arXiv subjects

Olga Fink

Publications and source records attributed to Olga Fink.

At least 127 records · Page 7Linked to original sources

arXiv2022

Continual Test-Time Domain Adaptation

Test-time domain adaptation aims to adapt a source pre-trained model to a target domain without using any source data. Existing works mainly consider the case where the target domain is static. However, real-world machine perception systems are running in non-stationary and continually changing environments where the target domain distribution can change over time. Existing methods, which are mostly based on self-training and entropy regularization, can suffer from these non-stationary environments. Due to the distribution shift over time in the target domain, pseudo-labels become unreliable. The noisy pseudo-labels can further lead to error accumulation and catastrophic forgetting. To tackle these issues, we propose a continual test-time adaptation approach~(CoTTA) which comprises two parts. Firstly, we propose to reduce the error accumulation by using weight-averaged and augmentation-averaged predictions which are often more accurate. On the other hand, to avoid catastrophic forgetting, we propose to stochastically restore a small part of the neurons to the source pre-trained weights during each iteration to help preserve source knowledge in the long-term. The proposed method enables the long-term adaptation for all parameters in the network. CoTTA is easy to implement and can be readily incorporated in off-the-shelf pre-trained models. We demonstrate the effectiveness of our approach on four classification tasks and a segmentation task for continual test-time adaptation, on which we outperform existing methods. Our code is available at \url{https://qin.ee/cotta}.

cs.CV↗

arXiv2022

Learnable Wavelet Packet Transform for Data-Adapted Spectrograms

Capturing high-frequency data concerning the condition of complex systems, e.g. by acoustic monitoring, has become increasingly prevalent. Such high-frequency signals typically contain time dependencies ranging over different time scales and different types of cyclic behaviors. Processing such signals requires careful feature engineering, particularly the extraction of meaningful time-frequency features. This can be time-consuming and the performance is often dependent on the choice of parameters. To address these limitations, we propose a deep learning framework for learnable wavelet packet transforms, enabling to learn features automatically from data and optimise them with respect to the defined objective function. The learned features can be represented as a spectrogram, containing the important time-frequency information of the dataset. We evaluate the properties and performance of the proposed approach by evaluating its improved spectral leakage and by applying it to an anomaly detection task for acoustic monitoring.

cs.SD↗

arXiv2022

A Prescriptive Dirichlet Power Allocation Policy with Deep Reinforcement Learning

Prescribing optimal operation based on the condition of the system and, thereby, potentially prolonging the remaining useful lifetime has a large potential for actively managing the availability, maintenance and costs of complex systems. Reinforcement learning (RL) algorithms are particularly suitable for this type of problems given their learning capabilities. A special case of a prescriptive operation is the power allocation task, which can be considered as a sequential allocation problem, where the action space is bounded by a simplex constraint. A general continuous action-space solution of such sequential allocation problems has still remained an open research question for RL algorithms. In continuous action-space, the standard Gaussian policy applied in reinforcement learning does not support simplex constraints, while the Gaussian-softmax policy introduces a bias during training. In this work, we propose the Dirichlet policy for continuous allocation tasks and analyze the bias and variance of its policy gradients. We demonstrate that the Dirichlet policy is bias-free and provides significantly faster convergence, better performance and better hyperparameters robustness over the Gaussian-softmax policy. Moreover, we demonstrate the applicability of the proposed algorithm on a prescriptive operation case, where we propose the Dirichlet power allocation policy and evaluate the performance on a case study of a set of multiple lithium-ion (Li-I) battery systems. The experimental results show the potential to prescribe optimal operation, improve the efficiency and sustainability of multi-power source systems.

cs.LG↗

arXiv2021

Safe multi-agent deep reinforcement learning for joint bidding and maintenance scheduling of generation units

This paper proposes a safe reinforcement learning algorithm for generation bidding decisions and unit maintenance scheduling in a competitive electricity market environment. In this problem, each unit aims to find a bidding strategy that maximizes its revenue while concurrently retaining its reliability by scheduling preventive maintenance. The maintenance scheduling provides some safety constraints which should be satisfied at all times. Satisfying the critical safety and reliability constraints while the generation units have an incomplete information of each others' bidding strategy is a challenging problem. Bi-level optimization and reinforcement learning are state of the art approaches for solving this type of problems. However, neither bi-level optimization nor reinforcement learning can handle the challenges of incomplete information and critical safety constraints. To tackle these challenges, we propose the safe deep deterministic policy gradient reinforcement learning algorithm which is based on a combination of reinforcement learning and a predicted safety filter. The case study demonstrates that the proposed approach can achieve a higher profit compared to other state of the art methods while concurrently satisfying the system safety constraints.

