Tag
The paper proposes 'Early-Bird Decoding,' a framework to accelerate diffusion large language models by using learnable block sizes and parallel sampling, achieving significant throughput improvements without modifying pretrained weights.
The paper proposes Representation-based Masked Diffusion Model (RMDM), which leverages text representations to improve parallel token updates in masked diffusion models, enhancing generation quality especially in few-step sampling.
This paper identifies an anchor collapse phenomenon in agentic search where parallel trajectories converge due to similar initial queries, and proposes DivInit, a training-free method that samples diverse initial queries to improve multi-hop question answering performance.
This paper reformulates language generation as a stochastic optimal control problem, addressing limitations of autoregressive and diffusion models, and proposes a closed-loop diffusion method in latent control space using Flow Matching, achieving high-fidelity generation and efficient parallel sampling.