Zihao Xu

Zihao Xu, Guang-Yuan Hao, Hao He et al.

Previous studies have shown that leveraging domain index can significantly boost domain adaptation performance (arXiv:2007.01807, arXiv:2202.03628). However, such domain indices are not always available. To address this challenge, we first provide a formal definition of domain index from the probabilistic perspective, and then propose an adversarial variational Bayesian framework that infers domain indices from multi-domain data, thereby providing additional insight on domain relations and improving domain adaptation performance. Our theoretical analysis shows that our adversarial variational Bayesian framework finds the optimal domain index at equilibrium. Empirical results on both synthetic and real data verify that our model can produce interpretable domain indices which enable us to achieve superior performance compared to state-of-the-art domain adaptation methods. Code is available at https://github.com/Wang-ML-Lab/VDI.

Xinkai Ma, Zhiqi Bai, Dingling Zhang et al.

Deep Research Agents have shown strong capability in multi-step information retrieval, reasoning, and long-form report generation, but existing benchmarks and systems remain predominantly text-centric, with limited evaluation of whether visual elements are factually reliable and well aligned with the surrounding analysis. To address this gap, we introduce TVIR (Text--Visual Interleaved Report Generation), which includes TVIR-Bench, a benchmark of 100 expert-curated multimodal deep research tasks that require visual elements to serve specific analytical sub-goals, and TVIR-Agent, a hierarchical multi-agent framework that serves as a strong baseline for constructing outlines, retrieving images, generating charts with traceable sources, and composing reports through context-aware sequential writing. We further develop a dual-path evaluation framework that combines Textual Assessment and Visual Assessment. Experiments across nine deep research systems show that TVIR-Agent achieves strong overall performance, underscoring the importance of explicit multimodal design and evaluation for evidence-driven report generation.

Tianyi Liu, Zihao Xu, Hao He et al.

Domain adaptation aims to mitigate distribution shifts among different domains. However, traditional formulations are mostly limited to categorical domains, greatly simplifying nuanced domain relationships in the real world. In this work, we tackle a generalization with taxonomy-structured domains, which formalizes domains with nested, hierarchical similarity structures such as animal species and product catalogs. We build on the classic adversarial framework and introduce a novel taxonomist, which competes with the adversarial discriminator to preserve the taxonomy information. The equilibrium recovers the classic adversarial domain adaptation's solution if given a non-informative domain taxonomy (e.g., a flat taxonomy where all leaf nodes connect to the root node) while yielding non-trivial results with other taxonomies. Empirically, our method achieves state-of-the-art performance on both synthetic and real-world datasets with successful adaptation. Code is available at https://github.com/Wang-ML-Lab/TSDA.

Zihao Xu, Yi Liu, Gelei Deng et al.

Security concerns for large language models (LLMs) have recently escalated, focusing on thwarting jailbreaking attempts in discrete prompts. However, the exploration of jailbreak vulnerabilities arising from continuous embeddings has been limited, as prior approaches primarily involved appending discrete or continuous suffixes to inputs. Our study presents a novel channel for conducting direct attacks on LLM inputs, eliminating the need for suffix addition or specific questions provided that the desired output is predefined. We additionally observe that extensive iterations often lead to overfitting, characterized by repetition in the output. To counteract this, we propose a simple yet effective strategy named CLIP. Our experiments show that for an input length of 40 at iteration 1000, applying CLIP improves the ASR from 62% to 83%

Renlong Hang, Zihao Xu, Jiuwei Zhao et al.

