PyTorchKR Papers

최신 논문

총 1,324편
태그·기간 등 상세 필터

카테고리 탐색

로그인 후 이 태그 구독

arXiv:2011.10568

Learn to Bind and Grow Neural Structures

Task-incremental learning involves the challenging problem of learning new tasks continually, without forgetting past knowledge. Many approaches address the problem by expanding the structure of a shared neural network as tasks arrive, but struggle to grow optimally, without losing past knowledge. We present a new framework, Learn to Bind and Grow, which learns a neural architecture for a new task incrementally, either by binding with layers of a similar task or by expanding layers which are more likely to conflict between tasks. Central to our approach is a novel, interpretable, parameterization of the shared, multi-task architecture space, which then enables computing globally optimal architectures using Bayesian optimization. Experiments on continual learning benchmarks show that our framework performs comparably with earlier expansion based approaches and is able to flexibly compute multiple optimal solutions with performance-size trade-offs.

arXiv:2011.10216

Sequential Targeting: an incremental learning approach for data imbalance in text classification

Classification tasks require a balanced distribution of data to ensure the learner to be trained to generalize over all classes. In real-world datasets, however, the number of instances vary substantially among classes. This typically leads to a learner that promotes bias towards the majority group due to its dominating property. Therefore, methods to handle imbalanced datasets are crucial for alleviating distributional skews and fully utilizing the under-represented data, especially in text classification. While addressing the imbalance in text data, most methods utilize sampling methods on the numerical representation of the data, which limits its efficiency on how effective the representation is. We propose a novel training method, Sequential Targeting(ST), independent of the effectiveness of the representation method, which enforces an incremental learning setting by splitting the data into mutually exclusive subsets and training the learner adaptively. To address problems that arise within incremental learning, we apply elastic weight consolidation. We demonstrate the effectiveness of our method through experiments on simulated benchmark datasets (IMDB) and data collected from NAVER.

arXiv:2011.09554

A Cognitive Approach based on the Actionable Knowledge Graph for supporting Maintenance Operations

In the era of Industry 4.0, cognitive computing and its enabling technologies (Artificial Intelligence, Machine Learning, etc.) allow to define systems able to support maintenance by providing relevant information, at the right time, retrieved from structured companies' databases, and unstructured documents, like technical manuals, intervention reports, and so on. Moreover, contextual information plays a crucial role in tailoring the support both during the planning and the execution of interventions. Contextual information can be detected with the help of sensors, wearable devices, indoor and outdoor positioning systems, and object recognition capabilities (using fixed or wearable cameras), all of which can collect historical data for further analysis. In this work, we propose a cognitive system that learns from past interventions to generate contextual recommendations for improving maintenance practices in terms of time, budget, and scope. The system uses formal conceptual models, incremental learning, and ranking algorithms to accomplish these objectives.

arXiv:2011.05734

A CNN-based Feature Space for Semi-supervised Incremental Learning in Assisted Living Applications

A Convolutional Neural Network (CNN) is sometimes confronted with objects of changing appearance ( new instances) that exceed its generalization capability. This requires the CNN to incorporate new knowledge, i.e., to learn incrementally. In this paper, we are concerned with this problem in the context of assisted living. We propose using the feature space that results from the training dataset to automatically label problematic images that could not be properly recognized by the CNN. The idea is to exploit the extra information in the feature space for a semi-supervised labeling and to employ problematic images to improve the CNN's classification model. Among other benefits, the resulting semi-supervised incremental learning process allows improving the classification accuracy of new instances by 40% as illustrated by extensive experiments.

