The paper introduces HACK GPs, a method that treats kernel selection for Gaussian Processes as an online learning problem with expert advice. Each candidate kernel is viewed as a GP expert, and a distribution over these experts is updated online using AdaHedge based on a loss that reflects both function fit and task alignment. Two variants—Mixture of Gaussians and categorical sampling—are presented, with theoretical guarantees that the weight concentrates on the best kernel under a loss‑gap condition, and empirical results show robust performance across Bayesian optimization, level set estimation, and Bayesian active learning compared to standard kernels and simple ensembles.
By Kavin Aravindan, Mani Tej Sriram, Gautam Dasarathy, Tejas Bodas
arXiv:2606. 04807v1 Announce Type: new Abstract: Mitigating social bias in Large Language Models (LLMs) presents a distinct alignment challenge: unlike verifiable tasks, bias lacks a single ground truth, creating a high-variance, subjective reward landscape.
By Saket Reddy, Ke Yang, ChengXiang Zhai
arXiv:2602. 14696v2 Announce Type: replace Abstract: Instruction fine-tuning of large language models (LLMs) often involves selecting a subset of instruction training data from a large candidate pool, using a small query set from the target task.
By Nihal V. Nayak, Paula Rodriguez-Diaz, Neha Hulkund, Sara Beery, David Alvarez-Melis
Mitigating social bias in Large Language Models (LLMs) presents a distinct alignment challenge: unlike verifiable tasks, bias lacks a single ground truth, creating a high-variance, subjective reward landscape. Previous preference-based fine-tuning methods have major trade-offs: Direct Preference Optimization (DPO) is limited by the lack of exploration inherent in offline training, while Proximal Policy Optimization (PPO) can lead to training instability due to potentially unreliable critic estimates.
arXiv:2608. 19790v1 Announce Type: new Abstract: Discovering materials with desirable properties often requires searching large candidate spaces while experimental or computational evaluations remain costly.
By Dino-Rober Demir, Florian Le Bronnec, Rio Yokota
The paper introduces Gradient-based Sample Selection Bayesian Optimization (GSSBO), a method that builds the Gaussian process surrogate on a strategically chosen subset of samples rather than the full dataset. By using gradient information to eliminate redundant points while keeping diversity and representativeness, GSSBO achieves sublinear regret bounds and reduces the cubic computational cost of standard BO. Experiments on synthetic and real-world tasks show that this approach maintains comparable optimization performance while significantly cutting GP fitting time and resource usage.
By Qiyu Wei, Haowei Wang, Zirui Cao, Songhao Wang, Richard Allmendinger, Mauricio A \'Alvarez
arXiv:2606. 18650v1 Announce Type: new Abstract: As Large Language Model (LLM) datasets scale to trillions of tokens, data selection has emerged as a critical frontier to filter out uninformative noise and construct adaptive learning trajectories.
By Jiaxing Wang, Deping Xiang, Jin Xu, Zirui Liu, Zicheng Zhang, Guoqiang Gong, Jun Fang, Chao Liu, Pengzhang Liu, Tongxuan Liu, Ke Zhang, Qixia Jiang
arXiv:2609.37076v1 Announce Type: new
Abstract: Large language models trained on vast corpora inherently risk memorizing harmful content that may later re-emerge in their outputs. To mitigate this is...
By Puning Yang, Qizhou Wang, Junchi Yu, Bo Han, Xiuying Chen
The paper introduces Power‑Law Entropy Search (PLES), a computational‑cost‑aware acquisition function that uses multi‑fidelity Bayesian optimization to efficiently estimate optimal hyperparameter scaling laws for large language model training. PLES focuses on reducing the overall uncertainty of scaling law estimates rather than optimizing a single objective, selecting configurations that maximize uncertainty reduction per unit computational cost. Experiments on synthetic benchmarks, surrogate models, and real LLM pre‑training runs show that PLES converges to accurate scaling laws using less than one‑tenth of the computational budget required by conventional grid search and other baselines.
By Zhiliang Chen, Sebastian Ament, David Eriksson, Maximilian Balandat, Eytan Bakshy, Jihao Andreas Lin
arXiv:2502. 01226v4 Announce Type: replace Abstract: Gaussian process (GP) bandits provide a powerful framework for performing blackbox optimization of unknown functions.
By Jack Sandberg, Morteza Haghir Chehreghani
arXiv:2601. 07094v2 Announce Type: replace-cross Abstract: Bayesian optimization (BO) iteratively fits a Gaussian process (GP) surrogate to accumulated evaluations and selects new queries via an acquisition function.
By Jiguang Li, Hengrui Luo
GUPO: Gradient Uncertainty-aware Policy Optimization for Post-Training Large Language Models proposes a new method for fine‑tuning LLMs after training. The approach models each group gradient as a random variable, estimates its probability distribution, and uses Dirichlet‑based gradient uncertainty to weight each group’s contribution during policy updates. Experiments on multiple benchmarks show that this uncertainty‑aware aggregation improves the effectiveness of post‑training policy optimization.
By Peizheng Guo, Jianqi Zhang, Xingyu Zhang, Yun Fan, Jiahuan Zhou, Changwen Zheng, Wenwen Qiang