Tropical cyclones (TCs) rank among the most destructive natural hazards, yet their forecasting faces fundamental trade-offs: numerical weather prediction (NWP) models are computationally prohibitive and struggle to leverage historical data, while existing deep learning (DL)-based intelligent models are variable-specific and deterministic, which fail to generalize across different forecasting variables. Here we present CycloneMAE, a scalable multi-task forecasting model that learns transferable TC representations from multi-modal data using a TC structure-aware masked autoencoder. By coupling a discrete probabilistic gridding mechanism with a pre-train/fine-tune paradigm, CycloneMAE simultaneously delivers deterministic forecasts and probability distributions. Evaluated across five global ocean basins, CycloneMAE outperforms leading NWP systems in pressure and wind forecasting up to 120 hours and in track forecasting up to 24 hours. Attribution analysis via integrated gradients reveals physically interpretable learning dynamics: short-term forecasts rely predominantly on the internal core convective structure from satellite imagery, whereas longer-term forecasts progressively shift attention to external environmental factors. Our framework establishes a scalable, probabilistic, and interpretable pathway for operational TC forecasting.

Haizhou Shi, Zihao Xu, Hengyi Wang et al.

The recent success of large language models (LLMs) trained on static, pre-collected, general datasets has sparked numerous research directions and applications. One such direction addresses the non-trivial challenge of integrating pre-trained LLMs into dynamic data distributions, task structures, and user preferences. Pre-trained LLMs, when tailored for specific needs, often experience significant performance degradation in previous knowledge domains -- a phenomenon known as "catastrophic forgetting". While extensively studied in the continual learning (CL) community, it presents new manifestations in the realm of LLMs. In this survey, we provide a comprehensive overview of the current research progress on LLMs within the context of CL. This survey is structured into four main sections: we first describe an overview of continually learning LLMs, consisting of two directions of continuity: vertical continuity (or vertical continual learning), i.e., continual adaptation from general to specific capabilities, and horizontal continuity (or horizontal continual learning), i.e., continual adaptation across time and domains (Section 3). We then summarize three stages of learning LLMs in the context of modern CL: Continual Pre-Training (CPT), Domain-Adaptive Pre-training (DAP), and Continual Fine-Tuning (CFT) (Section 4). Then we provide an overview of evaluation protocols for continual learning with LLMs, along with the current available data sources (Section 5). Finally, we discuss intriguing questions pertaining to continual learning for LLMs (Section 6). The full list of papers examined in this survey is available at https://github.com/Wang-ML-Lab/llm-continual-learning-survey.

Zihao Xu, Xuan Tang, Yufei Shi et al.

In continual learning, the learner learns multiple tasks in sequence, with data being acquired only once for each task. Catastrophic forgetting is a major challenge to continual learning. To reduce forgetting, some existing rehearsal-based methods use episodic memory to replay samples of previous tasks. However, in the process of knowledge integration when learning a new task, this strategy also suffers from catastrophic forgetting due to an imbalance between old and new knowledge. To address this problem, we propose a novel replay strategy called Manifold Expansion Replay (MaER). We argue that expanding the implicit manifold of the knowledge representation in the episodic memory helps to improve the robustness and expressiveness of the model. To this end, we propose a greedy strategy to keep increasing the diameter of the implicit manifold represented by the knowledge in the buffer during memory management. In addition, we introduce Wasserstein distance instead of cross entropy as distillation loss to preserve previous knowledge. With extensive experimental validation on MNIST, CIFAR10, CIFAR100, and TinyImageNet, we show that the proposed method significantly improves the accuracy in continual learning setup, outperforming the state of the arts.

Zhuowei Li, Zihao Xu, Ligong Han et al.

In-context Learning (ICL) empowers large language models (LLMs) to swiftly adapt to unseen tasks at inference-time by prefixing a few demonstration examples before queries. Despite its versatility, ICL incurs substantial computational and memory overheads compared to zero-shot learning and is sensitive to the selection and order of demonstration examples. In this work, we introduce Implicit In-context Learning (I2CL), an innovative paradigm that reduces the inference cost of ICL to that of zero-shot learning with minimal information loss. I2CL operates by first generating a condensed vector representation, namely a context vector, extracted from the demonstration examples. It then conducts an inference-time intervention through injecting a linear combination of the context vector and query activations back into the model's residual streams. Empirical evaluation on nine real-world tasks across three model architectures demonstrates that I2CL achieves few-shot level performance at zero-shot inference cost, and it exhibits robustness against variations in demonstration examples. Furthermore, I2CL facilitates a novel representation of task-ids, enhancing task similarity detection and fostering effective transfer learning. We also perform a comprehensive analysis and ablation study on I2CL, offering deeper insights into its internal mechanisms. Code is available at https://github.com/LzVv123456/I2CL.

