SSB: Simple but Strong Baseline for Boosting Performance of Open-Set Semi-Supervised Learning

Neural models for distantly supervised relation extraction (DS-RE) encode each sentence in an entity-pair bag separately. These are then aggregated for bag-level relation prediction. Since, at encoding time, these approaches do not allow information to flow from other sentences in the bag, we believe that they do not utilize the available bag data to the fullest. In response, we explore a simple baseline approach (PARE) in which all sentences of a bag are concatenated into a passage of sentences, and encoded jointly using BERT. The contextual embeddings of tokens are aggregated using attention with the candidate relation as query -- this summary of whole passage predicts the candidate relation. We find that our simple baseline solution outperforms existing state-of-the-art DS-RE models in both monolingual and multilingual DS-RE datasets.

Gathering labeled data is one of the most time-consuming and expensive tasks in supervised machine learning. In practical applications, there are usually quite limited labeled training samples but abundant unlabeled data that is easy to collect. Semi-supervised learning and active learning are two important techniques for learning a discriminative classification model when labeled data is scarce. However, unlabeled data with significant noises and outliers cannot be well exploited and usually worsen the performance of semi-supervised learning and the performance of active learning also needs a powerful initial classifier learned from the quite limited labeled training data. In order to solve the above issues, in this paper we proposed a novel model of semi-supervised dictionary active learning (SSDAL), which aims to integrate semi-supervised learning and active learning to effectively use all the training data. In particular, two criterions based on estimated class possibility are designed to select the unlabeled data with confident class estimation for semi-supervised learning and the informative unlabeled data for active learning, respectively. Extensive experiments are conducted to show the superior performance of our method in classification applications, e.g., handwritten digit recognition, face recognition and large-scale image classification.

Abstract3D multi-object tracking (MOT) has witnessed numerous novel benchmarks and approaches in recent years, especially those under the “tracking-by-detection” paradigm. Despite their progress and usefulness, an in-depth analysis of their strengths and weaknesses is not yet available. In this paper, we summarize current 3D MOT methods into a unified framework by decomposing them into four constituent parts: pre-processing of detection, association, motion model, and life cycle management. We then ascribe the failure cases of existing algorithms to each component and investigate them in detail. Based on the analyses, we propose corresponding improvements which lead to a strong yet simple baseline: SimpleTrack. Comprehensive experimental results on Waymo Open Dataset and nuScenes demonstrate that our final method could achieve new state-of-the-art results with minor modifications. Furthermore, we take additional steps and rethink whether current benchmarks authentically reflect the ability of algorithms for real-world challenges. We delve into the details of existing benchmarks and find some intriguing facts. Finally, we analyze the distribution and causes of remaining failures in SimpleTrack and propose future directions for 3D MOT. Our code is at https://github.com/tusen-ai/SimpleTrack.Keywords3D multi-object trackingAutonomous driving

Using an ensemble of neural networks with consistency regularization is effective for improving performance and stability of deep learning, compared to the case of a single network. In this paper, we present a semi-supervised Deep Coupled Ensemble (DCE) model, which contributes to ensemble learning and classification landmark exploration for better locating the final decision boundaries in the learnt latent space. First, multiple complementary consistency regularizations are integrated into our DCE model to enable the ensemble members to learn from each other and themselves, such that training experience from different sources can be shared and utilized during training. Second, in view of the possibility of producing incorrect predictions on a number of difficult instances, we adopt class-wise mean feature matching to explore important unlabeled instances as classification landmarks, on which the model predictions are more reliable. Minimizing the weighted conditional entropy on unlabeled data is able to force the final decision boundaries to move away from important training data points, which facilitates semi-supervised learning. Ensemble members could eventually have similar performance due to consistency regularization, and thus only one of these members is needed during the test stage, such that the efficiency of our model is the same as the non-ensemble case. Extensive experimental results demonstrate the superiority of our proposed DCE model over existing state-of-the-art semi-supervised learning methods.

