the cut. A commit, close, reveal authenticator instantiates the framework and obtains a multi-user lifetime QROM bound.
Computer vision is increasingly used to automate recognition tasks in large ecological datasets, but more complex tasks such as multi-object tracking continue to pose challenges. As researchers seek to incorporate vision models in ecology workflows, various lines of research have explored how to make imperfect predictions useful through human-in-the-loop processes. We propose a new approach to working with imperfect tracking predictions through an interactive prediction correction workflow taking place as a conversation with a multimodal large language model, which we tailor to a sonar fish tracking dataset as an initial proof of concept. We investigate the performance of the tool, Molmo2Fish, across guided and unguided tasks, correcting its own predicted tracks and external tracks. We find that Molmo2Fish achieves high performance on fish tracking and track correction tasks, but there is still much room to improve on incorporating natural language guidance. The code and data are publicly available at https://github.com/tidalove/molmo2fish.
Systemic change across the digital research landscape is required to reduce the environmental impact of digital research, but while many researchers and technical professionals are motivated to act, they often lack the skills required to translate motivation into lasting organisational change. We present an open-access course that teaches the foundations of advocacy and organisational change to researchers, research software engineers, and research technical professionals. Structured around the UNICEF five-step advocacy cycle, the course covers stakeholder analysis, power mapping, coalition building, storytelling, framing and messaging, and evaluation. It is grounded in the UK policy landscape, including the Concordat for Environmental Sustainability and the UKRI Environmental Sustainability Strategy, and uses a fictional case study to make concepts concrete. The course is designed as a dual-layer resource in which workshop slides and extended self-study notes are contained in a single source material. A short pilot was delivered at the NetDRIVE Net Zero DRI Summer School in June 2026. The course is freely available under a CC BY 4.0 licence.
Protein language models are widely used priors for protein sequence design, and a growing body of work controls them at inference time as an alternative to fine-tuning. Such guidance faces a dilemma: mild enough to preserve natural activation statistics, it barely moves the property; strong enough to move it, the generations become progressively harder to fold. We show the failure has a specific and cheaply detectable signature, an off-manifold collapse of the model's own representations. Guided activations fall toward a region statistically indistinguishable from random amino-acid input, and the sequences degenerate to low complexity, yet the property oracle being optimized can still score these generations as a success. The optimized oracle can therefore fail to witness the collapse and, for solubility, can actively reward it, whereas structure and composition expose the failure. Because the failure is already visible in a finished candidate, we detect it at the output rather than modify the generator. We introduce a cheap density prior over natural protein activations and keep only the candidates that remain typical under it, a training-free post-hoc step we call Mahalanobis filtering. At matched guidance settings it improves both the property score and the structural plausibility of the sequences it keeps at negligible cost, without touching the generator, and transfers across different guidance methods. We release the activation statistic at https://huggingface.co/Shuibai12138/off-manifold-collapse-plm
Repository-level repair benchmarks still center on a few mainstream languages, leaving systems languages such as Odin largely untested. We present OdinEval, a reproducible benchmark built from documented defects in public Odin repositories. Each instance binds an issue to base and fix commits, a gold patch, an issue-specific regression test, a historical toolchain, and execution records. Admission requires the test to fail on the base revision and pass after the gold fix. When no usable developer test exists, a black-box test is reviewed independently by three instances of the same model, executed in both historical states, and revised from recorded feedback under a versioned Test Writing Skill. We evaluate six language models on 168 filtered instances under one shared protocol. Kimi-K3 records the highest Resolved score at 66.7%, while Qwen3.8-Max has the highest Repro score at 96.4%. The release includes frozen data, source archives, containers, validators, model patches, and audit manifests.
