I want this
Machine learning, deep learning, and AI should be a part of the toolkits of most bioinformaticians. And as they would participate in our society in every facaulty, as a person who’s living in this era, it won’t be disadvantegeous to know more, even not the frontiers, but just high-end applications.
But for years I never had chances for such project experiences. Then this year I was brave enough to visit abroad during PhD. For that I need a host. Renouned labs in bioinformatics each has its strength as the PI’s first recognized techniques. Manifolds, Embeddings, Graphs, ecosystem, … The first fellowship I was applying was one from Bayer, which limit the hosts within Germany. At that point, I was simply looking for renouned labs within Germany, not considering fields, bioinforamtician, AI era, thing. But Rodrigo moved to UK? and not answering emails, Fabian has too many facaulties that I didn’t try. It might be gpt-4o at that time suggested Carsten to me. He’s the head of AI for Health in the same department as Fabian.
I thought, what a fate. For a fellowship encoraging multi-disciplinary research in medical field. I am affiliated to a medical institute, Carsten is AI for health. But, problems:
- What to do
- Carsten assigned me to HistoGPT fine-tuning projects, all tumor related
- For my current supervisor’s approval, tumor is not her subject
- Contacts
- Documents from Carsten only prepared after deadline
- Recommendation letter, my main supervisor didn’t agree on getting from my co-supervisor
- Contracts
- HR not involved yet
These foreseed the troubles later on. Status are:
- I didn’t get that Bayer fellowship
- I got two from Sweden, all requires PhD project related, i.e., non-tumor research
- No visiting contract, because of the censorship: my USTC background
- Because of no contract, when not sure if I can be onsite, accommodation cancellation without refund
- The time resources, and projects of the stay are not optimal
But I still come, regardless of all these. By observers, I am determined. Deep in my mind, I was worried. The winning drives are
- I didn’t take the 1 year TA contract to HongKong because of Qiu’s disencounragement in third year of my master’s and that destroyed my life.
- The AI experience that I want.
People
Only after being inside that I realized here is indeed a place where cool science occurs. This mostly owe to people, not the infrastructures. Computational Health Center takes one building. The building has everything inside for doing computational health research, and nothing else. The building appears prefabricated. The offices are open, everyone except PIs shall book a desk every other week. Each desk is standing and equipped with a monitor. There’s one small kitchen on each floor and a terrace (just some wooden chairs and desks beside by the grassland) at the entrance floor. The terrace is the place for coffee breaks and Thursday after-work drink-together. Central air-conditioner and automatic curtain system is equipped. One more note: It’s located close to Alliaz Arena, you see it by window, across a grassland.
What else would a computational health center need! People come here 2 days a week, as everyone can work remotely. You have good network connection, standing desk for long-time screen work.
With all these set-up, a much inferior building vs KI’s buildings (which happen to be the biggest scandel in construction), I still feel very confortable and inspired. Mostly because of the researchers here, all are in the best tire.
The PIs, like Fabian, while affliated to multiple institutes and universities, this site is actually his main site and most of the members are here. Also other renouned people graduated from his lab. And these PIs, all are very smart, quick-thinkers and quick-spearkers. The recent stars in computational biology across the glob are willing to come here. The first week I was here, Fabian invited the first author of the DNA to expression work from US. The second week, Carsten invited a causal bayesian researcher from Canada. Then a colleague who knows about that argued with the speaker over lunch for 2h. Yes, except for PIs, the researchers at each level are so passionate about science and research. In Carsten’s lab, most people are from outside Germany, Turkish, Iran, India, … There’re no other reasons that people come to this construction site, living Munich life with high cost of life, than being intrigued by science and research here. These colleagues are smart and high-level, which also refreshed my prejudice of these populations.
Projects
First observation is the PI’s attitute towards the projects in the group.
In every group there are multifold projects with different sizes in terms of outcome, importance, resources. Carsten cares each project, from the grandest to the smallest. Life won’t be too miserable if one gets a small project in the group. He’s also interested in the real science behind projects, not just the outcomes.
Another observation is the extension of one project: that’s how people started career, get fame and power. We see this happend, for those big names in any field. They had one high impact research, then they were reached out to give talks, have collaborations with academia and industry, they can start their own business. That’s the case for the PhD student same year in the group now. Not all collarative projects are well designed and will end up great, as many of them are just leveraging his already published work. Nevertheless, this still shows that this work is impactful and paves the way of the group and the first author’s success.
