Tiankai Yang

Ruiyao Xu, Mihir Parmar, Tiankai Yang et al.

Learning from preference-based feedback has become an effective approach for aligning LLMs across diverse tasks. However, high-quality human-annotated preference data remains expensive and scarce. Existing methods address this challenge through either self-rewarding, which scales by using purely AI-generated labels but risks unreliability, or active learning, which ensures quality through oracle annotation but cannot fully leverage unlabeled data. In this paper, we present CoAct, a novel framework that synergistically combines self-rewarding and active learning through strategic human-AI collaboration. CoAct leverages self-consistency to identify both reliable self-labeled data and samples that require oracle verification. Additionally, oracle feedback guides the model to generate new instructions within its solvable capability. Evaluated on three reasoning benchmarks across two model families, CoAct achieves average improvements of +13.25% on GSM8K, +8.19% on MATH, and +13.16% on WebInstruct, consistently outperforming all baselines.

Sihan Chen, Zhuangzhuang Qian, Wingchun Siu et al.

Outlier detection (OD), also known as anomaly detection, is a critical machine learning (ML) task with applications in fraud detection, network intrusion detection, clickstream analysis, recommendation systems, and social network moderation. Among open-source libraries for outlier detection, the Python Outlier Detection (PyOD) library is the most widely adopted, with over 8,500 GitHub stars, 25 million downloads, and diverse industry usage. However, PyOD currently faces three limitations: (1) insufficient coverage of modern deep learning algorithms, (2) fragmented implementations across PyTorch and TensorFlow, and (3) no automated model selection, making it hard for non-experts. To address these issues, we present PyOD Version 2 (PyOD 2), which integrates 12 state-of-the-art deep learning models into a unified PyTorch framework and introduces a large language model (LLM)-based pipeline for automated OD model selection. These improvements simplify OD workflows, provide access to 45 algorithms, and deliver robust performance on various datasets. In this paper, we demonstrate how PyOD 2 streamlines the deployment and automation of OD models and sets a new standard in both research and industry. PyOD 2 is accessible at [https://github.com/yzhao062/pyod](https://github.com/yzhao062/pyod). This study aligns with the Web Mining and Content Analysis track, addressing topics such as the robustness of Web mining methods and the quality of algorithmically-generated Web data.

Shawn Li, Huixian Gong, Hao Dong et al.

Out-of-distribution (OOD) detection is essential for ensuring the robustness of machine learning models by identifying samples that deviate from the training distribution. While traditional OOD detection has primarily focused on single-modality inputs, such as images, recent advances in multimodal models have demonstrated the potential of leveraging multiple modalities (e.g., video, optical flow, audio) to enhance detection performance. However, existing methods often overlook intra-class variability within in-distribution (ID) data, assuming that samples of the same class are perfectly cohesive and consistent. This assumption can lead to performance degradation, especially when prediction discrepancies are uniformly amplified across all samples. To address this issue, we propose Dynamic Prototype Updating (DPU), a novel plug-and-play framework for multimodal OOD detection that accounts for intra-class variations. Our method dynamically updates class center representations for each class by measuring the variance of similar samples within each batch, enabling adaptive adjustments. This approach allows us to amplify prediction discrepancies based on the updated class centers, thereby improving the model's robustness and generalization across different modalities. Extensive experiments on two tasks, five datasets, and nine base OOD algorithms demonstrate that DPU significantly improves OOD detection performance, setting a new state-of-the-art in multimodal OOD detection, with improvements of up to 80 percent in Far-OOD detection. To facilitate accessibility and reproducibility, our code is publicly available on GitHub.

Tiankai Yang, Yi Nian, Xinyuan Li et al.

Aligning large language models with human preferences must balance two competing goals: responding helpfully to legitimate requests and reliably refusing harmful ones. Most preference-based safety alignment methods collapse safety into a single scalar that is applied uniformly to every preference pair. The result is a model that looks safe on average but stays relatively unsafe on a minority of harm categories. We cast safety alignment as a per-category constrained optimization problem and derive Cat-DPO, a direct-preference-optimization algorithm with a separate adaptive safety margin for each harm category. The margin tightens when the model still produces unsafe responses on a category and relaxes once the model catches up, so the training signal tracks each category's current difficulty rather than averaging under one global rate. Across two LLM backbones and six preference-learning baselines, Cat-DPO improves aggregate helpfulness and harmlessness and compresses per-category safety variance and the best-to-worst gap, offering a drop-in per-category refinement of direct preference safety alignment.

Yuangang Li, Jiaqi Li, Zhuo Xiao et al.