Single-image rephotography aims to synthesize new shots of a scene from a single reference image with specified viewpoints, focal lengths, and photographic effects, which are intrinsically coupled in imaging. Existing methods typically treat these factors separately and struggle under joint control: novel-view synthesis may introduce geometric distortions under focal-length changes, while super-resolution and instruction-guided editing remain confined to 2D and cannot reliably extend detail restoration or appearance control to novel viewpoints. We attribute these limitations to imperfect single-image 3D reconstruction and the sampling limit of continuous focal-length enlargement. To reduce projection bias from geometric errors, we use implicitly transformed foundation-model features for robust target-view guidance. We further formulate focal-length enlargement as a geometry-guided super-resolution problem and exploit generative detail priors to recover details lost during sparse 3D resampling. Built on this 3D-aware generative backbone, we lift photographic-effect control from 2D filtering to 3D-aware appearance editing, preserving content consistency across viewpoints and focal lengths. These components form ReX-Shot, a geometry- and camera-grounded generative framework for single-image rephotography. To our knowledge, ReX-Shot is the first unified framework to jointly control viewpoint, focal length, and parameterized photographic effects from a single image. Experiments show that ReX-Shot outperforms representative baselines across all three controls while enabling near-real-time interactive rephotography.
Whether distinct neural architectures develop common collective dynamics remains an open question. Recent analysis of Transformer language models revealed a nearly flat, weakly infrared-enhanced time-scale density of states (TDOS) associated with near-marginal long-memory dynamics. Here we test whether a closely related organization emerges in Mamba, whose selective state-space dynamics provides a fundamentally different microscopic mechanism. Mamba allows relaxation dynamics to be resolved at three levels: the intrinsic spectrum of the learned state-space generator, its input-conditioned selective rescaling, and the collective TDOS of the complete block measured from its Jacobian. These spectra are not identical: selective dynamics and the remaining block transformations substantially reorganize the microscopic relaxation hierarchy. Nevertheless, the full block develops a reproducible slow-mode continuum whose infrared sector becomes progressively better resolved with increasing sequence length. Cumulative analysis yields $ρ(λ)\simλ^β$, with the long-sequence Mamba exponent stabilizing near $β_{\rm M}\simeq-0.17$. The corresponding memory dynamics follows $K(t)\sim t^{-(1+β)}$, close to the marginal $1/t$ regime. Despite fundamentally different microscopic dynamics, Transformer full-block spectra exhibit closely related infrared organization, with representative exponents of order $β_{\rm Tr}\sim-0.1$. These results separate explicit state-space memory from collective infrared organization and show that distinct sequence architectures can develop closely related near-marginal slow-mode dynamics. They extend infrared collective organization beyond Transformers and provide an independent test of the dynamical structure described by Cognitive Field Theory.
Uniformly allocating inference reasoning budgets to LLMs is expensive and prone to over-thinking penalties; especially in document tasks where visual layouts drive complexity. To address this, we introduce BudgetDoc, the first multimodal benchmark providing explicit supervision for model-budget-performance trade-offs across three document tasks. Using BudgetDoc, we train DRB (Document-Reasoning Balancer), an approx. 1B-parameter pre-flight estimator (SigLIP-2 + Qwen3-0.6B) that predicts ordinal model performance across budget levels, achieving a 0.753 weighted F1. When dynamically allocating reasoning budgets across five frontier models and three datasets, DRB matches or improves F1 scores compared to always-maximum-budget baselines in 9 of 15 configurations while drastically reducing cost. Finally, preliminary evaluations demonstrate DRB's potential to generalize to cross-model selection.
Compact object detectors are suitable for resource-constrained visual perception, but their limited representation capacity creates an accuracy gap relative to large models. Conventional detector distillation often relies on prediction-level supervision or a single feature-alignment target, such as response, distribution, correlation, or frequency-domain matching. Frequency-Decoupled Cross-Attention Knowledge Distillation (FD-CanKD) is presented as a detector-oriented framework that transfers teacher knowledge at three complementary levels: head-level prediction supervision, relation-level non-local context transfer, and frequency-level component-selective alignment. Student features first aggregate teacher-side spatial context through cross-attention-based relation transfer, after which frequency-aware alignment preserves complementary structural and detail-sensitive cues. Under controlled Microsoft Common Objects in Context (COCO) experiments, fixed 50-epoch from-scratch comparisons show that FD-CanKD remains competitive with representative detector knowledge distillation baselines. Post-distillation continued fine-tuning further produces a stronger refinement-ready student than detector-only fine-tuning, reaching 48.87 mean average precision (mAP) at intersection-over-union thresholds from 0.50 to 0.95 (mAP50:95), 65.84 mAP50, and 53.40 mAP75 after 20 additional epochs. All distillation modules are removed after training, leaving the deployed student unchanged at 19.7M parameters. The framework is instantiated and evaluated in a controlled YOLOv12 teacher-student setting as a representative compact-detector case study.