arXiv:2011.01844공식 코드 · Papers with Code

A Comprehensive Study of Class Incremental Learning Algorithms for Visual Tasks

The ability of artificial agents to increment their capabilities when confronted with new data is an open challenge in artificial intelligence. The main challenge faced in such cases is catastrophic forgetting, i.e., the tendency of neural networks to underfit past data when new ones are ingested. A first group of approaches tackles forgetting by increasing deep model capacity to accommodate new knowledge. A second type of approaches fix the deep model size and introduce a mechanism whose objective is to ensure a good compromise between stability and plasticity of the model. While the first type of algorithms were compared thoroughly, this is not the case for methods which exploit a fixed size model. Here, we focus on the latter, place them in a common conceptual and experimental framework and propose the following contributions: (1) define six desirable properties of incremental learning algorithms and analyze them according to these properties, (2) introduce a unified formalization of the class-incremental learning problem, (3) propose a common evaluation framework which is more thorough than existing ones in terms of number of datasets, size of datasets, size of bounded memory and number of incremental states, (4) investigate the usefulness of herding for past exemplars selection, (5) provide experimental evidence that it is possible to obtain competitive performance without the use of knowledge distillation to tackle catastrophic forgetting and (6) facilitate reproducibility by integrating all tested methods in a common open-source repository. The main experimental finding is that none of the existing algorithms achieves the best results in all evaluated settings. Important differences arise notably if a bounded memory of past classes is allowed or not.

arXiv:2010.15378

Collaborative Method for Incremental Learning on Classification and Generation

Although well-trained deep neural networks have shown remarkable performance on numerous tasks, they rapidly forget what they have learned as soon as they begin to learn with additional data with the previous data stop being provided. In this paper, we introduce a novel algorithm, Incremental Class Learning with Attribute Sharing (ICLAS), for incremental class learning with deep neural networks. As one of its component, we also introduce a generative model, incGAN, which can generate images with increased variety compared with the training data. Under challenging environment of data deficiency, ICLAS incrementally trains classification and the generation networks. Since ICLAS trains both networks, our algorithm can perform multiple times of incremental class learning. The experiments on MNIST dataset demonstrate the advantages of our algorithm.

arXiv:2010.15277공식 코드 · Papers with Code

Class-incremental learning: survey and performance evaluation on image classification

For future learning systems, incremental learning is desirable because it allows for: efficient resource usage by eliminating the need to retrain from scratch at the arrival of new data; reduced memory usage by preventing or limiting the amount of data required to be stored -- also important when privacy limitations are imposed; and learning that more closely resembles human learning. The main challenge for incremental learning is catastrophic forgetting, which refers to the precipitous drop in performance on previously learned tasks after learning a new one. Incremental learning of deep neural networks has seen explosive growth in recent years. Initial work focused on task-incremental learning, where a task-ID is provided at inference time. Recently, we have seen a shift towards class-incremental learning where the learner must discriminate at inference time between all classes seen in previous tasks without recourse to a task-ID. In this paper, we provide a complete survey of existing class-incremental learning methods for image classification, and in particular, we perform an extensive experimental evaluation on thirteen class-incremental methods. We consider several new experimental scenarios, including a comparison of class-incremental methods on multiple large-scale image classification datasets, an investigation into small and large domain shifts, and a comparison of various network architectures.

arXiv:2010.08657공식 코드 · Papers with Code

Class-incremental Learning with Pre-allocated Fixed Classifiers

In class-incremental learning, a learning agent faces a stream of data with the goal of learning new classes while not forgetting previous ones. Neural networks are known to suffer under this setting, as they forget previously acquired knowledge. To address this problem, effective methods exploit past data stored in an episodic memory while expanding the final classifier nodes to accommodate the new classes. In this work, we substitute the expanding classifier with a novel fixed classifier in which a number of pre-allocated output nodes are subject to the classification loss right from the beginning of the learning phase. Contrarily to the standard expanding classifier, this allows: (a) the output nodes of future unseen classes to firstly see negative samples since the beginning of learning together with the positive samples that incrementally arrive; (b) to learn features that do not change their geometric configuration as novel classes are incorporated in the learning model. Experiments with public datasets show that the proposed approach is as effective as the expanding classifier while exhibiting novel intriguing properties of the internal feature representation that are otherwise not-existent. Our ablation study on pre-allocating a large number of classes further validates the approach.

arXiv:2010.08020

On the Exploration of Incremental Learning for Fine-grained Image Retrieval

In this paper, we consider the problem of fine-grained image retrieval in an incremental setting, when new categories are added over time. On the one hand, repeatedly training the representation on the extended dataset is time-consuming. On the other hand, fine-tuning the learned representation only with the new classes leads to catastrophic forgetting. To this end, we propose an incremental learning method to mitigate retrieval performance degradation caused by the forgetting issue. Without accessing any samples of the original classes, the classifier of the original network provides soft "labels" to transfer knowledge to train the adaptive network, so as to preserve the previous capability for classification. More importantly, a regularization function based on Maximum Mean Discrepancy is devised to minimize the discrepancy of new classes features from the original network and the adaptive network, respectively. Extensive experiments on two datasets show that our method effectively mitigates the catastrophic forgetting on the original classes while achieving high performance on the new classes.

