Current methods for training convolutional neural networks depend on large amounts of labeled samples for supervised training. In this paper we present an approach for training a convolutional neural network using only unlabeled data. We train the network to discriminate between a set of surrogate classes. Each surrogate class is formed by applying a variety of transformations to a randomly sampled'seed' image patch. We find that this simple feature learning algorithm is surprisingly successful when applied to visual object recognition. The feature representation learned by our algorithm achieves classification results matching or outperforming the current state-of-the-art for unsupervised learning on several popular datasets (STL-10, CIFAR-10, Caltech-101).

Neural Information Processing Systems http://nips.cc/

Feature transformation is to derive a new feature set from original features to augment the AI power of data. In many science domains such as material performance screening, while feature transformation can model material formula interactions and compositions and discover performance drivers, supervised labels are collected from expensive and lengthy experiments. This issue motivates an Unsupervised Feature Transformation Learning (UFTL) problem. Prior literature, such as manual transformation, supervised feedback guided search, and PCA, either relies on domain knowledge or expensive supervised feedback, or suffers from large search space, or overlooks non-linear feature-feature interactions. UFTL imposes a major challenge on existing methods: how to design a new unsupervised paradigm that captures complex feature interactions and avoids large search space? To fill this gap, we connect graph, contrastive, and generative learning to develop a measurement-pretrain-finetune paradigm for UFTL. For unsupervised feature set utility measurement, we propose a feature value consistency preservation perspective and develop a mean discounted cumulative gain like unsupervised metric to evaluate feature set utility. For unsupervised feature set representation pretraining, we regard a feature set as a feature-feature interaction graph, and develop an unsupervised graph contrastive learning encoder to embed feature sets into vectors. For generative transformation finetuning, we regard a feature set as a feature cross sequence and feature transformation as sequential generation. We develop a deep generative feature transformation model that coordinates the pretrained feature set encoder and the gradient information extracted from a feature set utility evaluator to optimize a transformed feature generator.

In recent years, deep learning approaches have gained significant interest as a way of building hierarchical representations from unlabeled data. However, to our knowledge, these deep learning approaches have not been extensively studied for auditory data. In this paper, we apply convolutional deep belief networks to audio data and empirically evaluate them on various audio classification tasks. For the case of speech data, we show that the learned features correspond to phones/phonemes. In addition, our feature representations trained from unlabeled audio data show very good performance for multiple audio classification tasks. We hope that this paper will inspire more research on deep learning approaches applied to a wide range of audio recognition tasks.

Despite its great success, backpropagation has certain limitations that necessitate the investigation of new learning methods. In this study, we present a biologically plausible local learning rule that improves upon Hebb's well-known proposal and discovers unsupervised features by local competitions among neurons. This simple learning rule enables the creation of a forward learning paradigm called activation learning, in which the output activation (sum of the squared output) of the neural network estimates the likelihood of the input patterns, or "learn more, activate more" in simpler terms. For classification on a few small classical datasets, activation learning performs comparably to backpropagation using a fully connected network, and outperforms backpropagation when there are fewer training samples or unpredictable disturbances. Additionally, the same trained network can be used for a variety of tasks, including image generation and completion. Activation learning also achieves state-of-the-art performance on several real-world datasets for anomaly detection. This new learning paradigm, which has the potential to unify supervised, unsupervised, and semi-supervised learning and is reasonably more resistant to adversarial attacks, deserves in-depth investigation. The backpropagation algorithm [1] has driven the recent success of machine learning in tasks such as speech and image recognition [2], language processing, image and music creation, playing human games [3], etc. Many scientists argue that the backpropagation algorithm, despite being a highly effective tool for training neural networks by minimizing specific loss functions, is different from the rules governing human learning [4]-[7]. One limitation of backpropagation is that the features learned by minimizing a particular loss function tend to be task-specific. This makes it difficult for the trained models to perform generic tasks, necessitates a large amount of labeled data, and renders them vulnerable to adversarial attacks [8]. Inspired by the brain, which is believed to learn in a predominantly unsupervised fashion [9], [10], we intend to create a new learning paradigm that enables forward unsupervised training of neural networks based on a simple local learning rule while achieving comparable performance to backpropagation. The fundamental idea is that, when every neuron in a layer competes to activate while presenting distinct features, the network transmits the maximum amount of information to the next layer, and learning is enforced. Hebbian plasticity is a local correlation-based learning rule proposed by Hebb that has been supported by experimental evidences such as long-term potentiation and depression [5], [11]. It is simply phrased as'cells that fire together wire together.' H. Zhou is with the School of Information Science and Engineering, Shandong University, Qingdao, Shandong, 266237 China.

