PyTorchKR Papers

최신 논문

총 2,786편
태그·기간 등 상세 필터

카테고리 탐색

로그인 후 이 태그 구독

arXiv:2209.03603

nVFNet-RDC: Replay and Non-Local Distillation Collaboration for Continual Object Detection

Continual Learning (CL) focuses on developing algorithms with the ability to adapt to new environments and learn new skills. This very challenging task has generated a lot of interest in recent years, with new solutions appearing rapidly. In this paper, we propose a nVFNet-RDC approach for continual object detection. Our nVFNet-RDC consists of teacher-student models, and adopts replay and feature distillation strategies. As the 1st place solutions, we achieve 55.94% and 54.65% average mAP on the 3rd CLVision Challenge Track 2 and Track 3, respectively.

arXiv:2209.03182공식 코드 · Papers with Code

On the Effectiveness of Compact Biomedical Transformers

Language models pre-trained on biomedical corpora, such as BioBERT, have recently shown promising results on downstream biomedical tasks. Many existing pre-trained models, on the other hand, are resource-intensive and computationally heavy owing to factors such as embedding size, hidden dimension, and number of layers. The natural language processing (NLP) community has developed numerous strategies to compress these models utilising techniques such as pruning, quantisation, and knowledge distillation, resulting in models that are considerably faster, smaller, and subsequently easier to use in practice. By the same token, in this paper we introduce six lightweight models, namely, BioDistilBERT, BioTinyBERT, BioMobileBERT, DistilBioBERT, TinyBioBERT, and CompactBioBERT which are obtained either by knowledge distillation from a biomedical teacher or continual learning on the Pubmed dataset via the Masked Language Modelling (MLM) objective. We evaluate all of our models on three biomedical tasks and compare them with BioBERT-v1.1 to create efficient lightweight models that perform on par with their larger counterparts. All the models will be publicly available on our Huggingface profile at https://huggingface.co/nlpie and the codes used to run the experiments will be available at https://github.com/nlpie-research/Compact-Biomedical-Transformers.

arXiv:2209.02370공식 코드 · Papers with Code

Continual Learning, Fast and Slow

According to the Complementary Learning Systems (CLS) theory~\cite{mcclelland1995there} in neuroscience, humans do effective \emph{continual learning} through two complementary systems: a fast learning system centered on the hippocampus for rapid learning of the specifics, individual experiences; and a slow learning system located in the neocortex for the gradual acquisition of structured knowledge about the environment. Motivated by this theory, we propose \emph{DualNets} (for Dual Networks), a general continual learning framework comprising a fast learning system for supervised learning of pattern-separated representation from specific tasks and a slow learning system for representation learning of task-agnostic general representation via Self-Supervised Learning (SSL). DualNets can seamlessly incorporate both representation types into a holistic framework to facilitate better continual learning in deep neural networks. Via extensive experiments, we demonstrate the promising results of DualNets on a wide range of continual learning protocols, ranging from the standard offline, task-aware setting to the challenging online, task-free scenario. Notably, on the CTrL~\cite{veniat2020efficient} benchmark that has unrelated tasks with vastly different visual images, DualNets can achieve competitive performance with existing state-of-the-art dynamic architecture strategies~\cite{ostapenko2021continual}. Furthermore, we conduct comprehensive ablation studies to validate DualNets efficacy, robustness, and scalability. Code will be made available at \url{https://github.com/phquang/DualNet}.

arXiv:2209.02112공식 코드 · Papers with Code

Class-Incremental Learning via Knowledge Amalgamation

Catastrophic forgetting has been a significant problem hindering the deployment of deep learning algorithms in the continual learning setting. Numerous methods have been proposed to address the catastrophic forgetting problem where an agent loses its generalization power of old tasks while learning new tasks. We put forward an alternative strategy to handle the catastrophic forgetting with knowledge amalgamation (CFA), which learns a student network from multiple heterogeneous teacher models specializing in previous tasks and can be applied to current offline methods. The knowledge amalgamation process is carried out in a single-head manner with only a selected number of memorized samples and no annotations. The teachers and students do not need to share the same network structure, allowing heterogeneous tasks to be adapted to a compact or sparse data representation. We compare our method with competitive baselines from different strategies, demonstrating our approach's advantages.