Anomaly detection (AD) is an important machine learning task with applications in fraud detection, content moderation, and user behavior analysis. However, AD is relatively understudied in a natural language processing (NLP) context, limiting its effectiveness in detecting harmful content, phishing attempts, and spam reviews. We introduce NLP-ADBench, the most comprehensive NLP anomaly detection (NLP-AD) benchmark to date, which includes eight curated datasets and 19 state-of-the-art algorithms. These span 3 end-to-end methods and 16 two-step approaches that adapt classical, non-AD methods to language embeddings from BERT and OpenAI. Our empirical results show that no single model dominates across all datasets, indicating a need for automated model selection. Moreover, two-step methods with transformer-based embeddings consistently outperform specialized end-to-end approaches, with OpenAI embeddings outperforming those of BERT. We release NLP-ADBench at https://github.com/USC-FORTIS/NLP-ADBench, providing a unified framework for NLP-AD and supporting future investigations.

Shawn Li, Jiashu Qu, Yuxiao Zhou et al.

Vision-Language Models (VLMs) have advanced multi-modal tasks like image captioning, visual question answering, and reasoning. However, they often generate hallucinated outputs inconsistent with the visual context or prompt, limiting reliability in critical applications like autonomous driving and medical imaging. Existing studies link hallucination to statistical biases, language priors, and biased feature learning but lack a structured causal understanding. In this work, we introduce a causal perspective to analyze and mitigate hallucination in VLMs. We hypothesize that hallucination arises from unintended direct influences of either the vision or text modality, bypassing proper multi-modal fusion. To address this, we construct a causal graph for VLMs and employ counterfactual analysis to estimate the Natural Direct Effect (NDE) of vision, text, and their cross-modal interaction on the output. We systematically identify and mitigate these unintended direct effects to ensure that responses are primarily driven by genuine multi-modal fusion. Our approach consists of three steps: (1) designing structural causal graphs to distinguish correct fusion pathways from spurious modality shortcuts, (2) estimating modality-specific and cross-modal NDE using perturbed image representations, hallucinated text embeddings, and degraded visual inputs, and (3) implementing a test-time intervention module to dynamically adjust the model's dependence on each modality. Experimental results demonstrate that our method significantly reduces hallucination while preserving task performance, providing a robust and interpretable framework for improving VLM reliability. To enhance accessibility and reproducibility, our code is publicly available at https://github.com/TREE985/Treble-Counterfactual-VLMs.

Yiming Tang, Yi Fan, Chenxiao Yu et al.

The integration of large language models (LLMs) into information retrieval systems introduces new attack surfaces, particularly for adversarial ranking manipulations. We present $\textbf{StealthRank}$, a novel adversarial attack method that manipulates LLM-driven ranking systems while maintaining textual fluency and stealth. Unlike existing methods that often introduce detectable anomalies, StealthRank employs an energy-based optimization framework combined with Langevin dynamics to generate StealthRank Prompts (SRPs)-adversarial text sequences embedded within item or document descriptions that subtly yet effectively influence LLM ranking mechanisms. We evaluate StealthRank across multiple LLMs, demonstrating its ability to covertly boost the ranking of target items while avoiding explicit manipulation traces. Our results show that StealthRank consistently outperforms state-of-the-art adversarial ranking baselines in both effectiveness and stealth, highlighting critical vulnerabilities in LLM-driven ranking systems. Our code is publicly available at $\href{https://github.com/Tangyiming205069/controllable-seo}{here}$.

Tiankai Yang, Jiate Li, Yi Nian et al.

LLM-based agents increasingly operate across repeated sessions, maintaining task states to ensure continuity. In many deployments, a single agent serves multiple users within a team or organization, reusing a shared knowledge layer across user identities. This shared persistence expands the failure surface: information that is locally valid for one user can silently degrade another user's outcome when the agent reapplies it without regard for scope. We refer to this failure mode as unintentional cross-user contamination (UCC). Unlike adversarial memory poisoning, UCC requires no attacker; it arises from benign interactions whose scope-bound artifacts persist and are later misapplied. We formalize UCC through a controlled evaluation protocol, introduce a taxonomy of three contamination types, and evaluate the problem in two shared-state mechanisms. Under raw shared state, benign interactions alone produce contamination rates of 57--71%. A write-time sanitization is effective when shared state is conversational, but leaves substantial residual risk when shared state includes executable artifacts, with contamination often manifesting as silent wrong answers. These results indicate that shared-state agents need artifact-level defenses beyond text-level sanitization to prevent silent cross-user failures.

Li Li, Peilin Cai, Ryan A. Rossi et al.