eess.SY↗

arXiv2021

Fully Learnable Deep Wavelet Transform for Unsupervised Monitoring of High-Frequency Time Series

High-Frequency (HF) signals are ubiquitous in the industrial world and are of great use for monitoring of industrial assets. Most deep learning tools are designed for inputs of fixed and/or very limited size and many successful applications of deep learning to the industrial context use as inputs extracted features, which is a manually and often arduously obtained compact representation of the original signal. In this paper, we propose a fully unsupervised deep learning framework that is able to extract a meaningful and sparse representation of raw HF signals. We embed in our architecture important properties of the fast discrete wavelet transformation (FDWT) such as (1) the cascade algorithm, (2) the conjugate quadrature filter property that links together the wavelet, the scaling and transposed filter functions, and (3) the coefficient denoising. Using deep learning, we make this architecture fully learnable: both the wavelet bases and the wavelet coefficient denoising are learnable. To achieve this objective, we propose a new activation function that performs a learnable hard-thresholding of the wavelet coefficients. With our framework, the denoising FDWT becomes a fully learnable unsupervised tool that does neither require any type of pre- nor post-processing, nor any prior knowledge on wavelet transform. We demonstrate the benefits of embedding all these properties on three machine-learning tasks performed on open source sound datasets. We perform an ablation study of the impact of each property on the performance of the architecture, achieve results well above baseline and outperform other state-of-the-art methods.

cs.LG↗

arXiv2021

Integrating Expert Knowledge with Domain Adaptation for Unsupervised Fault Diagnosis

Data-driven fault diagnosis methods often require abundant labeled examples for each fault type. On the contrary, real-world data is often unlabeled and consists of mostly healthy observations and only few samples of faulty conditions. The lack of labels and fault samples imposes a significant challenge for existing data-driven fault diagnosis methods. In this paper, we aim to overcome this limitation by integrating expert knowledge with domain adaptation in a synthetic-to-real framework for unsupervised fault diagnosis. Motivated by the fact that domain experts often have a relatively good understanding on how different fault types affect healthy signals, in the first step of the proposed framework, a synthetic fault dataset is generated by augmenting real vibration samples of healthy bearings. This synthetic dataset integrates expert knowledge and encodes class information about the fault types. However, models trained solely based on the synthetic data often do not perform well because of the distinct distribution difference between the synthetically generated and real faults. To overcome this domain gap between the synthetic and real data, in the second step of the proposed framework, an imbalance-robust domain adaptation~(DA) approach is proposed to adapt the model from synthetic faults~(source) to the unlabeled real faults~(target) which suffer from severe class imbalance. The framework is evaluated on two unsupervised fault diagnosis cases for bearings, the CWRU laboratory dataset and a real-world wind-turbine dataset. Experimental results demonstrate that the generated faults are effective for encoding fault type information and the domain adaptation is robust against the different levels of class imbalance between faults.

eess.SP↗

arXiv2021

Domain Adaptive Semantic Segmentation with Self-Supervised Depth Estimation

Domain adaptation for semantic segmentation aims to improve the model performance in the presence of a distribution shift between source and target domain. Leveraging the supervision from auxiliary tasks~(such as depth estimation) has the potential to heal this shift because many visual tasks are closely related to each other. However, such a supervision is not always available. In this work, we leverage the guidance from self-supervised depth estimation, which is available on both domains, to bridge the domain gap. On the one hand, we propose to explicitly learn the task feature correlation to strengthen the target semantic predictions with the help of target depth estimation. On the other hand, we use the depth prediction discrepancy from source and target depth decoders to approximate the pixel-wise adaptation difficulty. The adaptation difficulty, inferred from depth, is then used to refine the target semantic segmentation pseudo-labels. The proposed method can be easily implemented into existing segmentation frameworks. We demonstrate the effectiveness of our approach on the benchmark tasks SYNTHIA-to-Cityscapes and GTA-to-Cityscapes, on which we achieve the new state-of-the-art performance of $55.0\%$ and $56.6\%$, respectively. Our code is available at \url{https://qin.ee/corda}.

