On Tuesday evening, I submit my grant for InnovateUK's July Round of Smart Grants through a grant writing partner service.
Idea: The research literature uncovers a bunch of new techniques for understanding animals, but these academics aren't going for the animals that we care most about.
Product: Give dog owners a fun and engaging way to deepen their relationship with man's best friend.
Take one-minute vertical video of your dog and analyse for behavioural cues. Build your streak, share your moments, and look back on memories and anniversaries with your 3D scanned in-app dog avatar (Additional outfits sold separately).
Tech: Apply contemporary machine learning approaches from the research literature to build an app to listen to and understand dogs.
Team: I'm a Machine Learning Engineer with 4 years experience. My co-founder is a senior application developer. We have someone with common sense to handle operations. We've lived together in AI Safety adjacent grouphouses.
Execution: TBD. Grant funding is for £495k for 12 months of industrial research including Community Dataset Creation, ML development, App development, Pre-User Testing.
--
Additional Info for context.
Grant writing partner tells me the success rate for applications without them is 5-10% and with them is 55-60%. 70% of the previous round's winners used this service.
https://apply-for-innovation-funding.service.gov.uk/competition/1982/overview/e5d935d6-0690-4f1d-b442-4fb868be8384
1,000🏅 Top traders
| # | Name | Total profit |
|---|---|---|
| 1 | Ṁ47 | |
| 2 | Ṁ44 | |
| 3 | Ṁ40 | |
| 4 | Ṁ4 |
Feedback fyi
- ### Assessor 1
The proposal presents an impossible task of canine understanding, and confuses this with training needs and anthropomorphism. As pet dogs have a huge array of contexts to develop within, cataloging them all for machine learning will not, at best, yield useful results. At worst, deeper confusion between dog and humans will occur. The responsibility bond between dog and human is unique because the nuances of personality is the key to comprehension, not understanding. Aside, face to face dog training is a requirement of owning a dog. Without this, only chaos occurs.
- ### Assessor 2
The significant scale of this application does not match the speculative nature of the project proposed, mainly because the proposed value proposition is not demonstrable. The project is high risk and the proposed value it represents is not supported by evidence. In particular, benefits to dog owners in terms of improved behaviour and a reduction in consequential costs at the NHS are speculative.
- ### Assessor 3
What is not explained is how this work compares with and is better than the multitude of dog training programmes that are available, many at zero cost. The personal touch also is a strong aspect of current dog training programmes especially at local neighbourhood level. No evidence is presented if the dog owning community is willing to pay for such an on line App.
https://www.reddit.com/r/smallbusinessuk/s/YZFq4H2Mqe
Assuming this reddit user claiming to be an Innovate UK assessor speaks true about the number of grantify applications per round, we've got a ~9% chance.
I won't bet on this market, but there may be a bunch of free mana available here.