Haau-Sing Li 李 效丞

PhD candidate @ELLISforEurope @UKPLab @sardine_lab_it. CodeAI. Prev @tiktok_us @NYUDataScience, @RenminUniv.

[Haau-Sing Li 李 效丞 @LHaausing]

1/12 📢New paper alert “DOCE: Finding the Sweet Spot for Execution-Based Code Generation”

The work is a group effort with @psanfernandes, @IGurevych, and @andre_t_martins.

@UKPLab @Lisbon_ELLIS @deep_spin

📰: https://t.co/VRQZNYXsyR

[Yaxin Du @DorothyDDU]

LoopCoder-v2 is out 🚀

Loop Transformers reuse the same block for recurrent hidden-state refinement — letting models “think” more without simply stacking more layers.

We study how many loops are actually worth it in Parallel Loop Transformers.

📜https://t.co/6iKMPelqQ2

[Haau-Sing Li 李 效丞 @LHaausing]

@Xianbao_QIAN Impressive performance when deployed for HTML/SVG services even when just on a 14B model.

[Haau-Sing Li 李 效丞 @LHaausing]

@SinclairWang1 We received very similar comments! We put "software engineering" and "analysis" there for keywords, with justifications already in the paper (intro and more), yet "lacks novelty" still appears for all reviewers. Guess a good rebuttal has to go right? @COLM_conf 🤔

[Haau-Sing Li 李 效丞 @LHaausing]

Also huge shoutout to my amazing supervisors @IGurevych and @andre_t_martins for their support and wisdom.

Wouldn’t be possible without you guys, and more exciting things will come.

[Haau-Sing Li 李 效丞 @LHaausing]

Sharing that I will join @tiktok_us AI Innovation Center @BytedanceTalk for research internship.

Will still work on Code AI advised by the amazing @sivil_taram. Hope there’ll be luck to us bringing the community nice deliverables!

Execution can and will talk.

Happy 2025!

[Andre Martins @andre_t_martins]

2) We have a spotlight poster in the main conference, “Reranking Laws for Language Generation: A Communication-Theoretic Perspective” with @tozefarinhas and @LHaausing (Thursday Dec 12 11:00-14:00).

[Haau-Sing Li 李 效丞 @LHaausing]

@a_stadt Was going to say I can review one but seems it’s going well!☺️

[Haau-Sing Li 李 效丞 @LHaausing]

@apsdehal Do you provide visa sponsorship? Would be nice to work with you as I think I work on things super related to what you do but I need sponsorship.

[Haau-Sing Li 李 效丞 @LHaausing]

@gan_chuang DOCE: Finding the Sweet Spot for Execution-Based Code Generation (the first unified inference framework for code gen, I presume strongly correlated to reasoning :) ) https://t.co/VRQZNYXsyR

[Haau-Sing Li 李 效丞 @LHaausing]

@LucileSaulnier @MistralAI Hi! I applied for the PhD research internship, can I message you for a discussion about that?

[Haau-Sing Li 李 效丞 @LHaausing]

@mialon_gregoire @MekalaDheeraj Hi Grégoire! I'm interested in this position and have submitted my application. We recently have papers published (at NeurIPS 2024 spotlight)/in submission (https://t.co/sV50i9eF5Y, https://t.co/RyDMehKB6T), see if that interests you!

[Haau-Sing Li 李 效丞 @LHaausing]

#NeurIPS2024 spotlight accepted. Congrats all!
@tozefarinhas @andre_t_martins

[Haau-Sing Li 李 效丞 @LHaausing]

@natolambert 🫤

[Haau-Sing Li 李 效丞 @LHaausing]

@natolambert Quite a lot of times I ever tried to look for seemingly nice reward models/synthetic data as well, the only GitHub page I found is the repo of e.g. vLLM

[Haau-Sing Li 李 效丞 @LHaausing]

@KempeLab @AIatMeta I have sent an follow-up email to your NYU email address, hope it reaches you well despite potential bothering~ :D

[Haau-Sing Li 李 效丞 @LHaausing]

@KempeLab @AIatMeta Hi Julia, I work exactly on test-time compute+reasoning and am super interested in this position, can I write you a follow-up email or DM?

[Haau-Sing Li 李 效丞 @LHaausing]

@ssgrn Interested, and super related to what I’m working on! DM sent

[Haau-Sing Li 李 效丞 @LHaausing]

@natolambert During inference specifically

[Haau-Sing Li 李 效丞 @LHaausing]

@natolambert A shameless plug, but yeah we also found execution feedback quite important https://t.co/KRvmH6Mere

[Haau-Sing Li 李 效丞 @LHaausing]

@mrdrozdov We have some more exploration using more up-to-date LLMs, with more methods. A shameless plug but we also got quite some interesting findings https://t.co/KRvmH6Mere