Proposing a Challenge for TIG

This guide provides a step-by-step process for proposing a new Challenge to The Innovation Game (TIG).

Follow the steps below to propose a new Challenge and optionally nominate a Challenge Owner — for consideration and potential adoption by The Innovation Game.

Challenge Owners

The role of Challenge Owners is to produce and update challenges to ensure the resulting innovations remain maximally useful for real-world applications and/or are competitive with SOTA (state-of-the-art) algorithms. They have the following responsibilities:

• Challenge Design: With support from TIG Labs, Challenge Owners use their expert understanding of the problem and industry knowledge to design the challenge.

• Challenge Maintenance: After a challenge is launched, Challenge Owners are responsible for maintaining/updating the challenge to ensure the innovation it drives stays maximally relevant for real-world applications.

• Quarterly Algorithm Performance Reports: Challenge Owners are required to provide quarterly reports on how the best algorithms submitted perform against standard benchmarks representing real-world scenarios.

• Engage their Network: Leverage their network to encourage the participation of world-class researchers, either by inviting them to contribute as Innovators or by involving them in discussions around the eligibility of the algorithm for advanced rewards.

Read more about Challenge Owners. You can also check out the Original Challenge Owners Whitepaper

Designing a Challenge

A problem proposed as a Challenge should be important, asymmetric, independent and allow random instance generation.

• Importance: The problem should address a significant question in science or technology. Importance can mean practical relevance or fundamental intellectual merit. It doesn’t have to be immediately profitable or commercially applicable - a purely scientific challenge can qualify if its resolution would be a notable advance.

• Asymmetric: The challenge should be designed such that finding a solution is substantially harder than verifying one. Finding a solution for a given instance should require significantly more computational resources than checking whether a proposed solution is correct. This asymmetry is crucial for TIG’s economic model and is a strong prevention against Sybil attacks.

• Independent: The challenge should be sufficiently independent of other existing challenges in TIG. Ideally, a proposed challenge should be such that an optimization or algorithm that drastically improves performance on it would not also drastically improve performance on other active challenges.

• Random Instance Generation: The challenge should allow for the generation of unlimited pseudo-random instances of the problem, with adjustable parameters that affect difficulty. This allows Benchmarkers to continually compete on new instances and enables the challenge difficulty over time. Each instance should be representative of the real-world version of the problem to the extent possible, so the progress in TIG translates to real progress on the underlying problem.

Read the Original TIG Challenge Design Whitepaper for more information.

Challenge Checklist & Requirements

When proposing a new TIG Challenge, ensure that the following sections are clearly outlined and included in your challenge description:

Objectives: State the goal of the challenge. What constitutes a valid solution or a successful outcome? For example, is the aim to find any solution that satisfies certain conditions, or to find an optimal/minimal solution according to some metric?

Constraints: List any constraints or rules that solutions must obey. These could include input domain restrictions, resource or time limits for solvers, allowable techniques, or other conditions that define the legitimate solution space.

Scope: Clarify the boundaries of the problem. What is included in the challenge, and what is explicitly out of scope. It is often useful to mention not only what the challenge covers, but also what it deliberately does not cover. State any assumptions you are making about the problem or the environment (for instance, idealized conditions, glossing over certain complexities, etc.) that are relevant to how the challenge is formulated.

Background: (If applicable) Provide brief context on how this problem has been addressed previously. Mention any existing algorithms, known results, or prior competitions related to the problem. Citing relevant prior work or literature can help justify the challenge and show how it builds on or differs from past approaches.

Submitting your Proposal

Once you have prepared your proposal, you can submit it by completing the Γ-Grant Program Application. Make sure to read the program overview and guidelines before submitting.

What Happens Next

Review Process

Your Proposal for TIG Challenge and Challenge Owner will be reviewed by a Committee of relevant experts appointed by The Innovation Game. The Committee will consider:

• Whether the proposed Challenge is suitable for The Innovation Game.

• Whether your proposed Challenge Owner is suitable for the role.

Committee Interaction

There may be interaction between you and the Committee as they:

• Seek to understand your proposal.
• Provide guidance on modifications that may be needed prior to approval.
• Evaluate the proposal for adoption by The Innovation Game.

Decision Outcomes

If the Committee rejects your proposal:

• You will be informed of the decision.
• Reasons for rejection will be provided to you.
• The Committee has absolute discretion regarding approval or rejection.
• The decision of the Committee will be final.

If the Committee accepts your proposal:

• You will be informed of the acceptance.
• You will be invited to enter into an agreement with The Innovation Game.
• The agreement will be a TIG Challenge Developer Grant Agreement to cover development of your Challenge Design.