> ## Documentation Index
> Fetch the complete documentation index at: https://aegean.ai/llms.txt
> Use this file to discover all available pages before exploring further.

# What you are given

> The tools, starter and seed ontology you start the Physical AI kDB project with, and the 31 papers of the library.

1. The library: 30 papers on VLA models, plus one more paper marked "ingest last". The [library](#the-library) section lists them. Save each PDF as `kdb/papers/<arxiv-id>.pdf`, for example `kdb/papers/2306.03310.pdf`, because the seed ontology and our grading refer to the papers by those names.
2. The project repository, [pantelis/physical-ai-kdb](https://github.com/pantelis/physical-ai-kdb). It installs four of the command-line tools that we also run at grading time.
   * `okf-validate` checks that your kDB is a valid OKF v0.2 bundle and follows the [rules of the kDB format](https://aegean.ai/aiml-common/projects/nlp/physical-ai-kdb/kdb-format.md#the-rules). With your lake it also checks the rows, and it reports every page that changed outside a logged action.
   * `okf-ground` checks the evidence of every Term page and every row. A quoted sentence must be on its cited page and inside its `bbox`, and a table value must be inside its `bbox`. Every `bbox` must match a highlight in the annotated copy of the PDF. Every file that a row cites must be registered in your lake, and its stored bytes must still match its sha256.
   * `okf-diff` compares two versions of your kDB pages, as two directories or two git commits. With `--snapshots` it compares two snapshots of your lake instead.
   * `okf-seed` writes the seed ontology into your kDB (item 5).
3. A starter, in `starter/` of the same repository. It is the `kdb` command, with `lake init`, `ingest`, `ask`, `export` and `history`, and two Pydantic AI agents. Through kdb-tools it also installs three more commands. `okf-export` writes your lake into a folder with a manifest of hashes, and `okf-restore` checks every hash before it restores one. `okf-mcp` serves the checks and actions over MCP. [Agent design](https://aegean.ai/aiml-common/projects/nlp/physical-ai-kdb/system.md#agent-design) says what it does and what is left to you.
4. The same checks as a Python package, `kdb_tools`, so your agent can call them as functions instead of parsing command output. `validate()` and `ground()` return a report with one finding per problem. `kdb_tools.locate.find_quote(page, text)` returns the `bbox` of every place a quote appears on a PDF page. It uses the same matching as `okf-ground`, so your agent and the checker agree on where a quote is. The package also holds the actions that write your lake, in `kdb_tools.actions`.
5. A seed ontology in three parts. `robotics` covers tasks (such as pick-and-place, pushing and articulated object manipulation), benchmarks and their suites (LIBERO and its suites, LIBERO-Plus, CALVIN, SIMPLER), robots (such as the Franka Emika Panda and the WidowX 250), the seven LIBERO-Plus perturbation types and the vision-language-action model class. `datasets` covers how data is collected and what a model uses it for, for any kind of dataset. `vla-datasets` specializes `datasets` for robot learning data. Teleoperation is a kind of human-collected data, cross-embodiment pretraining is a kind of pretraining, and so on. `okf-seed kdb` writes one `Term` page per term under `ontology/<ontology>/` and one page per ontology that says what it extends. Each term is defined by a sentence it quotes from a library paper and highlights in your copy of that paper. The [ontology](https://aegean.ai/aiml-common/projects/nlp/physical-ai-kdb/ontology.md) section explains how to use it.
6. The VLA classes. `kdb lake init` writes them into your kDB as `Class` and `Property` pages, with the core ontology they build on, and binds each class to a table in your lake. [The kDB format](https://aegean.ai/aiml-common/projects/nlp/physical-ai-kdb/kdb-format.md) lists them.

