Unsupervised Scheme Research Articles

Exploiting hierarchical dependence structures for unsupervised rank fusion in information retrieval

The goal of rank fusion in information retrieval (IR) is to deliver a single output list from multiple search results. Improving performance by combining the outputs of various IR systems is a challenging task. A central point is the fact that many non-obvious factors are involved in the estimation of relevance, inducing nonlinear interrelations between the data. The ability to model complex dependency relationships between random variables has become increasingly popular in the realm of information retrieval, and the need to further explore these dependencies for data fusion has been recently acknowledged. Copulas provide a framework to separate the dependence structure from the margins. Inspired by the theory of copulas, we propose a new unsupervised, dynamic, nonlinear, rank fusion method, based on a nested composition of non-algebraic function pairs. The dependence structure of the model is tailored by leveraging query-document correlations on a per-query basis. We experimented with three topic sets over CLEF corpora fusing 3 and 6 retrieval systems, comparing our method against the CombMNZ technique and other nonlinear unsupervised strategies. The experiments show that our fusion approach improves performance under explicit conditions, providing insight about the circumstances under which linear fusion techniques have comparable performance to nonlinear methods.

Mill condition monitoring based on instantaneous identification of specific force coefficients under variable cutting conditions

Following the necessity of increased performance and availability requirements for manufacturing systems, research is becoming more and more attracted by monitoring solutions for cutting tools. In this paper, a robust unsupervised strategy for milling tool wear monitoring under variable process parameters and lubrication conditions is presented. The proposed method is completely unsupervised, thus not requiring any kind of training procedure, and is validated on different machine tools. The solution is based upon the online estimation of specific force coefficients (SFC) from instantaneous cutting forces in high-feed milling of Ti6Al4V workpiece. This avoids the need for continuously variable feed per tooth during cutting tests, necessitated for the application of reference literature approach. For this purpose, a novel high-feed mill mechanistic model was conceived and developed. Five run-to-failures were performed in different lubrication conditions – cryogenic and traditional lubrication – with different cutting speeds (50 m/min, 70 m/min and 125 m/min) on two different machine tools. Principal Component Regression was introduced in order to deal with the variability of the estimated coefficients. Self-starting tabular cusum control charts were implemented and demonstrated high accuracy and reliability in the prediction of notch wear phenomena as well as chipping of tool cutting edges for all the cases considered. The solution detected an out-of-control conditions ranging from 166μm to 499μm of maximum flank wear for the analysed tests. The mean prediction error with respect to the 600μm threshold is of −45% with a peak of −72%, whereas reference literature algorithms reach −57% and −66%, respectively. A sensitivity analysis of control chart threshold was performed with reference to the maximum flank wear at the detection point. In a supervised scenario, the threshold can be increased to obtain a less conservative approach: for instance, a mean prediction error of −41% was reached by doubling the threshold.

Exposing unseen GAN-generated image using unsupervised domain adaptation

Benefiting from the tremendous success of Generative Adversarial Networks (GANs), we are entering a world where it will be increasingly difficult for people to distinguish whether an image is synthetic, which puts new demands on digital image forensics. To resolve this issue, several fake image detectors based on supervised binary classification have been designed. However, current methods remain vulnerable when testing samples are generated by an unseen GAN model. In this work, an unsupervised domain adaptation strategy is introduced to achieve the detection of GAN-generated images in an unknown domain. A Self-Attention block and novel loss function have been constructed to optimize the domain adaptation process, thus obtaining a better generalization. Experimental results demonstrate that the proposed scheme obtains high detection accuracy on the generalization problem and is robust to common post-processing in reality, which shows that unsupervised methods can enhance the generalization ability of detection models to some extent.

An Efficient Feature Extraction Network for Unsupervised Hyperspectral Change Detection

Change detection (CD) in hyperspectral images has become a research hotspot in the field of remote sensing due to the extremely wide spectral range of hyperspectral images compared to traditional remote sensing images. It is challenging to effectively extract features from redundant high-dimensional data for hyperspectral change detection tasks due to the fact that hyperspectral data contain abundant spectral information. In this paper, a novel feature extraction network is proposed, which uses a Recurrent Neural Network (RNN) to mine the spectral information of the input image and combines this with a Convolutional Neural Network (CNN) to fuse the spatial information of hyperspectral data. Finally, the feature extraction structure of hybrid RNN and CNN is used as a building block to complete the change detection task. In addition, we use an unsupervised sample generation strategy to produce high-quality samples for network training. The experimental results demonstrate that the proposed method yields reliable detection results. Moreover, the proposed method has fewer noise regions than the pixel-based method.