Repository-level LLM agents are typically evaluated on projects whose tests run on the build host. It remains unclear whether their repairs survive the mobile build-install-launch-test boundary, where a missing SDK, offline device, or pre-assertion crash can be mistaken for a program failure. We present AppEval, a benchmark and native-toolchain evaluation framework for mobile application repair across HarmonyOS/ArkTS, iOS/Swift, and Android/Kotlin. Each task separates a hidden behavior test from the reference production fix and is accepted only when the same installed-app target reaches an assertion failure on the defective revision and passes after the fix; infrastructure failures remain a distinct outcome. A common schema maps this contract to each platform's build system, runtime, and test runner. The audited Android partition contains 200 accepted instrumentation tasks from 24 independently buildable repositories. On these tasks, five agents achieve Pass@1 between 22.00% and 90.50%, a 68.50-percentage-point spread under the same dynamic oracle. These results show that mobile repair performance depends strongly on the evaluated agent while demonstrating why runtime-aware acceptance is necessary for meaningful comparison. The quantitative findings in this paper are Android-specific; audited iOS and HarmonyOS results are required before drawing cross-platform generalization conclusions.
Multimodal large language models (MLLMs) are increasingly used as OCR systems in document and knowledge-processing pipelines, but their ability to faithfully read real handwriting remains underexplored. Existing OCR benchmarks focus largely on printed text or clean single-line inputs, leaving limited coverage of realistic handwritten OCR scenarios such as multilingual handwriting, writer errors, and structurally complex mathematical expressions. We introduce OmniHandwritingOCR, a diagnostic benchmark for evaluating MLLMs and OCR systems on handwritten OCR. It covers handwritten text recognition and handwritten mathematical expression recognition across six subtasks and twelve subsets, totaling 77.57K labeled images from public datasets and newly collected student writings. A key component is a difficulty-stratified multi-line formula corpus designed to test robustness under increasing structural complexity. We evaluate thirteen open- and closed-source systems with five complementary metrics under a unified protocol. Results show that current systems remain far from faithful transcription: performance drops sharply on complex multi-line formulas, model rankings vary across language and formula settings, and several generative models hallucinate plausible but visually unsupported corrections. OmniHandwritingOCR provides a challenging testbed for diagnosing language, content, structural, and visual-grounding failure modes of multimodal models in handwritten OCR scenarios.
Vision-based industrial anomaly detectors are calibrated on one distribution but may be deployed on another that differs in illumination, fixture placement, or sensor characteristics, sharply degrading an otherwise accurate detector. Adapting to the incoming lot is a natural response, but labeled anomalies are scarce. We therefore consider calibration using only a handful of verified-normal images available before scoring the rest of the lot. Existing fixes require backpropagation, detector-specific tuning, or choices about feature directions that few calibration samples cannot justify. We present SPARC, a few-shot calibration method that intercepts patch features between encoder and detector and removes a closed-form, spatially indexed estimate of deployment-time nuisance through per-cell subspace projection. It needs only $k \le 8$ verified-normal images and uses the algebraic saturation rank $r{=}k{-}1$ on the encoder's native patch grid. The correction requires no gradient or weight updates and works with memory-bank, density, prototype, and mutual detectors. On the shift-prone benchmarks, SPARC improves pooled Image AUROC and AU-PRO$_{0.3}$ for all seven detectors whose image scores depend on corrected patch features by $+13.8$ and $+3.5$ percentage points (pp), respectively; on benchmarks without engineered shift, the changes are small and mixed. Controls that give competing corrections the same calibration images attribute these gains to the per-cell subspace structure rather than the images alone. Further ablations support the saturation-rank choice and characterize sensitivity to backbone and calibration conditions.
