We investigate the design aspects of feature distillation methods achieving network compression and propose a novel feature distillation method in which the distillation loss is designed to make a synergy among various aspects: teacher transform, student transform, distillation feature position and distance function. Our proposed distillation loss includes a feature transform with a newly designed margin ReLU, a new distillation feature position, and a partial L2 distance function to skip redundant information giving adverse effects to the compression of student. In ImageNet, our proposed method achieves 21.65% of top-1 error with ResNet50, which outperforms the performance of the teacher network, ResNet152. Our proposed method is evaluated on various tasks such as image classification, object detection and semantic segmentation and achieves a significant performance improvement in all tasks. The code is available at https://sites.google.com/view/byeongho-heo/overhaul
Identifying the intent of a citation in scientific papers (e.g., background information, use of methods, comparing results) is critical for machine reading of individual publications and automated analysis of the scientific literature. We propose structural scaffolds, a multitask model to incorporate structural information of scientific papers into citations for effective classification of citation intents. Our model achieves a new state-of-the-art on an existing ACL anthology dataset (ACL-ARC) with a 13.3% absolute increase in F1 score, without relying on external linguistic resources or hand-engineered features as done in existing methods. In addition, we introduce a new dataset of citation intents (SciCite) which is more than five times larger and covers multiple scientific domains compared with existing datasets. Our code and data are available at: https://github.com/allenai/scicite.
Graph kernels based on the $1$-dimensional Weisfeiler-Leman algorithm and corresponding neural architectures recently emerged as powerful tools for (supervised) learning with graphs. However, due to the purely local nature of the algorithms, they might miss essential patterns in the given data and can only handle binary relations. The $k$-dimensional Weisfeiler-Leman algorithm addresses this by considering $k$-tuples, defined over the set of vertices, and defines a suitable notion of adjacency between these vertex tuples. Hence, it accounts for the higher-order interactions between vertices. However, it does not scale and may suffer from overfitting when used in a machine learning setting. Hence, it remains an important open problem to design WL-based graph learning methods that are simultaneously expressive, scalable, and non-overfitting. Here, we propose local variants and corresponding neural architectures, which consider a subset of the original neighborhood, making them more scalable, and less prone to overfitting. The expressive power of (one of) our algorithms is strictly higher than the original algorithm, in terms of ability to distinguish non-isomorphic graphs. Our experimental study confirms that the local algorithms, both kernel and neural architectures, lead to vastly reduced computation times, and prevent overfitting. The kernel version establishes a new state-of-the-art for graph classification on a wide range of benchmark datasets, while the neural version shows promising performance on large-scale molecular regression tasks.
In several domains, data objects can be decomposed into sets of simpler objects. It is then natural to represent each object as the set of its components or parts. Many conventional machine learning algorithms are unable to process this kind of representations, since sets may vary in cardinality and elements lack a meaningful ordering. In this paper, we present a new neural network architecture, called RepSet, that can handle examples that are represented as sets of vectors. The proposed model computes the correspondences between an input set and some hidden sets by solving a series of network flow problems. This representation is then fed to a standard neural network architecture to produce the output. The architecture allows end-to-end gradient-based learning. We demonstrate RepSet on classification tasks, including text categorization, and graph classification, and we show that the proposed neural network achieves performance better or comparable to state-of-the-art algorithms.
Question-answering systems and voice assistants are becoming major part of client service departments of many organizations, helping them to reduce the labor costs of staff. In many such systems, there is always natural language understanding module that solves intent classification task. This task is complicated because of its case-dependency - every subject area has its own semantic kernel. The state of art approaches for intent classification are different machine learning and deep learning methods that use text vector representations as input. The basic vector representation models such as Bag of words and TF-IDF generate sparse matrixes, which are becoming very big as the amount of input data grows. Modern methods such as word2vec and FastText use neural networks to evaluate word embeddings with fixed dimension size. As we are developing a question-answering system for students and enrollees of the Perm National Research Polytechnic University, we have faced the problem of user's intent detection. The subject area of our system is very specific, that is why there is a lack of training data. This aspect makes intent classification task more challenging for using state of the art deep learning methods. In this paper, we propose an approach of the questions embeddings representation based on calculation of Shannon entropy.The goal of the approach is to produce low dimensional question vectors as neural approaches do and to outperform related methods, described above in condition of small dataset. We evaluate and compare our model with existing ones using logistic regression and dataset that contains questions asked by students and enrollees. The data is labeled into six classes. Experimental comparison of proposed approach and other models revealed that proposed model performed better in the given task.