We present PersonaConvBench, a large-scale benchmark for evaluating personalized reasoning and generation in multi-turn conversations with large language models (LLMs). Unlike existing work that focuses on either personalization or conversational structure in isolation, PersonaConvBench integrates both, offering three core tasks: sentence classification, impact regression, and user-centric text generation across ten diverse Reddit-based domains. This design enables systematic analysis of how personalized conversational context shapes LLM outputs in realistic multi-user scenarios. We benchmark several commercial and open-source LLMs under a unified prompting setup and observe that incorporating personalized history yields substantial performance improvements, including a 198 percent relative gain over the best non-conversational baseline in sentiment classification. By releasing PersonaConvBench with evaluations and code, we aim to support research on LLMs that adapt to individual styles, track long-term context, and produce contextually rich, engaging responses.

Tiankai Yang, Junjun Liu, Wingchun Siu et al.

Anomaly detection (AD) is essential in areas such as fraud detection, network monitoring, and scientific research. However, the diversity of data modalities and the increasing number of specialized AD libraries pose challenges for non-expert users who lack in-depth library-specific knowledge and advanced programming skills. To tackle this, we present AD-AGENT, an LLM-driven multi-agent framework that turns natural-language instructions into fully executable AD pipelines. AD-AGENT coordinates specialized agents for intent parsing, data preparation, library and model selection, documentation mining, and iterative code generation and debugging. Using a shared short-term workspace and a long-term cache, the agents integrate popular AD libraries like PyOD, PyGOD, and TSLib into a unified workflow. Experiments demonstrate that AD-AGENT produces reliable scripts and recommends competitive models across libraries. The system is open-sourced to support further research and practical applications in AD.

Kunyi Zhang, Tiankai Yang, Ziming Ding et al.

Recently, the community has witnessed numerous datasets built for developing and testing state estimators. However, for some applications such as aerial transportation or search-and-rescue, the contact force or other disturbance must be perceived for robust planning and control, which is beyond the capacity of these datasets. This paper introduces a Visual-Inertial-Dynamical (VID) dataset, not only focusing on traditional six degrees of freedom (6-DOF) pose estimation but also providing dynamical characteristics of the flight platform for external force perception or dynamics-aided estimation. The VID dataset contains hardware synchronized imagery and inertial measurements, with accurate ground truth trajectories for evaluating common visual-inertial estimators. Moreover, the proposed dataset highlights rotor speed and motor current measurements, control inputs, and ground truth 6-axis force data to evaluate external force estimation. To the best of our knowledge, the proposed VID dataset is the first public dataset containing visual-inertial and complete dynamical information in the real world for pose and external force evaluation. The dataset: https://github.com/ZJU-FAST-Lab/VID-Dataset and related files: https://github.com/ZJU-FAST-Lab/VID-Flight-Platform are open-sourced.

Ziming Ding, Tiankai Yang, Kunyi Zhang et al.

Recently, quadrotors are gaining significant attention in aerial transportation and delivery. In these scenarios, an accurate estimation of the external force is as essential as the 6 degree-of-freedom (DoF) pose since it is of vital importance for planning and control of the vehicle. To this end, we propose a tightly-coupled Visual-Inertial-Dynamics (VID) system that simultaneously estimates the external force applied to the quadrotor along with the 6 DoF pose. Our method builds on the state-of-the-art optimization-based Visual-Inertial system, with a novel deduction of the dynamics and external force factor extended from VIMO. Utilizing the proposed dynamics and external force factor, our estimator robustly and accurately estimates the external force even when it varies widely. Moreover, since we explicitly consider the influence of the external force, when compared with VIMO and VINS-Mono, our method shows comparable and superior pose accuracy, even when the external force ranges from neglectable to significant. The robustness and effectiveness of the proposed method are validated by extensive real-world experiments and application scenario simulation. We will release an open-source package of this method along with datasets with ground truth force measurements for the reference of the community.

Tiankai Yang, Yi Nian, Shawn Li et al.

Anomaly detection (AD) is an important machine learning task with many real-world uses, including fraud detection, medical diagnosis, and industrial monitoring. Within natural language processing (NLP), AD helps detect issues like spam, misinformation, and unusual user activity. Although large language models (LLMs) have had a strong impact on tasks such as text generation and summarization, their potential in AD has not been studied enough. This paper introduces AD-LLM, the first benchmark that evaluates how LLMs can help with NLP anomaly detection. We examine three key tasks: (i) zero-shot detection, using LLMs' pre-trained knowledge to perform AD without tasks-specific training; (ii) data augmentation, generating synthetic data and category descriptions to improve AD models; and (iii) model selection, using LLMs to suggest unsupervised AD models. Through experiments with different datasets, we find that LLMs can work well in zero-shot AD, that carefully designed augmentation methods are useful, and that explaining model selection for specific datasets remains challenging. Based on these results, we outline six future research directions on LLMs for AD.

