Difference between revisions of "Talk:114: Computational Linguists"
m |
|||
| Line 8: | Line 8: | ||
Having multiple theories and models is literally what makes science science in any field, as mentioned above physics is the same. [[Special:Contributions/172.70.110.231|172.70.110.231]] 23:45, 11 October 2022 (UTC) | Having multiple theories and models is literally what makes science science in any field, as mentioned above physics is the same. [[Special:Contributions/172.70.110.231|172.70.110.231]] 23:45, 11 October 2022 (UTC) | ||
| + | |||
| + | Even though this comic was made in jest, I actually think old-school computational linguists' days are numbered, with their complicated theoretical models outperformed by softmax(QK^T/sqrt(d_k))V. "Every time I fire a linguist, the performance of the speech recognizer goes up." Perhaps computational linguists are the new easy target? [[User:Phlaxyr|Phlaxyr]] ([[User talk:Phlaxyr|talk]]) 03:36, 12 July 2023 (UTC) | ||
Revision as of 03:36, 12 July 2023
Why Ryan North? 108.233.253.211 21:49, 8 January 2013 (UTC)
First of all, Ryan North happened to specialize in computational linguistics in his masters degree. He was mentioned because he was a computational linguist. On a side note, Ryan North's father was called Randall (though he was not the Randall whose comics this wiki explains). This may have somehow influenced Ryan's name appearing on this list.--Commarchinin (talk) 12:44, 30 August 2013 (UTC)
- It may also be a joke on Ryan's webcomic name: "Dinosaur Comics" 121.44.164.207 09:52, 20 October 2013 (UTC)
Don't particle physicists have the same issue: string theorists, Bohmians, Many-worldsians, Copenhagen-interpretians, all-possible-pathians, etc.? Djbrasier (talk) 02:55, 1 March 2017 (UTC)
Having multiple theories and models is literally what makes science science in any field, as mentioned above physics is the same. 172.70.110.231 23:45, 11 October 2022 (UTC)
Even though this comic was made in jest, I actually think old-school computational linguists' days are numbered, with their complicated theoretical models outperformed by softmax(QK^T/sqrt(d_k))V. "Every time I fire a linguist, the performance of the speech recognizer goes up." Perhaps computational linguists are the new easy target? Phlaxyr (talk) 03:36, 12 July 2023 (UTC)
