Mofe Adeyemi retweeted
3 papers accepted at #EMNLP2024
1) YORÙLECT (Yoruba dialects) led by @orevaahia @aremuadeolajr arxiv.org/abs/2406.19564
2) MT Hallucination Detection led by @kenza_ily @gongas_laura arxiv.org/abs/2407.16470
3) MINERS led by @gentaiscool @ruochenz_ arxiv.org/abs/2406.07424
Mofe Adeyemi retweeted
This work led by @AdeyemiMofe was accepted in #ACL2024! 🎉🎉
Excited to share that our paper on “Zero-Shot Cross-Lingual Reranking with Large Language Models for Low-Resource Languages” has been accepted to #ACL2024 🎉✨Big thank you to @theYorubayesian @rpradeep42 @lintool 🎉
Mofe Adeyemi retweeted
We have two papers accepted to #SIGIR2024!
1) On Backbone and Training Regimes for Dense Retrieval in African Languages.
2) CIRAL: A Test Collection for CLIR Evaluations in African Languages led by @AdeyemiMofe
🚀🚀 We find that, though LLMs perform better in English reranking, there is still competitive performance in monolingual (African queries -> African passages) vs. cross-lingual (English queries -> African passages) scenarios, more so in using translations generated by the LLM.
The obvious question: How do the latest prompt-decoder LLMs for listwise reranking perform on low-resource languages? For four African languages (Hausa, Somali, Swahili, and Yoruba), @AdeyemiMofe @theYorubayesian @rpradeep42 provide the answer: arxiv.org/abs/2312.16159
Mofe Adeyemi retweeted
💥💥 Catch us at #FIRE2023 starting tomorrow!
- Track Overview: 10:15am IST, Dec 15th
- Track Session: 11:00am IST, Dec 16th
We wrapped up CIRAL earlier, and we say a big thank you thank to participating teams 💯🎉🎉
Our public leaderboard is available: ciralproject.github.io/#time…
Mofe Adeyemi retweeted
Starting with a meticulous audit of mC4's document sources, we curate a new dataset, Wura, which covers 16 African languages + English, French, Portuguese and Arabic. Wura contains 4.7M documents in African languages - more than 1.5x what mC4 contains!
huggingface.co/datasets/cast…
Mofe Adeyemi retweeted
We've been doing some research on scaling pretraining data and language models for African languages and I'm excited to share our research at EMNLP 2023!
Work done with @AdeyemiMofe @orevaahia @j___y_t @AbrahamOwos @davlanade and @lintool
Here's a primer:
1 /
Mofe Adeyemi retweeted
That's a wrap! The Waterloo (@UWCheritonCS) team had fun attending the ACL 2023 Conference in Toronto, Canada! #ACL2023NLP 🇨🇦
We would like to congratulate @ralph_tang @likicode @ZhiyingJ @lintool et al. for winning the Best Paper Award at ACL 2023!!🏆
Next stop is SIGIR 2023. – at Waterloo, Ontario
ALT The University of Waterloo Team in front of the ACL 2023 banner.
Mofe Adeyemi retweeted
📢 New Update!
The training set for Hausa and Yoruba are now released 🎉🎉 and are accessible here: huggingface.co/datasets/CIRA…
Looking forward to the retrieval systems that are built 💯⚒️ Guidelines for participation can be found here ciralproject.github.io/#part…
Mofe Adeyemi retweeted
Have you registered? Join iSTEMLabsAfrica #WaZoBia Q1 2022 Hackathon Day Event on the 12th March, 2022. Come, explore, and prototype solutions to existing problems. Get the opportunity to engage with speakers.
Mofe Adeyemi retweeted
It has been hard not to be emotional when thinking about this stranger and his wife.
71 years ago, they made vows to each other, among which I assume was the promise to love one another in sickness and in health.
As this man's memory fades, he holds onto that promise. 3/