cs.CV↗

arXiv2021

Canonical Polyadic Decomposition and Deep Learning for Machine Fault Detection

Acoustic monitoring for machine fault detection is a recent and expanding research path that has already provided promising results for industries. However, it is impossible to collect enough data to learn all types of faults from a machine. Thus, new algorithms, trained using data from healthy conditions only, were developed to perform unsupervised anomaly detection. A key issue in the development of these algorithms is the noise in the signals, as it impacts the anomaly detection performance. In this work, we propose a powerful data-driven and quasi non-parametric denoising strategy for spectral data based on a tensor decomposition: the Non-negative Canonical Polyadic (CP) decomposition. This method is particularly adapted for machine emitting stationary sound. We demonstrate in a case study, the Malfunctioning Industrial Machine Investigation and Inspection (MIMII) baseline, how the use of our denoising strategy leads to a sensible improvement of the unsupervised anomaly detection. Such approaches are capable to make sound-based monitoring of industrial processes more reliable.

stat.ML↗

arXiv2021

Artificial intelligence across company borders

Artificial intelligence (AI) has become a valued technology in many companies. At the same time, a substantial potential for utilizing AI \emph{across} company borders has remained largely untapped. An inhibiting factor concerns disclosure of data to external parties, which raises legitimate concerns about intellectual property rights, privacy issues, and cybersecurity risks. Combining federated learning with domain adaptation can provide a solution to this problem by enabling effective cross-company AI without data disclosure. In this Viewpoint, we discuss the use, value, and implications of this approach in a cross-company setting.

cs.CY↗

arXiv2021

Maintenance scheduling of manufacturing systems based on optimal price of the network

Goods can exhibit positive externalities impacting decisions of customers in socials networks. Suppliers can integrate these externalities in their pricing strategies to increase their revenue. Besides optimizing the prize, suppliers also have to consider their production and maintenance costs. Predictive maintenance has the potential to reduce the maintenance costs and improve the system availability. To address the joint optimization of pricing with network externalities and predictive maintenance scheduling based on the condition of the system, we propose a bi-level optimization solution based on game theory. In the first level, the manufacturing company decides about the predictive maintenance scheduling of the units and the price of the goods. In the second level, the customers decide about their consumption using an optimization approach in which the objective function depends on their consumption, the consumption levels of other customers who are connected through the graph, and the price of the network which is determined by the supplier. To solve the problem, we propose the leader-multiple-followers game where the supplier as a leader predicts the strategies of the followers. Then, customers as the followers obtain their strategies based on the leader's and other followers' strategies. We demonstrate the effectiveness of our proposed method on a simulated case study. The results demonstrate that knowledge of the social network graph results in an increased revenue compared to the case when the underlying social network graph is not known. Moreover, the results demonstrate that obtaining the predictive maintenance scheduling based on the proposed optimization approach leads to an increased profit compared to the baseline decision-making (perform maintenance at the degradation limit).

cs.MA↗

arXiv2021

Uncertainty-aware Remaining Useful Life predictor

Remaining Useful Life (RUL) estimation is the problem of inferring how long a certain industrial asset can be expected to operate within its defined specifications. Deploying successful RUL prediction methods in real-life applications is a prerequisite for the design of intelligent maintenance strategies with the potential of drastically reducing maintenance costs and machine downtimes. In light of their superior performance in a wide range of engineering fields, Machine Learning (ML) algorithms are natural candidates to tackle the challenges involved in the design of intelligent maintenance systems. In particular, given the potentially catastrophic consequences or substantial costs associated with maintenance decisions that are either too late or too early, it is desirable that ML algorithms provide uncertainty estimates alongside their predictions. However, standard data-driven methods used for uncertainty estimation in RUL problems do not scale well to large datasets or are not sufficiently expressive to model the high-dimensional mapping from raw sensor data to RUL estimates. In this work, we consider Deep Gaussian Processes (DGPs) as possible solutions to the aforementioned limitations. We perform a thorough evaluation and comparison of several variants of DGPs applied to RUL predictions. The performance of the algorithms is evaluated on the N-CMAPSS (New Commercial Modular Aero-Propulsion System Simulation) dataset from NASA for aircraft engines. The results show that the proposed methods are able to provide very accurate RUL predictions along with sensible uncertainty estimates, providing more reliable solutions for (safety-critical) real-life industrial applications.

