Similar User
@Roprajo
@AnetteMFrank
@HD_NLP
@dustin_wright37
@zhangzhuosheng
@CisLmu
@barbara_plank
@PSH_Lewis
@huyushi98
@dnnslmr
@DrogoKhal4
@shrutirij
@billyuchenlin
@qinyuan_ye
@ysu_nlp
Should I use Macro F1 or Accuracy? Why not Kappa? Why do some use this, and others that? What's actually evaluated here? 😵💫 Happy to share the final version of this paper on multi-class classification evaluation: direct.mit.edu/tacl/article/d… #machinelearning #nlproc #ml
There are also papers I came across that just says "F1" without saying whether it is "micro" or "macro" or something else. I stopped taking the actual results by face value and browse through generally to get a broader picture of methods these days.
Interesting and worrying paper on meta-eval of classifiers metrics (direct.mit.edu/tacl/article/d…). Eg, some papers use "macro F1" to refer to arithmetic mean, others use it for harmonic mean, others dont say what they mean by Macro F1
40,000 Downloads! Happy to see that Julius' super-efficient text factuality checker has seen a decent amount of usage in the last months! Free on Huggingface: huggingface.co/juliussteen/De… ACL paper: aclanthology.org/2023.acl-short… @juliusmsteen 👀 #nlproc #machinelearning
Looking for an emergency reviewer for ARR submission 🚨 The paper is on commonsense reasoning. If you have reviewed any *ACL conference and interested, please DM me. The review needs to be submitted in next 24 hours. #NLProc
💡This cool #EMNLP2024 paper on semantic parsing also shows why you should NOT use a heuristic when evaluating a system. #nlproc #machinelearning #datascience #evaluation #statistics aclanthology.org/2024.emnlp-mai…
✨New Paper✨ The Mystery of Compositional Generalization in Graph-based Generative Commonsense Reasoning #EMNLP2024 📜 aclanthology.org/2024.findings-… 🖼️ Poster Session F (Riverfront Hall), Nov 14 @ 10:30
I am looking for emergency reviewers for missing reviews in the "Low-resourced and Less Studied Languages" track at #COLING2025. Please let me know if you can help! #nlproc
Lazy Twitter: Do you know of any corpora or wordlists in Scandinavian languages (Danish, Norwegian, Swedish, Icelandic) with CEFR labels? #education #linguistics #nlp #PhDlife #phd
It doesn't seem much, but I am actually proud of this: I never missed a reviewing DL so far! Probably this won't hold at some point, but still... Remember: If you miss a review DL by a few days it's usually not a big deal, but think of dropping a note to ACs or editors.
PhD then: “We validate our method on a large-scale dataset of hundreds of images.” PhD now: “We validate our method on 10 different modalities, 15 domains, 20 scenarios, 25 tasks, and 200 languages, each with millions of testing examples”.
New blogpost: Funny evaluation quirks and pitfalls! Also included: A few tips on how to achieve a more meaningful and robust evaluation! juriopitz.com/2024/10/17/eva… #nlproc #machinelearning
First time at the #TPDL 2024 conference (tpdl2024.nuk.si) in beautiful #Ljubljana presenting work together with the #Fotostiftung #Graubünden about #multimodal #LLMs for #OCR, #storytelling and #NER (to appear in link.springer.com/book/978303172…). #archives #DigitalHumanities
How to evaluate when the class distribution is imbalanced? I wrote a paper on this problem, maybe it helps to inform the selection process 🙂direct.mit.edu/tacl/article/d…
Which of the following metrics would you prioritize for a highly imbalanced classification problem?
Explainability can also be obtaied with static embeddings 🙂 See my blogpost juriopitz.com/2024/04/04/exp…
Matryoshka embeddings are really cool! With similar objectives in mind, we partition an embedding into features that are interpretable, each binding a different aspect of text. Different low-dimensional subsets can then be selected for different tasks. arxiv.org/abs/2206.07023
United States Trends
- 1. Thanksgiving 2 Mn posts
- 2. Thankful 371 B posts
- 3. #IDEGEN N/A
- 4. $TURKEY 2.856 posts
- 5. #TSTheErasTour 3.162 posts
- 6. $CUTO 6.656 posts
- 7. #Picksgiving 3.769 posts
- 8. #pbdpodcast N/A
- 9. T-Pain 4.552 posts
- 10. #Gratitude 12,8 B posts
- 11. Optimus 31,4 B posts
- 12. Al Roker 1.192 posts
- 13. Jimmy Fallon 2.097 posts
- 14. Gobble Gobble 21,3 B posts
- 15. Kylie Minogue 1.320 posts
- 16. The Fed 54,8 B posts
- 17. Go Lions 2.882 posts
- 18. Hoda 1.571 posts
- 19. Feliz Día de Acción de Gracias 4.386 posts
- 20. Dear Abby N/A
Who to follow
-
Pratik Joshi
@Roprajo -
Anette Frank
@AnetteMFrank -
Heidelberg University NLP Group
@HD_NLP -
Dustin Wright
@dustin_wright37 -
Zhuosheng Zhang
@zhangzhuosheng -
CIS, LMU Munich
@CisLmu -
Barbara Plank
@barbara_plank -
Patrick Lewis
@PSH_Lewis -
Yushi Hu
@huyushi98 -
Dennis Ulmer 🦋
@dnnslmr -
Kai Zhang
@DrogoKhal4 -
Shruti Rijhwani
@shrutirij -
Bill Yuchen Lin 🤖
@billyuchenlin -
Qinyuan Ye
@qinyuan_ye -
Yu Su ✈️ #NeurIPS2024
@ysu_nlp
Something went wrong.
Something went wrong.