For this week's Stanford NLP Seminar, we have @_jasonwei telling us how finetuned language models are zero-shot learners. Join us over zoom tomorrow at 11 am PT. Registration: https://t.co/L840QYLydU, Abstract: https://t.co/N5ykUOySvm

One of the main reasons Brazilian public universities do not reach their full potential is the faculty hiring process: exams + publication points (based on random score tables). Departments are hiring people good at exams and at scoring publication points, not research.

Stories of data gone wrong are the best stories and also the ones we so rarely get a space to share. I'm very excited to be kicking off the next #DataMishapsNight with a few stories of horrendous mistakes, overlooked edge cases, and algorithms gone amok. Join us!

Machine learning Twitter: how do you pass a large list of arguments to your python training scripts? If you are happy with any other library please comment below.

Aside from the considerable time spent training and hparam hunting for many models myself, the costs of doing so aren't trivial. Cloud costs, local NAS, computers, GPU, etc costs are rising w/ more & larger datasets and recent pretraining adventures (CLIP, MAE, etc)

Happy Birthday @PyTorch! It's been amazing to see PyTorch go from a small research project to one of the most used and loved AI framework.

Wow, it happened!!! "Optimization for Deep Learning" is not a subtopic in ICML 2022! Thanks @icmlconf!