Robert Jenq

5 posts

Robert Jenq

Robert Jenq

@RobJenqMD

Katılım Şubat 2015
87 Takip Edilen58 Takipçiler
Robert Jenq
Robert Jenq@RobJenqMD·
@Dudakov_Lab That’s a good point… if we assume that the right answers are indeed right, then essentially this shows that an impartial summation of the entire literature often fails to identify the truth
English
0
0
0
34
Dudakov Lab
Dudakov Lab@Dudakov_Lab·
@RobJenqMD That’s fair, but is it worrying that the answers can’t easily be gleaned by an “intelligent” being with access and ability to “digest” the entire biomedical literature? (Admittedly intelligent and digest are doing a lot of work in that question!).
English
1
0
0
27
Dudakov Lab
Dudakov Lab@Dudakov_Lab·
Knowing almost nothing about the ins and outs, at first I’m actually surprised it’s only 50%. Although with so much “disagreement” in the literature, it is probably extremely difficult knowing how to weight any particular finding. Recency? Quantity/consensus? Review articles?
NEJM AI@NEJM_AI

A model named PubMedGPT 2.7B was trained on millions of scientific articles. On a dataset of exam prep questions for USMLE Step 1, the model answered 50% correctly. Read more about this autoregressive language model:

English
1
0
0
768
Robert Jenq
Robert Jenq@RobJenqMD·
@LastWeekTonight Such a great episode! John, can you follow this with an “apology” to China? In Mandarin, please?
English
0
0
1
0
Last Week Tonight
Last Week Tonight@LastWeekTonight·
Here’s last night’s story about Taiwan, and what the people who live there would like their future to look like: youtu.be/9Y18-07g39g
YouTube video
YouTube
English
39
239
951
0
Robert Jenq
Robert Jenq@RobJenqMD·
@JeffBezos how about identifying talented preschoolers from disadvantaged backgrounds and changing their lives by providing resources
English
0
0
0
0