arXiv:2209.01558공식 코드 · Papers with Code

Scalable Adversarial Online Continual Learning

Adversarial continual learning is effective for continual learning problems because of the presence of feature alignment process generating task-invariant features having low susceptibility to the catastrophic forgetting problem. Nevertheless, the ACL method imposes considerable complexities because it relies on task-specific networks and discriminators. It also goes through an iterative training process which does not fit for online (one-epoch) continual learning problems. This paper proposes a scalable adversarial continual learning (SCALE) method putting forward a parameter generator transforming common features into task-specific features and a single discriminator in the adversarial game to induce common features. The training process is carried out in meta-learning fashions using a new combination of three loss functions. SCALE outperforms prominent baselines with noticeable margins in both accuracy and execution time.

arXiv:2209.01556

Reinforced Continual Learning for Graphs

Graph Neural Networks (GNNs) have become the backbone for a myriad of tasks pertaining to graphs and similar topological data structures. While many works have been established in domains related to node and graph classification/regression tasks, they mostly deal with a single task. Continual learning on graphs is largely unexplored and existing graph continual learning approaches are limited to the task-incremental learning scenarios. This paper proposes a graph continual learning strategy that combines the architecture-based and memory-based approaches. The structural learning strategy is driven by reinforcement learning, where a controller network is trained in such a way to determine an optimal number of nodes to be added/pruned from the base network when new tasks are observed, thus assuring sufficient network capacities. The parameter learning strategy is underpinned by the concept of Dark Experience replay method to cope with the catastrophic forgetting problem. Our approach is numerically validated with several graph continual learning benchmark problems in both task-incremental learning and class-incremental learning settings. Compared to recently published works, our approach demonstrates improved performance in both the settings. The implementation code can be found at \url{https://github.com/codexhammer/gcl}.

arXiv:2209.01326

Continual Learning for Steganalysis

To detect the existing steganographic algorithms, recent steganalysis methods usually train a Convolutional Neural Network (CNN) model on the dataset consisting of corresponding paired cover/stego-images. However, it is inefficient and impractical for those steganalysis tools to completely retrain the CNN model to make it effective against both the existing steganographic algorithms and a new emerging steganographic algorithm. Thus, existing steganalysis models usually lack dynamic extensibility for new steganographic algorithms, which limits their application in real-world scenarios. To address this issue, we propose an accurate parameter importance estimation (APIE) based-continual learning scheme for steganalysis. In this scheme, when a steganalysis model is trained on the new image dataset generated by the new steganographic algorithm, its network parameters are effectively and efficiently updated with sufficient consideration of their importance evaluated in the previous training process. This approach can guide the steganalysis model to learn the patterns of the new steganographic algorithm without significantly degrading the detectability against the previous steganographic algorithms. Experimental results demonstrate the proposed scheme has promising extensibility for new emerging steganographic algorithms.

arXiv:2209.00591

Incremental Online Learning Algorithms Comparison for Gesture and Visual Smart Sensors

Tiny machine learning (TinyML) in IoT systems exploits MCUs as edge devices for data processing. However, traditional TinyML methods can only perform inference, limited to static environments or classes. Real case scenarios usually work in dynamic environments, thus drifting the context where the original neural model is no more suitable. For this reason, pre-trained models reduce accuracy and reliability during their lifetime because the data recorded slowly becomes obsolete or new patterns appear. Continual learning strategies maintain the model up to date, with runtime fine-tuning of the parameters. This paper compares four state-of-the-art algorithms in two real applications: i) gesture recognition based on accelerometer data and ii) image classification. Our results confirm these systems' reliability and the feasibility of deploying them in tiny-memory MCUs, with a drop in the accuracy of a few percentage points with respect to the original models for unconstrained computing platforms.

