74 citations · 130 across the 36 of their papers we have counts for
38 papers
Flex-Forcing: Towards a Unified Autoregressive and Bidirectional Video Diffusion Model
Xinyin Ma, Julius Berner, Chao Liu +3
Recent progress in large-scale generative models has substantially advanced video generation, yet existing methods remain constrained by a rigid inference paradigm. Bidirectional d…
dMoE: dLLMs with Learnable Block Experts
Sicheng Feng, Zigeng Chen, Gongfan Fang +2
Diffusion Large Language Models (dLLMs) have recently emerged as a promising alternative to autoregressive models, offering competitive performance while naturally supporting paral…
Q-ARVD: Quantizing Autoregressive Video Diffusion Models
Siao Tang, Xinyin Ma, Gongfan Fang +2
Autoregressive video diffusion models (ARVDs) have emerged as a promising architecture for streaming video generation, paving the way for real-time interactive video generation and…
Mix-Quant: Quantized Prefilling, Precise Decoding for Agentic LLMs
Haiquan Lu, Zigeng Chen, Gongfan Fang +2
LLM agents have recently emerged as a powerful paradigm for solving complex tasks through planning, tool use, memory retrieval, and multi-step interaction. However, these agentic w…
DMax: Aggressive Parallel Decoding for dLLMs
Zigeng Chen, Gongfan Fang, Xinyin Ma +2
We present DMax, a new paradigm for efficient diffusion language models (dLLMs). It mitigates error accumulation in parallel decoding, enabling aggressive decoding parallelism whil…
Invisible Safety Threat: Malicious Finetuning for LLM via Steganography
Guangnian Wan, Xinyin Ma, Gongfan Fang +1
Understanding and addressing potential safety alignment risks in large language models (LLMs) is critical for ensuring their safe and trustworthy deployment. In this paper, we high…
Self-Purification Mitigates Backdoors in Multimodal Diffusion Language Models
Guangnian Wan, Qi Li, Gongfan Fang +2
Multimodal Diffusion Language Models (MDLMs) have recently emerged as a competitive alternative to their autoregressive counterparts. Yet their vulnerability to backdoor attacks re…
dVoting: Fast Voting for dLLMs
Sicheng Feng, Zigeng Chen, Xinyin Ma +2
Diffusion Large Language Models (dLLMs) represent a new paradigm beyond autoregressive modeling, offering competitive performance while naturally enabling a flexible decoding proce…
In-Video Instructions: Visual Signals as Generative Control
Gongfan Fang, Xinyin Ma, Xinchao Wang
Large-scale video generative models have recently demonstrated strong visual capabilities, enabling the prediction of future frames that adhere to the logical and physical cues in…
MixReasoning: Switching Modes to Think
Haiquan Lu, Gongfan Fang, Xinyin Ma +2
Reasoning models enhance performance by tackling problems in a step-by-step manner, decomposing them into sub-problems and exploring long chains of thought before producing an answ…
dParallel: Learnable Parallel Decoding for dLLMs
Zigeng Chen, Gongfan Fang, Xinyin Ma +2
Diffusion large language models (dLLMs) have recently drawn considerable attention within the research community as a promising alternative to autoregressive generation, offering p…
SparseD: Sparse Attention for Diffusion Language Models
Zeqing Wang, Gongfan Fang, Xinyin Ma +2
While diffusion language models (DLMs) offer a promising alternative to autoregressive models (ARs), existing open-source DLMs suffer from high inference latency. This bottleneck i…
Thinkless: LLM Learns When to Think
Gongfan Fang, Xinyin Ma, Xinchao Wang
Reasoning Language Models, capable of extended chain-of-thought reasoning, have demonstrated remarkable performance on tasks requiring complex logical inference. However, applying…
ConciseHint: Boosting Efficient Reasoning via Continuous Concise Hints during Generation
Siao Tang, Xinyin Ma, Gongfan Fang +1
Recent advancements in large reasoning models (LRMs) like DeepSeek-R1 and OpenAI o1 series have achieved notable performance enhancements on complex reasoning tasks by scaling up t…
VeriThinker: Learning to Verify Makes Reasoning Model Efficient
Zigeng Chen, Xinyin Ma, Gongfan Fang +2
Large Reasoning Models (LRMs) excel at complex tasks using Chain-of-Thought (CoT) reasoning. However, their tendency to overthinking leads to unnecessarily lengthy reasoning chains…
dKV-Cache: The Cache for Diffusion Language Models
Xinyin Ma, Runpeng Yu, Gongfan Fang +1
Diffusion Language Models (DLMs) have been seen as a promising competitor for autoregressive language models. However, diffusion language models have long been constrained by slow…
Dimple: Discrete Diffusion Multimodal Large Language Model with Parallel Decoding
Runpeng Yu, Xinyin Ma, Xinchao Wang
In this work, we propose Dimple, the first Discrete Diffusion Multimodal Large Language Model (DMLLM). We observe that training with a purely discrete diffusion approach leads to s…
Efficient Reasoning Models: A Survey
Sicheng Feng, Gongfan Fang, Xinyin Ma +1
Reasoning models have demonstrated remarkable progress in solving complex and logic-intensive tasks by generating extended Chain-of-Thoughts (CoTs) prior to arriving at a final ans…
Introducing Visual Perception Token into Multimodal Large Language Model
Runpeng Yu, Xinyin Ma, Xinchao Wang
To utilize visual information, Multimodal Large Language Model (MLLM) relies on the perception process of its vision encoder. The completeness and accuracy of visual perception sig…
CoT-Valve: Length-Compressible Chain-of-Thought Tuning
Xinyin Ma, Guangnian Wan, Runpeng Yu +2
Chain-of-Thought significantly enhances a model's reasoning capability, but it also comes with a considerable increase in inference costs due to long chains. With the observation t…
Remix-DiT: Mixing Diffusion Transformers for Multi-Expert Denoising
Gongfan Fang, Xinyin Ma, Xinchao Wang
Transformer-based diffusion models have achieved significant advancements across a variety of generative tasks. However, producing high-quality outputs typically necessitates large…
TinyFusion: Diffusion Transformers Learned Shallow
Gongfan Fang, Kunjun Li, Xinyin Ma +1
Diffusion Transformers have demonstrated remarkable capabilities in image generation but often come with excessive parameterization, resulting in considerable inference overhead in…
Collaborative Decoding Makes Visual Auto-Regressive Modeling Efficient
Zigeng Chen, Xinyin Ma, Gongfan Fang +1
In the rapidly advancing field of image generation, Visual Auto-Regressive (VAR) modeling has garnered considerable attention for its innovative next-scale prediction approach. Thi…
LiteFocus: Accelerated Diffusion Inference for Long Audio Synthesis
Zhenxiong Tan, Xinyin Ma, Gongfan Fang +1
Latent diffusion models have shown promising results in audio generation, making notable advancements over traditional methods. However, their performance, while impressive with sh…