Taming the Sigmoid Bottleneck: Provably Argmaxable Sparse Multi-Label Classification

Documenti analoghi

• Fast and Expressive Multi-Token Prediction with Probabilistic Circuits
  di: Grivas, Andreas, et al.
  Pubblicazione: (2025)
• Contrastive Learning for Multi Label ECG Classification with Jaccard Score Based Sigmoid Loss
  di: Takahashi, Junichiro, et al.
  Pubblicazione: (2026)
• Activation Bottleneck: Sigmoidal Neural Networks Cannot Forecast a Straight Line
  di: Toller, Maximilian, et al.
  Pubblicazione: (2024)
• Taming Polysemanticity in LLMs: Provable Feature Recovery via Sparse Autoencoders
  di: Chen, Siyu, et al.
  Pubblicazione: (2025)
• Semi-Supervised Sparse Gaussian Classification: Provable Benefits of Unlabeled Data
  di: Azar, Eyar, et al.
  Pubblicazione: (2024)