Tsinghua Huang Gao team wins ICML 2026 Outstanding Paper Award, Test of Time Award given to classic algorithm A3C

On July 6, the top international machine learning conference ICML 2026 was held in Seoul, South Korea, and announced the annual award-winning papers. The paper (The Flexibility Trap: Rethinking the Value of Arbitrary Order in Diffusion Language Models) by Tsinghua University’s Huang Gao team in collaboration with Alibaba won the Outstanding Paper Award. The research reveals that the flexibility of arbitrary generation order in diffusion language models actually limits model potential in general reasoning tasks such as mathematics and programming, while abandoning arbitrary order and adopting traditional left-to-right generation not only simplifies the method but also significantly improves reasoning accuracy.

Another Outstanding Paper Award was received by MIT and Yale University, proposing a high-accuracy sampling algorithm for diffusion models (High-Accuracy Sampling for Diffusion Models and Log-Concave Distributions), achieving exponential optimization in the number of steps (or sampling complexity) required to reach target sampling accuracy.

This year’s grand prize also included a position paper (Position: The Alignment Community is Unintentionally Building a Censor’s Toolkit), jointly authored by researchers from Ludwig Maximilian University of Munich and independent researchers, pointing out that current AI alignment techniques carry dual-use risks and can easily be maliciously manipulated into information censorship tools.

The Test of Time Award went to Google DeepMind’s classic reinforcement learning algorithm published in 2016 (Asynchronous Methods for Deep Reinforcement Learning). The Asynchronous Advantage Actor-Critic (A3C) architecture proposed in this research significantly improved the training efficiency of deep reinforcement learning, ushering in an era of efficiently training agents using ordinary multi-core CPUs.