PyTorchKR Papers

최신 논문

총 11,809편
태그·기간 등 상세 필터

카테고리 탐색

로그인 후 이 태그 구독

arXiv:2012.01959

Towards Human Haptic Gesture Interpretation for Robotic Systems

Physical human-robot interactions (pHRI) are less efficient and communicative than human-human interactions, and a key reason is a lack of informative sense of touch in robotic systems. Interpreting human touch gestures is a nuanced, challenging task with extreme gaps between human and robot capability. Among prior works that demonstrate human touch recognition capability, differences in sensors, gesture classes, feature sets, and classification algorithms yield a conglomerate of non-transferable results and a glaring lack of a standard. To address this gap, this work presents 1) four proposed touch gesture classes that cover an important subset of the gesture characteristics identified in the literature, 2) the collection of an extensive force dataset on a common pHRI robotic arm with only its internal wrist force-torque sensor, and 3) an exhaustive performance comparison of combinations of feature sets and classification algorithms on this dataset. We demonstrate high classification accuracies among our proposed gesture definitions on a test set, emphasizing that neural net-work classifiers on the raw data outperform other combinations of feature sets and algorithms. The accompanying video is here: https://youtu.be/gJPVImNKU68

arXiv:2012.01573

A Study of Few-Shot Audio Classification

Advances in deep learning have resulted in state-of-the-art performance for many audio classification tasks but, unlike humans, these systems traditionally require large amounts of data to make accurate predictions. Not every person or organization has access to those resources, and the organizations that do, like our field at large, do not reflect the demographics of our country. Enabling people to use machine learning without significant resource hurdles is important, because machine learning is an increasingly useful tool for solving problems, and can solve a broader set of problems when put in the hands of a broader set of people. Few-shot learning is a type of machine learning designed to enable the model to generalize to new classes with very few examples. In this research, we address two audio classification tasks (speaker identification and activity classification) with the Prototypical Network few-shot learning algorithm, and assess performance of various encoder architectures. Our encoders include recurrent neural networks, as well as one- and two-dimensional convolutional neural networks. We evaluate our model for speaker identification on the VoxCeleb dataset and ICSI Meeting Corpus, obtaining 5-shot 5-way accuracies of 93.5% and 54.0%, respectively. We also evaluate for activity classification from audio using few-shot subsets of the Kinetics~600 dataset and AudioSet, both drawn from Youtube videos, obtaining 51.5% and 35.2% accuracy, respectively.

arXiv:2012.01641

Meta-Generating Deep Attentive Metric for Few-shot Classification

Learning to generate a task-aware base learner proves a promising direction to deal with few-shot learning (FSL) problem. Existing methods mainly focus on generating an embedding model utilized with a fixed metric (eg, cosine distance) for nearest neighbour classification or directly generating a linear classier. However, due to the limited discriminative capacity of such a simple metric or classifier, these methods fail to generalize to challenging cases appropriately. To mitigate this problem, we present a novel deep metric meta-generation method that turns to an orthogonal direction, ie, learning to adaptively generate a specific metric for a new FSL task based on the task description (eg, a few labelled samples). In this study, we structure the metric using a three-layer deep attentive network that is flexible enough to produce a discriminative metric for each task. Moreover, different from existing methods that utilize an uni-modal weight distribution conditioned on labelled samples for network generation, the proposed meta-learner establishes a multi-modal weight distribution conditioned on cross-class sample pairs using a tailored variational autoencoder, which can separately capture the specific inter-class discrepancy statistics for each class and jointly embed the statistics for all classes into metric generation. By doing this, the generated metric can be appropriately adapted to a new FSL task with pleasing generalization performance. To demonstrate this, we test the proposed method on four benchmark FSL datasets and gain surprisingly obvious performance improvement over state-of-the-art competitors, especially in the challenging cases, eg, improve the accuracy from 26.14% to 46.69% in the 20-way 1-shot task on miniImageNet, while improve the accuracy from 45.2% to 68.72% in the 5-way 1-shot task on FC100. Code is available: https://github.com/NWPUZhoufei/DAM.

