MDLMs
PulseAugur coverage of MDLMs — every cluster mentioning MDLMs across labs, papers, and developer communities, ranked by signal.
2 day(s) with sentiment data
-
Masked Diffusion Language Models outperform AR models for agentic RL
A new research paper introduces Masked Diffusion Language Models (MDLMs) as a superior alternative to autoregressive (AR) models for text-based world modeling in agentic reinforcement learning. MDLMs demonstrate enhance…
-
New Mask-Aware Policy Gradients boost DLM reasoning on benchmarks · 2 sources tracked
Researchers have developed a novel approach called Mask-Aware Policy Gradients to enhance reasoning capabilities in Diffusion Language Models (DLMs). This method addresses the challenge of applying reinforcement learnin…
-
New TIE framework enhances Masked Diffusion Language Model ensembling
Researchers have introduced Trajectory-based Iterative Ensembling (TIE), a new framework for combining the knowledge of Masked Diffusion Language Models (MDLMs). TIE focuses on the unique decoding dynamics of MDLMs, obs…
-
New methods enhance MDLMs with improved padding and knowledge ensembling · 6 sources tracked
Researchers have introduced two novel approaches for Masked Diffusion Language Models (MDLMs). The first, VoidPadding, decouples the roles of end-of-sequence ([EOS]) tokens for semantic termination and padding, using a …
-
Diffusion models for graph-to-text generation prioritize entities
Researchers have analyzed the generation process of masked diffusion language models (MDLMs) for graph-to-text generation, finding they prioritize entities before relational words and structural tokens. A new method, la…