Wang Wei, Tiankai Yang, Hongjie Chen et al.

Model selection is a critical step in time series forecasting, traditionally requiring extensive performance evaluations across various datasets. Meta-learning approaches aim to automate this process, but they typically depend on pre-constructed performance matrices, which are costly to build. In this work, we propose to leverage Large Language Models (LLMs) as a lightweight alternative for model selection. Our method eliminates the need for explicit performance matrices by utilizing the inherent knowledge and reasoning capabilities of LLMs. Through extensive experiments with LLaMA, GPT and Gemini, we demonstrate that our approach outperforms traditional meta-learning techniques and heuristic baselines, while significantly reducing computational overhead. These findings underscore the potential of LLMs in efficient model selection for time series forecasting.

Wang Wei, Tiankai Yang, Hongjie Chen et al.

Efficient use of large language models (LLMs) is critical for deployment at scale: without adaptive routing, systems either overpay for strong models or risk poor performance from weaker ones. Selecting the right LLM for each query is fundamentally an online decision problem: models differ in strengths, prices fluctuate, and users value accuracy and cost differently. Yet most routers are trained offline with labels for all candidate models, an assumption that breaks in deployment, where only the outcome of the chosen model is observed. We bridge this gap with BaRP, a Bandit-feedback Routing with Preferences approach that trains under the same partial-feedback restriction as deployment, while supporting preference-tunable inference: operators can dial the performance/cost trade-off at test time without retraining. Framed as a contextual bandit over prompt features and a user preference vector, our method simulates an online feedback setting during training and adapts its routing decisions to each new prompt, rather than depending on full-information offline supervision. Comprehensive experiments show that our method consistently outperforms strong offline routers by at least 12.46% and the largest LLM by at least 2.45%, and generalizes robustly for unseen tasks.

Yuehan Qin, Li Li, Defu Cao et al.

Out-of-distribution (OOD) robustness is a critical challenge for modern machine learning systems, particularly as they increasingly operate in multimodal settings involving inputs like video, audio, and sensor data. Currently, many OOD detection methods have been proposed, each with different designs targeting various distribution shifts. A single OOD detector may not prevail across all the scenarios; therefore, how can we automatically select an ideal OOD detection model for different distribution shifts? Due to the inherent unsupervised nature of the OOD detection task, it is difficult to predict model performance and find a universally Best model. Also, systematically comparing models on the new unseen data is costly or even impractical. To address this challenge, we introduce M3OOD, a meta-learning-based framework for OOD detector selection in multimodal settings. Meta learning offers a solution by learning from historical model behaviors, enabling rapid adaptation to new data distribution shifts with minimal supervision. Our approach combines multimodal embeddings with handcrafted meta-features that capture distributional and cross-modal characteristics to represent datasets. By leveraging historical performance across diverse multimodal benchmarks, M3OOD can recommend suitable detectors for a new data distribution shift. Experimental evaluation demonstrates that M3OOD consistently outperforms 10 competitive baselines across 12 test scenarios with minimal computational overhead.

Yuman Gao, Yingjian Wang, Xingguang Zhong et al.

This letter presents a complete framework Meeting-Merging-Mission for multi-robot exploration under communication restriction. Considering communication is limited in both bandwidth and range in the real world, we propose a lightweight environment presentation method and an efficient cooperative exploration strategy. For lower bandwidth, each robot utilizes specific polytopes to maintains free space and super frontier information (SFI) as the source for exploration decision-making. To reduce repeated exploration, we develop a mission-based protocol that drives robots to share collected information in stable rendezvous. We also design a complete path planning scheme for both centralized and decentralized cases. To validate that our framework is practical and generic, we present an extensive benchmark and deploy our system into multi-UGV and multi-UAV platforms.

Neng Pan, Ruibin Zhang, Tiankai Yang et al.

In recent years, several progressive works promote the development of aerial tracking. One of the representative works is our previous work Fast-tracker which is applicable to various challenging tracking scenarios. However, it suffers from two main drawbacks: 1) the over simplification in target detection by using artificial markers and 2) the contradiction between simultaneous target and environment perception with limited onboard vision. In this paper, we upgrade the target detection in Fast-tracker to detect and localize a human target based on deep learning and non-linear regression to solve the former problem. For the latter one, we equip the quadrotor system with 360 degree active vision on a customized gimbal camera. Furthermore, we improve the tracking trajectory planning in Fast-tracker by incorporating an occlusion-aware mechanism that generates observable tracking trajectories. Comprehensive real-world tests confirm the proposed system's robustness and real-time capability. Benchmark comparisons with Fast-tracker validate that the proposed system presents better tracking performance even when performing more difficult tracking tasks.