---
title: Mamba: Linear-Time Sequence Modeling with Selective State Spaces
url: https://www.ml-quant.com/papers/arxiv/2312.00752/
site: ML-Quant (https://www.ml-quant.com)
updated: 2026-09-26
license: Summaries CC BY 4.0; links go to the original sources
index: https://www.ml-quant.com/llms.txt
identifier: arXiv:2312.00752
source_url: https://arxiv.org/pdf/2312.00752.pdf
featured: 2024-06-05
citations: 9205
topic: ML & AI Methods
---


# Mamba: Linear-Time Sequence Modeling with Selective State Spaces

Sequence Modeling: Mamba, a neural network architecture that doesn't use attention or MLP blocks, provides faster inference and better performance in language, audio, and genomics than Transformers.

- Source: https://arxiv.org/pdf/2312.00752.pdf
- Identifier: arXiv:2312.00752
- Released: 2023-12-01
- First featured: Quant Letter No. 52 (2024-06-05): https://www.ml-quant.com/issues/2024-06-05/
- Citations (Semantic Scholar): 9205
- Published in: not yet
- Topic: ML & AI Methods

## Related

- [Transformers are SSMs: Generalized Models and Efficient Algorithms Through Structured State Space Duality](https://www.ml-quant.com/papers/arxiv/2405.21060/): The research identifies a link between state-space models and Transformers in deep learning, leading to the creation of a faster language modeling architecture, Mamba-2.
- [Graph Mamba: Towards Learning on Graphs with State Space Models](https://www.ml-quant.com/papers/arxiv/2402.08678/): Graph Mamba Networks, a new type of Graph Neural Networks, have been introduced, which achieve excellent performance in various benchmark datasets despite lower computational cost.
- [Jamba-1.5: Hybrid Transformer-Mamba Models at Scale](https://www.ml-quant.com/papers/arxiv/2408.12570/): Transformer-Mamba Models: Jamba-1.5 is a new large language model with enhanced conversational and instruction-following capabilities, featuring a unique quantization technique for cost-effective inference.
- [Transformer Circuit Faithfulness Metrics are not Robust](https://www.ml-quant.com/papers/arxiv/2407.08734/): The authors explore the difficulties in evaluating the performance of neural network 'circuits', emphasizing the sensitivity of current methods to changes in the ablation methodology and the need for clearer claims about circuits.
- [Transaction Fraud Detection via Spatial-Temporal-Aware Graph Transformer](https://www.ml-quant.com/papers/arxiv/2307.05121/): A new graph neural network, STA-GT, is proposed for transaction fraud detection, effectively learning and incorporating spatial-temporal and global information.
- [MLBased CT Saturation Detection](https://www.ml-quant.com/papers/ssrn/4979169/): The study introduces a machine learning method for detecting saturation in Current Transformers using artificial neural networks and long short-term memory networks, enhancing anomaly detection in complex datasets.