HOW discussion: How can ACCELNET support collaboration to close knowledge gaps?

[crossley]

Please share your feedback on this group discussion by replying in this thread. Your contributions are important as we aim to identify gaps in knowledge, pinpoint collaborators who can help us bridge these gaps, and explore how sustainable collaboration within the ARCANI network can address these issues. We will incorporate the thoughts and opinions expressed here into a white paper focused on these critical questions.

[ookwrd]

Pairing a post-summer school grant writing activity, with some senior people joining/advising the writing – can be very efficient because there is a shared language and motivation and time to think and write.

[Somayeh-h]

Workshops like the Telluride Neuromorphic workshop which is a combination of talks (for the first week) then working towards a project for the next three weeks, and doing activities outside such as hiking to help having discussions in different settings would be great. Link: https://sites.google.com/view/telluride-2024/home?authuser=0

[ookwrd]

More cheese (serious idea about science backchanneling – Bruno or Pierrick can explain)

[HoverflyLab]

Follow-up meeting (could be online) to share how people have used the knowledge gained/connections made - forces you to not just get immediately submerged by your every day job.

[Daspraelon]

What team building would really work?

[ABBarron]

Promote more local gatherings that are well-supported, rather than spending on airfares. We often don't know who's doing related work in our area. We end up finding this out by meeting in faraway places. We could use a local network map to identify interested parties. Then support these with gatherings that aren't necessarily about the science to begin with, but just break ice and bread. Use the saved airfare costs for interesting settings. This could help establish the kind of friendship/collaborations that are more sustainable through time because they're local.

Yes! This is coming up! A global-local model so we can keep activity running locally and less frequently pull a larger team together.

[HoverflyLab]

we need more Andy Barrons (amazing ability to see connections where others see silos)

[gmfricke]

Need to fund LLM Andrew emulator to make interdisciplinary connections.

[stenti]

Inviting funders to meetings

[jarmarshall]

Invite funders to meetings (Virginia)

[ABBarron]

models of successful collaborative networks Mathematical Biosciences networks - NSF funded Marie Curie training networks NESCENT - 3 meetings over 18 months? global scholarships - split across institutions

[ABBarron]

supporting ECRs - TWCF priority. ECR co-uplift. There is more value to ECRs promoting each other than a mentor doing it. More sustained impact and ongoing ECR support.

[ABBarron]

mentoring of assoc prof level - helping a new leader establish. Buy out a course and help a new emerging person get their first grant to get their wings up. Catalyse. A regranting scheme - give an emerging leader a capacity to seed small projects and lead their own network.

[ABBarron]

WRITE CATAYLYTIC AND HIGH RISK HIGH REWARD INTO GRANTS. Booyah!

[ABBarron]

DI and other communities needs to fledge independent of the initial funding sources.

[fgashby]

At UC Santa Barbara, we started an interdisciplinary PhD program about 15 years ago that we called Dynamical Neuroscience. That program, which still exists, includes faculty from 7 or 8 campus departments and is broad enough to house all of the research described at this workshop under its umbrella. One of our greatest challenges though was to fund graduate students. This was largely because we used existing campus coursework to build the program, and the departments that offered those courses controlled the course TA assignments. For this reason, we didn't have any TAs to offer as funding for our grad students. Instead, all students were funded via fellowship or individual faculty grants. It would have made a huge difference if we could have funded even one student a year via some kind of ARCANI fellowship. A viable program requires enough students to form a cohort, and especially in the beginning, we were barely able to do this. Even one more student a year would have been an enormous advantage to the whole program.

[crossley]

Ultimately may not be a lot of money. This was a key idea to emerge from my table 1 discussions. The consequence of this may be that many of our more traditional ideas -- e.g., longer term residential lab visits, postdoc exchanges, etc -- may not scale effectively within our eventual budget.

One clever way to sidestep some of those small budget concerns may be to attach to larger international conferences (e.g., Society for Neuroscience) with satellite conferences. Since so many of us will be there already, this could ease costs. There is also the opportunity / possibility of deliberately making some of these satellites topically distinct yet complimentary to the main theme of the meeting, which would hopefully help expose more people to new fields and inspire interdisciplinary collaboration (e.g., machine learning / AI satellite at SFN).

Another intriguing idea that was floated is -- instead of a large meeting of many researchers of all career advancement -- to go with the dark smokey room of secret decision making (I'm thinking the X-files here). Silliness aside, the idea is that five of the worlds most powerful and well-funded researchers may be more effective for the greater collective than the greater collective itself. Perhaps some benevolent dictator or trickle down hopes and dreams here.

Another key idea that landed strongly for me is the idea that a key limitation / gap is the absence of suitable incentive structure for research and development that genuinely fills that gap between natural and artificial intelligence. The idea would be for ARCANI to provide such an incentive structure by investing / funding a sort of competition in which teams compete not to create the best robot or AI overall (in an engineering optimality sense), but rather compete to create the most biologically compatible / plausible agent. I suppose I'm thinking something along the lines of this:

https://www.brain-score.org/

[ABBarron]

ensuring equity of involvement. It's really important in invitations to ask if people need any help to come. That should be gender neutral but women can be disproportionately impacted by social constraints. ECR women may not feel able to ask for support and it is easy to offer an extra hotel room or whatever might help.