arXiv subjects

Wei Qiao

Publications and source records attributed to Wei Qiao.

11 recordsLinked to original sources

arXiv2026

Agile Deliberation: Concept Deliberation for Subjective Visual Classification

From content moderation to content curation, applications requiring vision classifiers for visual concepts are rapidly expanding. Existing human-in-the-loop approaches typically assume users begin with a clear, stable concept understanding to be able to provide high-quality supervision. In reality, users often start with a vague idea and must iteratively refine it through "concept deliberation", a practice we uncovered through structured interviews with content moderation experts. We operationalize the common strategies in deliberation used by real content moderators into a human-in-the-loop framework called "Agile Deliberation" that explicitly supports evolving and subjective concepts. The system supports users in defining the concept for themselves by exposing them to borderline cases. The system does this with two deliberation stages: (1) concept scoping, which decomposes the initial concept into a structured hierarchy of sub-concepts, and (2) concept iteration, which surfaces semantically borderline examples for user reflection and feedback to iteratively align an image classifier with the user's evolving intent. Since concept deliberation is inherently subjective and interactive, we painstakingly evaluate the framework through 18 user sessions, each 1.5h long, rather than standard benchmarking datasets. We find that Agile Deliberation achieves 7.5% higher F1 scores than automated decomposition baselines and more than 3% higher than manual deliberation, while participants reported clearer conceptual understanding and lower cognitive effort.

cs.AI↗

arXiv2026

Sentient: Detecting APTs Via Capturing Indirect Dependencies and Behavioral Logic

Advanced Persistent Threats (APTs) are difficult to detect due to their complexity and stealthiness. To mitigate such attacks, many approaches model entities and their relationship using provenance graphs to detect the stealthy and persistent characteristics of APTs. However, existing detection methods suffer from the flaws of missing indirect dependencies, noisy complex scenarios, and missing behavioral logical associations, which make it difficult to detect complex scenarios and effectively identify stealthy threats. In this paper, we propose Sentient, an APT detection method that combines pre-training and intent analysis. It employs a graph transformer to learn structural and semantic information from provenance graphs to avoid missing indirect dependencies. We mitigate scenario noise by combining global and local information. Additionally, we design an Intent Analysis Module (IAM) to associate logical relationships between behaviors. Sentient is trained solely on easily obtainable benign data to detect malicious behaviors that deviate from benign behavioral patterns. We evaluated Sentient on three widely-used datasets covering real-world attacks and simulated attacks. Notably, compared to six state-of-the-art methods, Sentient achieved an average reduction of 44% in false positive rate(FPR) for detection.

cs.CR↗

arXiv2025

ProvX: Generating Counterfactual-Driven Attack Explanations for Provenance-Based Detection

Provenance graph-based intrusion detection systems are deployed on hosts to defend against increasingly severe Advanced Persistent Threat. Using Graph Neural Networks to detect these threats has become a research focus and has demonstrated exceptional performance. However, the widespread adoption of GNN-based security models is limited by their inherent black-box nature, as they fail to provide security analysts with any verifiable explanations for model predictions or any evidence regarding the model's judgment in relation to real-world attacks. To address this challenge, we propose ProvX, an effective explanation framework for exlaining GNN-based security models on provenance graphs. ProvX introduces counterfactual explanation logic, seeking the minimal structural subset within a graph predicted as malicious that, when perturbed, can subvert the model's original prediction. We innovatively transform the discrete search problem of finding this critical subgraph into a continuous optimization task guided by a dual objective of prediction flipping and distance minimization. Furthermore, a Staged Solidification strategy is incorporated to enhance the precision and stability of the explanations. We conducted extensive evaluations of ProvX on authoritative datasets. The experimental results demonstrate that ProvX can locate critical graph structures that are highly relevant to real-world attacks and achieves an average explanation necessity of 51.59\%, with these metrics outperforming current SOTA explainers. Furthermore, we explore and provide a preliminary validation of a closed-loop Detection-Explanation-Feedback enhancement framework, demonstrating through experiments that the explanation results from ProvX can guide model optimization, effectively enhancing its robustness against adversarial attacks.

