AI for code; Functional programmer; {Natural | Programming} language enthusiast; AI/ML/PL. Tweets are my own
Joined August 2014
- Tweets160
- Following741
- Followers359
- Likes1.3K
Tongfei Chen retweeted
📢 Exciting News in NLP Metrics! 📊
🔍 There are a lot of different metrics for structured prediction tasks for NLP. Or are there? Our EMNLP 2023 paper suggests they might not be so different at all! Plus, we've got a great new Python library to help you make your own metrics! 🚀
Wrote a little package that visualizes directed acyclic graphs (DAGs) in the terminal.
PyPI: pypi.org/project/py-dagviz/ GitHub: github.com/ctongfei/py-dagvi…
9
Tongfei Chen retweeted
I wrote up some ideas in github.com/namedtensor/notat…, some co-developed w/ @kitsing20.
Tongfei Chen retweeted
Are dialogue and event understandings similar? We (w/ @ctongfei @ben_vandurme at @jhuclsp) draw connections between two fields and propose to solve the argument linking problem using a simple architecture.
Our talk: slideslive.com/38939698 (11/20 11:15 EST)
#EMNLP2020 #CODI
Two papers accepted to #acl2020nlp: 1. Uncertain natural language inference (w/ @ZhengpingJiang @azpoliak @KeisukeS_ @ben_vandurme) 2. Hierarchical entity typing via multi-level learning to rank (w/ @YunmoChen @ben_vandurme). Preprints and code will follow :-)
We release Espresso: A fast end-to-end neural automatic speech recognition (ASR) toolkit, based on #fairseq & @PyTorch, with Yiming Wang, me, Hainan Xu, @_sding, @VioletNPeng et al. Code: github.com/freewym/espresso; Paper: arxiv.org/abs/1909.08723
Stoked that our paper received best paper award at #NLP4ConvAI 😄 See you in Florence! (adding @ChetanNNaik)
Our paper "Improving Long Distance Slot Carryover in Spoken Dialogue Systems" with me, Chetan Naik, Hua He, @Pushpendre89 and Lambert Mathias is online at arxiv.org/abs/1906.01149. #NLP4ConvAI @ACL2019_Italy
Tongfei Chen retweeted
Tongfei Chen retweeted
Learning to Rank for Plausible Plausibility
Zhongyang Li, Tongfei Chen (@ctongfei) Benjamin Van Durme (@ben_vandurme) #ACL2019
Tongfei Chen retweeted
Check out our work on using query rewriting for dialogue state tracking at #naacl2019
Paper: arxiv.org/abs/1903.05164
Dataset: github.com/alexa/alexa-datas…
Our paper "Improving Long Distance Slot Carryover in Spoken Dialogue Systems" with me, Chetan Naik, Hua He, @Pushpendre89 and Lambert Mathias is online at arxiv.org/abs/1906.01149. #NLP4ConvAI @ACL2019_Italy
Tongfei Chen retweeted
Amazing paper from @Pushpendre89 and @ctongfei that re-frames DST as query reformulation arxiv.org/abs/1903.05164. Combined with the key insight from @jasonwu0731 in arxiv.org/abs/1905.08743 means we have scalable, ontology free intent tracking for dialogue
Tongfei Chen retweeted
Edward Hu (recently graduated @jhuclsp ugrad) is presenting
Improved Lexically Constrained Decoding for Translation and Monolingual Rewriting
by J. Edward Hu @HudaKhay Ryan Culkin, Patrick Xia @ctongfei @mjpost & Benjamin Van Durme
Poster at 3 Monday
aclweb.org/anthology/N19-109…
Tongfei Chen retweeted
If your paper is arXived but you don't tweet it…
Proved a No Free Lunch theorem 🚫🍱 for clustering tasks, using our recently advocated metric. No best algo! With @ctongfei and @_t_sne, who survived my quibbles about \mathcal vs \mathbb \vs \mathfrak!
arxiv.org/abs/1903.10092
Just realized the composition of finite state transducers (FSTs) is just a monadic operation -- weighted FSTs compose because any semiring gives rise to a monad on maps: Semiring[R] => Monad[[A] => Map[A, R]]; and WFSTs form their own category.
Tongfei Chen retweeted
I'm at #AAAI19 presenting our work on storytelling. We proposed a hierarchically generation strategy plan-and-write arxiv.org/abs/1811.05701
Please play with our demo cwc-story.isi.edu Guaranteed fun with the interactive mode. :)
Friends at Honolulu: let’s catch up! #nlproc