Emotion recognition is involved in several real-world applications. With an increase in available modalities, automatic understanding of emotions is being performed more accurately. The success in Multimodal Emotion Recognition (MER), primarily relies on the supervised learning paradigm. However, data annotation is expensive, time-consuming, and as emotion expression and perception depends on several factors (e.g., age, gender, culture) obtaining labels with a high reliability is hard. Motivated by these, we focus on unsupervised feature learning for MER. We consider discrete emotions, and as modalities text, audio and vision are used. Our method, as being based on contrastive loss between pairwise modalities, is the first attempt in MER literature. Our end-to-end feature learning approach has several differences (and advantages) compared to existing MER methods: i) it is unsupervised, so the learning is lack of data labelling cost; ii) it does not require data spatial augmentation, modality alignment, large number of batch size or epochs; iii) it applies data fusion only at inference; and iv) it does not require backbones pre-trained on emotion recognition task. The experiments on benchmark datasets show that our method outperforms several baseline approaches and unsupervised learning methods applied in MER. Particularly, it even surpasses a few supervised MER state-of-the-art.

Unsupervised semantic segmentation aims to discover and localize semantically meaningful categories within image corpora without any form of annotation. To solve this task, algorithms must produce features for every pixel that are both semantically meaningful and compact enough to form distinct clusters. Unlike previous works which achieve this with a single end-to-end framework, we propose to separate feature learning from cluster compactification. Empirically, we show that current unsupervised feature learning frameworks already generate dense features whose correlations are semantically consistent. This observation motivates us to design STEGO (Self-supervised Transformer with Energy-based Graph Optimization), a novel framework that distills unsupervised features into high-quality discrete semantic labels.

In recent years, deep learning approaches have gained significant interest as a way of building hierarchical representations from unlabeled data. However, to our knowledge, these deep learning approaches have not been extensively studied for auditory data. In this paper, we apply convolutional deep belief networks to audio data and empirically evaluate them on various audio classification tasks. For the case of speech data, we show that the learned features correspond to phones/phonemes. In addition, our feature representations trained from unlabeled audio data show very good performance for multiple audio classification tasks.

Abstract--Automatic recognition of human activities from time-series sensor data (referred to as HAR) is a growing area of research in ubiquitous computing. Most recent research in the field adopts supervised deep learning paradigms to automate extraction of intrinsic features from raw signal inputs and addresses HAR as a multi-class classification problem where detecting a single activity class within the duration of a sensory data segment suffices. However, due to the innate diversity of human activities and their corresponding duration, no data segment is guaranteed to contain sensor recordings of a single activity type. In this paper, we express HAR more naturally as a set prediction problem where the predictions are sets of ongoing activity elements with unfixed and unknown cardinality. For the first time, we address this problem by presenting a novel HAR approach that learns to output activity sets using deep neural networks. Moreover, motivated by the limited availability of annotated HAR datasets as well as the unfortunate immaturity of existing unsupervised systems, we complement our supervised set learning scheme with a prior unsupervised feature learning process that adopts convolutional auto-encoders to exploit unlabeled data.The empirical experiments on two widely adopted HAR datasets demonstrate the substantial improvement of our proposed methodology over the baseline models. I. INTRODUCTION With the proliferation of low-cost sensing technologies as well as the fast advancements in machine learning techniques, automatic human activity recognition (HAR) using wearable sensors has emerged as a key research area in ubiquitous computing [1]-[4].In this problem, high-level activity information is acquired through analyzing low-level sensor recordings with the goal of providing proactive assistance to users.

In this paper, we present a technique for unsupervised learning of visual representations. Specifically, we train a model for foreground and background classification task, in the process of which it learns visual representations. Foreground and background patches for training come af- ter mining for such patches from hundreds and thousands of unlabelled videos available on the web which we ex- tract using a proposed patch extraction algorithm. With- out using any supervision, with just using 150, 000 unla- belled videos and the PASCAL VOC 2007 dataset, we train a object recognition model that achieves 45.3 mAP which is close to the best performing unsupervised feature learn- ing technique whereas better than many other proposed al- gorithms. The code for patch extraction is implemented in Matlab and available open source at the following link .