An unsupervised fusion network for boosting denoising performance

While many efforts have been devoted to addressing image denoising and achieve continuously improving results during the past few decades, it is fair to say that no a stand-alone method is consistently better than others. Nonetheless, many existing denoising methods, each having a different denoising capability, can yield various but complementary denoised images with respect to specific local areas. To effectively exploit the complementarity and diversity among the denoised images obtained with different denoisers, in this work we fuse them to produce an overall better result, which is fundamental to achieve robust and competitive denoising performance especially for complex scenes. A framework called deep fusion network (DFNet) is proposed to generate a consistent estimation about the final denoised image, taking advantage of the complementarity of denoisers and suppressing the bias. Specifically, given a noisy image, we first exploit a set of representative image denoisers to denoise it respectively, and obtain the corresponding initial denoised images. Then these initial denoised images are concatenated and fed into the proposed DFNet, and the proposed DFNet seeks to adjust its network parameters to produce the fused image (as the final denoised image) with an unsupervised training strategy through minimizing the carefully designed loss function. The experimental results show that our approach outperforms the stand-alone methods as well as the ones using combination strategy by large margin both in objective and subjective evaluations. Compared to the those methods that are relatively close to our strategy, the proposed DFNet is extensible and parameter free, which means it can cope with a variable number of different denoisers and avoid the manual intervention during the fusion process. The proposed DFNet has greater flexibility and better practicality.

Unsupervised Radar Target Detection under Complex Clutter Background Based on Mixture Variational Autoencoder

The clutter background in modern radar target detection is complex and changeable. The performance of classical detectors based on parametric statistical modeling methods is often degraded due to model mismatch. Existing data-driven deep learning methods require cumbersome and expensive annotations. Furthermore, the performance of the detection network is severely degraded when the detection scene changes, since the trained network with the data from one scene is not suitable for another scene with different data distribution. To this end, it is crucial to develop an unsupervised detection method that can finely model complex and changing clutter scenes. This problem is challenging yet rewarding because it completely eliminates the cost of obtaining cumbersome annotations. In this paper, we introduce GM-CVAE, a novel unsupervised Gaussian Mixture Variational Autoencoder with a one-dimensional Convolutional neural network approach to finely model complex and changing clutter. Furthermore, we develop an unsupervised narrow-band radar target detection strategy based on reconstructed likelihood. Comprehensive experiments are carried out to show that the proposed method realizes the refined modeling of clutter and guarantees superior detection performance in the simulated complex clutter environment. Compared with baselines, the proposed method shows better performance.

SimCVD: Simple Contrastive Voxel-Wise Representation Distillation for Semi-Supervised Medical Image Segmentation.

Automated segmentation in medical image analysis is a challenging task that requires a large amount of manually labeled data. However, most existing learning-based approaches usually suffer from limited manually annotated medical data, which poses a major practical problem for accurate and robust medical image segmentation. In addition, most existing semi-supervised approaches are usually not robust compared with the supervised counterparts, and also lack explicit modeling of geometric structure and semantic information, both of which limit the segmentation accuracy. In this work, we present SimCVD, a simple contrastive distillation framework that significantly advances state-of-the-art voxel-wise representation learning. We first describe an unsupervised training strategy, which takes two views of an input volume and predicts their signed distance maps of object boundaries in a contrastive objective, with only two independent dropout as mask. This simple approach works surprisingly well, performing on the same level as previous fully supervised methods with much less labeled data. We hypothesize that dropout can be viewed as a minimal form of data augmentation and makes the network robust to representation collapse. Then, we propose to perform structural distillation by distilling pair-wise similarities. We evaluate SimCVD on two popular datasets: the Left Atrial Segmentation Challenge (LA) and the NIH pancreas CT dataset. The results on the LA dataset demonstrate that, in two types of labeled ratios (i.e., 20% and 10%), SimCVD achieves an average Dice score of 90.85% and 89.03% respectively, a 0.91% and 2.22% improvement compared to previous best results. Our method can be trained in an end-to-end fashion, showing the promise of utilizing SimCVD as a general framework for downstream tasks, such as medical image synthesis, enhancement, and registration.

