My dad died 30 years ago after it spread to his liver. He was in the study at duke for what is today the standard treatment line for immunotherapy. I hope he was able to help. He died 9 months after his diagnosis.
Enjoy the time you have with him now and be grateful for all the years you've had. Nothing is going to make the loss of a parent "OK" and it will hurt no matter if you see it coming like you are or it's a surprise.
Immune therapy is hit or miss. Friends lung cancer was exquisitely sensitive to it. Stage 4 to no sign of disease. Another friend with melanoma immune therapy didn't seem to help at all.
Melanoma is usually completely resistant to traditional cancer drugs. So that sometimes immune therapy works is very welcome.
Sadly it didn't. And it also lead to extreme bleeding in the brain metastasis that we had to stop it. Switched to BRAF/MEK inhibitors and that looked good for a few months until it didn't anymore.
For every train, there is a fixed number of tickets per price category. So sometimes, you can still find cheap tickets ("super sparpreis") a day before because thag specific train didn't have many bookings:)
Aw hell they picked up on that stupid American capitalist crap?
I worked in Germany in 2005 and back then everything was fixed price per kilometer for each train class, and you could get rail passes of sorts and get on whatever the hell trains you want during their validity. I'd take train roundtrips after work just to watch sunsets.
The price concept discussed above applies to long-distance trains. Local trains are still different, especially if you have the "Deutschlandticket" (germany ticket) you can just hop on any local train you like.
The scope is a bit different. The study uses an LLM to interpret pose estimation data and describe the behavior in each frame. The output is text which can be used to create embeddings of behavior. As someone who works in ethology, that's a clever (but maybe expensive) idea.
I think the author could use something similar. With multi-person pose estimation models.
Since its text, especially text with links to other articles, there is no need for tags.
If I had a clue how to do this (sorry, just a neuroscientist), I would probably create "communities" of pages on a network graph and weight the traversal across the graph network based on pages that the person liked (or spend X time on before).
I wish this treatment was available a few years ago.