AI & ML
SADL: On the Demand-Supply Geometry of Representation Collapse
arXiv cs.RO0 views
arXiv:2611.21596v1 Announce Type: new Abstract: In Joint-Embedding Predictive Architectures (JEPAs), a well-constrained projector output does not by itself prevent representation collapse. We find that, from initialization, the encoder representation is dominated by a component shared by different inputs. A batch-normalized projector removes this component, so regularization after the projector cannot correct it, and the differences between representations shrink until they collapse. We propose SADL (Supply And Demand Loss), which keeps the original training objective and constrains the encoder directly when needed. SADL compares the representation differences that the projector output demands with the differences the encoder can retain and pass on to the projector, and adds a constraint on the encoder only when supply falls short of demand. It works with different Demand Regularizers and introduces no auxiliary networks, additional trainable parameters, or architectural changes. In single-task and jointly trained four-task LeWorldModel (LeWM) and INTACT experiments, SADL lowers the collapse rate over ten random seeds from 50% to 0%. After one epoch of training, encoder effective rank is 3.4-6.4 times that of the corresponding baseline without SADL, and latent prediction error falls to 1/20-1/303 of it. Without additional tuning, SADL stays robust as its loss weight spans orders of magnitude, and with several Demand Regularizers INTACT exceeds 95% mean search-free control success. On four image datasets, SADL also consistently improves accuracy and stability across random seeds. These results show that matching the representation differences a regularizer demands with the differences the encoder can supply leads to faster and more stable representation learning in control and image classification.
Source:https://arxiv.org/abs/2611.21596PreviousLazy-MOSAIC: Accelerating Skill-Centric Manipulation Planning with Learned Skill-Effect World ModelsNextScaleVLA: Scaling Vision-Language-Action Models to 552 Billion Parameters
Related posts
Pebblebee Alert Dispatch lets an actual human respond to your emergency, triggered by Find Hub trackers
09to5Google

Will Wright’s next game uses AI to map your memories
0The Verge
Microsoft Built Its Own Coding Model, and the Price Is the Point
0dev.to

El diario económico coreano Maeil Business cuenta 12 primeros puestos de Darwin en Hugging Face: cuéntalos tú con Python
0dev.to
Can Peer Review Survive AI-Generated Text? [pdf]
0Hacker News
Toward Planning Representations That Separate Task Structure from Visual Content
0arXiv cs.RO