Zihao Zhang

Zihao Zhang, Zonghai Yao, Huixue Zhou et al.

Large language models have shown success as a tutor in education in various fields. Educating patients about their clinical visits plays a pivotal role in patients' adherence to their treatment plans post-discharge. This paper presents EHRTutor, an innovative multi-component framework leveraging the Large Language Model (LLM) for patient education through conversational question-answering. EHRTutor first formulates questions pertaining to the electronic health record discharge instructions. It then educates the patient through conversation by administering each question as a test. Finally, it generates a summary at the end of the conversation. Evaluation results using LLMs and domain experts have shown a clear preference for EHRTutor over the baseline. Moreover, EHRTutor also offers a framework for generating synthetic patient education dialogues that can be used for future in-house system training.

Yuxuan Guo, Shaohui Peng, Jiaming Guo et al.

Building open agents has always been the ultimate goal in AI research, and creative agents are the more enticing. Existing LLM agents excel at long-horizon tasks with well-defined goals (e.g., `mine diamonds' in Minecraft). However, they encounter difficulties on creative tasks with open goals and abstract criteria due to the inability to bridge the gap between them, thus lacking feedback for self-improvement in solving the task. In this work, we introduce autonomous embodied verification techniques for agents to fill the gap, laying the groundwork for creative tasks. Specifically, we propose the Luban agent target creative building tasks in Minecraft, which equips with two-level autonomous embodied verification inspired by human design practices: (1) visual verification of 3D structural speculates, which comes from agent synthesized CAD modeling programs; (2) pragmatic verification of the creation by generating and verifying environment-relevant functionality programs based on the abstract criteria. Extensive multi-dimensional human studies and Elo ratings show that the Luban completes diverse creative building tasks in our proposed benchmark and outperforms other baselines ($33\%$ to $100\%$) in both visualization and pragmatism. Additional demos on the real-world robotic arm show the creation potential of the Luban in the physical world.

Zihao Zhang, Fei Liu

While LLMs excel at open-ended reasoning, they often struggle with cost-sensitive planning, either treating all actions as having equal cost or failing to stay within strict budgets. In this paper, we introduce Cost-Augmented Monte Carlo Tree Search (CATS), a novel approach that brings explicit cost-awareness into LLM-guided planning. Tight cost constraints push the planner to quickly identify infeasible solutions, while looser constraints encourage optimization for minimal cost. We benchmark top LLMs such as GPT-4.1, Claude-3.7-Sonnet, and DeepSeek-R1, against our CATS planner to evaluate their performance in cost-sensitive scenarios. Our experiments suggest that raw LLMs such as GPT-4.1 often falter under tight budgets, whereas CATS consistently delivers strong performance, achieving higher task success rates and better cost efficiency. CATS provides an effective solution for budget-aware decision-making by combining the reasoning power of LLMs with structured search.

Vijeta Deshpande, Minhwa Lee, Zonghai Yao et al.

Prior research on Twitter (now X) data has provided positive evidence of its utility in developing supplementary health surveillance systems. In this study, we present a new framework to surveil public health, focusing on mental health (MH) outcomes. We hypothesize that locally posted tweets are indicative of local MH outcomes and collect tweets posted from 765 neighborhoods (census block groups) in the USA. We pair these tweets from each neighborhood with the corresponding MH outcome reported by the Center for Disease Control (CDC) to create a benchmark dataset, LocalTweets. With LocalTweets, we present the first population-level evaluation task for Twitter-based MH surveillance systems. We then develop an efficient and effective method, LocalHealth, for predicting MH outcomes based on LocalTweets. When used with GPT3.5, LocalHealth achieves the highest F1-score and accuracy of 0.7429 and 79.78\%, respectively, a 59\% improvement in F1-score over the GPT3.5 in zero-shot setting. We also utilize LocalHealth to extrapolate CDC's estimates to proxy unreported neighborhoods, achieving an F1-score of 0.7291. Our work suggests that Twitter data can be effectively leveraged to simulate neighborhood-level MH outcomes.

Haihan Gao, Rui Zhang, Qi Yi et al.

