@MattNiessner This is a skill the student learns *during* their PhD and requiring it from an applicant makes as much sense requiring a track record at NeurIPS.
This may revolutionize data science: we introduce TabPFN, a new tabular data classification method that takes 1 second & yields SOTA performance (better than hyperparameter-optimized gradient boosting in 1h). Current limits: up to 1k data points, 100 features, 10 classes. 🧵1/6
Some ML research areas are extremely crowded. Working there is a race. But it's not about "how to benefit science" but "who gets the credit." This is the furthest from being a scientist I can imagine. How did we get here?
Dělat to, co je správné – i když to může znamenat tu nejvyšší oběť. Výsadkáři, kteří provedli úspěšný útok na Heydricha, takovou odvahu měli. Jejich čin byl v roce 1942 světlem v temnotě nacistické diktatury. (1/2)
Annoying to see a recent "Review on Multispectral Photoacoustic Analysis of Cancer: Thyroid and Breast" in @MetabolitesMDPI special issue very much failing to mention our second-generation #MSOT imaging of breast cancer https://t.co/Ff2HfX4G3n
Our new paper on optoacoustic imaging of metastatic cervical lymph nodes is out! High-resolution handheld #MSOT showing again its capabilities on an interesting clinical problem. Great collaboration with @JasperVonk and the teams on both sides @MunichImaging & UMCG
@RockyMtnMommy @a33855036@FrankBraconi@NickGBreeze Wouldn't it be great to become also the most powerful country in reporting measurements that everyone else can understand? :D