Fully Test-Time Adaptation (TTA), which aims at adapting models to data drifts, has recently attracted wide interest. Numerous tricks and techniques have been proposed to ensure robust learning on arbitrary streams of unlabeled data. However, assessing the true impact of each individual technique and obtaining a fair comparison still constitutes a significant challenge. To help consolidate the community's knowledge, we present a categorization of selected orthogonal TTA techniques, including small batch normalization, stream rebalancing, reliable sample selection, and network confidence calibration. We meticulously dissect the effect of each approach on different scenarios of interest. Through our analysis, we shed light on trade-offs induced by those techniques between accuracy, the computational power required, and model complexity. We also uncover the synergy that arises when combining techniques and are able to establish new state-of-the-art results.
University of Modena and Reggio Emilia（摩德纳-雷焦艾米利亚大学）
In Continual Learning, a Neural Network is trained on a stream of data whose distribution shifts over time. Under these assumptions, it is especially challenging to improve on classes appearing later in the stream while remaining accurate on previous ones. This is due to the infamous problem of catastrophic forgetting, which causes a quick performance degradation when the classifier focuses on learning new categories. Recent literature proposed various approaches to tackle this issue, often resorting to very sophisticated techniques. In this work, we show that naive rehearsal can be patched to…
Pattern Recognition Lab（模式识别实验室）；Computer Sciences（计算机科学（学院））；Friedrich-Alexander-Universität Erlangen-Nürnberg（埃尔朗根-纽伦堡大学）；Institute of Veterinary Pathology（兽医病理学研究所）；Freie Universität Berlin（柏林自由大学）；Institute of Neuropathology（神经病理学研究所）
For histopathological tumor assessment, the count of mitotic figures per area is an important part of prognostication. Algorithmic approaches - such as for mitotic figure identification - have significantly improved in recent times, potentially allowing for computer-augmented or fully automatic screening systems in the future. This trend is further supported by whole slide scanning microscopes becoming available in many pathology labs and could soon become a standard imaging tool. For an application in broader fields of such algorithms, the availability of mitotic figure data sets of sufficien…
Retail Product Image Classification is an important Computer Vision and Machine Learning problem for building real world systems like self-checkout stores and automated retail execution evaluation. In this work, we present various tricks to increase accuracy of Deep Learning models on different types of retail product image classification datasets. These tricks enable us to increase the accuracy of fine tuned convnets for retail product image classification by a large margin. As the most prominent trick, we introduce a new neural network layer called Local-Concepts-Accumulation (LCA) layer whi…
Beijing University of Posts and Telecommunications（北京邮电大学）；China Mobile Research Institute（中国移动研究院）
Recently, out-of-distribution (OOD) generalization has attracted attention to the robustness and generalization ability of deep learning based models, and accordingly, many strategies have been made to address different aspects related to this issue. However, most existing algorithms for OOD generalization are complicated and specifically designed for certain dataset. To alleviate this problem, nicochallenge-2022 provides NICO++, a large-scale dataset with diverse context information. In this paper, based on systematic analysis of different schemes on NICO++ dataset, we propose a simple but ef…
Bosch Center for Artificial Intelligence（博世人工智能中心）；University of Freiburg（弗莱堡大学）
While neural architecture search methods have been successful in previous years and led to new state-of-the-art performance on various problems, they have also been criticized for being unstable, being highly sensitive with respect to their hyperparameters, and often not performing better than random search. To shed some light on this issue, we discuss some practical considerations that help improve the stability, efficiency and overall performance.
Northeastern University（东北大学）；The Chinese University of Hong Kong（香港中文大学）；NiuTrans Research（小牛翻译研究院）
Improving Transformer efficiency has become increasingly attractive recently. A wide range of methods has been proposed, e.g., pruning, quantization, new architectures and etc. But these methods are either sophisticated in implementation or dependent on hardware. In this paper, we show that the efficiency of Transformer can be improved by combining some simple and hardware-agnostic methods, including tuning hyper-parameters, better design choices and training strategies. On the WMT news translation tasks, we improve the inference efficiency of a strong Transformer system by 3.80X on CPU and 2.…
Accurate face landmark localization is an essential part of face recognition, reconstruction and morphing. To accurately localize face landmarks, we present our heatmap regression approach. Each model consists of a MobileNetV2 backbone followed by several upscaling layers, with different tricks to optimize both performance and inference cost. We use five naïve face landmarks from a publicly available face detector to position and align the face instead of using the bounding box like traditional methods. Moreover, we show by adding random rotation, displacement and scaling -- after alignment --…
Adversarial training (AT) is one of the most effective strategies for promoting model robustness. However, recent benchmarks show that most of the proposed improvements on AT are less effective than simply early stopping the training procedure. This counter-intuitive fact motivates us to investigate the implementation details of tens of AT methods. Surprisingly, we find that the basic settings (e.g., weight decay, training schedule, etc.) used in these methods are highly inconsistent. In this work, we provide comprehensive evaluations on CIFAR-10, focusing on the effects of mostly overlooked t…
