[Vote] Elder Voting (Election #3)

The poll is open! Select one of three candidates for election to the Council of Elders:


  • bor (bor1)
  • jefferythewind
  • yifanxie (yxbot)


All votes must come from active Numerai participants and votes are blind and log-weighted by stake size. Results are blind until poll closes. The poll will close at 8pm ET on July 5, 2022: NumerBay - The Numerai Community Marketplace


The transition to TC staking is probably the largest challenge we have right now as a modeling community. I would be interested in getting people sponsored that have ideas how to help people in tackling the TC challenge that with sponsoring from the CoE or with the process of retroactive reimbursements.

In general I am interested in getting people to share more about their models, and I tend to be quite forthcoming when people ask myself (if only because I am running something so cumbersome and tailor-made, it’s not easy to copy). And there is a smallish group as well that shares things in the datascience channel. If I can prod more people into sharing, all the better.

And it looks right now like the meetup in London will involve some sharing. Great! Now, how do we get more of that to the whole community?

  • In response to a question from hb: Besides chatting, I was under the impression that the workload from CoE is about 1 1/2- 2 hours a week. That is what I have available as well.

  • In response to a question from jrb: Would love to be part of the twitter spaces, but the current timeslot doesn’t work for me. one-two hours later (depending on summer-time shenanigans) would be perfect. I miss the social part of the weekly office hours with arbitrage!, although the daily thing was too often for me to keep track of.


Fellow Numeratis:

Just to be clear, this isn’t some kind of “campaign message” :slight_smile: the following is something I really want to do - to bring some of my favourite ML/DS communities closer together, and do greater things together. it IS my main idea going into the CoE membership voting, but I would also just like to share it so that we can perhaps do something together. So, here we go:

My main proposal would be to facilitate “targeted activities” to Extend awareness of Numerai to the wider ML/DS community , and by doing this also facilitate high-quality content sharing between us.

Having been to NumeraiConf, and having talked to the team and other participants, I believe this is the most critical topic shared by the team and participants. Wider engagement and participation are critical for more model diversity for the metamodel, more talents participating leading to better idea sharing and mutual learning, and better recognition of the project and the NMR token.

Meanwhile, I believe that the Numerai dataset, with its clean data, steady & improving training and validation sets, and readily productionalised process & toolkits, is probably the most ideal open dataset around for data scientists of all levels to learn all kinds of tricks (feat. engineering, selection, dim. reduction, representation learning, stacking, ensembling, etc), so there is clearly something we should work more to make the wider ML/DS community to be aware of.

At the same time, platforms like Kaggle are offering less and less interesting tabular competition, and this seems to be an unsolvable problem ever since they got acquired by Google. So even just targeting 10 million Kaggle registered users, there is a lot to gain both for recognition for the Numerai projects, and get more people to start building models around the numerai dataset

Concrete Actions
Here are some of the things I am confident we can achieve with the sponsorship from the CoE, some of them needed to be better thought through, but chiefly I would like us to bring numerai to wider DS/ML community, get them to try it out, and then make it easy & straight forward for them to join the tournaments properly.

  • Assemble a technical task force to create high-quality and “ready to run” technical notebooks on subjects that are related to Numerai, and share them widely. one of the best examples is this by @perfect_fit - look at the number of upvotes there! we can also create tutorial series like this to cover a wide range of end-to-ent stuff

  • Launch CoE-hosted numerai-related competitions, such that they are run by the numerai community (i.e. not by the team), and can come in different iterations. My first idea would be using Kaggle’s well-established community competition platform (free storage & compute & comp. hosting) where we can host numerai datasets (V2/V3), incrementally update (V4), provide baseline examples, set our metrics, and encourage both Numeratis and non-numeratis to compete side-by-side by individuals or by teams. The CoE can give out non-cash kudo to the winners, and the best content sharers - Cash not allowed as prize for community challenges, giving NMR as prize might be controversial given certain event in the distant past (but why not) :slight_smile: With well-oiled APIs on both sides (Kaggle & Numerai) we can even set up our own LB and do a TC optimisation challenge! just need to think through how to set such a competition up.

Anyway, these are my main thoughts for now, nice to have this opportunity to share it here, and hope it doesn’t sound too outlandish - but this is a DAO right? so it shouldn’t be lacking ambition :slight_smile: happy to discuss further to hammer out the details regardless the voting results!


What other time do we have but the present.

I love what @bor and @yxbot have written, and I agree. Community outreach and even the youth involved in a creative way is probably a great thing to do.

A little background about myself. I have a background in software development and systems design after having done quite a bit of living in Asia and Europe and I now back closer to homee in Montreal and a student again in Math and Stats. I’m focusing on ML ideas for finance. In the past I’ve spent countless hours with market data trying to design algos. I discovered Numerai and always thought this was the best project happening. I’ve learned a lot and continue to learn. I am a student after all.

The feedback from the Numerai community has been overwhelming recently. It is one of the few projects where I feel I’ve gotten back at least as much as I’ve put in. Everyone that hears about it is just like, wow that’s an awesome project. I still work desperately every week on models to thaw that frozen TC. I support open source projects and sharing. I’ve shared recent work on the #datascience channel but will try to create better content on the forums going forward. I consider myself a student of the game. I am also involved in a project using ML methods at high-frequency, which could be an interesting area to expand the repertoire of Numerai operations.

I intend to remember to attend the weekly fireside chats. I also wouldn’t mind working on an academic-style event here in Montreal. A lot of untapped DS here.

For me its always been fun here (at Numerai), respectful and nice, and I would like to keep it that way. It’s gotta be fun. It’s gotta be healthy. I think it is both of those things. I think we are all ready to continue to grow and expand.