The Loss Does Not See the Basis, but Adam Does
arXiv:2608.05136v1 Announce Type: new Abstract: Gradient descent on a factored model W = UV^op is implicitly biased toward low-rank solutions, while Adam, starting from the same small initialization,
Knowledge catalogue
arXiv:2608.05136v1 Announce Type: new Abstract: Gradient descent on a factored model W = UV^op is implicitly biased toward low-rank solutions, while Adam, starting from the same small initialization,
arXiv:2608.04864v1 Announce Type: cross Abstract: We introduce the neural echo as a tool for understanding the behavior of neural networks. It generalizes the model-based concepts of impulse responses
arXiv:2608.04611v1 Announce Type: cross Abstract: Frontier coding models now match or exceed strong human reference points on programming benchmarks, yet benchmark success does not imply maintainable
arXiv:2608.04570v1 Announce Type: new Abstract: Personalized LLMs with persistent memory are increasingly deployed, yet the faithfulness of their user models remains unexamined. We study over-inferenc
arXiv:2608.04432v1 Announce Type: cross Abstract: On two-sided content platforms, symmetric two-sided isolation (assigning matched fractions of creators and viewers to isolated treatment and control s
arXiv:2608.04285v1 Announce Type: new Abstract: Neurosymbolic AI systems that integrate machine learning and symbolic reasoning are rapidly gaining attention. They complement the data-intensive statis
arXiv:2608.04686v1 Announce Type: new Abstract: We study distributionally robust PAC learning for the 0--1-loss, where adversarial perturbations of the data distribution are constrained by a Cressie--
arXiv:2508.12480v3 Announce Type: replace Abstract: The ability to cooperate with unknown partners is a central challenge in cooperative AI and widely studied in the form of zero-shot coordination (ZS
arXiv:2608.04424v1 Announce Type: new Abstract: Existing document understanding benchmarks have largely focused on locating page elements, yet real-world document intelligence requires models to reaso
arXiv:2608.04222v1 Announce Type: cross Abstract: Turbulence is a central testbed for machine learning on physical dynamics because its governing laws are known exactly. However, most existing studies
arXiv:2608.04436v1 Announce Type: new Abstract: Text-to-image (T2I) models can produce visually compelling images, yet they remain limited on open-world tasks that require complex semantic understandi
arXiv:2603.18409v2 Announce Type: replace Abstract: Current document chunking methods for Retrieval-Augmented Generation (RAG) typically linearize text. This forced linearization strips away intrinsic
arXiv:2411.06024v1 Announce Type: cross Abstract: The exponential growth in protein-related databases and scientific literature, combined with increasing demands for efficient biological information r
arXiv:2601.22124v2 Announce Type: replace Abstract: Large language models (LLMs) are increasingly adapted for medical applications, but most are trained using data from a single institution because pr
arXiv:2608.04732v1 Announce Type: cross Abstract: Safe actor-critic control often treats barrier filtering, uncertainty estimation, and experience replay as separate modules, even though each changes
arXiv:2608.05139v1 Announce Type: new Abstract: Long-horizon reasoning in recent LLMs demands that the model switch between distinct skills inside a reasoning chain, such as first doing a math derivat
arXiv:2608.04011v1 Announce Type: cross Abstract: This article examines the enduring epistemic and methodological crisis of traditional legal practice in light of the opportunities and constraints int
arXiv:2608.04840v1 Announce Type: cross Abstract: Verifying the authenticity of satellite imagery has become increasingly critical given advances in generative artificial intelligence. Highly realisti
arXiv:2608.04260v1 Announce Type: new Abstract: Metaphorical language remains a major challenge for multilingual natural language processing because successful interpretation and translation require r
arXiv:2608.05000v1 Announce Type: new Abstract: Vision offers a critical axis for advancing foundation models, driving a shift towards natively unified multimodal pretraining. Despite this momentum, t
arXiv:2608.04377v1 Announce Type: cross Abstract: Hypergraph neural networks (HGNNs) have demonstrated remarkable capabilities in processing complex higher-order relationships. However, their performa
arXiv:2502.15952v3 Announce Type: replace Abstract: Recent works exploring the training dynamics of homogeneous neural network weights under gradient flow with small initialization have established th
arXiv:2608.04955v1 Announce Type: new Abstract: Multi-stage boundary representation (B-Rep) generation leverages intermediate wireframes to synthesize CAD models. However, geometric and topological ri
arXiv:2608.04759v1 Announce Type: cross Abstract: Although Multimodal Large Language Models (MLLMs) have made substantial progress, their spatial reasoning may still produce intermediate judgments inc
arXiv:2608.04697v1 Announce Type: new Abstract: Operational hazard analysis of aviation system operations must consider interactions among weather, ATC actions, airspace constraints, aircraft operatio
arXiv:2608.04879v1 Announce Type: cross Abstract: Vision Transformers (ViTs) achieve strong image-recognition performance, but their parameter count grows linearly with depth when each block is indepe
arXiv:2608.04405v1 Announce Type: cross Abstract: Long-context large language models (LLMs) are increasingly deployed in real-world applications, yet self-attention remains a major efficiency bottlene
arXiv:2608.04015v1 Announce Type: new Abstract: With the increase in digitized resources of Classical Latin texts and modern breakthroughs of Large Language Models (LLMs), I contribute to ancient lang
arXiv:2608.04230v1 Announce Type: cross Abstract: Deep learning models for scientific spatio-temporal downscaling often minimize reconstruction error while failing to preserve physically meaningful mu