This paper describes the system submitted by ANA Team for the SemEval-2019 Task 3: EmoContext. We propose a novel Hierarchical LSTMs for Contextual Emotion Detection (HRLCE) model. It classifies the emotion of an utterance given its conversational context. The results show that, in this task, our HRCLE outperforms the most recent state-of-the-art text classification framework: BERT. We combine the results generated by BERT and HRCLE to achieve an overall score of 0.7709 which ranked 5th on the final leader board of the competition among 165 Teams.
WiFi human sensing has achieved great progress in indoor localization, activity classification, etc. Retracing the development of these work, we have a natural question: can WiFi devices work like cameras for vision applications? In this paper We try to answer this question by exploring the ability of WiFi on estimating single person pose. We use a 3-antenna WiFi sender and a 3-antenna receiver to generate WiFi data. Meanwhile, we use a synchronized camera to capture person videos for corresponding keypoint annotations. We further propose a fully convolutional network (FCN), termed WiSPPN, to estimate single person pose from the collected data and annotations. Evaluation on over 80k images (16 sites and 8 persons) replies aforesaid question with a positive answer. Codes have been made publicly available at https://github.com/geekfeiw/WiSPPN.
Insufficient or even unavailable training data of emerging classes is a big challenge of many classification tasks, including text classification. Recognising text documents of classes that have never been seen in the learning stage, so-called zero-shot text classification, is therefore difficult and only limited previous works tackled this problem. In this paper, we propose a two-phase framework together with data augmentation and feature augmentation to solve this problem. Four kinds of semantic knowledge (word embeddings, class descriptions, class hierarchy, and a general knowledge graph) are incorporated into the proposed framework to deal with instances of unseen classes effectively. Experimental results show that each and the combination of the two phases achieve the best overall accuracy compared with baselines and recent approaches in classifying real-world texts under the zero-shot scenario.
In information retrieval (IR) and related tasks, term weighting approaches typically consider the frequency of the term in the document and in the collection in order to compute a score reflecting the importance of the term for the document. In tasks characterized by the presence of training data (such as text classification) it seems logical that the term weighting function should take into account the distribution (as estimated from training data) of the term across the classes of interest. Although `supervised term weighting' approaches that use this intuition have been described before, they have failed to show consistent improvements. In this article we analyse the possible reasons for this failure, and call consolidated assumptions into question. Following this criticism we propose a novel supervised term weighting approach that, instead of relying on any predefined formula, learns a term weighting function optimised on the training set of interest; we dub this approach \emph{Learning to Weight} (LTW). The experiments that we run on several well-known benchmarks, and using different learning methods, show that our method outperforms previous term weighting approaches in text classification.
We are concerned with the vulnerability of computer vision models to distributional shifts. We formulate a combinatorial optimization problem that allows evaluating the regions in the image space where a given model is more vulnerable, in terms of image transformations applied to the input, and face it with standard search algorithms. We further embed this idea in a training procedure, where we define new data augmentation rules according to the image transformations that the current model is most vulnerable to, over iterations. An empirical evaluation on classification and semantic segmentation problems suggests that the devised algorithm allows to train models that are more robust against content-preserving image manipulations and, in general, against distributional shifts.
