Learning Through Noise: Why Subliminal Learning Works and When It Fails
Max Planck Institute for Dynamics and Self-Organization(马克斯·普朗克动态与自组织研究所);Faculty of Physics, Institute for the Dynamics of Complex Systems, University of Göttingen(哥廷根大学物理系,复杂系统动力学研究所)
Vincent C. Brockers, Roman D. Ventzke, Valentin Neuhaus, Belén Hidalgo-Ogalde, Viola Priesemann
In the context of artificial neural networks, subliminal learning refers to the transfer of task-relevant knowledge or unintended biases from teacher to student models through distillation on task-unrelated input$\unicode{x2013}$output pairs. Prior explanations tie this effect to shared or closely matched teacher$\unicode{x2013}$student initialization. We show that a closely matched initialization is not necessary. Instead, subliminal learning is governed by compatible output heads. Using a controlled MNIST setting, we split outputs into an auxiliary head (for auxiliary, task-unrelated noise signals) and a class head (for classification) to demonstrate subliminal learning occurs$\unicode{x2014}$even when we randomly initialize hidden layers and remove layers, add new layers, or change the architecture (MLP-to-CNN). Compatible auxiliary heads enable transfer of a recoverable teacher signal, bringing the student's representations closer to the teacher's. When the class heads remain compatible as well, students trained only on task-unrelated noise can approach, and in favorable regimes match, teacher-level task performance. Our setting enables us to develop a theory that explains the mechanism of subliminal learning and to derive upper bounds on when subliminal learning fails. Together, our results turn subliminal learning from a surprising transfer effect into a theoretically grounded mechanism with predictable limits.
Why Low-Precision Transformer Training Fails: An Analysis on Flash Attention
Department of Electronic Engineering, Tsinghua University(清华大学电子工程系);Beijing National Research Center for Information Science and Technology(北京信息科学与技术国家研究中心);State Key laboratory of Space Network and Communications(空间网络与通信国家重点实验室)
Explainable Student Performance Prediction With Personalized Attention for Explaining Why A Student Fails
School of Computer Science (National Pilot Software Engineering School)(计算机科学学院(国家示范性软件工程学院));Beijing University of Posts and Telecommunications(北京邮电大学)
Curse of Attention: A Kernel-Based Perspective for Why Transformers Fail to Generalize on Time Series Forecasting and Beyond
The University of Hong Kong(香港大学);University of Wisconsin-Madison(威斯康星大学麦迪逊分校);The Simons Institute for the Theory of Computing(西蒙斯理论计算研究所);UC Berkeley(加州大学伯克利分校);Sun Yat-sen University(中山大学)
Catastrophic Compositional Generation: Why Vanilla Diffusion Models Fail to Extrapolate
Machine Learning Department, Carnegie Mellon University(卡内基梅隆大学机器学习系);Valence Labs;Department of Computer Science, University of Manchester(曼彻斯特大学计算机科学系)
Why Rectified Power Unit Networks Fail and How to Improve It: An Effective Field Theory Perspective
School of Computational Sciences(计算科学学院);Korea Institute for Advanced Study(韩国高级研究院);Department of Mathematical Science(数学科学系);Seoul National University(首尔国立大学)
Why Reasoning Fails to Plan: A Planning-Centric Analysis of Long-Horizon Decision Making in LLM Agents
University of Notre Dame(诺丁汉大学);Stanford University(斯坦福大学);Purdue University(普渡大学);University of Edinburgh(爱丁堡大学);University of Oxford(牛津大学);Yale University(耶鲁大学)
Why Pretraining Fails to Share Cross-Lingual Knowledge
Weizmann Institute of Science(魏茨曼科学研究所);Bar-Ilan University(巴伊兰大学);Johns Hopkins University(约翰斯·霍普金斯大学);A*STAR(新加坡科技研究局);University of Washington(华盛顿大学);MIT(麻省理工学院);MIT-IBM Watson AI Lab(MIT-IBM沃森人工智能实验室)