arXiv:2608.04606v1 Announce Type: new Abstract: Understanding complex surgical scenes requires recognizing multiple interdependent entities, such as instruments, actions, and targets, while maintainin
arXiv:2608.04175v1 Announce Type: new Abstract: Edge platforms used for aerial observation must interpret aircraft imagery under limited memory, limited compute, and intermittent connectivity. This se
arXiv:2608.04317v1 Announce Type: cross Abstract: Autonomous cyber defense systems based on Deep Reinforcement Learning (DRL) have attracted significant research attention, yet remain evaluated almost
arXiv:2608.04154v1 Announce Type: cross Abstract: Mapping paddy rice from very-high-resolution imagery in mountainous and hilly regions is difficult because terrain alters optical appearance and incre
arXiv:2608.04460v1 Announce Type: cross Abstract: The quantitative analysis of 3D neuronal morphologies requires capturing both graph topology and spatial geometry. Current message-passing Graph Neura
arXiv:2608.04174v1 Announce Type: new Abstract: Time series data are ubiquitous in practical applications, where classification (TSC) and extrinsic regression (TSER) have emerged as essential tasks fo
arXiv:2608.04453v1 Announce Type: cross Abstract: Online HD map construction is critical to prediction and planning in autonomous driving. We find that existing physical attacks against online map con
arXiv:2607.19322v2 Announce Type: replace Abstract: Rubric-based evaluation of open-ended generation faces a fundamental tension between expressiveness and reliability. Authoring a faithful rubric req
arXiv:2608.04429v1 Announce Type: new Abstract: Backlit and low-light images often suffer from severe exposure imbalance or global underexposure, presenting significant challenges for both visual perc
arXiv:2608.04949v1 Announce Type: cross Abstract: Unified Multimodal Relation Extraction (UMRE) aims to identify intra-modal and cross-modal relations between textual entities and visual objects. Howe
arXiv:2508.04780v2 Announce Type: replace-cross Abstract: The increasing frequency of extreme weather events, such as hurricanes, highlights the urgent need for efficient and equitable power system re
arXiv:2608.04173v1 Announce Type: new Abstract: Deep neural networks (DNNs) deployed on resource-constrained neuromorphic hardware face three concurrent challenges: the need for model compression thro
arXiv:2409.02228v2 Announce Type: replace-cross Abstract: When language models (LMs) are trained to forget (or 'unlearn'') a skill, how precisely does their behavior change? We study the behavior of t
arXiv:2608.04055v1 Announce Type: cross Abstract: The minimum change principle provides an information-theoretic characterization of the Bayes reversal channel in classical probability theory and has
arXiv:2608.04701v1 Announce Type: new Abstract: The abundance of casually captured monocular videos and images on social media provides a valuable source for immersive content creation, where generati
arXiv:2608.04935v1 Announce Type: new Abstract: Recent work has shown that a simple linear probe on frozen representations from modern vision foundation models (VFMs) can achieve state-of-the-art AIGI
arXiv:2608.04201v1 Announce Type: new Abstract: State estimation for nonlinear dynamical systems is commonly performed with the Unscented Kalman filter (UKF), which propagates the state moments throug
arXiv:2608.04882v1 Announce Type: new Abstract: We introduce a variational approach to a finite-temperature continuous-spin perceptron trained on a Gaussian mixture. The model allows for a broad class
arXiv:2608.04623v1 Announce Type: new Abstract: Zero-shot Skeleton Action Recognition (ZSAR) remains ambiguous when unseen actions share similar skeleton joint dynamics but differ in objects or scene
arXiv:2608.04902v1 Announce Type: new Abstract: Video-to-audio (V2A) generation extends image-to-audio generation (I2A) by introducing consecutive frames that provide essential temporal cues for audio
arXiv:2608.04170v1 Announce Type: cross Abstract: AI co-scientists can generate fluent materials-science hypotheses, but fluency does not show that an answer preserves a scientifically meaningful mech
arXiv:2608.04557v1 Announce Type: new Abstract: High-fidelity 3D MRI synthesis requires both globally coherent anatomy and fine-grained voxel-level detail. Although latent diffusion makes volumetric g
arXiv:2608.05069v1 Announce Type: cross Abstract: Video Anomaly Detection (VAD) is inherently challenging due to the scarcity of anomalies and the large visual variability in surveillance footage, inc
arXiv:2606.23462v2 Announce Type: replace-cross Abstract: Scientists do not, by profession, wage war. Yet warfare's vocabulary consistently appears in their abstracts. To quantify the extent to which
arXiv:2608.04562v1 Announce Type: new Abstract: Agent skills are increasingly optimized by automated feedback loops, producing long structured artifacts whose internal value remains unclear. We study
arXiv:2608.04714v1 Announce Type: cross Abstract: Benchmark scores are reported as properties of a model, yet the inference framework used to produce them, such as HuggingFace, vLLM, or Ollama, are co
arXiv:2608.04591v1 Announce Type: cross Abstract: Large language models (LLMs) are often asked whether something is absent from a record, list, or retrieved context. Yet non-observation licenses a neg
arXiv:2608.04820v1 Announce Type: new Abstract: Face inpainting with diffusion models has recently achieved impressive visual quality, yet preserving identity fidelity under significant occlusion and
arXiv:2608.04893v1 Announce Type: cross Abstract: Multi-agent LLM systems relay key--value caches instead of text and credit their gains to exchanged ``latent thoughts''. That credit is a claim about
arXiv:2608.04448v1 Announce Type: cross Abstract: Diffusion transformers deliver strong image generation, but their training cost grows superlinearly with resolution. Recent work justifies training or
arXiv:2510.20351v3 Announce Type: replace-cross Abstract: Large language models (LLMs) are increasingly exposed to data contamination, i.e., performance gains driven by prior exposure of test datasets