arXiv:2208.14307

Beyond Supervised Continual Learning: a Review

Continual Learning (CL, sometimes also termed incremental learning) is a flavor of machine learning where the usual assumption of stationary data distribution is relaxed or omitted. When naively applying, e.g., DNNs in CL problems, changes in the data distribution can cause the so-called catastrophic forgetting (CF) effect: an abrupt loss of previous knowledge. Although many significant contributions to enabling CL have been made in recent years, most works address supervised (classification) problems. This article reviews literature that study CL in other settings, such as learning with reduced supervision, fully unsupervised learning, and reinforcement learning. Besides proposing a simple schema for classifying CL approaches w.r.t. their level of autonomy and supervision, we discuss the specific challenges associated with each setting and the potential contributions to the field of CL in general.

arXiv:2208.13723공식 코드 · Papers with Code

Bayesian Continual Learning via Spiking Neural Networks

Among the main features of biological intelligence are energy efficiency, capacity for continual adaptation, and risk management via uncertainty quantification. Neuromorphic engineering has been thus far mostly driven by the goal of implementing energy-efficient machines that take inspiration from the time-based computing paradigm of biological brains. In this paper, we take steps towards the design of neuromorphic systems that are capable of adaptation to changing learning tasks, while producing well-calibrated uncertainty quantification estimates. To this end, we derive online learning rules for spiking neural networks (SNNs) within a Bayesian continual learning framework. In it, each synaptic weight is represented by parameters that quantify the current epistemic uncertainty resulting from prior knowledge and observed data. The proposed online rules update the distribution parameters in a streaming fashion as data are observed. We instantiate the proposed approach for both real-valued and binary synaptic weights. Experimental results using Intel's Lava platform show the merits of Bayesian over frequentist learning in terms of capacity for adaptation and uncertainty quantification.

arXiv:2208.12037공식 코드 · Papers with Code

Symbolic Replay: Scene Graph as Prompt for Continual Learning on VQA Task

VQA is an ambitious task aiming to answer any image-related question. However, in reality, it is hard to build such a system once for all since the needs of users are continuously updated, and the system has to implement new functions. Thus, Continual Learning (CL) ability is a must in developing advanced VQA systems. Recently, a pioneer work split a VQA dataset into disjoint answer sets to study this topic. However, CL on VQA involves not only the expansion of label sets (new Answer sets). It is crucial to study how to answer questions when deploying VQA systems to new environments (new Visual scenes) and how to answer questions requiring new functions (new Question types). Thus, we propose CLOVE, a benchmark for Continual Learning On Visual quEstion answering, which contains scene- and function-incremental settings for the two aforementioned CL scenarios. In terms of methodology, the main difference between CL on VQA and classification is that the former additionally involves expanding and preventing forgetting of reasoning mechanisms, while the latter focusing on class representation. Thus, we propose a real-data-free replay-based method tailored for CL on VQA, named Scene Graph as Prompt for Symbolic Replay. Using a piece of scene graph as a prompt, it replays pseudo scene graphs to represent the past images, along with correlated QA pairs. A unified VQA model is also proposed to utilize the current and replayed data to enhance its QA ability. Finally, experimental results reveal challenges in CLOVE and demonstrate the effectiveness of our method. The dataset and code will be available at https://github.com/showlab/CLVQA.

arXiv:2208.09734공식 코드 · Papers with Code

A Multi-Head Model for Continual Learning via Out-of-Distribution Replay

This paper studies class incremental learning (CIL) of continual learning (CL). Many approaches have been proposed to deal with catastrophic forgetting (CF) in CIL. Most methods incrementally construct a single classifier for all classes of all tasks in a single head network. To prevent CF, a popular approach is to memorize a small number of samples from previous tasks and replay them during training of the new task. However, this approach still suffers from serious CF as the parameters learned for previous tasks are updated or adjusted with only the limited number of saved samples in the memory. This paper proposes an entirely different approach that builds a separate classifier (head) for each task (called a multi-head model) using a transformer network, called MORE. Instead of using the saved samples in memory to update the network for previous tasks/classes in the existing approach, MORE leverages the saved samples to build a task specific classifier (adding a new classification head) without updating the network learned for previous tasks/classes. The model for the new task in MORE is trained to learn the classes of the task and also to detect samples that are not from the same data distribution (i.e., out-of-distribution (OOD)) of the task. This enables the classifier for the task to which the test instance belongs to produce a high score for the correct class and the classifiers of other tasks to produce low scores because the test instance is not from the data distributions of these classifiers. Experimental results show that MORE outperforms state-of-the-art baselines and is also naturally capable of performing OOD detection in the continual learning setting.

