Michael White
@mwhite14850
Followers
385
Following
498
Media
6
Statuses
188
Professor, Dept Linguistics, OSU; NLG, dialogue, semantics, etc (incl Shoggoth wrangling); notorious teller of dad jokes
Joined June 2014
We have updated our Arxiv with the #ACL2024NLP camera-ready version! Besides tree search, our work shares a few insights with recent work on analyzing self-correction (e.g., https://t.co/dMhpiH9tsq by @Dongwei__Jiang, https://t.co/r9dW6YapWR by @RyoKamoi): (1/4)
arxiv.org
Can LLMs consistently improve their previous outputs for better results? For this to be true, LLMs would need to be better at discriminating among previously-generated alternatives, than...
LLM planning methods, such as tree search, are critical for complex problem solving, but their practical utility can depend on the discriminator used with them. Check out our new findings: https://t.co/u1LnYtcb4u (1/6)
2
11
28
New #ACL2024 paper that essentially argues that tree search is not the panacea for LLM planning. For tree search to be useful, it needs a strong (90%) discriminator to rank the hypotheses in the search frontier. However, for many problems discrimination is no easier than
Our paper is accepted at #ACL2024NLP main conference! Check out our code and models:
2
10
40
LLM planning methods, such as tree search, are critical for complex problem solving, but their practical utility can depend on the discriminator used with them. Check out our new findings: https://t.co/u1LnYtcb4u (1/6)
6
44
187
Q* from OpenAI and tree-of-thought reasoning triggered a lot of enthusiasm on augmenting LLMs' reasoning/planning capabilities with search. But is search really the panacea for LLMs? Answer from our new study @osunlp: Not quite yet. TLDR: For advanced planning methods like tree
LLM planning methods, such as tree search, are critical for complex problem solving, but their practical utility can depend on the discriminator used with them. Check out our new findings: https://t.co/u1LnYtcb4u (1/6)
5
85
344
I've developed a handy way of streamlining our @emnlpmeeting reviewer criticisms on social media
5
18
81
What if a language model of code is not pre-trained on the programming language you are working with? Our #ACL2023 paper ( https://t.co/GmX8gjXw7L) shows that you can wrap it with something the LM knows! @osunlp (1/4)
arxiv.org
Despite recent progress in text-to-SQL parsing, current semantic parsers are still not accurate enough for practical use. In this paper, we investigate how to build automatic text-to-SQL error...
1
14
18
My lab is looking for a postdoc to work on some exciting LLM and Multimodal work. Please fill this simple form if you are interested: https://t.co/dG0YM6OF8A Colleagues and friends, please retweet to spread the words.
17
111
337
Congrats, Cory! What great news for Stanford.
I can't believe I'm saying this, but here goes: I'm going to be a professor! At Stanford Linguistics (starting fall 2024)!
0
0
2
I'm thrilled that EMNLP 2023 has a track for Human-Centered NLP. This is how I've been self-identifying of late, which has led to a number of confusions and conversations. I'm not an authority, but thought it might be useful to share this experience. I'd love your thoughts! 1/7
7
34
222
GPT-4 and its ilk are awesome for rapid prototyping and one-offs, but at the end of the day, enterprises will deploy far smaller distilled models in production. Here's my contrarian take -
28
148
719
Congrats to @_dmh , @chenghua_lin , and Javier GonzĂĄlez Corbelle (student), who have been elected to the SIGGEN board. They will join @evanmiltenburg and @raquelHervas . @EhudReiter , @sepajma , and @MirunaClinciu have completed their terms and are stepping down.
0
3
18
Congrats to our students for winning the 3rd place in the first @AmazonScience Alexa Prize TaskBot challenge! @ShijieChen98 @RonZiruChen Xiang Deng, Ashley Lewis @LingboMo Sam Stevens @zhenwang9102 @xiangyue96 Tianshu Zhang Advisor: @hhsun1. also supported @ysu_nlp @mwhite14850
@osunlp TacoBot team is excited to win the third place in the 1st Alexa Prize TaskBot Challenge. Many congratulations to the top two performers (GRILLBot from University of Glasgow and Twiz from FCT NOVA) and the other two finalists (QuakerBot from UPenn and GauchoBot from UCSB).
1
1
13
Check out this Voices of Excellence podcast where I was interviewed about my research â with masterful editing, I may have even made some sense! Great to talk about tons of collaborative work, even if my collaborators were too numerous to name! @osunlp
0
1
5
Come see our in-person poster about a new dataset for transparent, interactive semantic parsing this afternoon! #acl2022nlp #acl2022 #ACLinDublin #NLProc
Co-authors (Ash Lewis and Mike White @mwhite14850 ) are attending the conference in person. Our poster sessions will be: May 24, 15:15 (on-site) & May 25 7:30 (virtual), Dublin time. Please feel free to reach out via email at any time. đđ
0
0
8
Ah, my colleague found it in the instructions for authorsâitâs A0
0
0
2
um, is anyone else having trouble finding out what size posters should be for ACL in Dublin? #acl2022nlp @aclmeeting
1
0
1
Visiting Assistant Professor position in computational semantics in my department! Deadline May 15 https://t.co/SgWu9zNc4R
0
2
6
@emilymbender @stevenbjohnson âResist the urge to be impressedâ seems like the single most important piece of advice we can give to the public and teach our students. It acknowledges the fascination of LLMs and still encourages critical thinking.
0
8
64