Motivation: Radiomics refers to the high-throughput mining of quantitative features from radiographic images. It is a promising field in that it may provide a non-invasive solution for screening and classification. Standard machine learning classification and feature selection techniques, however, tend to display inferior performance in terms of (the stability of) predictive performance. This is due to the heavy multicollinearity present in radiomic data. We set out to provide an easy-to-use approach that deals with this problem. Results: We developed a four-step approach that projects the original high-dimensional feature space onto a lower-dimensional latent-feature space, while retaining most of the covariation in the data. It consists of (i) penalized maximum likelihood estimation of a redundancy filtered correlation matrix. The resulting matrix (ii) is the input for a maximum likelihood factor analysis procedure. This two-stage maximum-likelihood approach can be used to (iii) produce a compact set of stable features that (iv) can be directly used in any (regression-based) classifier or predictor. It outperforms other classification (and feature selection) techniques in both external and internal validation settings regarding survival in squamous cell cancers.
In this paper, we consider the problem of malware detection and classification based on image analysis. We convert executable files to images and apply image recognition using deep learning (DL) models. To train these models, we employ transfer learning based on existing DL models that have been pre-trained on massive image datasets. We carry out various experiments with this technique and compare its performance to that of an extremely simple machine learning technique, namely, k-nearest neighbors (\kNN). For our k-NN experiments, we use features extracted directly from executables, rather than image analysis. While our image-based DL technique performs well in the experiments, surprisingly, it is outperformed by k-NN. We show that DL models are better able to generalize the data, in the sense that they outperform k-NN in simulated zero-day experiments.
In many research fields, the sizes of the existing datasets vary widely. Hence, there is a need for machine learning techniques which are well-suited for these different datasets. One possible technique is the self-organizing map (SOM), a type of artificial neural network which is, so far, weakly represented in the field of machine learning. The SOM's unique characteristic is the neighborhood relationship of the output neurons. This relationship improves the ability of generalization on small datasets. SOMs are mostly applied in unsupervised learning and few studies focus on using SOMs as supervised learning approach. Furthermore, no appropriate SOM package is available with respect to machine learning standards and in the widely used programming language Python. In this paper, we introduce the freely available Supervised Self-organizing maps (SuSi) Python package which performs supervised regression and classification. The implementation of SuSi is described with respect to the underlying mathematics. Then, we present first evaluations of the SOM for regression and classification datasets from two different domains of geospatial image analysis. Despite the early stage of its development, the SuSi framework performs well and is characterized by only small performance differences between the training and the test datasets. A comparison of the SuSi framework with existing Python and R packages demonstrates the importance of the SuSi framework. In future work, the SuSi framework will be extended, optimized and upgraded e.g. with tools to better understand and visualize the input data as well as the handling of missing and incomplete data.
Adversarial perturbations dramatically decrease the accuracy of state-of-the-art image classifiers. In this paper, we propose and analyze a simple and computationally efficient defense strategy: inject random Gaussian noise, discretize each pixel, and then feed the result into any pre-trained classifier. Theoretically, we show that our randomized discretization strategy reduces the KL divergence between original and adversarial inputs, leading to a lower bound on the classification accuracy of any classifier against any (potentially whitebox) $\ell_\infty$-bounded adversarial attack. Empirically, we evaluate our defense on adversarial examples generated by a strong iterative PGD attack. On ImageNet, our defense is more robust than adversarially-trained networks and the winning defenses of the NIPS 2017 Adversarial Attacks & Defenses competition.
Object detectors are usually equipped with backbone networks designed for image classification. It might be sub-optimal because of the gap between the tasks of image classification and object detection. In this work, we present DetNAS to use Neural Architecture Search (NAS) for the design of better backbones for object detection. It is non-trivial because detection training typically needs ImageNet pre-training while NAS systems require accuracies on the target detection task as supervisory signals. Based on the technique of one-shot supernet, which contains all possible networks in the search space, we propose a framework for backbone search on object detection. We train the supernet under the typical detector training schedule: ImageNet pre-training and detection fine-tuning. Then, the architecture search is performed on the trained supernet, using the detection task as the guidance. This framework makes NAS on backbones very efficient. In experiments, we show the effectiveness of DetNAS on various detectors, for instance, one-stage RetinaNet and the two-stage FPN. We empirically find that networks searched on object detection shows consistent superiority compared to those searched on ImageNet classification. The resulting architecture achieves superior performance than hand-crafted networks on COCO with much less FLOPs complexity.