arXiv:2208.08112

DLCFT: Deep Linear Continual Fine-Tuning for General Incremental Learning

Pre-trained representation is one of the key elements in the success of modern deep learning. However, existing works on continual learning methods have mostly focused on learning models incrementally from scratch. In this paper, we explore an alternative framework to incremental learning where we continually fine-tune the model from a pre-trained representation. Our method takes advantage of linearization technique of a pre-trained neural network for simple and effective continual learning. We show that this allows us to design a linear model where quadratic parameter regularization method is placed as the optimal continual learning policy, and at the same time enjoying the high performance of neural networks. We also show that the proposed algorithm enables parameter regularization methods to be applied to class-incremental problems. Additionally, we provide a theoretical reason why the existing parameter-space regularization algorithms such as EWC underperform on neural networks trained with cross-entropy loss. We show that the proposed method can prevent forgetting while achieving high continual fine-tuning performance on image classification tasks. To show that our method can be applied to general continual learning settings, we evaluate our method in data-incremental, task-incremental, and class-incremental learning problems.

arXiv:2208.06931

A Theory for Knowledge Transfer in Continual Learning

Continual learning of a stream of tasks is an active area in deep neural networks. The main challenge investigated has been the phenomenon of catastrophic forgetting or interference of newly acquired knowledge with knowledge from previous tasks. Recent work has investigated forward knowledge transfer to new tasks. Backward transfer for improving knowledge gained during previous tasks has received much less attention. There is in general limited understanding of how knowledge transfer could aid tasks learned continually. We present a theory for knowledge transfer in continual supervised learning, which considers both forward and backward transfer. We aim at understanding their impact for increasingly knowledgeable learners. We derive error bounds for each of these transfer mechanisms. These bounds are agnostic to specific implementations (e.g. deep neural networks). We demonstrate that, for a continual learner that observes related tasks, both forward and backward transfer can contribute to an increasing performance as more tasks are observed.

arXiv:2208.06568공식 코드 · Papers with Code

On the Limitations of Continual Learning for Malware Classification

Malicious software (malware) classification offers a unique challenge for continual learning (CL) regimes due to the volume of new samples received on a daily basis and the evolution of malware to exploit new vulnerabilities. On a typical day, antivirus vendors receive hundreds of thousands of unique pieces of software, both malicious and benign, and over the course of the lifetime of a malware classifier, more than a billion samples can easily accumulate. Given the scale of the problem, sequential training using continual learning techniques could provide substantial benefits in reducing training and storage overhead. To date, however, there has been no exploration of CL applied to malware classification tasks. In this paper, we study 11 CL techniques applied to three malware tasks covering common incremental learning scenarios, including task, class, and domain incremental learning (IL). Specifically, using two realistic, large-scale malware datasets, we evaluate the performance of the CL methods on both binary malware classification (Domain-IL) and multi-class malware family classification (Task-IL and Class-IL) tasks. To our surprise, continual learning methods significantly underperformed naive Joint replay of the training data in nearly all settings -- in some cases reducing accuracy by more than 70 percentage points. A simple approach of selectively replaying 20% of the stored data achieves better performance, with 50% of the training time compared to Joint replay. Finally, we discuss potential reasons for the unexpectedly poor performance of the CL techniques, with the hope that it spurs further research on developing techniques that are more effective in the malware classification domain.

arXiv:2208.05388

ATLAS: Universal Function Approximator for Memory Retention

Artificial neural networks (ANNs), despite their universal function approximation capability and practical success, are subject to catastrophic forgetting. Catastrophic forgetting refers to the abrupt unlearning of a previous task when a new task is learned. It is an emergent phenomenon that hinders continual learning. Existing universal function approximation theorems for ANNs guarantee function approximation ability, but do not predict catastrophic forgetting. This paper presents a novel universal approximation theorem for multi-variable functions using only single-variable functions and exponential functions. Furthermore, we present ATLAS: a novel ANN architecture based on the new theorem. It is shown that ATLAS is a universal function approximator capable of some memory retention, and continual learning. The memory of ATLAS is imperfect, with some off-target effects during continual learning, but it is well-behaved and predictable. An efficient implementation of ATLAS is provided. Experiments are conducted to evaluate both the function approximation and memory retention capabilities of ATLAS.

