@KnightZhang625

NLP Researcher in Tencent London. CS PhD in University of Strathclyde, Glasgow, . Master degree from University of Sheffield. Make the world better place.

 Glasgow, Scotland
knightzhang625.github.io
Joined April 2015
  • Tweets163
  • Following443
  • Followers138
20Media
#NeurIPS2022 Presenting our work “ELASTIC: Numerical Reasoning with Adaptive Symbolic Compiler” at NeurIPS 2022. Enjoying talk with nice people during poster session. @NeurIPSConf @YashMosh @NeuraSearchLab @StrathCIS @UniStrathclyde – at New Orleans, LA
2
10
Very happy to see our paper “Time is Not a Label” accepted to #EMNLP2026 Main Conference! 🎉 This project started during Waylon’s internship at LIGHTSPEED STUDIOS, Tencent, where I had the opportunity to mentor him. Congrats to Waylon and all collaborators!
3
1
16
357
A Semantic Speed Gate learns which relations are persistent and which evolve over time, allowing obsolete facts to be geometrically shadowed rather than overwritten or deleted. This gives us an append-only temporal memory mechanism that performs strongly.
6
Most agent memory systems treat time as metadata attached to a fact. We introduce RoMem, which instead models time directly as a continuous geometric rotation in representation space.
13
#GDC2026 is almost here! Get a first look at LightSpeed Studios’ Sessions! 🚀 We’re diving deep into AI-powered game tech and global IP design philosophy across 6 sessions. Join us in SF to explore the future of play! More updates soon! #LightSpeedStudios #GameDev
3
8
1,350
Hit 100 citations on Google Scholar! A small step compared to many PhDs, but a meaningful milestone for me—enough to fuel my drive to keep up with the stars.
2
237
I'm thrilled to share that the first paper I supervised as part of mentoring an intern, "CueTip: An Interactive and Explainable Physics-aware Pool Assistant," has been accepted to SIGGRAPH 2025! 🎉 arxiv.org/pdf/2501.18291?
1
1
128
more replies
Experiments demonstrate that CueTip delivers contextual, physics-aware shot suggestions and explanations without compromising performance. In some scenarios, it even enhances win rates.
1
47
We're also offering several internship opportunities in our London office—if you're interested in NLP and AI research, check them out and apply: linkedin.com/jobs/search/?cu…

21 Tencent Nlp Intern jobs in United States

Today’s top 21 Tencent Nlp Intern jobs in United States. Leverage your professional network, and get hired. New Tencent Nlp Intern jobs added daily.

linkedin.com
73
🚀Exciting to see how recent advancements like OpenAI’s O1/O3 & DeepSeek’s R1 are pushing the boundaries! Check out our latest survey on Complex Reasoning with LLMs. Analyzed over 300 papers to explore the progress. Paper: arxiv.org/pdf/2502.17419 Github: github.com/zzli2022/Awesome-…
2
62
9
156
12,515
I am thrilled to share that our paper "GeoEval: Benchmark for Evaluating LLMs and Multi-Modal Models on Geometry Problem-Solving" with my supervisor @YashMosh, has been accepted at #ACL2024 @NeuraSearchLab #NeuraSearch @UniStrathclyde @StrathCIS @aclmeeting
2
5
11
868
This benchmark facilitates a deeper investigation into the performance of 10 LLMs and MMs on solving geometry math problems.
1
246
Replying to @KnightZhang625 @YashMosh @NeuraSearchLab @UniStrathclyde @StrathCIS @aclmeeting
This paper introduces the GeoEval benchmark, a comprehensive collection that includes a main subset of 2000 problems, a 750 problem subset focusing on backward reasoning, an augmented subset of 2000 problems, and a hard subset of 300 problems.
1
213
I am excited to share that our paper "GOLD: Geometry Problem Solver with Natural Language Description" with my supervisor @YashMosh has been accepted at #NAACL2024 ! Preprint will be available soon. @NeuraSearchLab #NeuraSearch @UniStrathclyde @StrathCIS #NAACL2024
2
3
17
918
JiaxinZhang retweeted
MrBeast
@MrBeast22 Jan 2024
I’m gonna give 10 random people that repost this and follow me $25,000 for fun (the $250,000 my X video made) I’ll pick the winners in 72 hours
368,916
2,485,698
220,006
1,832,839
286,747,645
Introducing AlphaGeometry: an AI system that solves Olympiad geometry problems at a level approaching a human gold-medalist. 📐 It was trained solely on synthetic data and marks a breakthrough for AI in mathematical reasoning. 🧵 dpmd.ai/alphageometry
113
924
378
3,815
1,197,549
CS 25 has a great roadmap of LLM papers! 🙏 Transformers United has great guest lectures spanning the foundations of Large Language Models An underrated aspect of the course is the curated list of papers on every topic Perfect for your weekend reads: web.stanford.edu/class/cs25/
8
114
1
521
68,068
Delightful that my University @UniStrathclyde @StrathCIS join the great Turning University Network
The Alan Turing Institute@turinginst3 Apr 2023
Today we’ve launched the new Turing University Network, providing UK universities the opportunity to work with us & our network. It’s crucial to the success of our new strategy and we look forward to working closely with new universities. Find out more: bit.ly/3ZxPkKJ
4
955
This must be the most perspicacious explanation from ChatGPT about how RoI pooling works. Try to search online in both English and Chinese, and show me either intuitional or obscure answers. I have to admit that ChatGPT contains some kind of intelligence. #ChatGPT
1
239
Load more