Pinned Tweet
Evaluation of text classification systems becomes complex when we think beyond effectiveness and consider aspects such as explainability, fairness, and domain adaptability. Submit your evaluation approaches to our #NEATCLasS (neatclass-workshop.github.io) workshop by April 17.
Sheikh Muhammad Sarwar retweeted
Here is a list of contributions from the @manningcics Center for Intelligent Information Retrieval (CIIR) to @SIGIRConf 2024!
Multiple exciting papers on RAG, Generative Retrieval, Proactive Conversations, and LLM Personalization.
Pre-prints coming up, stay tuned... #SIGIR2024
Sheikh Muhammad Sarwar retweeted
LLMs have problems that are difficult to identify and explain. To understand them better, we need to study the relationship between LLMs and their pre-training data, and with the OLMo release, @allen_ai helps make this possible!
sciencenews.org/article/gene…
OLMo is here! And it’s 100% open.
It’s a state-of-the-art LLM and we are releasing it with all pre-training data and code. Let’s get to work on understanding the science behind LLMs. Learn more about the framework and how to access it here:
blog.allenai.org/olmo-open-l…
Sheikh Muhammad Sarwar retweeted
We are building a multi-lingual collaborative tool for international students in the US and would appreciate your insights on the challenges of multi-lingual collaboration. Please participate in our survey!
Retweets appreciated.
Hey everyone!
I've launched a brief and enjoyable survey on the fascinating world of multilingual collaboration. Your insights are invaluable in shaping this research! Click the link or scan the QRcode to participate. Much appreciated! 🤝🌐
#LSU #survey
forms.office.com/r/N4yAMfD2c…
Sheikh Muhammad Sarwar retweeted
Reminder - for the terrific interdisciplinary Text as Data conference, abstract submissions coming up - due Aug 4!
tada2023.org/
It's a great, small, non-archival conference to discuss emerging work with folks across social sciences, humanities, and computer science.
Sheikh Muhammad Sarwar retweeted
Super excited to share our paper “Detecting Harmful Content on Online Platforms: What Platforms Need vs. Where Research Efforts Go” was accepted to ACM Computing Surveys: dl.acm.org/doi/10.1145/36033…!
Sheikh Muhammad Sarwar retweeted
Work done with my inspiring colleagues: @preslav_nakov @mhardalov @zzz2aaa Vibha Nayak, Kyle Dent, Ameya Bhatawdekar, Yoan Dinkov @didizlatkova @gbouchar and @IAugenstein
#NLProc #PolComm #onlinesafety #contentmoderation
Sheikh Muhammad Sarwar retweeted
Really happy of how this edition of #NEATCLasS went!
Thanks to the organisers @bjoernross @zzz2aaa @RNavigli, the presenters, and the participants for the great discussion!
I still can't believe there were so many attendees and spontaneous questions in such a sunny afternoon 🏝️ – at Cyprus
Sheikh Muhammad Sarwar retweeted
So... late last night, a friend called me in tears
She just lost her copywriting job at a 1000+ person company
The culprit? ChatGPT.
This is a wakeup call to lawyers, accountants, consultants etc (read this to stay ahead):
Sheikh Muhammad Sarwar retweeted
Does your work involve classifying user-generated text (sentiment, hateful language, misinformation...)? How can we tell if our models are any good? Submit your research on this topic to the #NEATCLasS workshop by 17 April! neatclass-workshop.github.io… #ICWSM
Sheikh Muhammad Sarwar retweeted
A good display of how empirical and setting-dependent deep learning can still be, and what driving up performance looks like. In any setting it's not so much "here's how you can improve" but "here's the 10 things you should try". And why high experimental throughput is necessary.
Generation and Retrieval ar very different problems. One can use a retrieval pipeline to solve machine translation. Would anyone apply that?
In order to understand why ChatGPT can't replace Google Search, it's useful to understand the early days of web search and the role that PageRank played. 1/n
237
Sheikh Muhammad Sarwar retweeted
In the LLM-science discussion, I see a common misconception that science is a thing you do and that writing about it is separate and can be automated. I’ve written over 300 scientific papers and can assure you that science writing can’t be separated from science doing. Why? 1/18
Sheikh Muhammad Sarwar retweeted
I am looking for PhD students this year!
If you are interested in working on misinformation/polarization particularly in the global south, apply to work with me!
Rutgers iSchool is well ranked & has a great set of faculty working on similar problems.
DM/email me with questions
Sheikh Muhammad Sarwar retweeted
📢Looking for a PhD fellowship in #XAI #ML #NLProc #IR? We have a position available for a start in Spring 2023. The successful candidate will be supervised by my colleague Christina Lioma and me @DIKU_Institut
Read more & apply: candidate.hr-manager.net/App… #PhDposition
Sheikh Muhammad Sarwar retweeted
The future of language models is retrieval augmented. Atlas has better few-shot performance than PaLM, Gopher and GPT3 at a fraction of their parameter counts on tasks such as QA and fact checking that require background knowledge.
🚨We’ve been working on better retrieval-augmented models & thrilled to present Atlas, led by @gizacard @EXGRV & myself🚨
Atlas is a end2end pretrained "RAG"-like model, beats models 50x its size on fewshot QA, sets numerous SotA on knowledge-intensive NLP arxiv.org/abs/2208.03299
Sheikh Muhammad Sarwar retweeted
In the spirit of high-fiving papers I'm enjoying reading, here is a 🧵about a paper: "Two Contrasting Data Annotation Paradigms for Subjective NLP Tasks" (aclanthology.org/2022.naacl-…) by @paul_rottger @bertievidgen @dirk_hovy and Janet Pierrehumbert (can't find a Twitter!)


