arXiv:2208.04952공식 코드 · Papers with Code

Continual Prune-and-Select: Class-incremental learning with specialized subnetworks

The human brain is capable of learning tasks sequentially mostly without forgetting. However, deep neural networks (DNNs) suffer from catastrophic forgetting when learning one task after another. We address this challenge considering a class-incremental learning scenario where the DNN sees test data without knowing the task from which this data originates. During training, Continual-Prune-and-Select (CP&S) finds a subnetwork within the DNN that is responsible for solving a given task. Then, during inference, CP&S selects the correct subnetwork to make predictions for that task. A new task is learned by training available neuronal connections of the DNN (previously untrained) to create a new subnetwork by pruning, which can include previously trained connections belonging to other subnetwork(s) because it does not update shared connections. This enables to eliminate catastrophic forgetting by creating specialized regions in the DNN that do not conflict with each other while still allowing knowledge transfer across them. The CP&S strategy is implemented with different subnetwork selection strategies, revealing superior performance to state-of-the-art continual learning methods tested on various datasets (CIFAR-100, CUB-200-2011, ImageNet-100 and ImageNet-1000). In particular, CP&S is capable of sequentially learning 10 tasks from ImageNet-1000 keeping an accuracy around 94% with negligible forgetting, a first-of-its-kind result in class-incremental learning. To the best of the authors' knowledge, this represents an improvement in accuracy above 10% when compared to the best alternative method.

arXiv:2208.04287

Continual Reinforcement Learning with TELLA

Training reinforcement learning agents that continually learn across multiple environments is a challenging problem. This is made more difficult by a lack of reproducible experiments and standard metrics for comparing different continual learning approaches. To address this, we present TELLA, a tool for the Test and Evaluation of Lifelong Learning Agents. TELLA provides specified, reproducible curricula to lifelong learning agents while logging detailed data for evaluation and standardized analysis. Researchers can define and share their own curricula over various learning environments or run against a curriculum created under the DARPA Lifelong Learning Machines (L2M) Program.

arXiv:2208.03609

Continual Learning for Tumor Classification in Histopathology Images

Recent years have seen great advancements in the development of deep learning models for histopathology image analysis in digital pathology applications, evidenced by the increasingly common deployment of these models in both research and clinical settings. Although such models have shown unprecedented performance in solving fundamental computational tasks in DP applications, they suffer from catastrophic forgetting when adapted to unseen data with transfer learning. With an increasing need for deep learning models to handle ever changing data distributions, including evolving patient population and new diagnosis assays, continual learning models that alleviate model forgetting need to be introduced in DP based analysis. However, to our best knowledge, there is no systematic study of such models for DP-specific applications. Here, we propose CL scenarios in DP settings, where histopathology image data from different sources/distributions arrive sequentially, the knowledge of which is integrated into a single model without training all the data from scratch. We then established an augmented dataset for colorectal cancer H&E classification to simulate shifts of image appearance and evaluated CL model performance in the proposed CL scenarios. We leveraged a breast tumor H&E dataset along with the colorectal cancer to evaluate CL from different tumor types. In addition, we evaluated CL methods in an online few-shot setting under the constraints of annotation and computational resources. We revealed promising results of CL in DP applications, potentially paving the way for application of these methods in clinical practice.

arXiv:2208.04227

A Multi-label Continual Learning Framework to Scale Deep Learning Approaches for Packaging Equipment Monitoring

Continual Learning aims to learn from a stream of tasks, being able to remember at the same time both new and old tasks. While many approaches were proposed for single-class classification, multi-label classification in the continual scenario remains a challenging problem. For the first time, we study multi-label classification in the Domain Incremental Learning scenario. Moreover, we propose an efficient approach that has a logarithmic complexity with regard to the number of tasks, and can be applied also in the Class Incremental Learning scenario. We validate our approach on a real-world multi-label Alarm Forecasting problem from the packaging industry. For the sake of reproducibility, the dataset and the code used for the experiments are publicly available.