Wen-Ding Li
@xu3kev
Followers
3K
Following
96K
Media
27
Statuses
525
LLM for code and reasoning. PhD student at Cornell. Previously Student Researcher at @google. Previously intern at @theteamatx.
Joined March 2013
#ICLR25 Presenting our work on leveraging LLM code generation methods for ARC-AGI today at 3pm, Hall 3 + 2B, Poster 232! Come check out our paper, winner of the 2024 ARC-AGI paper competition!
New ARC-AGI paper . @arcprize w/ fantastic collaborators @xu3kev @HuLillian39250 @ZennaTavares @evanthebouncy @BasisOrg .For few-shot learning: better to construct a symbolic hypothesis/program, or have a neural net do it all, ala in-context learning?.
1
4
20
Final 48 hours! โฐ Got interesting results about coding agents? Come share your work with the community and engage with our incredible speaker lineup spanning academia to industry!.
๐จ FINAL CALL: Only 2 days left to submit to the ๐ป๐๐๐ก ๐๐๐๐ฃ๐๐๐๐ ๐๐ ๐ฃ โ๐ ๐๐ ๐๐ ๐ฅ๐๐ ๐ธ๐๐๐๐ฅ๐๐ ๐ผ๐ฃ๐ (DL4C) workshop at NeurIPS2025 !. ๐Deadline: Aug 27th, 11:59PM UTC-12. Amazing speaker lineup including experts from CMU, UC Berkeley, Replit, poolside,
0
0
3
Please join DL4C workshop as a reviewer if you have experience on deep learning for code and especially coding agents!.
๐ Join the Program Committee for Deep Learning for Code (DL4C) Workshop at NeurIPS2025!. ๐ฌ We would be honored to have you join us as a reviewer!.โฐ Review Period: Aug 29 - Sep 12.๐ Efficient process with concise review forms to respect your time โก. Ready to contribute? Fill
0
0
7
RT @minimario1729: new deepseek release almost on-par with o3 (high) on livecodebench ๐ฒ๐
0
73
0
RT @DL4Code: What a fantastic 3rd #DL4C at @iclr_conf! Our deepest thanks to all speakers, presenters, attendees, and every community membeโฆ.
0
3
0
Thanks to everyone who joined us at the 3rd @DL4Code workshop at #ICLR2025! Deeply appreciate all the speakers, presenters, participants, and co-organizers who made it such a success. It was great seeing everyone in person and so much fun chatting about your amazing work!.
Incredibly proud of our 3rd @DL4Code workshop at #ICLR2025 in Singapore bringing the deep learning for code community together and connecting in-person! Huge thanks to our speakers, presenters, and everyone who made #DL4C a massive success. Here's a photo of the dedicated
0
0
6
RT @DL4Code: Starting our afternoon session with @minimario1729 talking about challenges in AI for SE like library and API version updates!โฆ.
0
3
0
0
3
0
Come join us for the Deep Learning for Code workshop today at Room 218!.Exciting line up of talks โ from AI agents to the future of software engineering!.
The workshop is happening now featuring our first invited speaker @dan_fried from @CarnegieMellon / @metaai talking about "Inducing Functions to Improve LLM Agents".
0
1
12
RT @DL4Code: The workshop is happening now featuring our first invited speaker @dan_fried from @CarnegieMellon / @metaai talking about "Inโฆ.
0
2
0
Just arrived in Singapore ๐ธ๐ฌ for #ICLR25 to present several papers on LLMs for code generation! Excited to reconnect with old friends and meet new onesโhappy to chat about all things LLMs, codegen, and reasoning. Also, come check out our LiveCodeBench poster this afternoon!.
Heading to ICLR to present LiveCodeBench at the Friday afternoon poster session and Challenges and Paths Towards AI For SWE at various ICLR workshops. Looking forward to connecting with folks working on Code and LLM Evaluation! (DMs open). PS: Check out our recent v6 update ๐
0
1
16
RT @StringChaos: Heading to ICLR to present LiveCodeBench at the Friday afternoon poster session and Challenges and Paths Towards AI For SWโฆ.
0
10
0
RT @EchoShao8899: SWE jobs have long been a battleground for autonomous AI. I was a bit surprised when @minimario1729 reached out, hoping tโฆ.
0
4
0
๐ค Has AI really solved software engineering?. Our roadmap paper explores where AI excelsโand where it still struggles. Challenges in continuous learning, adapting to new codebases, long-horizon planning, human-AI collaboration, etc., suggest thereโs more work ahead. Check it.
๐ข Excited to share our new paper: Challenges and Paths Towards AI for SWE. We discuss:.๐ ๏ธ 6 sub-tasks needed for SWE.๐ค 9 challenges of today's AI in SWE.๐ฎ 9 future directions to address the challenges. w/ collaborators from MIT, Berkeley, Cornell, Stanford, and UPenn. โฌ๏ธ (1/n)
1
1
12
RT @StringChaos: Check out evaluations for QwQ-Max-Preview on LiveCodeBench where it performs on par with o1-medium๐!! .
0
31
0
We extended the deadline to Wednesday. Thereโs still time to submit your deep learning for code work!.
We are extending the submission deadline by 48 hours (now Feb 12th EoD AoE)! Hope this gives you a bit more time to polish your paper :).@iclr_conf #DL4C #ICLR2025.
0
0
3
RT @DL4Code: โ๏ธTwo days left until the deadline! Looking forward to your submissions!.@iclr_conf #DL4C #ICLR2025.
0
25
0