Lin Xiaoqiang
@xiaoqiang_98
RS Intern @Meta, ML/AI PhD student @ National University of Singapore
ID: 2304604969
22-01-2014 10:17:58
62 Tweet
48 Followers
123 Following
Visit our poster on prompt optimization using your INSTINCT ICML Conference #ICML2024 in Poster Session 6 Thu 25 Jul 1:30PM Hall C 4-9 #711. #LLM #LLMs #PromptEngineering xqlin98.github.io/INSTINCT/
The ICML Conference #ICML2024 work of Lin Xiaoqiang Xinyi Xu Wu Zhaoxuan et Al. presents distributionally robust #DataValuation without a known validation distribution. #DataCentricAI Paper: openreview.net/forum?id=mbBeh… Visit us at Poster Session 3 Wed 24 Jul 11:30AM Hall C 4-9 #2402
Oral presentation by Lin Xiaoqiang on prompt optimization with human feedback ICML Conference #ICML2024 Workshop on Models of Human Feedback for AI Alignment.
Federated #ZerothOrderOptimization at ICML Conference #ICML2024 Workshop on Differentiable Almost Everything!
The #EMNLP2024 EMNLP 2025 (main) work of Gregory Lau Xinyuan Dao Hieu Chen Jiangwei chuan-sheng proposes the Waterfall framework for robust & scalable text #Watermarking to protect IP/copyrights of text for #LLM data provenance. 1st NLP conf paper! arxiv.org/abs/2407.04411
The #EMNLP2024 EMNLP 2025 (findings) position paper of Xinyi Xu Wu Zhaoxuan Rui Qiao Arun Verma Pang Wei Koh et al. proposes a data-centric viewpoint of AI research, focusing on #LLM #LLMs. Check it out @ arxiv.org/abs/2406.14473
👏👏👏 Hu.Wenyang Wu Zhaoxuan @ZCODE Zhuanghua Liu Lin Xiaoqiang Xinyi Xu patrick jaillet Dai Zhongxiang qphong Zijian Zhou et al. Accepted NeurIPS Conference #NeurIPS2024: #DataValuation #DataCentricAI #LLM #BayesianOptimization #InContextLearning #PromptEngineering #ActiveLearning
What demo eg's work best for #LLMs? Do they fit your instructions? 🤔 Introducing autoprompt opt that helps u craft prompts wif ✨EASE✨. Sit back, relax ☕️ 🚀 EASE jointly optimizes instructions & demo's for better LLM perf NeurIPS Conference #NeurIPS2024 poster session 5 east #2804
Our #PromptOptimization ZOPO connects prompt generation & representation, and uses a Neural Tangent Kernel-based #GaussianProcess in zeroth-order optimization to efficiently search for well-performing local optima NeurIPS Conference #NeurIPS2024 poster session 5 east #2803 (spotlight)
👏👏👏 pham nguyen Rachael Sim qphong Zhiliang Chen Xinyuan Lin Xiaoqiang Xinyi Xu patrick jaillet Dai Zhongxiang Arun Verma Rui Qiao Wu Zhaoxuan Angel Pang Wei Koh Apivich H. (Kaotoo) Gregory Lau Bingchen Wang et al. Accepted ICLR 2026 #ICLR2025 AAAI #AAAI2025
🚨🤔 How to efficiently learn from human preference feedback with latent non-linear reward function in an online setting? 💡🧠 The ICLR 2026 work of Arun Verma Dai Zhongxiang Lin Xiaoqiang patrick jaillet uses a neural network to estimate the unknown latent reward function from human
Day 2 (early afternoon) of being a scholar/researcher 🩷 I was invited to give a talk on Efficient Prompt Optimization for Adaptive LLMs: Instruction and Exemplar Tuning via Neural Bandits (published in ICML Conference #ICML2024: x.com/bryanklow/stat… and NeurIPS Conference #NeurIPS2024:
👏👏👏 Ruth qphong Hu.Wenyang Shu Yao Rui Qiao Angel Pang Wei Koh et al. Accepted ICML Conference #ICML2025: #DataValuation #DataCentricAI #DataSelection #LLM #LLMs #BayesianOptimization #FederatedLearning
🚀 Smarter Data Meet NICE ICML Conference – a new way to pick training data that actually improves #LLM task performance (not just loss)! ✨ This is a joint work with some really NICE folks I know 🤣😅😂 : Angel Lin Xiaoqiang Rui Qiao Pang Wei Koh Chuan Sheng Foo. (1) 🎯 From
Inspired by the success of inverse problems in uncovering fundamental scientific laws, our position paper, which is accepted to EMNLP 2025 #EMNLP2025 (findings), argues that inverse problems can be used to efficiently uncover underlying scaling laws for #LLMs that help the