Learning Long- and Short-Term Temporal Patterns for ML-Driven Fault Management in Optical Communication Networks

The deployment of 5G and network slicing has challenged the current network management requirements, triggering the need for programmable and software-driven architectures. Thus, automated real-time fault management for self-managed networks with machine learning and artificial intelligence at the forefront has become necessary. This is especially the case of optical communication systems, accountable for most of the data traffic worldwide. This study introduces the application of a novel failure detection and localization framework capable of forecasting failures in optical systems based on an unsupervised learning strategy. In this approach, the Long- and Short-term Time-series Network (LSTNet) is exploited for modeling the normal behavior of optical systems. Then, failure conditions are properly forecast without explicitly training the model for such cases, easing the data acquisition process. Later, forecast values and actual measurements from optical equipment are used to derive an outlier detection method to detect and locate failures to improve the decision-making process at the network orchestrator. Laboratory experiments comparing the proposed approach with the Recurrent and Long Short-Term Memory models in terms of failure detection and forecasting performance show that using the LSTNet reduces the mean squared errors in 95% for unseen data, indicating robustness and suitability for real-world environments.

Breaking the resolution barrier: A low-to-high network for large-scale high-resolution land-cover mapping using low-resolution labels

Large-scale high-resolution land-cover mapping is a way to comprehend the Earth’s surface and resolve the ecological and resource challenges facing humanity. High-resolution (≤1 m) remotely sensed images can now be captured more easily, with wider coverage, as sensors and satellites develop. Nevertheless, the synchronous renewal of land-cover maps is still challenging when using the common land-cover mapping methods, due to the requirement for high-resolution land-cover labels. Abundant low-resolution (∼30 m) land-cover products are available for use as alternative label sources, but the resolution gap between these products and the growing volume of high-resolution imagery is a barrier yet to be overcome. In this paper, to break through this obstacle, we propose a low-to-high network (L2HNet) to automatically generate high-resolution land-cover maps from high-resolution images by taking only low-resolution land-cover products as the training labels, thus getting rid of the requirement for finely labeled samples during the large-scale map updating process. Firstly, to obtain the mapping results with rich details, we propose a resolution-preserving (RP) backbone that contains parallel multi-scale convolutional layers for extracting the high-resolution features from the images. Furthermore, to settle the label noise issue caused by the mismatched resolution, a confident area selection (CAS) module and a low-to-high (L2H) loss function, with weak and unsupervised strategies, are designed for obtaining reliable supervision information from the coarse labels. The experimental results obtained for six administrative states located in the Chesapeake Bay watershed of the United States show that L2HNet outperforms several of the state-of-the-art methods and the mainstream land-cover mapping methods in creating 1-m resolution land-cover maps by taking 30-m resolution land-cover products as training labels. As a further application, L2HNet was also adopted to produce the first 1-m resolution land-cover map with level II classification hierarchy for the entire state of Maryland in the United States, which covers an area of about 33,872 km2. The land-cover map of Maryland is publicly available at http://hipag.whu.edu.cn/L2HNet.html.

Dynamic strain anomalies detection at Stromboli before 2019 vulcanian explosions using machine learning

Identifying and characterizing the dynamics of explosive activity is impelling to build tools for hazard assessment at open-conduit volcanoes: machine learning techniques are now a feasible choice. During the summer of 2019, Stromboli experienced two paroxysmal eruptions that occurred in two different volcanic phases, which gave us the possibility to conceive and test an early-warning algorithm on a real use case: the paroxysm on July, 3 was clearly preceded by smaller and less perceptible changes in the volcano dynamics, while the second paroxysm, on August 28 concluded the eruptive phase. Among the changes observed in the weeks preceding the July paroxysm one of the most significant is represented by the shape variation of the ordinary minor explosions, filtered in the very long period (VLP 2–50 s) band, recorded by the Sacks-Evertson strainmeter installed near the village of Stromboli. Starting from these observations, the usage of two independent methods (an unsupervised machine learning strategy and a cross-correlation algorithm) to classify strain transients falling in the ultra long period (ULP 50–200 s) frequency band, allowed us to validate the robustness of the approach. This classification leads us to establish a link between VLP and ULP shape variation forms and volcanic activity, especially related to the unforeseen 3 July 2019 paroxysm. Previous warning times used to precede paroxysms at Stromboli are of a few minutes only. For paroxysmal events occurring outside any long-lasting eruption, the initial success of our approach, although applied only to the few available examples, could permit us to anticipate this time to several days by detecting medium-term strain anomalies: this could be crucial for risk mitigation by prohibiting access to the summit. Our innovative analysis of dynamic strain may be used to provide an early-warning system also on other open conduit active volcanoes.

