Semih Yagcioglu

Director of AI at @Apziva

[Semih Yagcioglu @semihyagcioglu]

I am excited to share that our CompAct dataset for understanding sequential compositional generalization in multimodal models is now available for public access👇

CompAct dataset consists of sequences of actions with audio, image and text modalities describing each action in the sequence and specifically curated for understanding compositional generalization problem.

👉 https://t.co/I7D1g3OWxP

[Semih Yagcioglu @semihyagcioglu]

1. 🧵🎉 Excited to share that our paper "Sequential Compositional Generalization in Multimodal Models" is accepted as a long paper at #NAACL2024! 🌟 We'll be presenting our findings in Mexico City this June (@naaclmeeting). Dive into the full details here 👇

- Paper: https://t.co/v6K4nsfOWM
- Site: https://t.co/I7D1g3OWxP
- Code & Dataset: Coming soon 🙌

[Semih Yagcioglu @semihyagcioglu]

I am at #naacl2024 in Mexico City this week and will be presenting our work on “Sequential Compositional Generalization in Multimodal Models” tomorrow.

If you’re interested in #multimodality and compositional generalization looking forward to meeting you tomorrow to discuss our work and our findings.

The session will be in Don Alberto 2 at 9:36 am. #nlpproc @naaclmeeting

[Semih Yagcioglu @semihyagcioglu]

1. 🧵🎉 Excited to share that our paper "Sequential Compositional Generalization in Multimodal Models" is accepted as a long paper at #NAACL2024! 🌟 We'll be presenting our findings in Mexico City this June (@naaclmeeting). Dive into the full details here 👇

- Paper: https://t.co/v6K4nsfOWM
- Site: https://t.co/I7D1g3OWxP
- Code & Dataset: Coming soon 🙌

[Desmond Elliott @delliott]

We will present two papers at #NAACL2024 this week. I'm looking forward to meeting people interested in multimodality!

CompAct by @semihyagcioglu+ on Wed at 09:36 in Don Alberto 2 https://t.co/ko55RasBbx

PAELLA by @RitaRamos1+ on Wed at 11:00 in Poster 7 https://t.co/9szB463TB2

[Semih Yagcioglu @semihyagcioglu]

👉 For further information and updates please check our project website: https://t.co/LPaWmkl9Tf

[Semih Yagcioglu @semihyagcioglu]

1. 🧵🎉 Excited to share that our paper "Sequential Compositional Generalization in Multimodal Models" is accepted as a long paper at #NAACL2024! 🌟 We'll be presenting our findings in Mexico City this June (@naaclmeeting). Dive into the full details here 👇

- Paper: https://t.co/v6K4nsfOWM
- Site: https://t.co/I7D1g3OWxP
- Code & Dataset: Coming soon 🙌

[Semih Yagcioglu @semihyagcioglu]

8. 🎉 CompAct is the output of a huge joint collaborative effort spanning 3 research groups. Kudos to all! @ospanbatyr, @erkuterdem, @aykuterdemml, @delliott, @denizyuret

[Semih Yagcioglu @semihyagcioglu]

Excited to share that our paper where we explore training dynamics for compositional generalization has been accepted to EMNLP Findings 2023!#NLProc #NLP #EMNLP2023

[Semih Yagcioglu @semihyagcioglu]

Being present is a super power.

If you can really be there you sort of become invincible.

[Semih Yagcioglu @semihyagcioglu]

Office for the day ☀️

[Semih Yagcioglu @semihyagcioglu]

@ospanbatyr @erkuterdem @aykuterdemml Çok teşekkürler Batur!

[J. AI Research-JAIR @JAIR_Editor]

New Article: "Neural Natural Language Generation: A Survey on Multilinguality, Multimodality, Controllability and Learning" by Erdem et al., https://t.co/LkvzOSELY1