Tracking the Mamba explosion, one arXiv link at a time

What it does

This repository is a curated, chronologically organized index of papers on State Space Models—specifically the Mamba architecture and its offshoots—spanning computer vision, medical imaging, robotics, and language. It collects arXiv preprints, peer-reviewed conference papers, and associated code links in one long, scrollable list. The maintainers also host their own survey paper, presentation slides, and a video tutorial from COLM 2024, effectively turning the repo into a mini reading room for the field.

The interesting bit

The list reveals just how aggressively the research community is retrofitting Mamba into every imaginable pipeline: 3D segmentation, event-based tracking, medical report generation, and even diffusion models. Rather than limiting itself to core sequence modeling, the bibliography tracks cross-domain adoption, making the breadth of the SSM land grab visible in one place.

Key highlights

• Organizes papers by year and application area, from vision and video to medical imaging and 3D occupancy.
• Bundles the authors’ own survey, downloadable slides, and a COLM 2024 oral video tutorial.
• Links to source code alongside papers when available, saving a separate GitHub search.
• Covers niche crossovers like Mamba-meets-GNN and hybrid Transformer-Mamba architectures.

Caveats

• The README is a raw list with minimal commentary or synthesis; expect to do your own triage.
• Inclusion criteria are unclear, and the list mixes peer-reviewed work with fresh arXiv preprints without consistent labeling.
• Navigation is entirely manual scrolling; there are no filters, tags, or search beyond GitHub’s native tools.

Verdict

Researchers and engineers trying to stay current with the Mamba/SSM wave will find this a useful starting point. If you need runnable frameworks or ranked benchmarks, look elsewhere—this is strictly a reading list.