arXiv:2012.01604

Going Beyond Classification Accuracy Metrics in Model Compression

With the rise in edge-computing devices, there has been an increasing demand to deploy energy and resource-efficient models. A large body of research has been devoted to developing methods that can reduce the size of the model considerably without affecting the standard metrics such as top-1 accuracy. However, these pruning approaches tend to result in a significant mismatch in other metrics such as fairness across classes and explainability. To combat such misalignment, we propose a novel multi-part loss function inspired by the knowledge-distillation literature. Through extensive experiments, we demonstrate the effectiveness of our approach across different compression algorithms, architectures, tasks as well as datasets. In particular, we obtain up to $4.1\times$ reduction in the number of prediction mismatches between the compressed and reference models, and up to $5.7\times$ in cases where the reference model makes the correct prediction; all while making no changes to the compression algorithm, and minor modifications to the loss function. Furthermore, we demonstrate how inducing simple alignment between the predictions of the models naturally improves the alignment on other metrics including fairness and attributions. Our framework can thus serve as a simple plug-and-play component for compression algorithms in the future.

arXiv:2012.02149

Approximate kNN Classification for Biomedical Data

We are in the era where the Big Data analytics has changed the way of interpreting the various biomedical phenomena, and as the generated data increase, the need for new machine learning methods to handle this evolution grows. An indicative example is the single-cell RNA-seq (scRNA-seq), an emerging DNA sequencing technology with promising capabilities but significant computational challenges due to the large-scaled generated data. Regarding the classification process for scRNA-seq data, an appropriate method is the k Nearest Neighbor (kNN) classifier since it is usually utilized for large-scale prediction tasks due to its simplicity, minimal parameterization, and model-free nature. However, the ultra-high dimensionality that characterizes scRNA-seq impose a computational bottleneck, while prediction power can be affected by the "Curse of Dimensionality". In this work, we proposed the utilization of approximate nearest neighbor search algorithms for the task of kNN classification in scRNA-seq data focusing on a particular methodology tailored for high dimensional data. We argue that even relaxed approximate solutions will not affect the prediction performance significantly. The experimental results confirm the original assumption by offering the potential for broader applicability.

arXiv:2012.01506공식 코드 · Papers with Code

Few-Shot Classification with Feature Map Reconstruction Networks

In this paper we reformulate few-shot classification as a reconstruction problem in latent space. The ability of the network to reconstruct a query feature map from support features of a given class predicts membership of the query in that class. We introduce a novel mechanism for few-shot classification by regressing directly from support features to query features in closed form, without introducing any new modules or large-scale learnable parameters. The resulting Feature Map Reconstruction Networks are both more performant and computationally efficient than previous approaches. We demonstrate consistent and substantial accuracy gains on four fine-grained benchmarks with varying neural architectures. Our model is also competitive on the non-fine-grained mini-ImageNet and tiered-ImageNet benchmarks with minimal bells and whistles.

arXiv:2012.01469공식 코드 · Papers with Code

Fair Attribute Classification through Latent Space De-biasing

Fairness in visual recognition is becoming a prominent and critical topic of discussion as recognition systems are deployed at scale in the real world. Models trained from data in which target labels are correlated with protected attributes (e.g., gender, race) are known to learn and exploit those correlations. In this work, we introduce a method for training accurate target classifiers while mitigating biases that stem from these correlations. We use GANs to generate realistic-looking images, and perturb these images in the underlying latent space to generate training data that is balanced for each protected attribute. We augment the original dataset with this perturbed generated data, and empirically demonstrate that target classifiers trained on the augmented dataset exhibit a number of both quantitative and qualitative benefits. We conduct a thorough evaluation across multiple target labels and protected attributes in the CelebA dataset, and provide an in-depth analysis and comparison to existing literature in the space.

arXiv:2011.04545

Augmenting transferred representations for stock classification

Stock classification is a challenging task due to high levels of noise and volatility of stocks returns. In this paper we show that using transfer learning can help with this task, by pre-training a model to extract universal features on the full universe of stocks of the S$\&$P500 index and then transferring it to another model to directly learn a trading rule. Transferred models present more than double the risk-adjusted returns than their counterparts trained from zero. In addition, we propose the use of data augmentation on the feature space defined as the output of a pre-trained model (i.e. augmenting the aggregated time-series representation). We compare this augmentation approach with the standard one, i.e. augmenting the time-series in the input space. We show that augmentation methods on the feature space leads to $20\%$ increase in risk-adjusted return compared to a model trained with transfer learning but without augmentation.

