What is the General AI Challenge?

Original
Current AI solutions, deemed 'narrow AI', are essentially task-specific - they learn to do one, well-defined task extremely well. A 'general AI' system, on the other hand, will be capable of “learning how to learn”, much in the same way a human does. While a technology like this would serve as a tool to optimize our science and technology, with limitless benefits to our wellbeing and health, there are still quite a few hurdles to overcome before general AI becomes a reality.

We spoke to Olga Afanasjeva, COO of GoodAI and Director of the General AI Challenge, which launches on 15 February! Read on to learn more about getting involved, and what the Challenge hopes to achieve.

Can you give us an overview of the General AI Challenge?
In an effort to get to beneficial general-purpose AI, artificial intelligence R&D company GoodAI has put together the General AI challenge, along with a prize fund of $5mil meant for the champions of the multi-year Challenge’s many rounds.

GoodAI has also scheduled several open-invite kickoff meetups all over the globe, including Tokyo (15 February) and New York City (22 February), both organized by their partner Araya. Organizers will underline the goals of the Challenge, and also introduce the warm-up round and its criteria. Those who come to the meetups will have the chance to network and form development teams.

What do you hope to achieve?
GoodAI wants to incentivize talent to focus on developing general purpose AI that is both beneficial and safe. General AI is humanity’s most promising frontier and the ultimate leverage to expedite advancements in science and technology, tackle diseases, and learn more about our universe. The Challenge is essentially a “citizen science” approach to the research needed to reach that new frontier.

The Challenge can also be seen as an inclusive alliance among innovators from around the world, bringing people together to focus on one major goal. It provides each team or individual with the freedom to solve the outlined research problems using their own preferred methods.

Who's getting involved?
Along with a star-studded advisory panel, the organizers of the Challenge have also partnered with Microsoft Czech Republic and Slovakia and NVIDIA to help empower talented researchers and programmers who would otherwise lack computational resources for their experiments. To help develop their solutions, participants worldwide will be provided with free access to Microsoft Azure cloud space. In addition to monetary prizes, participants will get a chance to win a cutting-edge NVIDIA GPU.

Along with technological partners, the Organizers are also proud to welcome partners who have an indispensable role in engaging the global community, such as Araya, RE•WORK, and local partners - Czech Technical University, SlovakStartup, and 0100 Campus startup hub.

What's the first round of the Challenge?
In the first round of the General AI Challenge, participants will be designing an AI that is able to learn in a way similar to humans: by developing useful heuristics and learning strategies, and being able to reuse those strategies when learning new skills in new domains. Essentially, it would be capable of “learning how to learn”, allowing it to innovate solutions to new, complex tasks in less time.

Organizers will provide participants with a set of training tasks implemented in CommAI-Env, a simple bit-based communication environment. These tasks were designed to test and facilitate the development of the gradual learning ability. They are based on the CommAI-mini set, recently proposed by Baroni et al, 2017.

The champions of the warm-up round will be free to share their solution with the world under an open license, if they decide to do so.

In the warm-up round, GoodAI is offering a total of $50,000 in prize money:

●     $15,000 for the AI agent that performs best in an objective test (a set of evaluation tasks)

●     $10,000 for the best idea (selected by the Challenge Advisory Board).

●     The remaining prize money will go to the 2nd- and 3rd-place winners in both categories.

How can people get involved?
If you’re interested in taking part, you’ve got a lot of options:

  1. Get involved in the first round of the Challenge  by registering here on 15 Feb
  2. Join the Challenge forum if you just want to brainstorm or contribute ideas
  3. Jump in at a later round - there are more technical and theoretical rounds to come!
  4. Stay tuned to the General AI Challenge by following @ai_challenge on Twitter and our Facebook page

View the General AI Challenge website here & get involved!

Neural Networks Deep Learning Machine Learning Computing Systems AI Machine Intelligence


0 Comments

    As Featured In

    Original
    Original
    Original
    Original
    Original
    Original

    Partners & Attendees

    Intel.001
    Nvidia.001
    Graphcoreai.001
    Ibm watson health 3.001
    Facebook.001
    Acc1.001
    Rbc research.001
    Twentybn.001
    Forbes.001
    Maluuba 2017.001
    Mit tech review.001
    Kd nuggets.001