Summary of A Diagonal Structured State Space Model on Loihi 2 For Efficient Streaming Sequence Processing, by Svea Marie Meyer et al.
A Diagonal Structured State Space Model on Loihi 2 for Efficient Streaming Sequence Processing
by Svea Marie Meyer, Philipp Weidel, Philipp Plank, Leobardo Campos-Macias, Sumit Bam Shrestha, Philipp Stratmann, Mathis Richter
First submitted to arxiv on: 23 Sep 2024
Categories
- Main: Machine Learning (cs.LG)
- Secondary: Artificial Intelligence (cs.AI); Emerging Technologies (cs.ET); Neural and Evolutionary Computing (cs.NE)
GrooveSquid.com Paper Summaries
GrooveSquid.com’s goal is to make artificial intelligence research accessible by summarizing AI papers in simpler terms. Each summary below covers the same AI paper, written at different levels of difficulty. The medium difficulty and low difficulty versions are original summaries written by GrooveSquid.com, while the high difficulty version is the paper’s original abstract. Feel free to learn from the version that suits you best!
| Summary difficulty | Written by | Summary |
|---|---|---|
| High | Paper authors | High Difficulty Summary Read the original abstract here |
| Medium | GrooveSquid.com (original content) | Medium Difficulty Summary The Deep State-Space Models (SSM) have achieved state-of-the-art performance in long-range sequence modeling tasks, but their recurrent structure limits efficient processing on Graphics Processing Units (GPUs). To overcome this limitation, researchers developed an efficient token-by-token inference of the SSM S4D on Intel’s Loihi 2 neuromorphic processor. This paper compares the performance of S4D on Loihi 2 with recurrent and convolutional implementations on Jetson Orin Nano in various benchmarks, including sMNIST, psMNIST, and sCIFAR. The results show that while Jetson performed better in offline processing mode, Loihi 2 outperformed during token-by-token processing, consuming significantly less energy, latency, and time while maintaining a higher throughput. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary The paper uses special models to understand long sequences of information. These models are good at this task, but they can’t be easily used on computers because they take too much memory and time. To solve this problem, the researchers created a way to use these models on special processors that are designed for certain types of computing. They compared how well this worked with other ways of using the model and found that it was much faster and more efficient. |
Keywords
» Artificial intelligence » Inference » Token