arXiv:2010.08503

Classification of Manifest Huntington Disease using Vowel Distortion Measures

Huntington disease (HD) is a fatal autosomal dominant neurocognitive disorder that causes cognitive disturbances, neuropsychiatric symptoms, and impaired motor abilities (e.g., gait, speech, voice). Due to its progressive nature, HD treatment requires ongoing clinical monitoring of symptoms. Individuals with the gene mutation which causes HD may exhibit a range of speech symptoms as they progress from premanifest to manifest HD. Differentiating between premanifest and manifest HD is an important yet understudied problem, as this distinction marks the need for increased treatment. Speech-based passive monitoring has the potential to augment clinical assessments by continuously tracking manifestation symptoms. In this work we present the first demonstration of how changes in connected speech can be measured to differentiate between premanifest and manifest HD. To do so, we focus on a key speech symptom of HD: vowel distortion. We introduce a set of vowel features which we extract from connected speech. We show that our vowel features can differentiate between premanifest and manifest HD with 87% accuracy.

arXiv:2009.14422

Small Drone Classification with Light CNN and New Micro-Doppler Signature Extraction Method Based on A-SPC Technique

As the threats of small drones increase, not only the detection but also the classification of small drones has become important. Many recent studies have applied an approach to utilize the micro-Doppler signature (MDS) for the small drone classification by using frequency modulated continuous wave (FMCW) radars. In this letter, we propose a novel method to extract the MDS images of the small drones with the FMCW radar. Moreover, we propose a light convolutional neural network (CNN) whose structure is straightforward, and the number of parameters is quite small for fast classification. The proposed method contributes to increasing the classification accuracy by improving the quality of MDS images. We classified the small drones with the MDS images extracted by the conventional method and the proposed method through the proposed CNN. The experimental results showed that the total classification accuracy was increased by 10.00 % due to the proposed method. The total classification accuracy was recorded at 97.14 % with the proposed MDS extraction method and the proposed light CNN.

arXiv:2009.11426

A Research Review on Detection and Classification of Power Quality Disturbances caused by Integration of Renewable Energy Sources

With the increased interest in integrating renewable energy sources (RES) such as wind power and solar into the power systems owing to their zero greenhouse gas emissions and the involvement of power converters for integration in grid, the detection, classification and mitigation of power quality events has become indispensable. For employing an appropriate mitigation technique, it is a pre-requisite to correctly classify the various types of disturbances in power quality. This paper, therefore, presents a detailed research reviews on detection and classification of power quality disturbances caused by the integration of renewable energy sources and associated works present in literature till date. Attempts are also made to highlight the current and future issues involved in the detection, classification and mitigation of PQ disturbances. Best efforts have been made to make this paper serve as a full-fledged reference for the future work in this field. A list of 230 research publications on the subject is also appended for quick reference.

arXiv:2008.12487

Classification of Imagined Speech Using Siamese Neural Network

Imagined speech is spotlighted as a new trend in the brain-machine interface due to its application as an intuitive communication tool. However, previous studies have shown low classification performance, therefore its use in real-life is not feasible. In addition, no suitable method to analyze it has been found. Recently, deep learning algorithms have been applied to this paradigm. However, due to the small amount of data, the increase in classification performance is limited. To tackle these issues, in this study, we proposed an end-to-end framework using Siamese neural network encoder, which learns the discriminant features by considering the distance between classes. The imagined words (e.g., arriba (up), abajo (down), derecha (right), izquierda (left), adelante (forward), and atr\'as (backward)) were classified using the raw electroencephalography (EEG) signals. We obtained a 6-class classification accuracy of 31.40% for imagined speech, which significantly outperformed other methods. This was possible because the Siamese neural network, which increases the distance between dissimilar samples while decreasing the distance between similar samples, was used. In this regard, our method can learn discriminant features from a small dataset. The proposed framework would help to increase the classification performance of imagined speech for a small amount of data and implement an intuitive communication system.