Zihao Xu, John Harvill, Ziwei Fan et al.

Large Language Models (LLMs) incur significant computational and memory costs when processing long prompts, as full self-attention scales quadratically with input length. Token compression aims to address this challenge by reducing the number of tokens representing inputs. However, existing prompt-compression approaches primarily operate in token space and overlook inefficiencies in the latent embedding space. In this paper, we propose K-Token Merging, a latent-space compression framework that merges each contiguous block of K token embeddings into a single embedding via a lightweight encoder. The compressed sequence is processed by a LoRA-adapted LLM, while generation remains in the original vocabulary. Experiments on structural reasoning (Textualized Tree), sentiment classification (Amazon Reviews), and code editing (CommitPackFT) show that K-Token Merging lies on the Pareto frontier of performance vs. compression, achieving up to 75% input length reduction with minimal performance degradation.

Zihao Xu, Xiao Cheng, Ruijie Meng et al.

LLM API calls are becoming a ubiquitous program construct, yet they create a boundary that no existing program analysis can cross: runtime values enter a natural-language prompt, undergo opaque processing inside the LLM, and re-emerge as code, SQL, JSON, or text that the program consumes. Every analysis that tracks data across function boundaries, including taint analysis, program slicing, dependency analysis, and change-impact analysis, relies on dataflow summaries of callee behavior. LLM calls have no such summaries, breaking all of these analyses at what we call the NL/PL boundary. We present the first information flow method to bridge this boundary. Grounded in quantitative information flow theory, our taxonomy defines 24 labels along two orthogonal dimensions: information preservation level (from lexically preserved to fully blocked) and output modality (natural language, structured format, executable artifact). We label 9,083 placeholder-output pairs from 4,154 real-world Python files and validate reliability with Cohen's $κ= 0.82$ and near-complete coverage (0.01\% unclassifiable). We demonstrate the taxonomy's utility on two downstream applications: (1)~a two-stage taint propagation pipeline combining taxonomy-based filtering with LLM verification achieves $F_1 = 0.923$ on 353 expert-annotated pairs, with cross-language validation on six real-world OpenClaw prompt injection cases further confirming effectiveness; (2)~taxonomy-informed backward slicing reduces slice size by a mean of 15\% in files containing non-propagating placeholders. Per-label analysis reveals that four blocked labels account for nearly all non-propagating cases, providing actionable filtering criteria for tool builders.

Zihao Xu, Hao He, Guang-He Lee et al.

Existing domain adaptation methods tend to treat every domain equally and align them all perfectly. Such uniform alignment ignores topological structures among different domains; therefore it may be beneficial for nearby domains, but not necessarily for distant domains. In this work, we relax such uniform alignment by using a domain graph to encode domain adjacency, e.g., a graph of states in the US with each state as a domain and each edge indicating adjacency, thereby allowing domains to align flexibly based on the graph structure. We generalize the existing adversarial learning framework with a novel graph discriminator using encoding-conditioned graph embeddings. Theoretical analysis shows that at equilibrium, our method recovers classic domain adaptation when the graph is a clique, and achieves non-trivial alignment for other types of graphs. Empirical results show that our approach successfully generalizes uniform alignment, naturally incorporates domain information represented by graphs, and improves upon existing domain adaptation methods on both synthetic and real-world datasets. Code will soon be available at https://github.com/Wang-ML-Lab/GRDA.

Youjia Fu, Zihao Xu, Junsong Fu et al.

