paper-with-me

Papers

SelFusion: Self-distillation for Diffusion Language Models

2026-08-24 · Hyeongsoo Lim, Jinyoung Kim, Eunseo Seo, Minho Jang, Jiwon Yoon arxiv

Diffusion language models (DLMs) alleviate the inherent latency bottleneck of autoregressive (AR) large language models (LLMs), but their degraded generation quality limits practical applicability. Although knowledge distillation (KD) can be a promising direction for improving performance, we empirically find that naively applying conventional KD yields only marginal gains, or even degrades generation quality. Based on these observations, we propose a novel self-distillation framework for DLMs, namely SelFusion. To enable effective KD without an external teacher model, SelFusion performs two forward passes with different masking levels, defining the hard mode with a larger masking probability and the easy mode with a smaller masking probability. However, the easy mode is not always more accurate than the hard mode and can be overconfident on incorrect tokens. Thus, we introduce bidirectional KD between the two modes, which can dynamically determine the distillation direction based on token-level correctness. Experimental results on instruction-following tasks show that the proposed self-distillation substantially outperforms other KD methods with external LLM and DLM teachers. In many configurations, the student trained with SelFusion even surpasses the performance of the LLM teacher, providing a practical path toward improving DLM generation quality. Source code can be found at https://github.com/scai-research/SelFusion_official

📄 PDF Abstract BibTeX arXiv:2608.22898

Code (0)

등록된 구현이 없습니다.

Tasks

Knowledge Distillation

Similar Papers 제목 키워드 기반

VesselFusion: Diffusion Models for Vessel Centerline Extraction from 3D CT Images

2026-03-09 · Soichi Mita, Shumpei Takezaki, Ryoma Bise arxiv

Vessel centerline extraction from 3D CT images is an important task because it reduces annotation effort to build a model that estimates a vessel structure. It is challenging to estimate natural vessel structures since c…

Diffusion Self-Distillation for Zero-Shot Customized Image Generation

2024-11-27 · CVPR 2025 1 · Shengqu Cai, Eric Chan, Yunzhi Zhang, Leonidas Guibas 외

Text-to-image diffusion models produce impressive results but are frustrating tools for artists who desire fine-grained control. For example, a common use case is to create images of a specific instance in novel contexts…

Image GenerationLanguage ModelingLanguage Modelling

Diffusion Model is Effectively Its Own Teacher

2025-01-01 · CVPR 2025 1 · Xinyin Ma, Runpeng Yu, Songhua Liu, Gongfan Fang 외

In this paper, we introduce a novel self-distillation paradigm for improving the performance of diffusion models. Previous studies have shown that introducing a teacher to distill the diffusion model can enhance its …

model

GDSD: Reinforcement Learning as Guided Denoiser Self-Distillation for Diffusion Language Models

2026-05-28 · Xiaohang Tang, Keyue Jiang, Che Liu, Qifang Zhao 외 arxiv

Reinforcement learning (RL) can be used to improve the policy (denoiser) of diffusion large language models (dLLMs), while being hindered by the intractability of the policy likelihood. A dominant and efficient family of…

Reinforcement Learning

Diversity-Driven Generative Dataset Distillation Based on Diffusion Model with Self-Adaptive Memory

2025-05-26 · Mingzhuo Li, Guang Li, Jiafeng Mao, Takahiro Ogawa 외

Dataset distillation enables the training of deep neural networks with comparable performance in significantly reduced time by compressing large datasets into small and representative ones. Although the introduction of g…

Dataset DistillationDiversity