cs.LG↗

arXiv2021

Interpretable Detection of Partial Discharge in Power Lines with Deep Learning

Partial discharge (PD) is a common indication of faults in power systems, such as generators, and cables. These PD can eventually result in costly repairs and substantial power outages. PD detection traditionally relies on hand-crafted features and domain expertise to identify very specific pulses in the electrical current, and the performance declines in the presence of noise or of superposed pulses. In this paper, we propose a novel end-to-end framework based on convolutional neural networks. The framework has two contributions. First, it does not require any feature extraction and enables robust PD detection. Second, we devise the pulse activation map. It provides interpretability of the results for the domain experts with the identification of the pulses that led to the detection of the PDs. The performance is evaluated on a public dataset for the detection of damaged power lines. An ablation study demonstrates the benefits of each part of the proposed framework.

eess.SP↗

arXiv2021

Temporal signals to images: Monitoring the condition of industrial assets with deep learning image processing algorithms

The ability to detect anomalies in time series is considered highly valuable in numerous application domains. The sequential nature of time series objects is responsible for an additional feature complexity, ultimately requiring specialized approaches in order to solve the task. Essential characteristics of time series, situated outside the time domain, are often difficult to capture with state-of-the-art anomaly detection methods when no transformations have been applied to the time series. Inspired by the success of deep learning methods in computer vision, several studies have proposed transforming time series into image-like representations, used as inputs for deep learning models, and have led to very promising results in classification tasks. In this paper, we first review the signal to image encoding approaches found in the literature. Second, we propose modifications to some of their original formulations to make them more robust to the variability in large datasets. Third, we compare them on the basis of a common unsupervised task to demonstrate how the choice of the encoding can impact the results when used in the same deep learning architecture. We thus provide a comparison between six encoding algorithms with and without the proposed modifications. The selected encoding methods are Gramian Angular Field, Markov Transition Field, recurrence plot, grey scale encoding, spectrogram, and scalogram. We also compare the results achieved with the raw signal used as input for another deep learning model. We demonstrate that some encodings have a competitive advantage and might be worth considering within a deep learning framework. The comparison is performed on a dataset collected and released by Airbus SAS, containing highly complex vibration measurements from real helicopter flight tests. The different encodings provide competitive results for anomaly detection.

cs.LG↗

arXiv2021

Decision Support System for an Intelligent Operator of Utility Tunnel Boring Machines

In tunnel construction projects, delays induce high costs. Thus, tunnel boring machines (TBM) operators aim for fast advance rates, without safety compromise, a difficult mission in uncertain ground environments. Finding the optimal control parameters based on the TBM sensors' measurements remains an open research question with large practical relevance. In this paper, we propose an intelligent decision support system developed in three steps. First past projects performances are evaluated with an optimality score, taking into account the advance rate and the working pressure safety. Then, a deep learning model learns the mapping between the TBM measurements and this optimality score. Last, in real application, the model provides incremental recommendations to improve the optimality, taking into account the current setting and measurements of the TBM. The proposed approach is evaluated on real micro-tunnelling project and demonstrates great promises for future projects.

eess.SY↗

arXiv2020

Battery Model Calibration with Deep Reinforcement Learning

Lithium-Ion (Li-I) batteries have recently become pervasive and are used in many physical assets. To enable a good prediction of the end of discharge of batteries, detailed electrochemical Li-I battery models have been developed. Their parameters are typically calibrated before they are taken into operation and are typically not re-calibrated during operation. However, since battery performance is affected by aging, the reality gap between the computational battery models and the real physical systems leads to inaccurate predictions. A supervised machine learning algorithm would require an extensive representative training dataset mapping the observation to the ground truth calibration parameters. This may be infeasible for many practical applications. In this paper, we implement a Reinforcement Learning-based framework for reliably and efficiently inferring calibration parameters of battery models. The framework enables real-time inference of the computational model parameters in order to compensate the reality-gap from the observations. Most importantly, the proposed methodology does not need any labeled data samples, (samples of observations and the ground truth calibration parameters). Furthermore, the framework does not require any information on the underlying physical model. The experimental results demonstrate that the proposed methodology is capable of inferring the model parameters with high accuracy and high robustness. While the achieved results are comparable to those obtained with supervised machine learning, they do not rely on the ground truth information during training.