Overfitting in RL has become one of the main obstacles to applications in reinforcement learning(RL). Existing methods do not provide explicit semantic constrain for the feature extractor, hindering the agent from learning a unified cross-domain representation and resulting in performance degradation on unseen domains. Besides, abundant data from multiple domains are needed. To address these issues, in this work, we propose prompt-based visual alignment (PVA), a robust framework to mitigate the detrimental domain bias in the image for zero-shot policy transfer. Inspired that Visual-Language Model (VLM) can serve as a bridge to connect both text space and image space, we leverage the semantic information contained in a text sequence as an explicit constraint to train a visual aligner. Thus, the visual aligner can map images from multiple domains to a unified domain and achieve good generalization performance. To better depict semantic information, prompt tuning is applied to learn a sequence of learnable tokens. With explicit constraints of semantic information, PVA can learn unified cross-domain representation under limited access to cross-domain data and achieves great zero-shot generalization ability in unseen domains. We verify PVA on a vision-based autonomous driving task with CARLA simulator. Experiments show that the agent generalizes well on unseen domains under limited access to multi-domain data.

Zihao Zhang, Stefan Zohren

We design multi-horizon forecasting models for limit order book (LOB) data by using deep learning techniques. Unlike standard structures where a single prediction is made, we adopt encoder-decoder models with sequence-to-sequence and Attention mechanisms to generate a forecasting path. Our methods achieve comparable performance to state-of-art algorithms at short prediction horizons. Importantly, they outperform when generating predictions over long horizons by leveraging the multi-horizon setup. Given that encoder-decoder models rely on recurrent neural layers, they generally suffer from slow training processes. To remedy this, we experiment with utilising novel hardware, so-called Intelligent Processing Units (IPUs) produced by Graphcore. IPUs are specifically designed for machine intelligence workload with the aim to speed up the computation process. We show that in our setup this leads to significantly faster training times when compared to training models with GPUs.

Yinghui Li, Ruiyang Liu, ZiHao Zhang et al.

Deep neural models have hitherto achieved significant performances on numerous classification tasks, but meanwhile require sufficient manually annotated data. Since it is extremely time-consuming and expensive to annotate adequate data for each classification task, learning an empirically effective model with generalization on small dataset has received increased attention. Existing efforts mainly focus on transferring task-relevant knowledge from other similar data to tackle the issue. These approaches have yielded remarkable improvements, yet neglecting the fact that the task-irrelevant features could bring out massive negative transfer effects. To date, no large-scale studies have been performed to investigate the impact of task-irrelevant features, let alone the utilization of this kind of features. In this paper, we firstly propose Task-Irrelevant Transfer Learning (TIRTL) to exploit task-irrelevant features, which mainly are extracted from task-irrelevant labels. Particularly, we suppress the expression of task-irrelevant information and facilitate the learning process of classification. We also provide a theoretical explanation of our method. In addition, TIRTL does not conflict with those that have previously exploited task-relevant knowledge and can be well combined to enable the simultaneous utilization of task-relevant and task-irrelevant features for the first time. In order to verify the effectiveness of our theory and method, we conduct extensive experiments on facial expression recognition and digit recognition tasks. Our source code will be also available in the future for reproducibility.

Xu Sun, Zhenfeng Fan, Zihao Zhang et al.

Heatmap regression based face alignment has achieved prominent performance on static images. However, the stability and accuracy are remarkably discounted when applying the existing methods on dynamic videos. We attribute the degradation to random noise and motion blur, which are common in videos. The temporal information is critical to address this issue yet not fully considered in the existing works. In this paper, we visit the video-oriented face alignment problem in two perspectives: detection accuracy prefers lower error for a single frame, and detection consistency forces better stability between adjacent frames. On this basis, we propose a Jitter loss function that leverages temporal information to suppress inaccurate as well as jittered landmarks. The Jitter loss is involved in a novel framework with a fine-tuning ConvLSTM structure over a backbone replaceable network. We further demonstrate that accurate and stable landmarks are associated with different regions with overlaps in a canonical coordinate, based on which the proposed Jitter loss facilitates the optimization process during training. The proposed framework achieves at least 40% improvement on stability evaluation metrics while enhancing detection accuracy versus state-of-the-art methods. Generally, it can swiftly convert a landmark detector for facial images to a better-performing one for videos without retraining the entire model.