`okf-seed kdb` quotes ten of the papers, so it needs these files in `kdb/papers/` before it runs:

| arXiv id | Paper |
| - | - |
| 2112.03227 | CALVIN |
| 2306.03310 | LIBERO |
| 2307.15818 | RT-2 |
| 2308.12952 | BridgeData V2 |
| 2310.08864 | Open X-Embodiment |
| 2405.05941 | SIMPLER |
| 2405.12213 | Octo |
| 2410.24164 | π0 |
| 2502.19645 | OpenVLA-OFT |
| 2510.13626 | LIBERO-Plus |

LIBERO-Plus is the paper you ingest last. The seed quotes it for the perturbation types, so its PDF is in `kdb/papers/` from the start, but no row may cite it before Submission 2.

## The library

The library has 30 papers on vision-language-action (VLA) models, plus one paper to ingest last. Each paper links to its Semantic Scholar page and to one version of its PDF on arXiv. Download that version and no other. Several of these papers are still being revised, and a later version moves sentences and table values to other pages. The checks compare your sources with our copies of these versions. A correct source taken from another version therefore fails `okf-ground`.

The papers report on the same benchmarks (LIBERO, SIMPLER, CALVIN), compare the same models (RT-2, Octo, OpenVLA, π0) and train on the same datasets (Open X-Embodiment, BridgeData V2, DROID). Because of that overlap, the kDB can answer questions that span several papers.

