Thank you for the feedback that enhances the survey's comprehensiveness.

Tomorrow, I'll be presenting part of the tutorial at ICASSP 2024. The topics include speech LM and ways to prompt it and parameter-efficient learning to enable text LLM to understand speech. See you there!

Excited to share Speech Trident survey🔱 Check it out on GitHub
github.com/ga642381/speec…

I will be giving a talk as part of the IEEE ICASSP 2024 tutorial with Dr. Huck Yang Huck Yang, Dr. Pin-Yu Chen Pin-Yu Chen, Prof. Hung-yi Lee Hung-yi Lee (李宏毅), and Cheng-Han Chiang Cheng Han Chiang (姜成翰).

Can LLMs produce faithful explanations for fact-checks? We found that the answer is 🚨NO! 🚨. Excited to present the amazing work done by my talented mentee Kyungha Kim Sangyun and wonderful collaborators/advisor Hou Pong (Ken) Chan Manling Li @elgreco_winter

(1/n)

ACL announcement:
'The ACL Executive Committee has voted to significantly change ACL's approach to protecting anonymous peer review. The change is effective immediately. (1/4) #NLPRoc

Check out our latest work on a comprehensiveness study of trust and safety in LLMs. Comments and feedback are welcome!

Had a great time sharing our work at Generation Evaluation & Metrics Workshop in EMNLP 2024
I will also present this paper on Sunday morning’s Finding virtual poster session. Come and say hi 👋

Went to check out the venue of #EMNLP2023 today 🎰🎢
Interesting place😂
Exited to meet old and new friends 😁

The most ridiculous and funny way to delay a flight: it was “fueled too much to take off” and we stay still to “burn the fuel” for a good 45 mins…
What a good way to start the #EMNLP trip. 🤣
See everyone very soon in Singapore!

The video recording of my keynote talk IJCAIconf #IJCAI2023 on 'An Eye for AI: Towards Scientific Approaches for Evaluating and Improving Robustness and Safety of Foundation Models' is out. Believe it or not, it resonates with the recent executive order😉
ijcai-23.org/video/?vid=390…

(1/n) [A🧵on AI-text detector] With the increasing capability of large language models to generate human-like content, detecting AI-generated text becomes critical for the responsible use of LLMs, especially for addressing AI misuse (e.g. fake news & AI plagiarism) IBM Research

Excited to share our first video-language foundation model by Twelve Labs!
We reinforce previous findings (e.g. Flamingo) that the alignment between vision and language models is the key; we push it to a new limit with video-centric architecture and massive video-text corpora.