Obtaining large-scale annotated data for NLP tasks in the scientific domain is challenging and expensive. We release SciBERT, a pretrained language model based on BERT (Devlin et al., 2018) to address the lack of high-quality, large-scale labeled scientific data. SciBERT leverages unsupervised pretraining on a large multi-domain corpus of scientific publications to improve performance on downstream scientific NLP tasks. We evaluate on a suite of tasks including sequence tagging, sentence classification and dependency parsing, with datasets from a variety of scientific domains. We demonstrate statistically significant improvements over BERT and achieve new state-of-the-art results on several of these tasks. The code and pretrained models are available at https://github.com/allenai/scibert/.
We present a complete classification of all minimal problems for generic arrangements of points and lines completely observed by calibrated perspective cameras. We show that there are only 30 minimal problems in total, no problems exist for more than 6 cameras, for more than 5 points, and for more than 6 lines. We present a sequence of tests for detecting minimality starting with counting degrees of freedom and ending with full symbolic and numeric verification of representative examples. For all minimal problems discovered, we present their algebraic degrees, i.e. the number of solutions, which measure their intrinsic difficulty. It shows how exactly the difficulty of problems grows with the number of views. Importantly, several new minimal problems have small degrees that might be practical in image matching and 3D reconstruction.
Sentiment classification typically relies on a large amount of labeled data. In practice, the availability of labels is highly imbalanced among different languages, e.g., more English texts are labeled than texts in any other languages, which creates a considerable inequality in the quality of related information services received by users speaking different languages. To tackle this problem, cross-lingual sentiment classification approaches aim to transfer knowledge learned from one language that has abundant labeled examples (i.e., the source language, usually English) to another language with fewer labels (i.e., the target language). The source and the target languages are usually bridged through off-the-shelf machine translation tools. Through such a channel, cross-language sentiment patterns can be successfully learned from English and transferred into the target languages. This approach, however, often fails to capture sentiment knowledge specific to the target language, and thus compromises the accuracy of the downstream classification task. In this paper, we employ emojis, which are widely available in many languages, as a new channel to learn both the cross-language and the language-specific sentiment patterns. We propose a novel representation learning method that uses emoji prediction as an instrument to learn respective sentiment-aware representations for each language. The learned representations are then integrated to facilitate cross-lingual sentiment classification. The proposed method demonstrates state-of-the-art performance on benchmark datasets, which is sustained even when sentiment labels are scarce.
Image understanding heavily relies on accurate multi-label classification. In recent years, deep learning algorithms have become very successful for such tasks, and various commercial and open-source APIs have been released for public use. However, these APIs are often trained on different datasets, which, besides affecting their performance, might pose a challenge to their performance evaluation. This challenge concerns the different object-class dictionaries of the APIs' training dataset and the benchmark dataset, in which the predicted labels are semantically similar to the benchmark labels but considered different simply because they have different wording in the dictionaries. To face this challenge, we propose semantic similarity metrics to obtain richer understating of the APIs predicted labels and thus their performance. In this study, we evaluate and compare the performance of 13 of the most prominent commercial and open-source APIs in a best-of-breed challenge on the Visual Genome and Open Images benchmark datasets. Our findings demonstrate that, while using traditional metrics, the Microsoft Computer Vision, Imagga, and IBM APIs performed better than others. However, applying semantic metrics also unveil the InceptionResNet-v2, Inception-v3, and ResNet50 APIs, which are trained only with the simple ImageNet dataset, as challengers for top semantic performers.
Aspect-based sentiment analysis (ABSA), which aims to identify fine-grained opinion polarity towards a specific aspect, is a challenging subtask of sentiment analysis (SA). In this paper, we construct an auxiliary sentence from the aspect and convert ABSA to a sentence-pair classification task, such as question answering (QA) and natural language inference (NLI). We fine-tune the pre-trained model from BERT and achieve new state-of-the-art results on SentiHood and SemEval-2014 Task 4 datasets.