arXiv:2010.05154공식 코드 · Papers with Code

Lambda Learner: Fast Incremental Learning on Data Streams

One of the most well-established applications of machine learning is in deciding what content to show website visitors. When observation data comes from high-velocity, user-generated data streams, machine learning methods perform a balancing act between model complexity, training time, and computational costs. Furthermore, when model freshness is critical, the training of models becomes time-constrained. Parallelized batch offline training, although horizontally scalable, is often not time-considerate or cost-effective. In this paper, we propose Lambda Learner, a new framework for training models by incremental updates in response to mini-batches from data streams. We show that the resulting model of our framework closely estimates a periodically updated model trained on offline data and outperforms it when model updates are time-sensitive. We provide theoretical proof that the incremental learning updates improve the loss-function over a stale batch model. We present a large-scale deployment on the sponsored content platform for a large social network, serving hundreds of millions of users across different channels (e.g., desktop, mobile). We address challenges and complexities from both algorithms and infrastructure perspectives, and illustrate the system details for computation, storage, and streaming production of training data.

arXiv:2010.05063공식 코드 · Papers with Code

Adaptive Aggregation Networks for Class-Incremental Learning

Class-Incremental Learning (CIL) aims to learn a classification model with the number of classes increasing phase-by-phase. An inherent problem in CIL is the stability-plasticity dilemma between the learning of old and new classes, i.e., high-plasticity models easily forget old classes, but high-stability models are weak to learn new classes. We alleviate this issue by proposing a novel network architecture called Adaptive Aggregation Networks (AANets), in which we explicitly build two types of residual blocks at each residual level (taking ResNet as the baseline architecture): a stable block and a plastic block. We aggregate the output feature maps from these two blocks and then feed the results to the next-level blocks. We adapt the aggregation weights in order to balance these two types of blocks, i.e., to balance stability and plasticity, dynamically. We conduct extensive experiments on three CIL benchmarks: CIFAR-100, ImageNet-Subset, and ImageNet, and show that many existing CIL methods can be straightforwardly incorporated into the architecture of AANets to boost their performances.

arXiv:2010.04605공식 코드 · Papers with Code

Instance Weighted Incremental Evolution Strategies for Reinforcement Learning in Dynamic Environments

Evolution strategies (ES), as a family of black-box optimization algorithms, recently emerge as a scalable alternative to reinforcement learning (RL) approaches such as Q-learning or policy gradient, and are much faster when many central processing units (CPUs) are available due to better parallelization. In this paper, we propose a systematic incremental learning method for ES in dynamic environments. The goal is to adjust previously learned policy to a new one incrementally whenever the environment changes. We incorporate an instance weighting mechanism with ES to facilitate its learning adaptation, while retaining scalability of ES. During parameter updating, higher weights are assigned to instances that contain more new knowledge, thus encouraging the search distribution to move towards new promising areas of parameter space. We propose two easy-to-implement metrics to calculate the weights: instance novelty and instance quality. Instance novelty measures an instance's difference from the previous optimum in the original environment, while instance quality corresponds to how well an instance performs in the new environment. The resulting algorithm, Instance Weighted Incremental Evolution Strategies (IW-IES), is verified to achieve significantly improved performance on challenging RL tasks ranging from robot navigation to locomotion. This paper thus introduces a family of scalable ES algorithms for RL domains that enables rapid learning adaptation to dynamic environments.

arXiv:2010.02784공식 코드 · Papers with Code

A Multi-Task Incremental Learning Framework with Category Name Embedding for Aspect-Category Sentiment Analysis

