Xiaoying Zhang

Xiaoying Zhang, Junpu Chen, Hongning Wang et al.

Off-policy learning, referring to the procedure of policy optimization with access only to logged feedback data, has shown importance in various real-world applications, such as search engines, recommender systems, and etc. While the ground-truth logging policy, which generates the logged data, is usually unknown, previous work simply takes its estimated value in off-policy learning, ignoring both high bias and high variance resulted from such an estimator, especially on samples with small and inaccurately estimated logging probabilities. In this work, we explicitly model the uncertainty in the estimated logging policy and propose a Uncertainty-aware Inverse Propensity Score estimator (UIPS) for improved off-policy learning, with a theoretical convergence guarantee. Experiment results on synthetic and three real-world recommendation datasets demonstrate the advantageous sample efficiency of the proposed UIPS estimator against an extensive list of state-of-the-art baselines.

Xiao-Ying Zhang, Ming-Ming Wang

Quantum machine learning has shown advantages in many ways compared to classical machine learning. In machine learning, a difficult problem is how to learn a model with high robustness and strong generalization ability from a limited feature space. Combining multiple models as base learners, ensemble learning (EL) can effectively improve the accuracy, generalization ability, and robustness of the final model. The key to EL lies in two aspects, the performance of base learners and the choice of the combination strategy. Recently, quantum EL (QEL) has been studied. However, existing combination strategies in QEL are inadequate in considering the accuracy and variance among base learners. This paper presents a hybrid EL framework that combines quantum and classical advantages. More importantly, we propose an efficient combination strategy for improving the accuracy of classification in the framework. We verify the feasibility and efficiency of our framework and strategy by using the MNIST dataset. Simulation results show that the hybrid EL framework with our combination strategy not only has a higher accuracy and lower variance than the single model without the ensemble, but also has a better accuracy than the majority voting and the weighted voting strategies in most cases.

Xiaoying Zhang, Jean-Francois Ton, Wei Shen et al.

We introduce Adversarial Policy Optimization (AdvPO), a novel solution to the pervasive issue of reward over-optimization in Reinforcement Learning from Human Feedback (RLHF) for Large Language Models (LLMs). Over-optimization occurs when a reward model serves as an imperfect proxy for human preference, and RL-driven policy optimization erroneously exploits reward inaccuracies. In this paper, we begin by introducing a lightweight way to quantify uncertainties in rewards, relying solely on the last layer embeddings of the reward model, without the need for computationally expensive reward ensembles. AdvPO then addresses a distributionally robust optimization problem centred around the confidence interval of the reward model's predictions for policy improvement. Through comprehensive experiments on the Anthropic HH and TL;DR summarization datasets, we illustrate the efficacy of AdvPO in mitigating the overoptimization issue, consequently resulting in enhanced performance as evaluated through human-assisted evaluation.

Xiaoying Zhang, Da Peng, Yipeng Zhang et al.

Recent progress in (multimodal) large language models ((M)LLMs) has shifted focus from pre-training to inference-time computation and post-training optimization, largely due to concerns over the availability of high-quality human data. However, these strategies alone are insufficient to drive substantial model improvements. We argue that effective model advancement requires strong synergy among pre-training, inference-time computation, and post-training optimization. In this paper, we introduce Self-Improving cognition (SIcog), a self-learning framework for constructing next-generation foundation MLLMs by imparting multimodal knowledge and enhancing systematic cognitive capabilities through multimodal pre-training with self-generated data. Specifically, we propose Chain-of-Description for step-by-step visual understanding and integrate structured Chain-of-Thought (CoT) reasoning to support in-depth multimodal reasoning. SIcog first equips a base model with systematic perception and reasoning using minimal external supervision. The enhanced models then generate candidate image captions and CoT reasoning responses for unlabeled images and image-question pairs across diverse tasks, which are filtered through a semantic-similarity-guided self-consistency mechanism. These high-quality, self-generated samples enable large-scale multimodal pre-training, creating a self-improvement loop. Experiments demonstrate SIcog's effectiveness in developing MLLMs with enhanced multimodal cognition. Using only 213K self-generated pre-training samples, SIcog achieves significant improvements, including +3.6% on MMStar and +3.5% on AI2D, outperforming previous pre-training approaches. When combined with post-training techniques for CoT reasoning, SIcog yields +9% gains on MMVet and +8.5% on ScienceQA.

Haidong Xie, Jia Tan, Xiaoying Zhang et al.

Deep learning is applied to many complex tasks in the field of wireless communication, such as modulation recognition of spectrum waveforms, because of its convenience and efficiency. This leads to the problem of a malicious third party using a deep learning model to easily recognize the modulation format of the transmitted waveform. Some existing works address this problem directly using the concept of adversarial examples in the image domain without fully considering the characteristics of the waveform transmission in the physical world. Therefore, we propose a low-intercept waveform~(LIW) generation method that can reduce the probability of the modulation being recognized by a third party without affecting the reliable communication of the friendly party. Our LIW exhibits significant low-interception performance even in the physical hardware experiment, decreasing the accuracy of the state of the art model to approximately $15\%$ with small perturbations.