| # | Paper | Title | Year | arXiv PDF |
| - | - | - | - | - |
| 1 | [CALVIN](https://www.semanticscholar.org/paper/4be02694125b71876552900a53c85c47a2a83614) | CALVIN: A Benchmark for Language-Conditioned Policy Learning for Long-Horizon Robot Manipulation Tasks | 2021 | [2112.03227v4](https://arxiv.org/pdf/2112.03227v4) |
| 2 | [RT-1](https://www.semanticscholar.org/paper/fd1cf28a2b8caf2fe29af5e7fa9191cecfedf84d) | RT-1: Robotics Transformer for Real-World Control at Scale | 2022 | [2212.06817v2](https://arxiv.org/pdf/2212.06817v2) |
| 3 | [ACT (ALOHA)](https://www.semanticscholar.org/paper/91eb20f923ea3b0246868902aef4e9bea572b800) | Learning Fine-Grained Bimanual Manipulation with Low-Cost Hardware | 2023 | [2304.13705v1](https://arxiv.org/pdf/2304.13705v1) |
| 4 | [BridgeData V2](https://www.semanticscholar.org/paper/6a375be82efc01ec4ed73334655935a56ba82d38) | BridgeData V2: A Dataset for Robot Learning at Scale | 2023 | [2308.12952v3](https://arxiv.org/pdf/2308.12952v3) |
| 5 | [Diffusion Policy](https://www.semanticscholar.org/paper/bdba3bd30a49ea4c5b20b43dbd8f0eb59e9d80e2) | Diffusion Policy: Visuomotor Policy Learning via Action Diffusion | 2023 | [2303.04137v5](https://arxiv.org/pdf/2303.04137v5) |
| 6 | [LIBERO](https://www.semanticscholar.org/paper/98cfd7b1b29453c4e82536f5afdc6ddc58bbb1b3) | LIBERO: Benchmarking Knowledge Transfer for Lifelong Robot Learning | 2023 | [2306.03310v2](https://arxiv.org/pdf/2306.03310v2) |
| 7 | [Open X-Embodiment](https://www.semanticscholar.org/paper/ef7d31137ef06c5be8c2824ecc5af6ce3358cc8f) | Open X-Embodiment: Robotic Learning Datasets and RT-X Models | 2023 | [2310.08864v9](https://arxiv.org/pdf/2310.08864v9) |
| 8 | [RT-2](https://www.semanticscholar.org/paper/38939304bb760473141c2aca0305e44fbe04e6e8) | RT-2: Vision-Language-Action Models Transfer Web Knowledge to Robotic Control | 2023 | [2307.15818v1](https://arxiv.org/pdf/2307.15818v1) |
| 9 | [CogACT](https://www.semanticscholar.org/paper/54caf5927a6d656bb84d2a16e16fa2b0c5d6f522) | CogACT: A Foundational Vision-Language-Action Model for Synergizing Cognition and Action in Robotic Manipulation | 2024 | [2411.19650v1](https://arxiv.org/pdf/2411.19650v1) |
| 10 | [DROID](https://www.semanticscholar.org/paper/ee5070fe52fd17da9a89d3f342fb07cc9ae51afe) | DROID: A Large-Scale In-The-Wild Robot Manipulation Dataset | 2024 | [2403.12945v2](https://arxiv.org/pdf/2403.12945v2) |
| 11 | [Octo](https://www.semanticscholar.org/paper/1d2753d74025e7a71594506623be81f18b073adb) | Octo: An Open-Source Generalist Robot Policy | 2024 | [2405.12213v2](https://arxiv.org/pdf/2405.12213v2) |
| 12 | [OpenVLA](https://www.semanticscholar.org/paper/8f9ceb5ffad8e7a066dfc9d9aaa5153b714740ee) | OpenVLA: An Open-Source Vision-Language-Action Model | 2024 | [2406.09246v3](https://arxiv.org/pdf/2406.09246v3) |
| 13 | [RDT-1B](https://www.semanticscholar.org/paper/41ba8127b73abd32051a89764cb976095b1dee94) | RDT-1B: a Diffusion Foundation Model for Bimanual Manipulation | 2024 | [2410.07864v2](https://arxiv.org/pdf/2410.07864v2) |
| 14 | [RoboVLMs](https://www.semanticscholar.org/paper/7d638461965faa79cfd28d5f0a6fbb12fca41c9f) | What Matters in Building Vision-Language-Action Models for Generalist Robots | 2024 | [2412.14058v4](https://arxiv.org/pdf/2412.14058v4) |
| 15 | [SIMPLER](https://www.semanticscholar.org/paper/202cc3fcea588cee37ffabcb0b5e145567109a66) | Evaluating Real-World Robot Manipulation Policies in Simulation | 2024 | [2405.05941v1](https://arxiv.org/pdf/2405.05941v1) |
| 16 | [TinyVLA](https://www.semanticscholar.org/paper/dc62bc6536e9e3ad80242f10f44c046e4c7bd3d1) | TinyVLA: Toward Fast, Data-Efficient Vision-Language-Action Models for Robotic Manipulation | 2024 | [2409.12514v5](https://arxiv.org/pdf/2409.12514v5) |
| 17 | [TraceVLA](https://www.semanticscholar.org/paper/57431570f624069de2fa99429c7c3d3760271bc2) | TraceVLA: Visual Trace Prompting Enhances Spatial-Temporal Awareness for Generalist Robotic Policies | 2024 | [2412.10345v3](https://arxiv.org/pdf/2412.10345v3) |