arXiv:2008.06042

Image Processing Tools for Financial Time Series Classification

The application of deep learning to time series forecasting is one of the major challenges in present machine learning. We propose a novel methodology that combines machine learning and image processing methods to define and predict market states with intraday financial data. A wavelet transform is applied to the log-return of stock prices for both image extraction and denoising. A convolutional neural network then extracts patterns from denoised wavelet images to classify daily time series, i.e. a market state is associated with the binary prediction of the daily close price movement based on the wavelet image constructed from the price changes in the first hours of the day. This method overcomes the low signal-to-noise ratio problem in financial time series and gets a competitive prediction accuracy of the market states 'Up' and 'Down' of financial data as tested on the S&P 500.

arXiv:2008.04874

Classification of Radio Signals Using Truncated Gaussian Discriminant Analysis of Convolutional Neural Network-Derived Features

To improve the utility and scalability of distributed radio frequency (RF) sensor and communication networks, reduce the need for convolutional neural network (CNN) retraining, and efficiently share learned information about signals, we examined a supervised bootstrapping approach for RF modulation classification. We show that CNN-bootstrapped features of new and existing modulation classes can be considered as mixtures of truncated Gaussian distributions, allowing for maximumlikelihood-based classification of new classes without retraining the network. In this work, the authors observed classification performance using maximum likelihood estimation of CNNbootstrapped features to be comparable to that of a CNN trained on all classes, even for those classes on which the bootstrapping CNN was not trained. This performance was achieved while reducing the number of parameters needed for new class definition from over 8 million to only 200. Furthermore, some physical features of interest, not directly labeled during training, e.g. signal-to-noise ratio (SNR), can be learned or estimated from these same CNN-derived features. Finally, we show that SNR estimation accuracy is highest when classification accuracy is lowest and therefore can be used to calibrate a confidence in the classification.

arXiv:2008.04768

Constrained Active Classification Using Partially Observable Markov Decision Processes

In this work, we study the problem of actively classifying the attributes of dynamical systems characterized as a finite set of Markov decision process (MDP) models. We are interested in finding strategies that actively interact with the dynamical system and observe its reactions so that the attribute of interest is classified efficiently with high confidence. We present a decision-theoretic framework based on partially observable Markov decision processes (POMDPs). The proposed framework relies on assigning a classification belief (a probability distribution) to the attributes of interest. Given an initial belief, a confidence level over which a classification decision can be made, a cost bound, safe belief sets, and a finite time horizon, we compute POMDP strategies leading to classification decisions. We present three different algorithms to compute such strategies. The first algorithm computes the optimal strategy exactly by value iteration. To overcome the computational complexity of computing the exact solutions, we propose a second algorithm based on adaptive sampling and a third based on a Monte Carlo tree search to approximate the optimal probability of reaching a classification decision. We illustrate the proposed methodology using examples from medical diagnosis, security surveillance, and wildlife classification.

arXiv:2008.00431

Contact Classification in COVID-19 Tracing

The present paper addresses the task of reliably identifying critical contacts by using COVID-19 tracing apps. A reliable classification is crucial to ensure a high level of protection, and at the same time to prevent many people from being sent to quarantine by the app. Tracing apps are based on the capabilities of current smartphones to enable a broadest possible availability. Existing capabilities of smartphones include the exchange of Bluetooth Low Energy (BLE) signals and of audio signals, as well as the use of gyroscopes and magnetic sensors. The Bluetooth power measurements, which are often used today, may be complemented by audio ranging and attitude estimation in the future. Smartphones are worn in different ways, often in pockets and bags, which makes the propagation of signals and thus the classification rather unpredictable. Relying on the cooperation of users to wear their phones hanging from their neck would change the situation considerably. In this case the performance, achievable with BLE and audio measurements, becomes predictable. Our analysis identifies parameters that result in accurate warnings, at least within the scope of validity of the models. A significant reduction of the spreading of the disease can then be achieved by the apps, without causing many people to unduly go to quarantine. The present paper is the first of three papers which analyze the situation in some detail.

arXiv:2007.11877

Proposal for a Comprehensive (Crypto) Asset Taxonomy

Developments in the distributed ledger technology have led to new types of assets with a broad range of purposes. Although some classification frameworks for common instruments from traditional finance and some for these new, so called cryptographic assets already exist and are used, a holistic approach to integrate both worlds is missing. The present paper fills this research gap by identifying 14 attributes, each of which is assigned different characteristics, that can be used to classify all types of assets in a structured manner. Our proposed taxonomy which is an extension of existing classification frameworks, summarises these findings in a morphological box and is tested for practicability by classifying exemplary assets like cash and bitcoin. The final classification framework can help to ensure that the various stakeholders, such as investors or supervisors, have a consistent view of the different types of assets, and in particular of their characteristics, and also helps to establish standardised terminology.