Semi-supervised learning (SSL) relies on partial supervision information for prediction, where only a small set of samples are associated with labels. Performance of SSL is significantly degraded if the given labels are not reliable. Such problems arise in realistic applications such as web image search using noisy textual tags. This paper proposes a novel and efficient graph based SSL method with the unique capacity of pruning contradictory labels and inferring new labels through a bidirectional and alternating optimization process. The objective is to automatically identify the most suitable samples for manipulation, labeling or unlabeling, and meanwhile estimate a smooth classification function over a weighted graph. Different from other graph based SSL approaches, the proposed method employs a bivariate objective function and iteratively modifies label variables on both labeled and unlabeled samples. Starting from such a SSL setting, we present a relearning framework to improve the performance of base learner, particularly for the application of web image search. Besides the toy demonstration on artificial data, we evaluated the proposed method on Flickr image search with unreliable textual labels. Experimental results confirm the significant improvements of the method over the baseline text based search engine and the state-of-the-art SSL methods.

BasicTAD: An astounding RGB-Only baseline for temporal action detection

When the amount of labeled data are limited, semisupervised learning can improve the learner's performance by also using the often easily available unlabeled data. In particular, a popular approach requires the learned function to be smooth on the underlying data manifold. By approximating this manifold as a weighted graph, such graph-based techniques can often achieve state-of-the-art performance. However, their high time and space complexities make them less attractive on large data sets. In this paper, we propose to scale up graph-based semisupervised learning using a set of sparse prototypes derived from the data. These prototypes serve as a small set of data representatives, which can be used to approximate the graph-based regularizer and to control model complexity. Consequently, both training and testing become much more efficient. Moreover, when the Gaussian kernel is used to define the graph affinity, a simple and principled method to select the prototypes can be obtained. Experiments on a number of real-world data sets demonstrate encouraging performance and scaling properties of the proposed approach. It also compares favorably with models learned via l1 -regularization at the same level of model sparsity. These results demonstrate the efficacy of the proposed approach in producing highly parsimonious and accurate models for semisupervised learning.

This study explores a simple but strong baseline for person re-identification (ReID). Person ReID with deep neural networks has progressed and achieved high performance in recent years. However, many state-of-the-art methods design complex network structures and concatenate multi-branch features. In the literature, some effective training tricks briefly appear in several papers or source codes. The present study collects and evaluates these effective training tricks in person ReID. By combining these tricks, the model achieves 94.5% rank-1 and 85.9% mean average precision on Market1501 with only using the global features of ResNet50. The performance surpasses all existing global- and part-based baselines in person ReID. We propose a novel neck structure named as batch normalization neck (BNNeck). BNNeck adds a batch normalization layer after global pooling layer to separate metric and classification losses into two different feature spaces because we observe they are inconsistent in one embedding space. Extended experiments show that BNNeck can boost the baseline, and our baseline can improve the performance of existing state-of-the-art methods. Our codes and models are available at: https://github.com/michuanhaohao/reid-strong-baseline.

One-pass learning is one of the most popular and widely used online learning method for high-speed and large-scale streaming data. However, large-scale data usually has the problem of label scarcity, and the insufficient labeled data has a great negative impact on the performance of one-pass learning. The existing research on semi-supervised online learning has difficulty extending to one-pass learning due to the iterative nature of semi-supervised methods. In addition, the distribution of complex streaming data may also change over time. Just directly ignoring the change in data distribution will seriously affect the value and significance of learning. Therefore, one-pass learning under distribution shift when labels are scarce is a challenging task. In this paper, we propose a semi-supervised one-pass learning method with forgetting factor to handle the change of data distribution efficiently. In this method, we use an improved mini-batch label propagation algorithm to obtain pseudo-labels for the incoming unlabeled data. The forgetting factor is introduced to calculate the memory matrix for the labeled data and the pseudo-labeled data to deal with the data distribution change. We also introduce a semi-supervised confidence factor to balance the weight of labeled and pseudo-labeled data. Extensive experiments on streaming datasets demonstrate the effectiveness of our approach. Especially under the one-pass constraint, it has achieved far better performance than the existing semi-supervised online learning.

