AI GRANT

Get $5,000 for your open source AI project

Learn more

Sponsors

Stay In Touch

To be notified when grant recipients are chosen, and when applications open again later this year, sign up below.



Spring 2017 Finalists

The 45 finalists below were chosen from more than 450 applications. Ten grant recipients will be announced on May 27th.

  • Rowel Atienza, open-source dataset for Tagalog speech corpus.
  • Liam Atkinson, a neural network to generate puns.
  • Scott Hawley, Frank Baird, Tamara Baird, free heart & lung audio dataset.
  • Juan Carrasquilla, simulation of many-body quantum systems with neural networks.
  • Volkan Cirik, a language grounding dataset for Go games.
  • Chris Coughlin, aircraft damage detection.
  • Joshua Daniel Eisenberg, extracting narrative structure from long-form text.
  • Gabriel Goh, Distill.pub article on the Wasserstein metric.
  • Jeremy Raboff Gordon, augmenting Numenta's hierarchical temporal memory with neural timing.
  • Wah Loon Keng, Laura Graesser, an experimentation system for reinforcement learning using OpenAI Gym, Tensorflow, and Keras.
  • Logan Graham, a skunkworks AI team tackling social challenges.
  • Oliver Hennigh, predicting steady-state fluid flow using deep neural networks.
  • Dang Ha The Hien, Tensorflow library for object recognition on mobile devices.
  • Stuart Andrew Hunt, distributed neuroevolution using Elixir/Erlang and NEAT/HyperNEAT.
  • Hussein Jaafar, R-CNN model for human cell classification.
  • Daniel Kang, speeding up CNN evaluations over video.
  • Ji-Sung Kim, novel deep learning models with minimal racial/gender/socioeconomic biases.
  • Minjae David Kim, robust massively parallel GPU-enabled optimizer.
  • Bart Kolendowski, using machine learning to characterize cancer and predict outcomes from cancer treatments.
  • Kevin Kwok, a fast, cross-platform library for hardware-accelerated deep learning in the browser using WebGL.
  • Edward Lee, predicting survival times in lung cancer patients.
  • Eli Litwack, using computer vision to assist dyslexics in reading.
  • Stefanie Lück, replace manual microscopy with automated object detection and prediction.
  • Rahul Makhijani, identifying child pornography in videos with deep learning and no training data.
  • Colin Morris, named entity generation.
  • Brian Muhia, generating FastText skipgram models for Kenyan languages.
  • Natalia Mykhaylova, training datasets and source identification algorithms for sensor networks that improve public health.
  • Ojash Neopane, datasets and algorithms for autonomous surgery.
  • Liam Paull, duckietown.mit.edu, an open source reproducible advanced autonomy class, a standardized research project, and a robotics outreach effort.
  • Gerry Pesavento, recycling robot.
  • Lenka Pitonakova, creating an on-line platform for swarm robotics, where robot code and algorithm designs can be exchanged and collaborated on in an open-source fashion.
  • Jordi Pons Puig, Audio AI: the Freesound datasets project.
  • Nikhila Ravi, a comprehensive guide to building conversational agents from first principles.
  • Nancy Fulda and Daniel Ricks, expand and augment the Interactive Fiction Learning Environment so that it can be a useful tool for AI researchers.
  • Russell Kaplan and Christopher Sauer, natural language guided reinforcement learning.
  • Michael Schaarschmidt, TensorForce: practical deep reinforcement learning on top of TensorFlow.
  • Benjamin Schreck, computer-generated rap lyrics.
  • Sanuj Sharma, deep learning applied to biomedical image processing
  • Tim Shi, deep learning experiment framework.
  • Aditya/Adi Sidapara, a chatbot to report racially motivated violence and domestic abuse.
  • Patrick Slade, machine learning for motion recognition and trajectory generation of human movement for rehabilitation
  • Christof Stocker, data partitioning, preprocessing, generation, resampling, and cross validation library for Julia.
  • Manasi Vartak, ModelDB: a system to manage machine learning models.
  • Sean Vasquez, handwriting generation tool + creation of new dataset.
  • Natalie Widmann, detecting political and ideological reporting bias in newspapers.
  • Mark Wronkiewicz, simulating physiologically plausible human brain electromagnetic activity using GANs.

FAQ

Can anyone apply?

Yes – any age, any country, no credentials required.

What type of project qualifies?

Any project in AI, large or small, as long as you release your work under the MIT or Apache 2.0 license. If possible, any datasets employed should also be released.

As for what qualifies as "AI," I'm open-minded. Your project can be:

  • contributing to an existing open source project
  • creating a new framework or tool
  • researching a new technique
  • applying an existing algorithm to a new problem
  • creating or curating a free dataset that others can use
  • educating, explaining, or otherwise helping people learn new or existing techniques

…or anything else that feels like AI.

How will you select the winners?

I'm looking for smart people with interesting ideas that could be useful to the world. I'll pay extra attention to projects that seem like they won't get funded another way.

Some friends who know a lot about AI will help review applications (if you know a lot about AI and want to help review applications, email me!)

I have a job. Can I still apply?

Yes! Anyone can apply. You just need to make sure that your employer will let you release your work under the MIT or Apache 2.0 license.

Can teams apply?

Yes. Multiple people can apply together as a team. Please fill out a single application, but provide background information for each person in the team. You should designate a lead person to coordinate the interview as well as receive and distribute/spend the money.

How does it work if I win?

I'll give you 50% up front. I'll check in with you before sending the second half. Then I'll do a follow-up phone call and/or email at the end to hear how it went.

I'll publish a list of all the winners on aigrant.org after they've been selected, and maybe when the projects complete, or reach interesting milestones.

During the project, you don't need to make regular reports on your progress (though I'd be happy to hear from you).

How do I apply?

To apply, please fill out the form at aigrant.org.

I'll select 10 to 2045 finalists for a short interview to make sure you're a real person and to ask a few questions about your project. Winners will be chosen after the interviews are complete.

What's the deadline?

The key dates are:

  • April 30th - All applications must be received by 12pm Pacific.
  • May 14th - Finalists selected, all applicants notified.
  • May 27th - Final winners selected & announced.

Of course, these dates are subject to change if they turn out to be wildly wrong for some reason!

Can I apply multiple times?

Yes, you can apply as often as you like. You can only win once.

How do I get the money?

I can send you the money via PayPal, wire transfer, or bitcoin.

I love this idea and I want to help! Can I provide additional funding, hardware, GPU time, datasets, mentorship, or help reviewing applications?

Yes! 🙏 Thank you for being awesome! If you want to contribute in any way, please email me.

Are you doing this as part of a company?

Nope, this is just me. The money is an unrestricted personal gift. It's not an equity investment or loan, I won't own any of your intellectual property, and there's no contract to sign.

Will this be a regular thing?

Maybe! Let's see how it goes.

Where did you get the idea to do this?

This post is inspired by, and directly borrows many elements from Nadia Eghbal's amazing no-strings-attached grant program. Thanks, Nadia!