English
Related papers

Related papers: Revealing the Attention Floating Mechanism in Mask…

200 papers

While the Self-Attention mechanism in the Transformer model has proven to be effective in many domains, we observe that it is less effective in more diverse settings (e.g. multimodality) due to the varying granularity of each token and the…

Computer Vision and Pattern Recognition · Computer Science 2024-06-06 Wayner Barrios , SouYoung Jin

Masked diffusion models (MDMs) are a promising alternative to autoregressive models (ARMs), but they suffer from inherently much higher training variance. High variance leads to noisier gradient estimates and unstable optimization, so even…

Machine Learning · Computer Science 2026-05-22 Mengni Jia , Mengyu Zhou , Yihao Liu , Xiaoxi Jiang , Guanjun Jiang

Masked diffusion models (MDMs) are a potential alternative to autoregressive models (ARMs) for language generation, but generation quality depends critically on the generation order. Prior work either hard-codes an ordering (e.g., blockwise…

Machine Learning · Computer Science 2026-05-22 Chunsan Hong , Sanghyun Lee , Jong Chul Ye

Large diffusion vision-language models (LDVLMs) have recently emerged as a promising alternative to autoregressive models, enabling parallel decoding for efficient inference and leveraging bidirectional attention for global context. Despite…

Computer Vision and Pattern Recognition · Computer Science 2026-05-20 Sujung Hong , Chanyong Yoon , Seong Jae Hwang

The research explores the utilization of a deep learning model employing an attention mechanism in medical text mining. It targets the challenge of analyzing unstructured text information within medical data. This research seeks to enhance…

Computation and Language · Computer Science 2024-06-04 Lingxi Xiao , Muqing Li , Yinqiu Feng , Meiqi Wang , Ziyi Zhu , Zexi Chen

Long-context understanding is crucial for many NLP applications, yet transformers struggle with efficiency due to the quadratic complexity of self-attention. Sparse attention methods alleviate this cost but often impose static, predefined…

Computation and Language · Computer Science 2025-06-16 Hanzhi Zhang , Heng Fan , Kewei Sha , Yan Huang , Yunhe Feng

Transformer-based diffusion models have recently superseded traditional U-Net architectures, with multimodal diffusion transformers (MM-DiT) emerging as the dominant approach in state-of-the-art models like Stable Diffusion 3 and Flux.1.…

Computer Vision and Pattern Recognition · Computer Science 2025-08-12 Joonghyuk Shin , Alchan Hwang , Yujin Kim , Daneul Kim , Jaesik Park

Self-attention is the key mechanism of transformers, which are the essential building blocks of modern foundation models. Recent studies have shown that pure self-attention suffers from an increasing degree of rank collapse as depth…

Machine Learning · Computer Science 2024-11-04 Xinyi Wu , Amir Ajorlou , Yifei Wang , Stefanie Jegelka , Ali Jadbabaie

Channel estimation is one of the key issues in practical massive multiple-input multiple-output (MIMO) systems. Compared with conventional estimation algorithms, deep learning (DL) based ones have exhibited great potential in terms of…

Information Theory · Computer Science 2021-08-24 Jiabao Gao , Mu Hu , Caijun Zhong , Geoffrey Ye Li , Zhaoyang Zhang

Recent Speech Large Language Models~(LLMs) have achieved impressive capabilities in end-to-end speech interaction. However, the prevailing autoregressive paradigm imposes strict serial constraints, limiting generation efficiency and…

Computation and Language · Computer Science 2026-02-10 Ziyang Cheng , Yuhao Wang , Heyang Liu , Ronghua Wu , Qunshan Gu , Yanfeng Wang , Yu Wang

Block discrete diffusion language models factorize a sequence autoregressively over fixed-size positional blocks, decoupling within-block parallel denoising from across-block conditioning. We argue that this rigid partition wastes structure…

Computation and Language · Computer Science 2026-05-18 Yichen Zhu , Xiaoming Shi , Peng Zhao , Weiyu Chen , Debing Zhang , James Kwok

Autoregressive language models (ARMs) suffer from the reversal curse: after learning ''$A$ is $B$,'' they often fail on the reverse query ''$B$ is $A$.'' Masked diffusion language models (MDMs) exhibit this failure in a much weaker form,…

Artificial Intelligence · Computer Science 2026-05-13 Moongyu Jeon , Sangwoo Shin , BumJun Kim , Kyelim Lee , Albert No

Masked Diffusion Language Models (MDLMs) enable parallel token decoding, providing a promising alternative to the sequential nature of autoregressive generation. However, their iterative denoising process remains computationally expensive…

Computation and Language · Computer Science 2026-03-10 Younjoo Lee , Junghoo Lee , Seungkyun Dan , Jaiyoung Park , Jung Ho Ahn

In recent advancements in high-fidelity image generation, Denoising Diffusion Probabilistic Models (DDPMs) have emerged as a key player. However, their application at high resolutions presents significant computational challenges. Current…

Computer Vision and Pattern Recognition · Computer Science 2023-12-01 Jing Nathan Yan , Jiatao Gu , Alexander M. Rush

The success of Transformer language models is widely credited to their dot-product attention mechanism, which interweaves a set of key design principles: mixing information across positions (enabling multi-token interactions),…

Computation and Language · Computer Science 2025-10-14 Huiyin Xue , Nafise Sadat Moosavi , Nikolaos Aletras

Attention mechanisms represent a fundamental paradigm shift in neural network architectures, enabling models to selectively focus on relevant portions of input sequences through learned weighting functions. This monograph provides a…

Machine Learning · Computer Science 2026-01-08 Hasi Hays

The success of Transformer-based Language Models (LMs) stems from their attention mechanism. While this mechanism has been extensively studied in explainability research, particularly through the attention values obtained during the forward…

Computation and Language · Computer Science 2024-12-24 Shahar Katz , Lior Wolf

Diffusion language models (dLMs) have emerged as a promising paradigm that enables parallel, non-autoregressive generation, but their learning efficiency lags behind that of autoregressive (AR) language models when trained from scratch. To…

Memory units have been widely used to enrich the capabilities of deep networks on capturing long-term dependencies in reasoning and prediction tasks, but little investigation exists on deep generative models (DGMs) which are good at…

Machine Learning · Computer Science 2016-05-31 Chongxuan Li , Jun Zhu , Bo Zhang

Diffusion models, while increasingly adept at generating realistic images, are notably hindered by hallucinations -- unrealistic or incorrect features inconsistent with the trained data distribution. In this work, we propose Adaptive…

Computer Vision and Pattern Recognition · Computer Science 2025-02-25 Trevine Oorloff , Yaser Yacoob , Abhinav Shrivastava