Convolutional Neural Network for Phase-Only Hologram Optimization Based on the Point Source Method With the Holographic Viewing-Window

An optimization algorithm for phase-only hologram (POH) based on the point source method (PSM) with the holographic viewing-window (HVW) by using a convolutional neural network (CNN) is proposed. The network training adopts an unsupervised strategy that maps the target image to optimized constant phases (OCPs) corresponding to the wavefront of each point source instead of random constant phases (RCPs), and then the optimized POH can be obtained by wavefront superposition. The simulation and experimental results show that the speckle noise of the reconstruction is significantly suppressed compared with the initial random phase method (RPM), which demonstrates the feasibility of the proposed method and shows a strong generalization ability of the trained CNN. The HVW of a holographic near-eye display (NED) has a continuous observation range of about 6 mm in the horizontal direction without using additional optical expansion elements, and a complete high-quality image can be observed within this range.

Football Pitch Condition Analysis Based on k-Means Clustering

Football, a highly popular sport all over the world, requires that professional footballers practice it on a field of play in ideal conditions, which, among other things, includes the usage and maintenance of healthy natural grass. In this study, we present an unsupervised allocator strategy for image analysis of football pitches that uses k-means clustering and color comparison to assess whether a playing field is in good or bad condition. Our approach considers proportions of dominant RGB colors for automatized decision-making. We developed a prototype and tested it with a series of images; this paper offers a comparison between the findings of this test and our expected results.

Remaining useful life prediction based on transfer multi-stage shrinkage attention temporal convolutional network under variable working conditions

Many data-driven remaining useful life (RUL) prediction methods usually assume that the training and test data are independent and identically distributed. However, the different degradation trends of machines under variable working conditions can lead to problems with disparate distribution of degradation features and difficulties in obtaining the corresponding labels. To address the above problems, this paper proposed a RUL prediction method based on a transfer multi-stage shrinkage attention temporal convolutional network under variable working conditions. Firstly, a shrinkage attention module is designed by using the attention mechanism and shrinkage operation to eliminate the interference of irrelevant information and increase the focus on critical features. Secondly, a multi-stage shrinkage attention temporal convolution block based on a hybrid attention subnetwork and soft thresholding subnetwork is designed to efficiently learn the manifold structure of the input data to capture the degenerate information-rich deep features. Finally, an unsupervised domain adaptation strategy based on representation subspace distance and bases mismatch penalization is proposed to enhance the learning of cross-domain invariant features. The proposed method is experimentally studied on XJTU-SY and FEMTO datasets. The experimental results demonstrate that the effectiveness and accuracy of the proposed method in RUL prediction are higher than other methods.

SimIPU: Simple 2D Image and 3D Point Cloud Unsupervised Pre-training for Spatial-Aware Visual Representations

Pre-training has become a standard paradigm in many computer vision tasks. However, most of the methods are generally designed on the RGB image domain. Due to the discrepancy between the two-dimensional image plane and the three-dimensional space, such pre-trained models fail to perceive spatial information and serve as sub-optimal solutions for 3D-related tasks. To bridge this gap, we aim to learn a spatial-aware visual representation that can describe the three-dimensional space and is more suitable and effective for these tasks. To leverage point clouds, which are much more superior in providing spatial information compared to images, we propose a simple yet effective 2D Image and 3D Point cloud Unsupervised pre-training strategy, called SimIPU. Specifically, we develop a multi-modal contrastive learning framework that consists of an intra-modal spatial perception module to learn a spatial-aware representation from point clouds and an inter-modal feature interaction module to transfer the capability of perceiving spatial information from the point cloud encoder to the image encoder, respectively. Positive pairs for contrastive losses are established by the matching algorithm and the projection matrix. The whole framework is trained in an unsupervised end-to-end fashion. To the best of our knowledge, this is the first study to explore contrastive learning pre-training strategies for outdoor multi-modal datasets, containing paired camera images and LIDAR point clouds.