Recent advancements in transformer-based monocular 3D object detection techniques have exhibited exceptional performance in inferring 3D attributes from single 2D images. However, most existing methods rely on resource-intensive transformer architectures, which often lead to significant drops in computational efficiency and performance when handling long sequence data. To address these challenges and advance monocular 3D object detection technology, we propose an innovative network architecture, MonoMM, a Multi-scale \textbf{M}amba-Enhanced network for real-time Monocular 3D object detection. This well-designed architecture primarily includes the following two core modules: Focused Multi-Scale Fusion (FMF) Module, which focuses on effectively preserving and fusing image information from different scales with lower computational resource consumption. By precisely regulating the information flow, the FMF module enhances the model adaptability and robustness to scale variations while maintaining image details. Depth-Aware Feature Enhancement Mamba (DMB) Module: It utilizes the fused features from image characteristics as input and employs a novel adaptive strategy to globally integrate depth information and visual information. This depth fusion strategy not only improves the accuracy of depth estimation but also enhances the model performance under different viewing angles and environmental conditions. Moreover, the modular design of MonoMM provides high flexibility and scalability, facilitating adjustments and optimizations according to specific application needs. Extensive experiments conducted on the KITTI dataset show that our method outperforms previous monocular methods and achieves real-time detection.

Zihao Xu, Yi Liu, Gelei Deng et al.

Large Language Models (LLMS) have increasingly become central to generating content with potential societal impacts. Notably, these models have demonstrated capabilities for generating content that could be deemed harmful. To mitigate these risks, researchers have adopted safety training techniques to align model outputs with societal values to curb the generation of malicious content. However, the phenomenon of "jailbreaking", where carefully crafted prompts elicit harmful responses from models, persists as a significant challenge. This research conducts a comprehensive analysis of existing studies on jailbreaking LLMs and their defense techniques. We meticulously investigate nine attack techniques and seven defense techniques applied across three distinct language models: Vicuna, LLama, and GPT-3.5 Turbo. We aim to evaluate the effectiveness of these attack and defense techniques. Our findings reveal that existing white-box attacks underperform compared to universal techniques and that including special tokens in the input significantly affects the likelihood of successful attacks. This research highlights the need to concentrate on the security facets of LLMs. Additionally, we contribute to the field by releasing our datasets and testing framework, aiming to foster further research into LLM security. We believe these contributions will facilitate the exploration of security measures within this domain.

Yongen Zhao, Zihao Xu, Wenzhi Lu et al.

Wheeled-legged robots combine the energy efficiency of wheeled locomotion with the terrain adaptability of legged systems, making them promising platforms for agile mobility in complex and dynamic environments. However, enabling high-dynamic reflexive evasion against fast-moving obstacles remains challenging due to the hybrid morphology, mode coupling, and non-holonomic constraints of such platforms. In this work, we propose AWARE, Adaptive Wheeled-Legged Avoidance and Reflexive Evasion, a hierarchical reinforcement learning framework for high-dynamic obstacle avoidance in wheeled-legged robots. The proposed system naturally exhibits diverse emergent gaits and evasive behaviors, including forward lunge and lateral dodge, thereby leveraging the robot's hybrid morphology to enhance agility under highly dynamic threats. Extensive experiments in Isaac Lab simulation and real-world deployment on the M20 platform across diverse dynamic scenarios demonstrate that AWARE achieves robust and agile obstacle avoidance while revealing behaviorally distinct evasive strategies. These results highlight both the practical effectiveness of AWARE and the intrinsic reflexive agility of wheeled-legged robots.

Yiquan Wang, Jiaying Wang, Tin-Yeh Huang et al.

