NLP Lineage

1 December 2023 · Language modelling

Mamba

Mamba: Linear-Time Sequence Modeling with Selective State Spaces

A. Gu, T. Dao

Date
1 December 2023
What the date marks
arXiv submission
Kind
Architectural idea
Organisation
CMU / Princeton
Access
Apache 2.0
Other dates
COLM 2024
Threads
In the guided tour
Chapter 9. Making it run anywhere
Learning phases
1. Transformer architecture; 8. Training from scratch

Notable: no later milestone in the data is documented as descending from it. How levels are decided

Read the primary source (arxiv.org)

Full summary in progress. Until then, this page shows the milestone's record from the research data.

What it changed

Before it

Attention's quadratic cost was treated as the price of good sequence modelling.

The problem

State-space models scale linearly but could not match transformers on language.

What it introduced

Make the state-space parameters input-dependent, so the model can selectively remember or forget, with a hardware-aware scan.

Why it matters

The strongest challenger to attention so far. Frontier models remain transformers, so treat this as a live parallel branch.

What modern models inherit

An open question about whether attention is the final answer.

What came next

Mamba-2 and hybrid attention-SSM architectures.

Before and after

The milestones it descends from, and the milestones that descend from it. Work developed alongside it is not counted. Weak links are left out unless you include them.

Before: where it came from

Mamba has 13 documented ancestors across 6 chapters, reaching back to Recurrent networks in 1990. Showing its direct parents and the 5 turning points among them.

Mamba has 13 documented ancestors across 6 chapters, reaching back to Recurrent networks in 1990. Showing all of them.

Mamba has 15 recorded ancestors, including indirect ones, across 6 chapters, reaching back to Recurrent networks in 1990. Showing its direct parents and the 5 turning points among them.

Mamba has 15 recorded ancestors, including indirect ones, across 6 chapters, reaching back to Recurrent networks in 1990. Showing all of them.

  1. 2023MambaThis milestone.
  2. 2022FlashAttentionMamba was influenced by thisChapter 9. Making it run anywhere
  3. 2020BigBirdFlashAttention succeeded thisChapter 10. Reading more than you remember
  4. 2019Sparse Transformer Turning pointBigBird was influenced by thisChapter 4. Pretrain once, use everywhere
  5. 2017Transformer Turning pointFlashAttention built directly on thisChapter 3. Attention, then only attention
  6. 2016Layer normTransformer built directly on thisChapter 3. Attention, then only attention
  7. 2015Residual connectionsTransformer was influenced by thisChapter 3. Attention, then only attention
  8. 2015Luong attentionTransformer was influenced by thisChapter 3. Attention, then only attention
  9. 2014AdamTransformer was influenced by thisChapter 2. Words become vectors
  10. 2014Seq2seq Turning pointTransformer built directly on thisChapter 2. Words become vectors
  11. 2014Attention Turning pointTransformer built directly on thisChapter 3. Attention, then only attention
  12. 2014GRU / encoder-decoderAttention built directly on thisChapter 2. Words become vectors
  13. 1997LSTM Turning pointSeq2seq built directly on thisChapter 1. Language as probability
  14. 1990Recurrent networksLSTM was influenced by thisChapter 1. Language as probability
  1. 2023MambaThis milestone.
  2. 2022FlashAttentionMamba was influenced by thisChapter 9. Making it run anywhere
  3. 2020BigBirdFlashAttention succeeded thisChapter 10. Reading more than you remember
  4. 2019Sparse Transformer Turning pointBigBird was influenced by thisChapter 4. Pretrain once, use everywhere
  5. 2019Transformer-XLSparse Transformer was influenced by this (indirect influence)Chapter 4. Pretrain once, use everywhere
  6. 2017Transformer Turning pointFlashAttention built directly on thisChapter 3. Attention, then only attention
  7. 2017ConvS2STransformer was influenced by this (indirect influence)Chapter 3. Attention, then only attention
  8. 2016Layer normTransformer built directly on thisChapter 3. Attention, then only attention
  9. 2015Residual connectionsTransformer was influenced by thisChapter 3. Attention, then only attention
  10. 2015Luong attentionTransformer was influenced by thisChapter 3. Attention, then only attention
  11. 2014AdamTransformer was influenced by thisChapter 2. Words become vectors
  12. 2014Seq2seq Turning pointTransformer built directly on thisChapter 2. Words become vectors
  13. 2014Attention Turning pointTransformer built directly on thisChapter 3. Attention, then only attention
  14. 2014GRU / encoder-decoderAttention built directly on thisChapter 2. Words become vectors
  15. 1997LSTM Turning pointSeq2seq built directly on thisChapter 1. Language as probability
  16. 1990Recurrent networksLSTM was influenced by thisChapter 1. Language as probability

After: what it led to

No documented descendants are recorded for Mamba without weak links.

No documented descendants are recorded for Mamba without weak links.

No recorded descendants are recorded for Mamba.

No recorded descendants are recorded for Mamba.

Direct links and their sources

Where it came from

Influenced by

Developed alongside

What it led to

None recorded in the sources used.