arXiv:2006.11781

Wavelet Classification for Over-the-Air Non-Orthogonal Waveforms

Non-cooperative communications using non-orthogonal multicarrier signals are challenging since self-created inter carrier interference (ICI) exists, which would prevent successful signal classification. Deep learning (DL) can deal with the classification task without domain-knowledge at the cost of training complexity since neural network hyperparameters have to be extensively tuned. Previous work showed that a tremendously trained convolutional neural network (CNN) classifier can efficiently identify feature-diversity dominant signals while it failed when feature-similarity dominates. Therefore, a pre-processing strategy, which can amplify signal feature diversity is of great importance. This work applies single-level wavelet transform to manually extract time-frequency features from non-orthogonal signals. Composite statistical features are investigated and the wavelet enabled two-dimensional time-frequency feature grid is further simplified into a one-dimensional feature vector via proper statistical transform. The dimensionality reduced features are fed to an error-correcting output codes (ECOC) model, consisting of multiple binary support vector machine (SVM) learners, for multiclass signal classification. Low-cost experiments reveal 100% classification accuracy for feature-diversity dominant signals and 90% for feature-similarity dominant signals, which is nearly 28% accuracy improvement when compared with the CNN classification results.

arXiv:2006.11655

A Novel Method for ECG Signal Classification via One-Dimensional Convolutional Neural Network

This paper presents an end-to-end ECG signal classification method based on a novel segmentation strategy via 1D Convolutional Neural Networks (CNN) to aid the classification of ECG signals. The ECG segmentation strategy named R-R-R strategy (i.e., retaining ECG data between the R peaks just before and after the current R peak) for segmenting the original ECG data into segments in order to train and test the 1D CNN models. The novel strategy mimics physicians in scanning ECG to a greater extent, and maximizes the inherent information of ECG segments. The performance of the classification models for 5-class and 6-class are verified with ECG signals from 48 records of the MIT-BIH arrhythmia database. As the heartbeat types are divided into 5 classes (i.e., normal beat, left bundle branch block beat, right bundle branch block beat, ventricular ectopic beat, and paced beat) in the MIT-BIH, the best classification accuracy, the area under the curve (AUC), the sensitivity and the F1-score reach 99.24%, 0.9994, 0.99 and 0.99, respectively. As the heartbeat types are divided into 6 classes (i.e., normal beat, left bundle branch block beat, right bundle branch block beat, ventricular ectopic beat, paced beat and other beats) in the MIT-BIH, the beat classification accuracy, the AUC, the sensitivity, and the F1-score reach 97.02%, 0.9966, 0.97, and 0.97, respectively. Meanwhile, according to the recommended practice from the Association for Advancement of Medical Instrumentation (AAMI), the heartbeat types are divided into 5 classes (i.e., normal beat, supraventricular ectopic beats, ventricular ectopic beats, fusion beats, and unclassifiable beats), the beat classification accuracy, the sensitivity, and the F1-score reach 97.45%, 0.97, and 0.97, respectively. The experimental results show that the proposed method achieves better performance than the state-of-the-art methods.

arXiv:2006.11393

Unifying Few- and Zero-Shot Egocentric Action Recognition

Although there has been significant research in egocentric action recognition, most methods and tasks, including EPIC-KITCHENS, suppose a fixed set of action classes. Fixed-set classification is useful for benchmarking methods, but is often unrealistic in practical settings due to the compositionality of actions, resulting in a functionally infinite-cardinality label set. In this work, we explore generalization with an open set of classes by unifying two popular approaches: few- and zero-shot generalization (the latter which we reframe as cross-modal few-shot generalization). We propose a new set of splits derived from the EPIC-KITCHENS dataset that allow evaluation of open-set classification, and use these splits to show that adding a metric-learning loss to the conventional direct-alignment baseline can improve zero-shot classification by as much as 10%, while not sacrificing few-shot performance.