end-to-end

该论文提出一种基于预训练text-to-text模型和Read–Tag–Write解码策略的高效流式端到端逆文本正规范化（ITN）系统，在越南语数据集上达到与非流式基线相当的精度，同时满足实时延迟要求。

WeVisDoc is a two-stage data-centric framework for robust end-to-end document parsing that expands coverage and uses targeted diagnostics to improve performance, achieving state-of-the-art results on benchmarks.

VibeVoice-ASR-Streaming is an LLM-based end-to-end model for streaming speaker-attributed speech recognition, achieving state-of-the-art performance with released 1.5B and 7B model weights.

PriorEye introduces geospatial visual priors for end-to-end autonomous driving, enhancing anticipatory behavior and robustness through a dual-memory architecture, as presented at ECCV 2026.

DialectS2S is an end-to-end speech dialogue model for low-resource Chinese dialects, introducing a scalable data synthesis pipeline and a two-stage post-training strategy with self-aligned speech supervision. Experiments show improvements in dialect consistency, response quality, and intelligibility, with fully open-sourced models, data, and code.

SimWAM is a simple yet effective World Action Model for end-to-end autonomous driving that uses video generation purely as a training signal, achieving state-of-the-art 91.5 PDMS on NAVSIM while reducing inference latency.

PaDoc introduces a layout-grounded parallel decoding method for end-to-end document parsers, decoupling layout and content decoding to reduce decoding depth and improve throughput. It achieves state-of-the-art results on OmniDocBenchFull and significantly speeds up inference compared to sequential baselines.

Proposes LatentRM, a reward modeling framework that learns intermediate reasoning traces as discrete latent variables to explicitly maximize downstream scalar reward likelihood, improving preference modeling and policy alignment across in-distribution and OOD tasks.

A robot can be fast on average but still fail when the scene changes; the LingBot-VA 2.0 design addresses this by pairing foresight reasoning with regrounding, achieving a fourfold speedup while ensuring state-reality closeness.

This paper presents the first systematic study of federated training for SpeechLLM-based end-to-end ASR systems, evaluating on English and Italian tasks and achieving competitive word error rates with reduced communication costs.

WeNet is an open-source end-to-end speech recognition toolkit that unifies training code, pretrained models, and deployment runtimes in one project, supporting multiple model architectures and runtime backends.

This paper introduces a novel training criterion to reduce reliance on shortcuts in automated spoken English proficiency assessment systems, aiming to prevent test-takers from inflating scores without genuine improvement. Experiments on audio and text-based systems show reduced correlation with exploitable features, bringing automated scores closer to human judgment.

Kimi K3 demonstrates consistent improvement in end-to-end knowledge work, as evidenced by public benchmarks and internal reviews, indicating enhanced agent capabilities.

OvisOCR2 is a 0.8B end-to-end document-parsing VLM that tops the OmniDocBench leaderboard, outperforming pipeline OCR systems on real-world scanned medical documents with higher accuracy and efficiency.

OvisOCR2 is a 0.8B parameter end-to-end document parsing model that converts document page images to Markdown, achieving state-of-the-art scores on public benchmarks through a combination of supervised fine-tuning, reinforcement learning, and model fusion.

Announcing RT-DocLayout, the world's first layout analysis model capable of pixel-level multi-point polygon boxes for real-world documents, accepted to ECCV 2026. The paper argues that layout analysis remains critical despite end-to-end models, and presents an efficient 33M parameter model running at 132.1 FPS.

COALA is a robust framework for contextual biasing in automatic speech recognition (ASR) that uses a contrastive regularizer and biasing score estimation to improve recognition of domain-specific entities from large biasing lists. Experiments on LibriSpeech show consistent superior performance.

This paper presents a general neurosymbolic reasoning and learning methodology that modularly integrates Answer Set Programming with an Energy Based Model substrate, supporting joint optimization in continuous latent space and end-to-end training. It demonstrates applications on MNIST, CLEVR, and MOT benchmarks.

Researchers at UMA_Robots developed a single end-to-end neural network that operates robustly for hours with precision and adaptation to visual input.

LeVLJEPA is the first fully non-contrastive end-to-end vision-language pretraining method, achieving competitive performance with CLIP and SigLIP without negatives, temperature, or momentum encoder. It learns via cross-modal prediction with stop-gradient targets and per-modality distributional regularization, providing stronger dense semantic features for downstream tasks like VLM backbones and semantic segmentation.