Although Large Language Models (LLMs) encode rich factual knowledge in their parameters, reliably recalling and verifying such knowledge remains a key bottleneck in factual question answering. Existing end-to-end methods entangle knowledge elicitation with reasoning, making it difficult to determine whether correct answers arise from parametric knowledge or the input context. To address this challenge, we propose VAKE (Verifiable Activation of Parametric KnowledgE), a two-stage reinforcement-learning framework that externalizes latent parametric knowledge through explicit Priming and transfers the acquired elicitation capability to implicit Reasoning. Given a query and an insufficient retrieved subgraph, the Priming policy explicitly inserts bridging triples as verifiable evidence, with supervision provided by rewards derived from answers generated by a separate frozen model over the augmented subgraph. Building on the policy learned during Priming, the Reasoning stage trains the model to answer from the original input, testing whether the capability acquired through explicit knowledge elicitation transfers to implicit reasoning. Experiments across seven benchmarks and models from 3B to 14B show that VAKE consistently outperforms standard baselines, including when transferring directly from HotpotQA to OOD datasets. LLM-based evaluation further shows that over 80% of the inserted triples provide factual bridging knowledge not derivable from the retrieved context, while more than half elicit knowledge inaccessible through direct prompting. These results suggest that VAKE activates latent parametric knowledge rather than copying the input context or memorizing dataset-specific associations.
Training terminal agents requires scalable executable supervision, yet synthesizing high-quality terminal tasks remains challenging. Each task couples an instruction, an initialized environment, a reference solution, and an executable verifier; if these artifacts are generated from inconsistent assumptions, the resulting task may be unsolvable or incorrectly evaluated. Meanwhile, multi-stage synthesis can discard the goals, dependencies, state transitions, and procedural constraints encoded in the original sources. We present FACET (Fine-grained Agentic Construction of Executable Tasks), a framework that addresses both information preservation and cross-artifact consistency. FACET reconstructs related agent skills into coherent, information-rich scenarios, then realizes and repairs the execution environment before generating the final task artifacts. The resulting container state serves as shared grounding for the instruction, solution, and verifier, while execution-based validation and targeted repair correct artifact-specific failures without unnecessarily regenerating valid components. FACET produces complex terminal tasks with dense executable checks, and successful trajectories collected from these tasks provide effective, data-efficient supervision. Fine-tuning models across multiple scales consistently improves performance on Terminal-Bench 2.1, while analyses of alternative generation schemes support the importance of environment-grounded construction for task validity and solution-verifier alignment. These results establish source-intent preservation and shared executable-state grounding as key principles for scalable terminal-task synthesis.
Multimodal large language models (MLLMs) have shown strong potential for image quality assessment (IQA) by improving consistency between quality ratings and their underlying reasoning. However, most approaches supervise reasoning through human-provided ratings and rarely examine whether it faithfully reflects image quality. Rating accuracy alone does not ensure faithful reasoning; a shared reward also obscures supervision sources and may reinforce unfaithful reasoning when a correct rating occurs by chance. To improve the faithfulness and reliability of blind IQA, we aim to (1) decouple credit assignment for reasoning and rating and (2) provide verifiable supervision for faithful reasoning. We introduce MR-IQA-2, an actor-editor-judge framework that operationalizes reasoning-editing-reflection. The actor generates quality reasoning for an input image, and the editor revises the image according to the identified quality factors. A frozen judge compares the original and edited images and provides reflective supervision for the actor's reasoning. MR-IQA-2 further uses fine-grained credit assignment to decouple reasoning and rating supervision. Judge feedback supervises reasoning, whereas human ratings supervise the predicted rating. Masked token-specific updates distinguish these signals while preserving the causal relation from reasoning to rating. Across IQA benchmarks, MR-IQA-2 achieves competitive rating alignment with humans. Visual reflection also enables richer and more faithful visual understanding beyond rating, which may inform image-quality optimization and related downstream tasks. Code is available at https://github.com/RobinY99/MR-IQA-2.
