CLARU
Litian2002MIT

Robotics Papers Vector Database

Semantic search database over 63,381 academic papers from robotics, computer vision, and machine learning conferences (2023-2026) with BAAI/bge-m3 embeddings for neural information retrieval.

Downloads66

Technical Profile

Modalities
textembeddings
Data Format
LanceDB
License
MIT
Part of the Robotics Papers Vector Database family

Access

View on HuggingFace

Need custom text data?

Claru builds purpose-built datasets for any environment applications with dense human annotations and quality assurance.

Request a Sample Pack

Related Datasets

efficient-vla-extracts

Parser-derived artifacts from robotics and vision-language-action papers collected from arXiv, containing text extractions, structured parses, and figures for the efficient-vla-wiki project.

textimages
4K downloadsMay 2026See individual source papers

BEHAVIOR-1K 2025 Challenge Demos V-JEPA 2 ViT-G 130h Embeddings

Precomputed video feature representations for a 130-hour subsample of BEHAVIOR-1K 2025 challenge demonstrations extracted using V-JEPA 2 ViT-G encoder. This dataset enables faster downstream experimentation by avoiding repeated video decoding and encoder forward passes.

rgbembeddings
1K downloadsMay 2026MIT

Ego-Exo Manufacturing

A video dataset of real shoe manufacturing operations with temporally synchronized ego-centric (first-person) and exo-centric (overhead) video pairs, along with standalone sessions annotated with atomic action descriptions and expert commentary.

videotext
229 downloadsMar 2026apache-2.0

Scentience Agent Skills

Structured agent skill definitions for olfaction-aware robotics and AI, grounded in the Scentience OPU hardware and SDK ecosystem. Covers BLE device connectivity, real-time plume navigation, olfactory inertial odometry, and COLIP multimodal embeddings.

chemical_sensingolfactionimuble+1
90 downloadsMar 2026apache-2.0

BRIGHT+

An upgraded text retrieval benchmark designed to support reasoning-intensive retrieval across diverse domains including robotics, biology, economics, and more. It applies MARCUS, a multi-agent LLM-based pipeline, to clean and semantically chunk the original BRIGHT dataset, significantly improving signal-to-noise ratio and retrieval effectiveness.

text
70 downloadsJul 2025cc-by-4.0