High‐Performance Neuromorphic Computing Based on Ferroelectric Synapses with Excellent Conductance Linearity and Symmetry

AbstractArtificial synapses can boost neuromorphic computing to overcome the inherent limitations of von Neumann architecture. As a promising memristor candidate, ferroelectric tunnel junctions (FTJ) enable the authors to successfully emulate spike‐timing‐dependent synapses. However, the nonlinear and asymmetric synaptic weight update under repeated presynaptic stimulation hampers neuromorphic computing by favoring the runaway of synaptic weights during learning. Here, the authors demonstrate an FTJ whose conductivity varies linearly and symmetrically by judiciously combining ferroelectric domain switching and oxygen vacancy migration. The artificial neural network based on this FTJ‐synapse achieves classification accuracy of 96.7% during supervised learning, which is the closest to the maximum theoretical value of 98% achieved to date. This artificial synapse also demonstrates stable unsupervised learning in a noisy environment for its well‐balanced spike‐timing‐dependent plasticity response. The novel concept of controlling ionic migration in ferroelectric materials paves the way toward highly reliable and reproducible supervised and unsupervised learning strategies.

Deep learning based domain adaptation for mitochondria segmentation on EM volumes

Background and Objective: Accurate segmentation of electron microscopy (EM) volumes of the brain is essential to characterize neuronal structures at a cell or organelle level. While supervised deep learning methods have led to major breakthroughs in that direction during the past years, they usually require large amounts of annotated data to be trained, and perform poorly on other data acquired under similar experimental and imaging conditions. This is a problem known as domain adaptation, since models that learned from a sample distribution (or source domain) struggle to maintain their performance on samples extracted from a different distribution or target domain. In this work, we address the complex case of deep learning based domain adaptation for mitochondria segmentation across EM datasets from different tissues and species. Methods: We present three unsupervised domain adaptation strategies to improve mitochondria segmentation in the target domain based on (1) state-of-the-art style transfer between images of both domains; (2) self-supervised learning to pre-train a model using unlabeled source and target images, and then fine-tune it only with the source labels; and (3) multi-task neural network architectures trained end-to-end with both labeled and unlabeled images. Additionally, to ensure good generalization in our models, we propose a new training stopping criterion based on morphological priors obtained exclusively in the source domain. The code and its documentation are publicly available at https://github.com/danifranco/EM_domain_adaptation. Results: We carried out all possible cross-dataset experiments using three publicly available EM datasets. We evaluated our proposed strategies and those of others based on the mitochondria semantic labels predicted on the target datasets. Conclusions: The methods introduced here outperform the baseline methods and compare favorably to the state of the art. In the absence of validation labels, monitoring our proposed morphology-based metric is an intuitive and effective way to stop the training process and select in average optimal models.

An unsupervised deep learning ensemble model for anomaly detection in static attributed social networks

Due to its importance in several applications, including fraud and spammer detection, anomaly detection has emerged as a key challenge in social network analysis in recent years. By including both graph and node properties, aberrant nodes in static attributed networks can be identified with more precision. Anomaly detection algorithms based on deep learning have attracted a lot of attention because the classic detection methods have significant limitations in the face of the massive growth in data volume and dimensionality. Most studies in this area focus on developing supervised techniques for spotting outliers. However, supervised algorithms for anomaly detection cannot be properly used because there is insufficient labeled data in real-world problems. The fact that the anomalies themselves, as well as the ground truth and the class labels, are unknown, presents a substantial difficulty for unsupervised anomaly detection. Several researchers have developed unsupervised learning strategies for using unlabeled data to train detection models. However, because of the diversity in anomaly types, the inability of any one model to identify all possible anomalies is a major limitation of all such models. To better detect anomalies in social networks, we created an ensemble model that combines models utilizing autoencoders (AEs), variational autoencoders (VAEs), and generative adversarial networks (GANs). During the inference (detection) phase, we use a novel approach for assigning the weights to various models used in our ensemble model. Extensive tests on two popular datasets were done to demonstrate the usefulness of the proposed ensemble.: the BlogCatalog, and Flickr. Experiments have shown that our suggested model vastly outperforms the state-of-the-art approaches utilized individually in the ensemble.

Data-driven determination of number of discrete conformations in single-particle cryo-EM

Background and objectiveOne of the strengths of single-particle cryo-EM compared to other structural determination techniques is its ability to image heterogeneous samples containing multiple molecular species, different oligomeric states or distinct conformations. This is achieved using routines for in-silico 3D classification that are now well established in the field and have successfully been used to characterize the structural heterogeneity of important biomolecules. These techniques, however, rely on expert-user knowledge and trial-and-error experimentation to determine the correct number of conformations, making it a labor intensive, subjective, and difficult to reproduce procedure. MethodsWe propose an approach to address the problem of automatically determining the number of discrete conformations present in heterogeneous single-particle cryo-EM datasets. We do this by systematically evaluating all possible partitions of the data and selecting the result that maximizes the average variance of similarities measured between particle images and the corresponding 3D reconstructions. ResultsUsing this strategy, we successfully analyzed datasets of heterogeneous protein complexes, including: 1) in-silico mixtures obtained by combining closely related antibody-bound HIV-1 Env trimers and other important membrane channels, and 2) naturally occurring mixtures from diverse and dynamic protein complexes representing varying degrees of structural heterogeneity and conformational plasticity. ConclusionsThe availability of unsupervised strategies for 3D classification combined with existing approaches for fully automatic pre-processing and 3D refinement, represents an important step towards converting single-particle cryo-EM into a high-throughput technique.