Recently, diabetic retinopathy (DR) screening utilizing ultra-wide optical coherence tomography angiography (UW-OCTA) has been used in clinical practices to detect signs of early DR. However, developing a deep learning-based DR analysis system using UW-OCTA images is not trivial due to the difficulty of data collection and the absence of public datasets. By realistic constraints, a model trained on small datasets may obtain sub-par performance. Therefore, to help ophthalmologists be less confused about models' incorrect decisions, the models should be robust even in data scarcity settings. To …
Over the past few years, graph neural networks (GNN) and label propagation-based methods have made significant progress in addressing node classification tasks on graphs. However, in addition to their reliance on elaborate architectures and algorithms, there are several key technical details that are frequently overlooked, and yet nonetheless can play a vital role in achieving satisfactory performance. In this paper, we first summarize a series of existing tricks-of-the-trade, and then propose several new ones related to label usage, loss function formulation, and model design that can signifi…
Indian Institute of Science（印度科学学院）；Google Research（谷歌研究院）
Universal Domain Adaptation (UniDA) deals with the problem of knowledge transfer between two datasets with domain-shift as well as category-shift. The goal is to categorize unlabeled target samples, either into one of the "known" categories or into a single "unknown" category. A major problem in UniDA is negative transfer, i.e. misalignment of "known" and "unknown" classes. To this end, we first uncover an intriguing tradeoff between negative-transfer-risk and domain-invariance exhibited at different layers of a deep network. It turns out we can strike a balance between these two metrics at a …
The University of Texas at Austin（德克萨斯大学奥斯汀分校）；Rice University（莱斯大学）
Training deep graph neural networks (GNNs) is notoriously hard. Besides the standard plights in training deep architectures such as vanishing gradients and overfitting, it also uniquely suffers from over-smoothing, information squashing, and so on, which limits their potential power for encoding the high-order neighbor structure in large-scale graphs. Although numerous efforts are proposed to address these limitations, such as various forms of skip connections, graph normalization, and random dropping, it is difficult to disentangle the advantages brought by a deep GNN architecture from those …
VUNO, Inc.（VUNO公司）；Pukyong National University（釜庆国立大学）；Korea University（高丽大学）
While pre-trained language models (PLMs) have become a de-facto standard promoting the accuracy of text classification tasks, recent studies find that PLMs often predict over-confidently. Although various calibration methods have been proposed, such as ensemble learning and data augmentation, most of the methods have been verified in computer vision benchmarks rather than in PLM-based text classification tasks. In this paper, we present an empirical study on confidence calibration for PLMs, addressing three categories, including confidence penalty losses, data augmentations, and ensemble metho…
When there is a distributional shift between data used to train a predictive algorithm and current data, performance can suffer. This is known as the domain adaptation problem. Bootstrap aggregating, or bagging, is a popular method for improving stability of predictive algorithms, while reducing variance and protecting against over-fitting. This article proposes a domain adaptive bagging method coupled with a new iterative nearest neighbor sampler. The key idea is to draw bootstrap samples from the training data in such a manner that their distribution equals that of new testing data. The prop…
Laboratory for Scientific Computing（科学计算实验室）；Clare Hall（克莱尔学堂）
The aim of this work is to propose a meta-algorithm for automatic classification in the presence of discrete binary classes. Classifier learning in the presence of overlapping class distributions is a challenging problem in machine learning. Overlapping classes are described by the presence of ambiguous areas in the feature space with a high density of points belonging to both classes. This often occurs in real-world datasets, one such example is numeric data denoting properties of particle decays derived from high-energy accelerators like the Large Hadron Collider (LHC). A significant body of…
Deep Residual Error and Bag-of-Tricks Learning for Gravitational Wave Surrogate Modeling
Aristotle University of Thessaloniki（塞萨洛尼基亚里士多德大学）；Department of Informatics, Aristotle University of Thessaloniki（塞萨洛尼基亚里士多德大学信息学系）；Department of Physics, Aristotle University of Thessaloniki（塞萨洛尼基亚里士多德大学物理系）；GSI Helmholtz Center for Heavy Ion Research（亥姆霍兹重离子研究中心（GSI））
As one of standard approaches to train deep neural networks, dropout has been applied to regularize large models to avoid overfitting, and the improvement in performance by dropout has been explained as avoiding co-adaptation between nodes. However, when correlations between nodes are compared after training the networks with or without dropout, one question arises if co-adaptation avoidance explains the dropout effect completely. In this paper, we propose an additional explanation of why dropout works and propose a new technique to design better activation functions. First, we show that dropo…
missForestPredict -- Missing data imputation for prediction settings
KU Leuven（KU Leuven（荷语天主教鲁汶大学））；Leuven Unit for Health Technology Assessment Research (LUHTAR)（鲁汶卫生技术评估研究单元（LUHTAR））；School for Public Health and Primary Care, Maastricht University（马斯特里赫特大学公共卫生与基层保健学院）；Leiden University Medical Center（莱顿大学医学中心）
