New podcast episode! 📢
@l2k and @emilymbender dive into the problems with bigger and bigger language models, the difference between form and meaning, the limits of benchmarks, and the #BenderRule.
🎥:
They discuss 4 of Emily's papers ⬇️
1/5
@l2k and @emilymbender dive into the problems with bigger and bigger language models, the difference between form and meaning, the limits of benchmarks, and the #BenderRule.
🎥:
They discuss 4 of Emily's papers ⬇️
1/5
"On the Dangers of Stochastic Parrots: Can Language Models Be Too Big? 🦜" (Bender, Gebru et al. 2021)
Possible risks associated with bigger and bigger language models, and ways to mitigate those risks.
dl.acm.org/doi/pdf/10.114…
2/5
Possible risks associated with bigger and bigger language models, and ways to mitigate those risks.
dl.acm.org/doi/pdf/10.114…
2/5
"Climbing towards NLU: On Meaning, Form, and Understanding in the Age of Data" (Bender & Koller, 2020)
Why systems trained only on form (like language models) have no a priori way to learn meaning.
aclanthology.org/2020.acl-main.…
3/5
Why systems trained only on form (like language models) have no a priori way to learn meaning.
aclanthology.org/2020.acl-main.…
3/5
"AI and the Everything in the Whole Wide World Benchmark" (Raji et al., 2020)
How benchmarks have been taken out of context and are limited on their own
ml-retrospectives.github.io/neurips2020/ca…
4/5
How benchmarks have been taken out of context and are limited on their own
ml-retrospectives.github.io/neurips2020/ca…
4/5
"The #BenderRule: On Naming the Languages We Study and Why It Matters" (Bender 2019)
Why you should always state the name of the language(s) you're studying, even if it's "just" English.
thegradient.pub/the-benderrule…
5/5
Why you should always state the name of the language(s) you're studying, even if it's "just" English.
thegradient.pub/the-benderrule…
5/5
• • •
Missing some Tweet in this thread? You can try to
force a refresh