cs.CR↗

arXiv2025

Slot: Provenance-Driven APT Detection through Graph Reinforcement Learning

Advanced Persistent Threats (APTs) represent sophisticated cyberattacks characterized by their ability to remain undetected within the victim system for extended periods, aiming to exfiltrate sensitive data or disrupt operations. Existing detection approaches often struggle to effectively identify these complex threats, construct the attack chain for defense facilitation, or resist adversarial attacks. To overcome these challenges, we propose Slot, an advanced APT detection approach based on provenance graphs and graph reinforcement learning. Slot excels in uncovering multi-level hidden relationships, such as causal, contextual, and indirect connections, among system behaviors through provenance graph mining. By pioneering the integration of graph reinforcement learning, Slot dynamically adapts to new user activities and evolving attack strategies, enhancing its resilience against adversarial attacks. Additionally, Slot automatically constructs the attack chain according to detected attacks with clustering algorithms, providing precise identification of attack paths and facilitating the development of defense strategies. Evaluations with real-world datasets demonstrate Slot's outstanding accuracy, efficiency, adaptability, and robustness in APT detection, with most metrics surpassing state-of-the-art methods. Additionally, case studies conducted to assess Slot's effectiveness in supporting APT defense further establish it as a practical and reliable tool for cybersecurity protection.

cs.CR↗

arXiv2024

Zero-Shot Image Moderation in Google Ads with LLM-Assisted Textual Descriptions and Cross-modal Co-embeddings

We present a scalable and agile approach for ads image content moderation at Google, addressing the challenges of moderating massive volumes of ads with diverse content and evolving policies. The proposed method utilizes human-curated textual descriptions and cross-modal text-image co-embeddings to enable zero-shot classification of policy violating ads images, bypassing the need for extensive supervised training data and human labeling. By leveraging large language models (LLMs) and user expertise, the system generates and refines a comprehensive set of textual descriptions representing policy guidelines. During inference, co-embedding similarity between incoming images and the textual descriptions serves as a reliable signal for policy violation detection, enabling efficient and adaptable ads content moderation. Evaluation results demonstrate the efficacy of this framework in significantly boosting the detection of policy violating content.

cs.CV↗

arXiv2024

Winemaking: Extracting Essential Insights for Efficient Threat Detection in Audit Logs

Advanced Persistent Threats (APTs) are continuously evolving, leveraging their stealthiness and persistence to put increasing pressure on current provenance-based Intrusion Detection Systems (IDS). This evolution exposes several critical issues: (1) The dense interaction between malicious and benign nodes within provenance graphs introduces neighbor noise, hindering effective detection; (2) The complex prediction mechanisms of existing APTs detection models lead to the insufficient utilization of prior knowledge embedded in the data; (3) The high computational cost makes detection impractical. To address these challenges, we propose Winemaking, a lightweight threat detection system built on a knowledge distillation framework, capable of node-level detection within audit log provenance graphs. Specifically, Winemaking applies graph Laplacian regularization to reduce neighbor noise, obtaining smoothed and denoised graph signals. Subsequently, Winemaking employs a teacher model based on GNNs to extract knowledge, which is then distilled into a lightweight student model. The student model is designed as a trainable combination of a feature transformation module and a personalized PageRank random walk label propagation module, with the former capturing feature knowledge and the latter learning label and structural knowledge. After distillation, the student model benefits from the knowledge of the teacher model to perform precise threat detection. We evaluate Winemaking through extensive experiments on three public datasets and compare its performance against several state-of-the-art IDS solutions. The results demonstrate that Winemaking achieves outstanding detection accuracy across all scenarios and the detection time is 1.4 to 5.2 times faster than the current state-of-the-art methods.

cs.CR↗

arXiv2024

Advertiser Content Understanding via LLMs for Google Ads Safety

Ads Content Safety at Google requires classifying billions of ads for Google Ads content policies. Consistent and accurate policy enforcement is important for advertiser experience and user safety and it is a challenging problem, so there is a lot of value for improving it for advertisers and users. Inconsistent policy enforcement causes increased policy friction and poor experience with good advertisers, and bad advertisers exploit the inconsistency by creating multiple similar ads in the hope that some will get through our defenses. This study proposes a method to understand advertiser's intent for content policy violations, using Large Language Models (LLMs). We focus on identifying good advertisers to reduce content over-flagging and improve advertiser experience, though the approach can easily be extended to classify bad advertisers too. We generate advertiser's content profile based on multiple signals from their ads, domains, targeting info, etc. We then use LLMs to classify the advertiser content profile, along with relying on any knowledge the LLM has of the advertiser, their products or brand, to understand whether they are likely to violate a certain policy or not. After minimal prompt tuning our method was able to reach 95\% accuracy on a small test set.

