Retrieval & memory
Titans: Learning to Memorize at Test Time
Introduces a neural long-term memory component that learns during inference, alongside attention-based context processing.
Discovered 2026-10-21 · Sources checked 2026-10-21
This summary was generated with AI assistance.
Why it matters
Provides another approach to representing information from long sequences.
What changed
Investigates trainable memory beyond a fixed recurrent hidden state or bounded attention window.
Where it helps
- Long-sequence modeling
- Learned memory architectures
Evidence & reproduction
Reproduction not evaluated. A linked paper or repository does not establish independent reproduction.
Limitations & open questions
Pub-AI has not independently evaluated this method. Consult the linked sources for their evaluation setup and limitations.
Sources
- Titans: Learning to Memorize at Test Time ↗
arXiv · primary · Published 2024-12-31 · Checked 2026-10-21