| 18 | [π0](https://www.semanticscholar.org/paper/7e7e59d2e247d99954081080ddd5aae93d10b9e0) | π0: A Vision-Language-Action Flow Model for General Robot Control | 2024 | [2410.24164v4](https://arxiv.org/pdf/2410.24164v4) |
| 19 | [CoT-VLA](https://www.semanticscholar.org/paper/2d7f3a99e916fc80ff890d109699f9682253e66d) | CoT-VLA: Visual Chain-of-Thought Reasoning for Vision-Language-Action Models | 2025 | [2503.22020v1](https://arxiv.org/pdf/2503.22020v1) |
| 20 | [Dita](https://www.semanticscholar.org/paper/7aa2936c237c86f33ad413eeeabca6b123515c57) | Dita: Scaling Diffusion Transformer for Generalist Vision-Language-Action Policy | 2025 | [2503.19757v2](https://arxiv.org/pdf/2503.19757v2) |
| 21 | [FAST (π0-FAST)](https://www.semanticscholar.org/paper/3880f5bad862ba1b18f4f8ec060038b326b118ed) | FAST: Efficient Action Tokenization for Vision-Language-Action Models | 2025 | [2501.09747v1](https://arxiv.org/pdf/2501.09747v1) |
| 22 | [GR00T N1](https://www.semanticscholar.org/paper/731c50b0d6af4c1cb8d95f506541681ea487973b) | GR00T N1: An Open Foundation Model for Generalist Humanoid Robots | 2025 | [2503.14734v2](https://arxiv.org/pdf/2503.14734v2) |
| 23 | [NORA](https://www.semanticscholar.org/paper/0801ed93c26dcb72e2bddf756f971e10357942a2) | NORA: A Small Open-Sourced Generalist Vision Language Action Model for Embodied Tasks | 2025 | [2504.19854v1](https://arxiv.org/pdf/2504.19854v1) |
| 24 | [OpenVLA-OFT](https://www.semanticscholar.org/paper/5088b84507462bf725bb4898623d1dead4c6a206) | Fine-Tuning Vision-Language-Action Models: Optimizing Speed and Success | 2025 | [2502.19645v2](https://arxiv.org/pdf/2502.19645v2) |
| 25 | [RL4VLA](https://www.semanticscholar.org/paper/1d92c345ff78631aa5509c84e6ff91d446e2c712) | What Can RL Bring to VLA Generalization? An Empirical Study | 2025 | [2505.19789v4](https://arxiv.org/pdf/2505.19789v4) |
| 26 | [SimpleVLA-RL](https://www.semanticscholar.org/paper/bca7f4dd4db559d6772b470d9fe5391e3608cc8c) | SimpleVLA-RL: Scaling VLA Training via Reinforcement Learning | 2025 | [2509.09674v1](https://arxiv.org/pdf/2509.09674v1) |
| 27 | [SmolVLA](https://www.semanticscholar.org/paper/6ab4d113676d00e74b55e918fee4c7affaa8652f) | SmolVLA: A Vision-Language-Action Model for Affordable and Efficient Robotics | 2025 | [2506.01844v1](https://arxiv.org/pdf/2506.01844v1) |
| 28 | [SpatialVLA](https://www.semanticscholar.org/paper/10301766e5686bda76722ef2af1362213b934cc0) | SpatialVLA: Exploring Spatial Representations for Visual-Language-Action Model | 2025 | [2501.15830v5](https://arxiv.org/pdf/2501.15830v5) |
| 29 | [UniVLA](https://www.semanticscholar.org/paper/23ec5b0184fe48558f14401247a411d2a48bd75f) | UniVLA: Learning to Act Anywhere with Task-centric Latent Actions | 2025 | [2505.06111v3](https://arxiv.org/pdf/2505.06111v3) |
| 30 | [π0.5](https://www.semanticscholar.org/paper/1e2a82ef5e0325a26ca0344c4f8c256c52fae7ec) | π0.5: a Vision-Language-Action Model with Open-World Generalization | 2025 | [2504.16054v1](https://arxiv.org/pdf/2504.16054v1) |

Ingest last. Its PDF is in `kdb/papers/` from the start, because the seed ontology quotes it, but no row may cite it before Submission 2.

| # | Paper | Title | Year | arXiv PDF |
| - | - | - | - | - |
| 31 | [LIBERO-Plus](https://www.semanticscholar.org/paper/dfa5ba43d6df553bb3a8cf3fc1cf2d1210cc10c3) | LIBERO-Plus: In-depth Robustness Analysis of Vision-Language-Action Models | 2025 | [2510.13626v3](https://arxiv.org/pdf/2510.13626v3) |

***

Found an error on this page? [Report an issue](https://github.com/aegean-ai/eaia-feedback/issues/new?template=page-issue.yml\&page=/aiml-common/projects/nlp/physical-ai-kdb/inputs).


This documentation is built and hosted on [Mintlify](https://mintlify.com), a developer documentation platform.