Proactive interference (PI) is a documented failure mode in large language models in which retrieval of a repeatedly overwritten value degrades as prior overwrites accumulate, mirroring a classical phenomenon in human working memory. Post-training quantization (PTQ) is now the default deployment path for open-weight models, yet its effect on this failure mode has not been tested. We evaluate three precision levels (FP16, INT8, INT4/NF4, via bitsandbytes) across three architecturally distinct instruction-tuned models (Qwen2.5-7B-Instruct, Mistral-7B-Instruct-v0.3, Phi-3.5-mini-instruct), holding the retrieval task fixed. INT4 quantization significantly reduces accuracy under high interference in every model (e.g., from 81.0% to 68.3% for Qwen), confirmed by paired McNemar's tests ($p \le 2.6 \times 10^{-6}$) and a mixed-effects regression spanning all interference levels; INT8, often assumed safe, also carries a smaller but real penalty in two of three models. The effect is specific to semantically similar (word-type) distractors and reverses sign under a numeric control condition, and is mechanistically linked to a rise in same-key intrusion errors under INT4 (from 21.5% to 24.6% of trials, $p = 4.8 \times 10^{-7}$). A follow-up ablation shows the effect originates in the quantized transformer backbone rather than the output projection layer. These results suggest that bitsandbytes 4-bit quantization can impose an additional cost on applications relying on long, updatable, semantically dense contexts, even when aggregate benchmark accuracy appears largely unaffected. We release our code and tokenizer-verified vocabulary construction method at https://github.com/ShayanShahrabi/compress-and-forget
We study online service with one maximum-waiting-time charge per service batch. Requests arrive at points of a finite metric, and a mobile server pays for its movement and, for each service walk, the maximum waiting time among the requests served by that walk. We distinguish elective service, where an encountered request may be left pending, from automatic service, where every encounter serves it. Although the two semantics have different optimal schedule structures, we prove that their offline optimal values are equal. On a finite line and on an explicitly represented weighted tree, the common offline value is computable by polynomial-time dynamic programming, whereas exact optimization on arbitrary finite metrics is NP-hard.
For the online problem, we prove a metric-independent group-trajectory certificate lemma that charges spatially separated request groups to two parity classes of time windows. It yields deterministic polynomial-time competitive ratios 10 on a line, 12 on a weighted tree, and 20 on an arbitrary finite metric, under both service semantics. With an exact metric-Steiner-tree oracle, the general-metric ratio improves to 12. The polynomial algorithm uses a half-scaled running maximum of terminal-MST weights; the running maximum is necessary because terminal MST weight is not monotone under new arrivals. A fixed two-point line gives a deterministic visible-service lower bound of 3 for every metric class above. Finally, when request locations are hidden until visited, dyadic exploration is 84-competitive on a known finite line. This phenomenon is line-specific: one hidden request gives deterministic and randomized lower bounds 3 and 2 on a line, while a d-leaf unit star gives lower bounds 2d-1 and d.
A convolutional sequence labeler's receptive field is routinely treated as the extent of the model's usable context: it sets dilation schedules, bounds streaming horizons, and underwrites locality claims. However, we show that this can be false: when a normalization layer computes statistics from the current input along the sequence at inference, those statistics open a sequence-spanning path that bypasses the convolutional receptive field to provide global context. We derive this from the layer's Jacobian (the criterion needs no experiment), and what the path carries has a closed form. On a synthetic labeling process with computable optima, the global summary that a sequence-spanning normalization encodes already supplies almost all of what a larger receptive field would buy where labels come in long runs: a network reaching 9 positions comes within 0.009 of the whole-sequence optimum, against a near-chance bound for its reach. Closing the path, by taking the same statistics per position, multiplies what enlarging the receptive field is worth by up to an order of magnitude on simulated genomes at every difficulty level tested and on real 1000 Genomes haplotypes. The same path also confounds attribution: ablating a trained network's receptive-field-enlarging blocks severs part of the path, overstating their contribution 8.3-16.1-fold relative to retraining from scratch. The substitution of normalization for receptive field fades as labels switch more often. Where labels run long, neither the receptive-field justification nor the ablation is wrong about its numbers, but both credit the wrong component.
Showing 701–725 of 2035 papers
« Previous
Page 29 of 82
Next »