Knowledge catalogue
arXiv:2602.02214v3 Announce Type: replace Abstract: To achieve real-time interactive video generation, current methods distill pretrained bidirectional video diffusion models into few-step autoregress
arXiv:2605.21661v1 Announce Type: cross Abstract: Adapting pretrained diffusion models to downstream objectives such as inverse problems often requires expensive test-time guidance or optimization. We
arXiv:2510.23090v2 Announce Type: replace Abstract: Recent advances have investigated the use of pretrained large language models (LLMs) for time-series forecasting by aligning numerical inputs with L
arXiv:2605.22484v1 Announce Type: new Abstract: Vision-Language Models (VLMs) excel at tasks like zero-shot classification and cross-modal retrieval by mapping images and text to a shared space, but t
arXiv:2605.20268v1 Announce Type: cross Abstract: Real-world time series come with text: metadata, descriptions, news, reports. Yet time series foundation models process numerical sequences in isolati
arXiv:2605.20752v1 Announce Type: new Abstract: Vision-language-action (VLA) policies have advanced language-conditioned robotic manipulation by transferring semantic priors from pretrained vision-lan
arXiv:2605.20372v1 Announce Type: new Abstract: Multimodal semantic segmentation benefits remote sensing analysis by combining complementary information from different sensor modalities. In real-world
arXiv:2502.12120v3 Announce Type: replace-cross Abstract: Scaling laws guide the development of large language models (LLMs) by offering estimates for the optimal balance of model size, tokens, and co
arXiv:2602.24138v2 Announce Type: replace Abstract: Automated recognition of surgical phases and steps is a fundamental capability for intraoperative decision support, workflow automation, and skill a
arXiv:2605.21121v1 Announce Type: new Abstract: Single-image-to-3D generative models can now produce high-quality geometry, yet conditioning on a single view inevitably introduces ambiguity about unse
Today we release a study on decoupling the benefits of subword tokenization for language model training, by simulating each suspected benefit one at a time inside a 1.7B byte-level pretraining pipelin
arXiv:2605.21489v1 Announce Type: cross Abstract: Pretrained diffusion models serve as frozen teachers feeding downstream pipelines such as text-to-3D, single-step distillation, and data attribution.
arXiv:2605.19407v1 Announce Type: cross Abstract: We investigate data filtering for large model pretraining via new scaling studies that target the high compute, data-scarce regime. In spite of an app
arXiv:2605.19728v1 Announce Type: new Abstract: Foundation video models produce visually impressive results, but their use in embodied AI remains limited because they are primarily trained on natural
arXiv:2605.18807v1 Announce Type: cross Abstract: Encoder-decoder models offer substantial inference-time savings over decoder-only models, but their pretraining objectives suffer from sparse supervis
arXiv:2601.15014v2 Announce Type: replace-cross Abstract: We study in-context learning for nonparametric regression with alpha-Holder smooth regression functions, for some alpha>0. We prove that, with
arXiv:2602.05709v2 Announce Type: replace Abstract: Low-rank adaptation (LoRA) approximates the update of a pretrained weight matrix using the product of two low-rank matrices. However, standard LoRA
arXiv:2605.20134v1 Announce Type: new Abstract: Learning generalizable trajectory representations from raw GPS traces remains difficult because the data is continuous, noisy, and irregularly sampled.
arXiv:2605.19908v1 Announce Type: new Abstract: Authorship attribution models fine-tuned with the same pretrained encoder, data, and loss can differ four-fold in performance depending only on their sc
arXiv:2601.09722v2 Announce Type: replace-cross Abstract: In this work, we present an annotation framework that demonstrates how a multilingual LLM pretrained on a large corpus can be used as a teache
arXiv:2603.21787v2 Announce Type: replace Abstract: Event cameras are attractive for industrial robotics because they provide high temporal resolution, high dynamic range, and reduced motion blur. How
arXiv:2605.16865v1 Announce Type: new Abstract: Supervised fine-tuning (SFT) is widely used to inject new knowledge into language models, but it often degrades pretrained capabilities such as reasonin
arXiv:2605.17743v1 Announce Type: new Abstract: Continual test-time adaptation adapts a source-pretrained model to non-stationary, unlabeled target streams while retaining past competence, yet texture
arXiv:2605.18359v1 Announce Type: new Abstract: Large multimodal models (LMMs) inherit the self-attention mechanism of pretrained language backbones, yet standard attention can exhibit suboptimal allo
arXiv:2602.10134v2 Announce Type: replace-cross Abstract: Large language models (LLMs) are pretrained on corpora containing trillions of tokens and, therefore, inevitably memorize sensitive informatio
arXiv:2605.18232v1 Announce Type: cross Abstract: Somali is a Cushitic language of the Horn of Africa with ~25 million speakers, yet no documented dedicated Somali pretraining corpus with a companion
arXiv:2605.18383v1 Announce Type: new Abstract: We present TabH2O, a foundation model for tabular data that performs classification and regression in a single forward pass via in-context learning. Tab
arXiv:2602.10503v2 Announce Type: replace Abstract: Pretrained on large-scale and diverse datasets, VLA models demonstrate strong generalization and adaptability as general-purpose robotic policies. H
arXiv:2605.16810v1 Announce Type: new Abstract: We present a training-free framework for occluded text rendering with a pretrained FLUX.1-dev backbone. The task requires a model to render recognizable
arXiv:2605.15220v1 Announce Type: cross Abstract: Data mixing decides how to combine different sources or types of data and is a consequential problem throughout language model training. In pretrainin