(T)ACSA tasks, including aspect-category sentiment analysis (ACSA) and targeted aspect-category sentiment analysis (TACSA), aims at identifying sentiment polarity on predefined categories. Incremental learning on new categories is necessary for (T)ACSA real applications. Though current multi-task learning models achieve good performance in (T)ACSA tasks, they suffer from catastrophic forgetting problems in (T)ACSA incremental learning tasks. In this paper, to make multi-task learning feasible for incremental learning, we proposed Category Name Embedding network (CNE-net). We set both encoder and decoder shared among all categories to weaken the catastrophic forgetting problem. Besides the origin input sentence, we applied another input feature, i.e., category name, for task discrimination. Our model achieved state-of-the-art on two (T)ACSA benchmark datasets. Furthermore, we proposed a dataset for (T)ACSA incremental learning and achieved the best performance compared with other strong baselines.

arXiv:2009.09078

Beyond Social Media Analytics: Understanding Human Behaviour and Deep Emotion using Self Structuring Incremental Machine Learning

This thesis develops a conceptual framework considering social data as representing the surface layer of a hierarchy of human social behaviours, needs and cognition which is employed to transform social data into representations that preserve social behaviours and their causalities. Based on this framework two platforms were built to capture insights from fast-paced and slow-paced social data. For fast-paced, a self-structuring and incremental learning technique was developed to automatically capture salient topics and corresponding dynamics over time. An event detection technique was developed to automatically monitor those identified topic pathways for significant fluctuations in social behaviours using multiple indicators such as volume and sentiment. This platform is demonstrated using two large datasets with over 1 million tweets. The separated topic pathways were representative of the key topics of each entity and coherent against topic coherence measures. Identified events were validated against contemporary events reported in news. Secondly for the slow-paced social data, a suite of new machine learning and natural language processing techniques were developed to automatically capture self-disclosed information of the individuals such as demographics, emotions and timeline of personal events. This platform was trialled on a large text corpus of over 4 million posts collected from online support groups. This was further extended to transform prostate cancer related online support group discussions into a multidimensional representation and investigated the self-disclosed quality of life of patients (and partners) against time, demographics and clinical factors. The capabilities of this extended platform have been demonstrated using a text corpus collected from 10 prostate cancer online support groups comprising of 609,960 prostate cancer discussions and 22,233 patients.

arXiv:2009.10073

Contextual Bandits for adapting to changing User preferences over time

Contextual bandits provide an effective way to model the dynamic data problem in ML by leveraging online (incremental) learning to continuously adjust the predictions based on changing environment. We explore details on contextual bandits, an extension to the traditional reinforcement learning (RL) problem and build a novel algorithm to solve this problem using an array of action-based learners. We apply this approach to model an article recommendation system using an array of stochastic gradient descent (SGD) learners to make predictions on rewards based on actions taken. We then extend the approach to a publicly available MovieLens dataset and explore the findings. First, we make available a simplified simulated dataset showing varying user preferences over time and how this can be evaluated with static and dynamic learning algorithms. This dataset made available as part of this research is intentionally simulated with limited number of features and can be used to evaluate different problem-solving strategies. We will build a classifier using static dataset and evaluate its performance on this dataset. We show limitations of static learner due to fixed context at a point of time and how changing that context brings down the accuracy. Next we develop a novel algorithm for solving the contextual bandit problem. Similar to the linear bandits, this algorithm maps the reward as a function of context vector but uses an array of learners to capture variation between actions/arms. We develop a bandit algorithm using an array of stochastic gradient descent (SGD) learners, with separate learner per arm. Finally, we will apply this contextual bandit algorithm to predicting movie ratings over time by different users from the standard Movie Lens dataset and demonstrate the results.

arXiv:2009.05105

What am I allowed to do here?: Online Learning of Context-Specific Norms by Pepper

Social norms support coordination and cooperation in society. With social robots becoming increasingly involved in our society, they also need to follow the social norms of the society. This paper presents a computational framework for learning contexts and the social norms present in a context in an online manner on a robot. The paper utilizes a recent state-of-the-art approach for incremental learning and adapts it for online learning of scenes (contexts). The paper further utilizes Dempster-Schafer theory to model context-specific norms. After learning the scenes (contexts), we use active learning to learn related norms. We test our approach on the Pepper robot by taking it through different scene locations. Our results show that Pepper can learn different scenes and related norms simply by communicating with a human partner in an online manner.

