We are pleased to welcome Dr. Shuaiqi Yuan, who recently joined our group as a postdoctoral research scholar. Dr. Yuan holds a PhD in Safety and Security Science from Delft University of Technology in the Netherlands.
The demand for disruption-free fault diagnosis of mechanical equipment under a constantly changing operation environment poses a great challenge to the deployment of data-driven diagnosis models in practice. Extant continual learning-based diagnosis models suffer from consuming a large number of labeled samples to be trained for adapting to new diagnostic tasks and failing to account for the diagnosis of heterogeneous fault types across different machines. In this paper, we use a representative mechanical equipment – rotating machinery — as an example and develop an uncertainty-aware continual learning framework (UACLF) to provide a unified interface for fault diagnosis of rotating machinery under various dynamic scenarios: class continual scenario, domain continual scenario, and both. The proposed UACLF takes a three-step to tackle fault diagnosis of rotating machinery with homogeneous-heterogeneous faults under dynamic environments. In the first step, an inter-class classification loss function and an intra-class discrimination loss function are devised to extract informative feature representations from the raw vibration signal for fault classification. Next, an uncertainty-aware pseudo labeling mechanism is developed to select unlabeled fault samples that we are able to assign pseudo labels confidently, thus expanding the training samples for faults arising in the new environment. Thirdly, an adaptive prototypical feedback mechanism is used to enhance the decision boundary of fault classification and diminish the model misclassification rate. Experimental results on three datasets suggest that the proposed UACLF outperforms several alternatives in the literature on fault diagnosis of rotating machinery across various working conditions and different machines.
Large language models (LLMs) like GPT-4 or LlaMa 3 provide superior performance in complex human-like interactions. But they are costly, or too large for edge devices such as smartphones and harder to self-host, leading to security and privacy concerns. This paper introduces a novel interpretable knowledge distillation approach to enhance the performance of smaller, more economical LLMs that firms can self-host. We study this problem in the context of building a customer service agent aimed at achieving high customer satisfaction through goal-oriented dialogues. Unlike traditional knowledge distillation, where the “student” model learns directly from the “teacher” model’s responses via fine-tuning, our interpretable “strategy” teaching approach involves the teacher providing strategies to improve the student’s performance in various scenarios. This method alternates between a “scenario generation” step and a “strategies for improvement” step, creating a customized library of scenarios and optimized strategies for automated prompting. The method requires only black-box access to both student and teacher models; hence it can be used without manipulating model parameters. In our customer service application, the method improves performance, and the learned strategies are transferable to other LLMs and scenarios beyond the training set. The method’s interpretabilty helps safeguard against potential harms through human audit.
Establishing trustworthiness is fundamental for the responsible utilization of medical artificial intelligence (AI), particularly in cancer diagnostics, where misdiagnosis can lead to devastating consequences. However, there is currently a lack of systematic approaches to resolve the reliability challenges stemming from the model limitations and the unpredictable variability in the application domain. In this work, we address trustworthiness from two complementary aspects—data trustworthiness and model trustworthiness—in the task of subtyping non-small cell lung cancers using whole side images. We introduce TRUECAM, a framework that provides trustworthiness-focused, uncertainty-aware, end-to-end cancer diagnosis with model-agnostic capabilities by leveraging spectral-normalized neural Gaussian Process (SNGP) and conformal prediction (CP) to simultaneously ensure data and model trustworthiness. Specifically, SNGP enables the identification of inputs beyond the scope of trained models, while CP offers a statistical validity guarantee for models to contain correct classification. Systematic experiments performed on both internal and external cancer cohorts, utilizing a widely adopted specialized model and two foundation models, indicate that TRUECAM achieves significant improvements in classification accuracy, robustness, fairness, and data efficiency (i.e., selectively identifying and utilizing only informative tiles for classification). These highlight TRUECAM as a general wrapper framework around medical AI of different sizes, architectures, purposes, and complexities to enable their responsible use.
In recent years, deep learning-enabled systems have made remarkable progress, powering a surge in advanced intelligent applications. This growth and its real-world impact have been further amplified by the advent of large foundation models (e.g., LLM, Stable Diffusion). Yet, the rapid evolution of these AI systems often proceeds without comprehensive quality assurance and engineering support. This gap is evident in the integration of standards for quality, reliability, and safety assurance, as well as the need for mature toolchain support that provides systematic and explainable feedback of the development lifecycle. In this talk, I will present a high-level overview of our team’s ongoing initiatives to lay the groundwork for Trustworthy Assurance of AI Systems and its industrial applications, e.g., including (1) AI software testing and analysis, (2) our latest trustworthiness assurance efforts for AI-driven Cyber-physical systems with an emphasis on sim2real transition. (3) risk and safety assessment for large foundational models, including those akin to large language models, and vision transformers.
Typically, organizations lose around five percent of their revenue to fraud. In this presentation, we explore advanced AI techniques to address this issue. Drawing on our recent research, we begin by examining cost-sensitive fraud detection methods, such as CS-Logit which integrates the economic imbalances inherent in fraud detection into the optimization of AI models. We then move on to data engineering strategies that enhance the predictive capabilities of both the data and AI models through intelligent instance and feature engineering. We also delve into network data, showcasing our innovative research methods like Gotcha and CATCHM for effective data featurization. A significant focus is placed on Explainable AI (XAI), which demystifies high-performance AI models used in fraud detection, aiding in the development of effective fraud prevention strategies. We provide practical examples from various sectors including credit card fraud, anti-money laundering, insurance fraud, tax evasion, and payment transaction fraud. Furthermore, we discuss the overarching issue of model risk, which encompasses everything from data input to AI model deployment. Throughout the presentation, the speaker will thoroughly discuss his recent research, conducted in partnership with leading global financial institutions such as BNP Paribas Fortis, Allianz, ING, and Ageas.