This paper proposes a novel hybrid model, STGCN-LSTM, to forecast Olympic medal distributions by integrating the spatio-temporal relationships among countries and the long-term dependencies of national performance. The Spatial-Temporal Graph Convolution Network (STGCN) captures geographic and interactive factors-such as coaching exchange and socio-economic links-while the Long Short-Term Memory (LSTM) module models historical trends in medal counts, economic data, and demographics. To address zero-inflated outputs (i.e., the disparity between countries that consistently yield wins and those never having won medals), a Zero-Inflated Compound Poisson (ZICP) framework is incorporated to separate random zeros from structural zeros, providing a clearer view of potential breakthrough performances. Validation includes historical backtracking, policy shock simulations, and causal inference checks, confirming the robustness of the proposed method. Results shed light on the influence of coaching mobility, event specialization, and strategic investment on medal forecasts, offering a data-driven foundation for optimizing sports policies and resource allocation in diverse Olympic contexts.

Xiaoxiao He, Haizhou Shi, Ligong Han et al.

Cardiovascular disease (CVD) and cardiac dyssynchrony are major public health problems in the United States. Precise cardiac image segmentation is crucial for extracting quantitative measures that help categorize cardiac dyssynchrony. However, achieving high accuracy often depends on centralizing large datasets from different hospitals, which can be challenging due to privacy concerns. To solve this problem, Federated Learning (FL) is proposed to enable decentralized model training on such data without exchanging sensitive information. However, bandwidth limitations and data heterogeneity remain as significant challenges in conventional FL algorithms. In this paper, we propose a novel efficient and adaptive federate learning method for cardiac segmentation that improves model performance while reducing the bandwidth requirement. Our method leverages the low-rank adaptation (LoRA) to regularize model weight update and reduce communication overhead. We also propose a \mymethod{} aggregation technique to address data heterogeneity among clients. This technique adaptively penalizes the aggregated weights from different clients by comparing the validation accuracy in each client, allowing better generalization performance and fast local adaptation. In-client and cross-client evaluations on public cardiac MR datasets demonstrate the superiority of our method over other LoRA-based federate learning approaches.

Zihao Xu, Junchen Ding, Yiling Lou et al.

Large Language Models (LLMs) have achieved significant progress in language understanding and reasoning. Evaluating and analyzing their logical reasoning abilities has therefore become essential. However, existing datasets and benchmarks are often limited to overly simplistic, unnatural, or contextually constrained examples. In response to the growing demand, we introduce SmartyPat-Bench, a challenging, naturally expressed, and systematically labeled benchmark derived from real-world high-quality Reddit posts containing subtle logical fallacies. Unlike existing datasets and benchmarks, it provides more detailed annotations of logical fallacies and features more diverse data. To further scale up the study and address the limitations of manual data collection and labeling - such as fallacy-type imbalance and labor-intensive annotation - we introduce SmartyPat, an automated framework powered by logic programming-based oracles. SmartyPat utilizes Prolog rules to systematically generate logically fallacious statements, which are then refined into fluent natural-language sentences by LLMs, ensuring precise fallacy representation. Extensive evaluation demonstrates that SmartyPat produces fallacies comparable in subtlety and quality to human-generated content and significantly outperforms baseline methods. Finally, experiments reveal nuanced insights into LLM capabilities, highlighting that while excessive reasoning steps hinder fallacy detection accuracy, structured reasoning enhances fallacy categorization performance.

Junchen Ding, Penghao Jiang, Zihao Xu et al.

As large language models (LLMs) increasingly mediate ethically sensitive decisions, understanding their moral reasoning processes becomes imperative. This study presents a comprehensive empirical evaluation of 14 leading LLMs, both reasoning enabled and general purpose, across 27 diverse trolley problem scenarios, framed by ten moral philosophies, including utilitarianism, deontology, and altruism. Using a factorial prompting protocol, we elicited 3,780 binary decisions and natural language justifications, enabling analysis along axes of decisional assertiveness, explanation answer consistency, public moral alignment, and sensitivity to ethically irrelevant cues. Our findings reveal significant variability across ethical frames and model types: reasoning enhanced models demonstrate greater decisiveness and structured justifications, yet do not always align better with human consensus. Notably, "sweet zones" emerge in altruistic, fairness, and virtue ethics framings, where models achieve a balance of high intervention rates, low explanation conflict, and minimal divergence from aggregated human judgments. However, models diverge under frames emphasizing kinship, legality, or self interest, often producing ethically controversial outcomes. These patterns suggest that moral prompting is not only a behavioral modifier but also a diagnostic tool for uncovering latent alignment philosophies across providers. We advocate for moral reasoning to become a primary axis in LLM alignment, calling for standardized benchmarks that evaluate not just what LLMs decide, but how and why.