arXiv:2009.02147

A Practical Incremental Method to Train Deep CTR Models

Deep learning models in recommender systems are usually trained in the batch mode, namely iteratively trained on a fixed-size window of training data. Such batch mode training of deep learning models suffers from low training efficiency, which may lead to performance degradation when the model is not produced on time. To tackle this issue, incremental learning is proposed and has received much attention recently. Incremental learning has great potential in recommender systems, as two consecutive window of training data overlap most of the volume. It aims to update the model incrementally with only the newly incoming samples from the timestamp when the model is updated last time, which is much more efficient than the batch mode training. However, most of the incremental learning methods focus on the research area of image recognition where new tasks or classes are learned over time. In this work, we introduce a practical incremental method to train deep CTR models, which consists of three decoupled modules (namely, data, feature and model module). Our method can achieve comparable performance to the conventional batch mode training with much better training efficiency. We conduct extensive experiments on a public benchmark and a private dataset to demonstrate the effectiveness of our proposed method.

arXiv:2009.01803

Sparse Meta Networks for Sequential Adaptation and its Application to Adaptive Language Modelling

Training a deep neural network requires a large amount of single-task data and involves a long time-consuming optimization phase. This is not scalable to complex, realistic environments with new unexpected changes. Humans can perform fast incremental learning on the fly and memory systems in the brain play a critical role. We introduce Sparse Meta Networks -- a meta-learning approach to learn online sequential adaptation algorithms for deep neural networks, by using deep neural networks. We augment a deep neural network with a layer-specific fast-weight memory. The fast-weights are generated sparsely at each time step and accumulated incrementally through time providing a useful inductive bias for online continual adaptation. We demonstrate strong performance on a variety of sequential adaptation scenarios, from a simple online reinforcement learning to a large scale adaptive language modelling.

arXiv:2009.00919공식 코드 · Papers with Code

Continual Prototype Evolution: Learning Online from Non-Stationary Data Streams

Attaining prototypical features to represent class distributions is well established in representation learning. However, learning prototypes online from streaming data proves a challenging endeavor as they rapidly become outdated, caused by an ever-changing parameter space during the learning process. Additionally, continual learning does not assume the data stream to be stationary, typically resulting in catastrophic forgetting of previous knowledge. As a first, we introduce a system addressing both problems, where prototypes evolve continually in a shared latent space, enabling learning and prediction at any point in time. In contrast to the major body of work in continual learning, data streams are processed in an online fashion, without additional task-information, and an efficient memory scheme provides robustness to imbalanced data streams. Besides nearest neighbor based prediction, learning is facilitated by a novel objective function, encouraging cluster density about the class prototype and increased inter-class variance. Furthermore, the latent space quality is elevated by pseudo-prototypes in each batch, constituted by replay of exemplars from memory. As an additional contribution, we generalize the existing paradigms in continual learning to incorporate data incremental learning from data streams by formalizing a two-agent learner-evaluator framework. We obtain state-of-the-art performance by a significant margin on eight benchmarks, including three highly imbalanced data streams.

arXiv:2008.13517공식 코드 · Papers with Code

GraphSAIL: Graph Structure Aware Incremental Learning for Recommender Systems

Given the convenience of collecting information through online services, recommender systems now consume large scale data and play a more important role in improving user experience. With the recent emergence of Graph Neural Networks (GNNs), GNN-based recommender models have shown the advantage of modeling the recommender system as a user-item bipartite graph to learn representations of users and items. However, such models are expensive to train and difficult to perform frequent updates to provide the most up-to-date recommendations. In this work, we propose to update GNN-based recommender models incrementally so that the computation time can be greatly reduced and models can be updated more frequently. We develop a Graph Structure Aware Incremental Learning framework, GraphSAIL, to address the commonly experienced catastrophic forgetting problem that occurs when training a model in an incremental fashion. Our approach preserves a user's long-term preference (or an item's long-term property) during incremental model updating. GraphSAIL implements a graph structure preservation strategy which explicitly preserves each node's local structure, global structure, and self-information, respectively. We argue that our incremental training framework is the first attempt tailored for GNN based recommender systems and demonstrate its improvement compared to other incremental learning techniques on two public datasets. We further verify the effectiveness of our framework on a large-scale industrial dataset.