Opens profile photo
Follow
Click to Follow ruizhang_nlp
Rui Zhang
@ruizhang_nlp
Researcher in #NLProc | Assistant Professor
@PennStateEECS
State College, PAryanzhumich.github.ioJoined April 2013
890 Following
1,573 Followers

Rui Zhang’s Tweets

Pinned Tweet
Rui Zhang
@ruizhang_nlp
Can we have different reference summaries for one document🧐? Yes🤩! In this new benchmark 👉MACSum✍️ led by
@YusenZhangNLP
, we annotated multiple reference summaries by controlling a mix of attributes together (Length, Extractiveness, Specificity, Topic, Speaker)!
Quote Tweet
Yusen Zhang
@YusenZhangNLP
A new dataset for Controllable Summarization🙌 arxiv.org/pdf/2211.05041 Most research crafts pseudo datasets and focuses on controlling single attributes individually. We propose MACSum, the first human-annotated summarization dataset for controlling mixed attributes. [1/7]
Show this thread
Image
1
9
Rui Zhang Retweeted
Sergey Levine
@svlevine
I wrote a short article about some possibly untapped potential in the combination of RL and language models: sergeylevine.substack.com/p/offline-rl-a What if the full power of LMs will come not from their ability to be *generative* per se, but their ability to predict the behavior of people?
sergeylevine.substack.com
Offline RL and Large Language Models
What if the most significant capabilities of large language models have yet to be unlocked?
4
241
Show this thread
Rui Zhang Retweeted
Uri Alon
@urialon1
📢 New Paper: Program-aided Language models Prompting methods such as chain-of-thought (
@_jasonwei
) employ LLM for decomposing the problem into steps *and* solving each step. Instead, PaL decomposes the problem into *programmatic* steps and solves using a Python interpreter. 1/4
A diagram illustrating PAL: Given a mathematical reasoning question, "chain-of-thought" (left) generates intermediate reasoning steps of free-form text. In contrast, Program-aided Language models (PAL, right) generate intermediate steps and Python code. This shifts the role of running the reasoning steps from the language model to the Python interpreter. The final answer is obtained by running the generated reasoning chain. Chain-of-thought reasoning is highlighted in blue; PAL steps are highlighted in gray and pink; Python interpreter run is highlighted in black and green
13
487
Show this thread
Rui Zhang
@ruizhang_nlp
Have been thinking for a while how to use large language models to improve data-efficiency of RL...
Quote Tweet
Sergey Levine
@svlevine
NLP and offline RL are a perfect fit, enabling large language models to be trained to maximize rewards for tasks such as dialogue and text generation. We describe how ILQL can make this easy in our new paper: sea-snell.github.io/ILQL_site/ Code: github.com/Sea-Snell/Impl Thread ->
Show this thread
Embedded video
0:18
32.9K views
18
Rui Zhang Retweeted
SUKI 2022
@suki_2022
We just added a commitment button, so you can directly (re)submit your ARR paper links with reviews!
Image
Quote Tweet
SUKI 2022
@suki_2022
Image
Paper deadline extended by one week! We welcome archival and non-archival submissions to SUKI 2022 by April 15 through ARR (openreview.net/group?id=aclwe). You can also resubmit your NAACL 2022 papers with the link to ARR reviews. Check our website suki-workshop.github.io!
11
Rui Zhang Retweeted
MIA 2022
@mia_workshop
We are excited to announce that we have the Best Paper Award with a prize 💰 Please submit to MIA 2022 (#NAACL2022) by April 15! openreview.net/group?id=aclwe We'll also accept ARR-reviewed papers and previously submitted non-archival papers till Apr 29. [1/2]
Image
Eunsol Choi and 7 others
Quote Tweet
MIA 2022
@mia_workshop
Two weeks until the paper submission deadline of MIA 2022, collocated with #NAACL2022 in beautiful Seattle! 🌎 We'll cover diverse multilingual NLP topics, with a focus on how we can improve information access for everyone! Submit here 👉 by April 15: openreview.net/group?id=aclwe
Show this thread
1
4
Show this thread
Rui Zhang Retweeted
Tao Yu
@taoyds
UnifiedSKG (github.com/hkunlp/unified) is one of the shared tasks at SUKI! We provide strong but simple unified sota code and models for 21 tasks that involve structured knowledge. Also, there is another interesting shared task FinQA on financial data! Participations welcome!👇
Quote Tweet
SUKI 2022
@suki_2022
Hello World! Structured and Unstructured Knowledge Integration (SUKI) workshop at #NAACL2022 is welcoming submissions and shared task participations🙌! Papers due by April 8. Two shared tasks due by June 8 with cash awards🥰. Details are available 👉 suki-workshop.github.io
Show this thread
Image
16
Rui Zhang
@ruizhang_nlp
Call for Paper of our Multilingual Information Access (MIA) workshop at #NAACL2022! We welcome both research papers on diverse topics and shared task participations on Cross-lingual Open-Retrieval QA. All the details are available 👉 mia-workshop.github.io.
Quote Tweet
MIA 2022
@mia_workshop
Excited to announce Workshop on Multilingual Information Access (MIA)& our X-lingual QA shared task 🌎 at #NAACL2022! We encourage submissions on diverse aspects of cross-lingual knowledge-intensive NLP. mia-workshop.github.io/cfp.html Submit 👉 (by Apr 8): openreview.net/group?id=aclwe [1/3]
Show this thread
Image
7