Unsupervised Deep Learning Registration of Uterine Cervix Sequence Images.

Simple SummaryA sequence of images can be taken after the application of acetic acid during a colposcopic examination of the uterine cervix to capture the dynamic visual variations due to the aceto-whitening reaction on the cervical epithelium. Automated analyis of these changes require spatial alignment of the apparent change in the cervix location in the image sequence due to patient movement or imaging device repositioning. We developed a new self-supervised RGB-colored deep learning-based image registration method to automatically align the images that does not require a manually-provided reference standard. We also fine-tuned a transformer-based segmentation network to evaluate the result of our registration method which achieved 12.62% higher in Dice/IoU scores in cervix boundary detection compared to the unregistered images.During a colposcopic examination of the uterine cervix for cervical cancer prevention, one or more digital images are typically acquired after the application of diluted acetic acid. An alternative approach is to acquire a sequence of images at fixed intervals during an examination before and after applying acetic acid. This approach is asserted to be more informative as it can capture dynamic pixel intensity variations on the cervical epithelium during the aceto-whitening reaction. However, the resulting time sequence images may not be spatially aligned due to the movement of the cervix with respect to the imaging device. Disease prediction using automated visual evaluation (AVE) techniques using multiple images could be adversely impacted without correction for this misalignment. The challenge is that there is no registration ground truth to help train a supervised-learning-based image registration algorithm. We present a novel unsupervised registration approach to align a sequence of digital cervix color images. The proposed deep-learning-based registration network consists of three branches and processes the red, green, and blue (RGB, respectively) channels of each input color image separately using an unsupervised strategy. Each network branch consists of a convolutional neural network (CNN) unit and a spatial transform unit. To evaluate the registration performance on a dataset that has no ground truth, we propose an evaluation strategy that is based on comparing automatic cervix segmentation masks in the registered sequence and the original sequence. The compared segmentation masks are generated by a fine-tuned transformer-based object detection model (DeTr). The segmentation model achieved Dice/IoU scores of 0.917/0.870 and 0.938/0.885, which are comparable to the performance of our previous model in two datasets. By comparing our segmentation on both original and registered time sequence images, we observed an average improvement in Dice scores of 12.62% following registration. Further, our approach achieved higher Dice and IoU scores and maintained full image integrity compared to a non-deep learning registration method on the same dataset.

Visual-LiDAR SLAM Based on Unsupervised Multi-channel Deep Neural Networks

Recently, deep learning techniques have been applied to solve visual or light detection and ranging (LiDAR) simultaneous localization and mapping (SLAM) problems. Supervised deep learning SLAM methods need ground truth data for training, but collecting such data is costly and labour-intensive. Unsupervised training strategies have been adopted by some visual or LiDAR SLAM methods. However, these methods only exploit the potential of single-sensor modalities, which do not take the complementary advantages of LiDAR and visual data. In this paper, we propose a novel unsupervised multi-channel visual-LiDAR SLAM method (MVL-SLAM) which can fuse visual and LiDAR data together. Our SLAM system consists of an unsupervised multi-channel visual-LiDAR odometry (MVLO) component, a deep learning–based loop closure detection component, and a 3D mapping component. The visual-LiDAR odometry component adopts a multi-channel recurrent convolutional neural network (RCNN). Its input consists of front, left, and right view depth images generated from 360^{circ } 3D LiDAR data and RGB images. We use the features from a deep convolutional neural network (CNN) for the loop closure detection component. Our SLAM method does not require ground truth data for training and can directly construct environmental 3D maps from the 3D mapping component. Experiments conducted on the KITTI odometry dataset have shown the rotation and translation errors are lower than some of the other unsupervised methods, including UnMono, SfmLearner, DeepSLAM, and UnDeepVO. Experimental results show that our methods have good performance. By fusing visual and LiDAR data, MVL-SLAM has higher accuracy and robustness of the pose estimation compared with other single-modal SLAM systems.