In hyperspectral images, the creation of ground-truth data for supervised learning methods is costly in terms of computation cost and time. In addition, the number of labeled data and the quality of labeled training data affects the success of the classification. as a solution to this problem, a graph-based semi-supervised hyperspectral image classifier is proposed in this study. The system was developed on graphics processing unit (GPU) to get rid of the high processing cost of semi-supervised learning. In addition, subtractive clustering is proposed as a new approach to select labeled samples for semi-supervised learning. The results of the system tests with public data sets showed that the classification performance of semi-supervised learning can be close to supervised learning with a small number of labeled data.

As the foundation of global communication networks, optical fibers are vulnerable to various disruptive events, including mechanical damage, such as cuts, and malicious physical layer breaches, such as eavesdropping via fiber bending. Traditional monitoring methods often fail to identify subtle or novel anomalies, stimulating the proliferation of Machine Learning (ML) techniques for detection of threats before they cause significant harm. In this paper, we evaluate the performance of Semi-Supervised Learning (SSL) and Unsupervised Learning (USL) approaches for detecting various abnormal events, such as fiber bending and vibrations, by analyzing polarization signatures with minimal reliance on labeled data. We experimentally collect thirteen polarization signatures on three different types of fiber cable and process them using One-Class Support Vector Machine (OCSVM) as an SSL, and Density-Based Spatial Clustering of Applications with Noise (DBSCAN) as a USL algorithm for anomaly detection. We introduce tailored evaluation metrics designed to guide hyper-parameter tuning and capture generalization over different anomaly types, detection consistency, and robustness to false positives, enabling practical deployment of OCSVM and DBSCAN in optical fiber security. Our findings demonstrate DBSCAN as a strong contender to detect previously unseen threats in scenarios where labeled data are not available, despite some variability in performance between different scenarios, with F1 score values between 0.615 and 0.995. In contrast, OCSVM, trained on normal operating conditions, maintains high F1 scores of 0.98 to 0.998, demonstrating accurate detection of complex anomalies in optical networks.

In this paper, we propose a nonlinear metric learning framework to boost the performance of semi-supervised learning (SSL) algorithms. Constructed on top of Laplacian SVM (LapSVM), the proposed method learns a smooth nonlinear feature space transformation that makes the input data points more linearly separable. Coherent point drifting (CPD) is utilized as the geometric model with the consideration of its remarkable expressive power in generating sophisticated yet smooth deformations. Our framework has broad applicability, and it can be integrated with many other SSL classifiers than LapSVM. Experiments performed on synthetic and real world datasets show the effectiveness of our CPD-LapSVM over the state-of-the-art metric learning solutions in SSL.

SMGCL: Semi-supervised Multi-view Graph Contrastive Learning

Most existing work on multi-label learning focused on supervised learning which requires manual annotation samples that is labor-intensive, time-consuming and costly. To address such a problem, we present a novel method that incorporates active learning into the semi-supervised learning for multi-label image classification. What’s more, aiming at the curse of dimensionality existing in high-dimensional data, we explore a dimensionality reduction technique with non-negative sparseness constraint to extract a group of features that can completely describe the data and hence make the learning model more efficiently. Experimental results on common data sets validate that the proposed algorithm is relatively effective to improve the performance of the learner in multi-label classification, and the obtained learner is with reliability and robustness after data dimensionality using NNS-DR (Non-Negative Sparseness for Dimensionality Reduction).

Matrix factorization has demonstrated outstanding performance in machine learning. Recently, graph-based matrix factorization has gained widespread attention. However, graph-based methods are only suitable for handling small amounts of data. This paper proposes a fast semi-supervised learning method using only matrix factorization, which considers both global and local information. By introducing bipartite graphs into symmetric matrix factorization, the technique can handle large datasets effectively. It is worth noting that by utilizing tag information, the proposed symmetric matrix factorization becomes convex and unconstrained, i.e., the non-convex problem minx(1−x2)2 is transformed into a convex problem. This allows it to be optimized quickly using state-of-the-art unconstrained optimization algorithms. The computational complexity of the proposed method is O(nmd), which is much lower than that of the original symmetric matrix factorization, which is O(n2d), and even lower than that of other anchor-based methods, which is O(nmd+m2n+m3), where n represents the number of samples, d represents the number of features, and m≪n represents the number of anchors. The experimental results on multiple public datasets indicate that the proposed method achieves higher performance in less time.