cs.LG↗

arXiv2020

Distributed joint dynamic maintenance and production scheduling in manufacturing systems: Framework based on model predictive control and Benders decomposition

Scheduling the maintenance based on the condition, respectively the degradation level of the system leads to improved system's reliability while minimizing the maintenance cost. Since the degradation level changes dynamically during the system's operation, we face a dynamic maintenance scheduling problem. In this paper, we address the dynamic maintenance scheduling of manufacturing systems based on their degradation level. The manufacturing system consists of several units with a defined capacity and an individual dynamic degradation model, seeking to optimize their reward. The units sell their production capacity, while maintaining the systems based on the degradation state to prevent failures. The manufacturing units are jointly responsible for fulfilling the demand of the system. This induces a coupling constraint among the agents. Hence, we face a large-scale mixed-integer dynamic maintenance scheduling problem. In order to handle the dynamic model of the system and large-scale optimization, we propose a distributed algorithm using model predictive control (MPC) and Benders decomposition method. In the proposed algorithm, first, the master problem obtains the maintenance scheduling for all the agents, and then based on this data, the agents obtain their optimal production using the distributed MPC method which employs the dual decomposition approach to tackle the coupling constraints among the agents. The effectiveness of the proposed method is investigated on a case study.

cs.DC↗

arXiv2020

Unsupervised Transfer Learning for Anomaly Detection: Application to Complementary Operating Condition Transfer

Anomaly Detectors are trained on healthy operating condition data and raise an alarm when the measured samples deviate from the training data distribution. This means that the samples used to train the model should be sufficient in quantity and representative of the healthy operating conditions. But for industrial systems subject to changing operating conditions, acquiring such comprehensive sets of samples requires a long collection period and delay the point at which the anomaly detector can be trained and put in operation. A solution to this problem is to perform unsupervised transfer learning (UTL), to transfer complementary data between different units. In the literature however, UTL aims at finding common structure between the datasets, to perform clustering or dimensionality reduction. Yet, the task of transferring and combining complementary training data has not been studied. Our proposed framework is designed to transfer complementary operating conditions between different units in a completely unsupervised way to train more robust anomaly detectors. It differs, thereby, from other unsupervised transfer learning works as it focuses on a one-class classification problem. The proposed methodology enables to detect anomalies in operating conditions only experienced by other units. The proposed end-to-end framework uses adversarial deep learning to ensure alignment of the different units' distributions. The framework introduces a new loss, inspired by a dimensionality reduction tool, to enforce the conservation of the inherent variability of each dataset, and uses state-of-the art once-class approach to detect anomalies. We demonstrate the benefit of the proposed framework using three open source datasets.

cs.LG↗

arXiv2020

Fusing Physics-based and Deep Learning Models for Prognostics

Physics-based and data-driven models for remaining useful lifetime (RUL) prediction typically suffer from two major challenges that limit their applicability to complex real-world domains: (1) incompleteness of physics-based models and (2) limited representativeness of the training dataset for data-driven models. Combining the advantages of these two directions while overcoming some of their limitations, we propose a novel hybrid framework for fusing the information from physics-based performance models with deep learning algorithms for prognostics of complex safety-critical systems under real-world scenarios. In the proposed framework, we use physics-based performance models to infer unobservable model parameters related to a system's components health solving a calibration problem. These parameters are subsequently combined with sensor readings and used as input to a deep neural network to generate a data-driven prognostics model with physics-augmented features. The performance of the hybrid framework is evaluated on an extensive case study comprising run-to-failure degradation trajectories from a fleet of nine turbofan engines under real flight conditions. The experimental results show that the hybrid framework outperforms purely data-driven approaches by extending the prediction horizon by nearly 127\%. Furthermore, it requires less training data and is less sensitive to the limited representativeness of the dataset compared to purely data-driven approaches.

eess.SY↗