Anisotropy Is Inherent to Self-Attention in Transformers
원문 초록
The representation degeneration problem is a phenomenon that is widely observed among self-supervised learning methods based on Transformers. In NLP, it takes the form of anisotropy, a singular property of hidden representations which makes them unexpectedly close to each other in terms of angular distance (cosine-similarity). Some recent works tend to show that anisotropy is a consequence of optimizing the cross-entropy loss on long-tailed distributions of tokens. We show in this paper that anisotropy can also be observed empirically in language models with specific objectives that should not suffer directly from the same consequences. We also show that the anisotropy problem extends to Transformers trained on other modalities. Our observations suggest that anisotropy is actually inherent to Transformers-based models.
함께 살펴볼 논문
공통 태그를 근거로 연결했습니다. 실제 인용 관계나 논문의 우수성을 뜻하지 않습니다.
RetFiner: A Vision-Language Refinement Scheme for Retinal Foundation Models
Boosting Generative Adversarial Transferability with Self-supervised Vision Transformer Features
Post-training for Deepfake Speech Detection
Topology-Aware Modeling for Unsupervised Simulation-to-Reality Point Cloud Recognition
Efficient and Generalizable Speaker Diarization via Structured Pruning of Self-Supervised Models
Pushing the Performance of Synthetic Speech Detection with Kolmogorov-Arnold Networks and Self-Supervised Learning Models
토론
이 논문과 연결된 포럼 글이 아직 없습니다.