Related Work on Improving the Translation Agent

[yanzhangnlp]

Hello Andrew, thank you for your excellent contributions to the translation agent! I am reaching out to highlight our recent work, "TEaR: Improving LLM-based Machine Translation with Systematic Self-Refinement", which was released this February. Our research utilizes a self-refinement strategy along with a state-of-the-art GPT evaluation method to enhance translation agents, which may complement your work. For those interested, you can access our paper here: https://arxiv.org/abs/2402.16379 and the source code here: https://github.com/fzp0424/self_correct_mt.

[xiaoqiangcs]

From the paper and code provided by the author above, the approaches of these two projects are indeed very similar.

[AndrewHYC]

The content of this project and this paper seems to be very similar. The paper is well developed and worthy of citation.

[fzp0424]

Hi, Andrew and all collaborators on this project! It's great to see the implementation of such a fantastic concept bringing Agent into the Machine Translation field. In February, we also released an LLM-based Self-Refinement framework that aligns closely with your translation-agent project. We believe it may complement your work. For those interested, you can access our paper here: https://arxiv.org/abs/2402.16379 and the source code here: https://github.com/fzp0424/self_correct_mt.

[siddhantx0]

Andrew, i have a research paper on document similarity -- was wondering what you'd think about it. it's cryptocurrency document similarity--been working on it for 1+ year.

…

On Mon, Jun 24, 2024 at 10:02 PM fzppp ***@***.***> wrote: Hi, Andrew and all collaborators on this project! It's great to see the implementation of such a fantastic concept bringing Agent into the Machine Translation field. In February, we also released an LLM-based Self-Refinement framework that aligns closely with your translation-agent project. We believe it may complement your work. For those interested, you can access our paper here: https://arxiv.org/abs/2402.16379 and the source code here: https://github.com/fzp0424/self_correct_mt. — Reply to this email directly, view it on GitHub <#27 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AY7DCNEHRRSIEXJVW7PKNXLZJDMUFAVCNFSM6AAAAABJ2PYUL6VHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMZDCOBXHA3DENZTGQ> . You are receiving this because you are subscribed to this thread.Message ID: ***@***.***>

[fzp0424]

@andrewyng @j-dominguez9 @methanet @johnsanterre I hope you find our work of interest. Thank you.