arXiv:2605.15687v1 Announce Type: cross Abstract: Multimodal large language models (MLLMs) may memorize sensitive cross-modal information during pretraining, making machine unlearning (MU) crucial. Ex
arXiv:2603.00574v2 Announce Type: replace-cross Abstract: Adapting pretrained multi-modal models to evolving test-time distributions, known as multi-modal test-time adaptation, presents a significant
arXiv:2605.13904v1 Announce Type: cross Abstract: Brain encoder models predict cortical fMRI responses from the internal activations of pretrained vision and language networks, and are typically evalu
arXiv:2605.14406v1 Announce Type: cross Abstract: Large-scale pretraining on Earth observation imagery has yielded powerful representations of the natural and built environment. However, most existing
arXiv:2605.15156v1 Announce Type: cross Abstract: Large language models (LLMs) achieve strong performance across a wide range of tasks, but remain frozen after pretraining until subsequent updates. Ma
arXiv:2605.14063v1 Announce Type: new Abstract: Continual test-time adaptation (CTTA) updates a pretrained model online on an unlabeled, non-stationary stream while anchoring it to a frozen source che
arXiv:2605.14497v1 Announce Type: cross Abstract: Offline-to-online reinforcement learning harnesses the stability of offline pretraining and the flexibility of online fine-tuning. A key challenge lie
arXiv:2605.13382v1 Announce Type: new Abstract: While autoregressive (AR) Vision-Language-Action (VLA) models have demonstrated formidable reasoning capabilities in robotic tasks, their sequential dec
arXiv:2605.12705v1 Announce Type: new Abstract: How can we train models whose post-trained capabilities survive subsequent fine-tuning? Rather than focusing on downstream interventions to mitigate for
arXiv:2605.13152v1 Announce Type: cross Abstract: We introduce EvObj for unsupervised 3D instance segmentation that bridges the geometric domain gap between synthetic pretraining data and real-world p
arXiv:2605.13101v1 Announce Type: cross Abstract: Synthesis planning seeks an efficient sequence of chemical reactions that produce a target molecule. Typically, a pretrained single-step (autoregressi
arXiv:2605.13403v1 Announce Type: cross Abstract: Latent Action Models (LAMs) have emerged as an effective paradigm for handling heterogeneous datasets during Vision-Language-Action (VLA) model pretra
arXiv:2603.29475v2 Announce Type: replace Abstract: Survival analysis is crucial for many medical applications, but remains challenging for modern machine learning due to limited data, censoring, and
arXiv:2605.12756v1 Announce Type: cross Abstract: Large language models (LLMs) are pretrained by minimizing the cross-entropy loss for next-token prediction. In this paper, we study whether this optim
arXiv:2605.11691v1 Announce Type: new Abstract: Partial differential equations (PDEs) govern diverse physical phenomena, yet high-fidelity numerical solutions are computationally expensive and Machine
arXiv:2605.12379v1 Announce Type: new Abstract: Many reinforcement learning (RL) tasks have discrete action spaces, but most generative policy methods based on diffusion and flow matching are designed
arXiv:2605.12019v1 Announce Type: new Abstract: Human Activity Recognition (HAR) is a core task in pervasive computing systems, where models must operate under strict computational constraints while r
arXiv:2605.12009v1 Announce Type: new Abstract: We propose a subgraph importance estimation method for pretrained Graph Neural Networks (GNNs) on graph-level tasks, formulated as a linear Group Lasso
arXiv:2605.12431v1 Announce Type: new Abstract: Conventional gait de-identification methods often encounter an inherent trade-off: they either provide insufficient identity suppression or introduce sp
arXiv:2512.11114v2 Announce Type: replace Abstract: Balancing competing objectives is omnipresent across disciplines, from drug design to autonomous systems. Multi-objective Bayesian optimization is a
arXiv:2605.11881v1 Announce Type: new Abstract: The high-dimensional features extracted from large-scale unlabeled data via various pretrained models with diverse architectures are referred to as hete
arXiv:2605.11872v1 Announce Type: new Abstract: Orthogonal parameter-efficient fine-tuning (PEFT) adapts pretrained weights through structure-preserving multiplicative transformations, but existing me
arXiv:2605.12449v1 Announce Type: new Abstract: While self-supervised pretraining has reduced vision systems' reliance on synthetic data, simulation remains an indispensable tool for closed-loop optim
arXiv:2312.06950v3 Announce Type: replace-cross Abstract: Fully fine-tuning pretrained large-scale transformer models has become a popular paradigm for video-language modeling tasks, such as temporal
arXiv:2602.02408v4 Announce Type: replace Abstract: Model editing aims to correct errors in large, pretrained models without altering unrelated behaviors. While some recent works have edited vision-la
arXiv:2605.11659v1 Announce Type: new Abstract: Cross-Domain Few-Shot Learning (CDFSL) aims to adapt large-scale pretrained models to specialized target domains with limited samples, yet the few-shot
arXiv:2602.05902v2 Announce Type: replace-cross Abstract: Post-training quantization (PTQ) enables efficient deployment of large language models by mapping pretrained weights to low-bit formats withou
arXiv:2605.10337v1 Announce Type: new Abstract: Intracranial electrocorticography (ECoG) offers high-signal-to-noise access to cortical activity for brain-computer interfaces, yet limited per-patient
arXiv:2605.09081v1 Announce Type: cross Abstract: We introduce the first universal pretraining corpus for industrial time-series data: FactoryNet. 51M datapoints across 23k end-to-end task executions
arXiv:2605.09438v1 Announce Type: new Abstract: Many features in pretrained Transformers span multiple layers: they emerge through stages of inference, persist in the residual stream, or are built joi