On Controllable Sparse Alternatives to Softmax

Anirban Laha, Saneem Ahmed Chemmengath, Priyanka Agrawal, Mitesh Khapra, Karthik Sankaranarayanan, Harish G. Ramaswamy

Nov-17-2025, 21:38:19 GMT–Neural Information Processing Systems

Even though softmax is the most prevalent approach amongst them, it has a shortcoming in that its outputs are composed of only non-zeroes and is therefore ill-suited for producing sparse probability distributions as output.

artificial intelligence, machine learning, natural language, (18 more...)

Neural Information Processing Systems

Nov-17-2025, 21:38:19 GMT

Conferences PDF

Country:
- Europe > Portugal
  - Lisbon > Lisbon (0.04)
- North America
  - Canada > Quebec
    - Montreal (0.04)
  - United States
    - California > San Diego County
      - San Diego (0.04)
    - Massachusetts
      - Middlesex County > Cambridge (0.04)
      - Plymouth County > Hanover (0.04)

Technology:
- Information Technology > Artificial Intelligence
  - Machine Learning
    - Neural Networks (0.47)
    - Statistical Learning (0.77)
  - Natural Language (1.00)

Duplicate Docs Excel Report

Title
On Controllable Sparse Alternatives to Softmax
On Controllable Sparse Alternatives to Softmax

Similar Docs Excel Report more

Title	Similarity	Source
None found