Kalliopi Basioti, Pritish Sahu, Qingze Tony Liu et al.

Raven's Progressive Matrices (RPMs) is an established benchmark to examine the ability to perform high-level abstract visual reasoning (AVR). Despite the current success of algorithms that solve this task, humans can generalize beyond a given puzzle and create new puzzles given a set of rules, whereas machines remain locked in solving a fixed puzzle from a curated choice list. We propose Generative Visual Puzzles (GenVP), a framework to model the entire RPM generation process, a substantially more challenging task. Our model's capability spans from generating multiple solutions for one specific problem prompt to creating complete new puzzles out of the desired set of rules. Experiments on five different datasets indicate that GenVP achieves state-of-the-art (SOTA) performance both in puzzle-solving accuracy and out-of-distribution (OOD) generalization in 22 OOD scenarios. Compared to SOTA generative approaches, which struggle to solve RPMs when the feasible solution space increases, GenVP efficiently generalizes to these challenging setups. Moreover, our model demonstrates the ability to produce a wide range of complete RPMs given a set of abstract rules by effectively capturing the relationships between abstract rules and visual object properties.

Yang Zou, Zhaoshuai Qi, Yating Liu et al.

Object pose estimation, which plays a vital role in robotics, augmented reality, and autonomous driving, has been of great interest in computer vision. Existing studies either require multi-stage pose regression or rely on 2D-3D feature matching. Though these approaches have shown promising results, they rely heavily on appearance information, requiring complex input (i.e., multi-view reference input, depth, or CAD models) and intricate pipeline (i.e., feature extraction-SfM-2D to 3D matching-PnP). We propose AxisPose, a model-free, matching-free, single-shot solution for robust 6D pose estimation, which fundamentally diverges from the existing paradigm. Unlike existing methods that rely on 2D-3D or 2D-2D matching using 3D techniques, such as SfM and PnP, AxisPose directly infers a robust 6D pose from a single view by leveraging a diffusion model to learn the latent axis distribution of objects without reference views. Specifically, AxisPose constructs an Axis Generation Module (AGM) to capture the latent geometric distribution of object axes through a diffusion model. The diffusion process is guided by injecting the gradient of geometric consistency loss into the noise estimation to maintain the geometric consistency of the generated tri-axis. With the generated tri-axis projection, AxisPose further adopts a Triaxial Back-projection Module (TBM) to recover the 6D pose from the object tri-axis. The proposed AxisPose achieves robust performance at the cross-instance level (i.e., one model for N instances) using only a single view as input without reference images, with great potential for generalization to unseen-object level.

Yutong Xie, Haiyang Wang, Yan Hao et al.

In this paper, we propose a data-driven visual rhythm prediction method, which overcomes the previous works' deficiency that predictions are made primarily by human-crafted hard rules. In our approach, we first extract features including original frames and their residuals, optical flow, scene change, and body pose. These visual features will be next taken into an end-to-end neural network as inputs. Here we observe that there are some slight misaligning between features over the timeline and assume that this is due to the distinctions between how different features are computed. To solve this problem, the extracted features are aligned by an elaborately designed layer, which can also be applied to other models suffering from mismatched features, and boost performance. Then these aligned features are fed into sequence labeling layers implemented with BiLSTM and CRF to predict the onsets. Due to the lack of existing public training and evaluation set, we experiment on a dataset constructed by ourselves based on professionally edited Music Videos (MVs), and the F1 score of our approach reaches 79.6.