cs.IR↗

arXiv2024

Scaling Up LLM Reviews for Google Ads Content Moderation

Large language models (LLMs) are powerful tools for content moderation, but their inference costs and latency make them prohibitive for casual use on large datasets, such as the Google Ads repository. This study proposes a method for scaling up LLM reviews for content moderation in Google Ads. First, we use heuristics to select candidates via filtering and duplicate removal, and create clusters of ads for which we select one representative ad per cluster. We then use LLMs to review only the representative ads. Finally, we propagate the LLM decisions for the representative ads back to their clusters. This method reduces the number of reviews by more than 3 orders of magnitude while achieving a 2x recall compared to a baseline non-LLM model. The success of this approach is a strong function of the representations used in clustering and label propagation; we found that cross-modal similarity representations yield better results than uni-modal representations.

cs.IR↗

arXiv2023

Benchmarking Robustness to Adversarial Image Obfuscations

Automated content filtering and moderation is an important tool that allows online platforms to build striving user communities that facilitate cooperation and prevent abuse. Unfortunately, resourceful actors try to bypass automated filters in a bid to post content that violate platform policies and codes of conduct. To reach this goal, these malicious actors may obfuscate policy violating images (e.g. overlay harmful images by carefully selected benign images or visual patterns) to prevent machine learning models from reaching the correct decision. In this paper, we invite researchers to tackle this specific issue and present a new image benchmark. This benchmark, based on ImageNet, simulates the type of obfuscations created by malicious actors. It goes beyond ImageNet-$\textrm{C}$ and ImageNet-$\bar{\textrm{C}}$ by proposing general, drastic, adversarial modifications that preserve the original content intent. It aims to tackle a more common adversarial threat than the one considered by $\ell_p$-norm bounded adversaries. We evaluate 33 pretrained models on the benchmark and train models with different augmentations, architectures and training methods on subsets of the obfuscations to measure generalization. We hope this benchmark will encourage researchers to test their models and methods and try to find new approaches that are more robust to these obfuscations.

cs.CV↗

arXiv2021

Control Co-Design for Buoyancy-Controlled MHK Turbine: A Nested Optimization of Geometry and Spatial-Temporal Path Planning

Recent research progress has confirmed that using advanced controls can result in massive increases in energy capture for marine hydrokinetic (MHK) energy systems, including ocean current turbines (OCTs) and wave energy converters (WECs); however, to realize maximum benefits, the controls, power-take-off system, and basic structure of the device must all be co-designed from early stages. This paper presents an OCT turbine control co-design framework, accounting for the plant geometry and spatial-temporal path planning to optimize the performance. Developing a control co-design framework means that it is now possible to evaluate the effects of changing plant geometry on a level playing field when accounting for the OCT plant power optimization. The investigated framework evaluates the key design parameters, including the sizes of the generator, rotor, and variable buoyancy tank in the OCT system, and formulates these parameters' effect on the OCT model and harnessed power through defining a power-to-weight ratio, subject to the design and operational constraints. The control co-design is formulated as a nested optimization problem, where the outer loop optimizes the plant geometry and the inner loop accounts for the spatial-temporal path planning to optimize the harnessed power with respect to the linear model of the OCT system and ocean current uncertainties. Compared with a baseline design, results verify the efficacy of the proposed framework in co-designing an optimal OCT system to gain the maximum power-to-weight ratio.

eess.SY↗

arXiv2018

Continuous Circadian Phase Estimation Using Adaptive Notch Filter

Actigraphy has been widely used for the analysis of circadian rhythm. Current practice applies regression analysis to data from multiple days to estimate the circadian phase. This paper presents a filtering method for online processing of biometric data to estimate the circadian phase. We apply the proposed method on actigraphy data of fruit flies (Drosophila melanogaster).

eess.SY↗