The group for risk, reliability, and resilience informatics warmly welcomes the following PhD students to join the team:
Tao Wang, and Xinru Zhang.
The malfunction of deep learning systems in safety-critical applications (e.g., aerospace) will lead to devastating outcomes. Thus, how to ensure the trustworthiness of deep learning systems in high-stakes decision settings is an imperative problem to be tackled. This project aims at accommodating heterogeneous sources of risks pertaining to the input data of deep learning systems in the open world, multi-source uncertainty inherent in the model reliability for individual prediction as well as the coupled relationship between input data-related risk and the model reliability tailored to each individual prediction towards devising an uncertainty quantification-based method for trustworthiness modeling of deep learning systems. We believe that the proposed effort will make substantial contributions to the development of novel and effective theories and models for enhancing the trustworthiness of deep learning systems, offer new insights for the trustworthiness modeling of deep learning systems in the open environment, and boost the advancement of trustworthy deep learning systems.
The conventional aggregated performance measure (i.e., mean squared error) with respect to the whole dataset would not provide desired safety and quality assurance for each individual prediction made by a machine learning model in risk-sensitive regression problems. In this paper, we propose an informative indicator $\mathcal{R} \left(\bm{x} \right)$ to quantify model reliability for individual prediction (MRIP) for the purpose of safeguarding the usage of machine learning (ML) models in mission-critical applications. Specifically, we define the reliability of a ML model with respect to its prediction on each individual input $\bm{x}$ as the probability of the observed difference between the prediction of ML model and the actual observation falling within a small interval when the input $\bm{x}$ varies within a small range subject to a preset distance constraint, namely $\mathcal{R}(\bm{x}) = P(|y^* – \hat{y}^*| \leq \varepsilon | \bm{x}^* \in B(\bm{x}))$, where $y^*$ denotes the observed target value for the input $\bm{x}^*$, $\hat{y}^*$ denotes the model prediction for the input $\bm{x}^*$, and $\bm{x}^*$ is an input in the neighborhood of $\bm{x}$ subject to the constraint $B\left( \bm{x} \right) = \left\{ {\left. {{\bm{x}^*}} \right|\left\| {{\bm{x}^*} – \bm{x}} \right\| \le \delta } \right\}$. The developed MRIP indicator $\mathcal{R} \left(\bm{x} \right)$ provides a direct, objective, quantitative, and general-purpose measure of “reliability” or the probability of success of the ML model for each individual prediction by fully exploiting the local information associated with the input $\bm{x}$ and ML model. Next, to mitigate the intensive computational effort involved in MRIP estimation, we develop a two-stage ML-based framework to directly learn the relationship between $\bm{x}$ and its MRIP $\mathcal{R} \left( \bm{x} \right)$, thus enabling to provide the reliability estimate $\mathcal{R} \left( \bm{x} \right)$ for any unseen input instantly. Thirdly, we propose an information gain-based approach to help determine a threshold value pertaing to $\mathcal{R} \left( \bm{x} \right)$ in support of decision makings on when to accept or abstain from counting on the ML model prediction. Comprehensive computational experiments and quantitative comparisons with existing methods on a broad range of real-world datasets reveal that the developed ML-based framework for MRIP estimation shows a robust performance in improving the reliability estimate of individual prediction, and the MRIP indicator $\mathcal{R} \left( \bm{x} \right)$ thus provides an essential layer of safety net when adopting ML models in risk-sensitive environments.
Blind deconvolution (BD) has been demonstrated to be an efficacious approach for extracting bearing fault-specific features from vibration signals under strong background noise. Despite BD’s appealing feature in adaptability and mathematical interpretability, a significant challenge persists: \textit{How to effectively integrate BD with fault-diagnosing classifiers?} This issue is intricate to be tackled because the traditional BD method is solely designed for feature extraction with its own optimizer and objective function. When BD is combined with the downstream deep learning classifier, the different learning objectives easily get in conflict. To address this problem, this paper introduces classifier-guided BD (ClassBD) for joint learning of BD-based feature extraction and deep learning-based fault diagnosis. Towards this goal, we first develop a time and frequency neural BD that employs neural networks to implement conventional BD, thereby facilitating seamless integration of BD and the deep learning classifier for co-optimization of model parameters. In the neural BD, we incorporate two filters: i) a time domain quadratic filter to utilize quadratic convolutional networks for extracting periodic impulses; ii) a frequency domain linear filter composed of a fully-connected neural network to amplify discrete frequency components. Next, we develop a unified framework built upon a deep learning classifier to guide the learning of BD filters. In addition, we devise a physics-informed loss function composed of kurtosis, $l_2/l_4$ norm, and a cross-entropy loss to jointly optimize the BD filters and deep learning classifier. In so doing, the fault labels are fully exploited to direct BD to extract features in distinguishing classes amidst strong noise. To the best of our knowledge, this is the first of its kind that BD is successfully applied to bearing fault diagnosis. Experimental results from three different datasets highlight that ClassBD outperforms other state-of-the-art methods under noisy conditions. The source codes of this paper are available at https://github.com/asdvfghg/ClassBD.
