PyTorchKR Papers

최신 논문

총 2,786편
태그·기간 등 상세 필터

카테고리 탐색

로그인 후 이 태그 구독

arXiv:2112.14146공식 코드 · Papers with Code

Towards continual task learning in artificial neural networks: current approaches and insights from neuroscience

The innate capacity of humans and other animals to learn a diverse, and often interfering, range of knowledge and skills throughout their lifespan is a hallmark of natural intelligence, with obvious evolutionary motivations. In parallel, the ability of artificial neural networks (ANNs) to learn across a range of tasks and domains, combining and re-using learned representations where required, is a clear goal of artificial intelligence. This capacity, widely described as continual learning, has become a prolific subfield of research in machine learning. Despite the numerous successes of deep learning in recent years, across domains ranging from image recognition to machine translation, such continual task learning has proved challenging. Neural networks trained on multiple tasks in sequence with stochastic gradient descent often suffer from representational interference, whereby the learned weights for a given task effectively overwrite those of previous tasks in a process termed catastrophic forgetting. This represents a major impediment to the development of more generalised artificial learning systems, capable of accumulating knowledge over time and task space, in a manner analogous to humans. A repository of selected papers and implementations accompanying this review can be found at https://github.com/mccaffary/continual-learning.

arXiv:2112.13410

Generative Kernel Continual learning

Kernel continual learning by \citet{derakhshani2021kernel} has recently emerged as a strong continual learner due to its non-parametric ability to tackle task interference and catastrophic forgetting. Unfortunately its success comes at the expense of an explicit memory to store samples from past tasks, which hampers scalability to continual learning settings with a large number of tasks. In this paper, we introduce generative kernel continual learning, which explores and exploits the synergies between generative models and kernels for continual learning. The generative model is able to produce representative samples for kernel learning, which removes the dependence on memory in kernel continual learning. Moreover, as we replay only on the generative model, we avoid task interference while being computationally more efficient compared to previous methods that need replay on the entire model. We further introduce a supervised contrastive regularization, which enables our model to generate even more discriminative samples for better kernel-based classification performance. We conduct extensive experiments on three widely-used continual learning benchmarks that demonstrate the abilities and benefits of our contributions. Most notably, on the challenging SplitCIFAR100 benchmark, with just a simple linear kernel we obtain the same accuracy as kernel continual learning with variational random features for one tenth of the memory, or a 10.1\% accuracy gain for the same memory budget.

arXiv:2112.11944공식 코드 · Papers with Code

Continual learning of longitudinal health records

Continual learning denotes machine learning methods which can adapt to new environments while retaining and reusing knowledge gained from past experiences. Such methods address two issues encountered by models in non-stationary environments: ungeneralisability to new data, and the catastrophic forgetting of previous knowledge when retrained. This is a pervasive problem in clinical settings where patient data exhibits covariate shift not only between populations, but also continuously over time. However, while continual learning methods have seen nascent success in the imaging domain, they have been little applied to the multi-variate sequential data characteristic of critical care patient recordings. Here we evaluate a variety of continual learning methods on longitudinal ICU data in a series of representative healthcare scenarios. We find that while several methods mitigate short-term forgetting, domain shift remains a challenging problem over large series of tasks, with only replay based methods achieving stable long-term performance. Code for reproducing all experiments can be found at https://github.com/iacobo/continual

arXiv:2112.09153공식 코드 · Papers with Code

An Empirical Investigation of the Role of Pre-training in Lifelong Learning

The lifelong learning paradigm in machine learning is an attractive alternative to the more prominent isolated learning scheme not only due to its resemblance to biological learning but also its potential to reduce energy waste by obviating excessive model re-training. A key challenge to this paradigm is the phenomenon of catastrophic forgetting. With the increasing popularity and success of pre-trained models in machine learning, we pose the question: What role does pre-training play in lifelong learning, specifically with respect to catastrophic forgetting? We investigate existing methods in the context of large, pre-trained models and evaluate their performance on a variety of text and image classification tasks, including a large-scale study using a novel data set of 15 diverse NLP tasks. Across all settings, we observe that generic pre-training implicitly alleviates the effects of catastrophic forgetting when learning multiple tasks sequentially compared to randomly initialized models. We then further investigate why pre-training alleviates forgetting in this setting. We study this phenomenon by analyzing the loss landscape, finding that pre-trained weights appear to ease forgetting by leading to wider minima. Based on this insight, we propose jointly optimizing for current task loss and loss basin sharpness to explicitly encourage wider basins during sequential fine-tuning. We show that this optimization approach outperforms several state-of-the-art task-sequential continual learning algorithms across multiple settings, occasionally even without retaining a memory that scales in size with the number of tasks.

arXiv:2112.10021공식 코드 · Papers with Code

Continual Learning with Knowledge Transfer for Sentiment Classification

This paper studies continual learning (CL) for sentiment classification (SC). In this setting, the CL system learns a sequence of SC tasks incrementally in a neural network, where each task builds a classifier to classify the sentiment of reviews of a particular product category or domain. Two natural questions are: Can the system transfer the knowledge learned in the past from the previous tasks to the new task to help it learn a better model for the new task? And, can old models for previous tasks be improved in the process as well? This paper proposes a novel technique called KAN to achieve these objectives. KAN can markedly improve the SC accuracy of both the new task and the old tasks via forward and backward knowledge transfer. The effectiveness of KAN is demonstrated through extensive experiments.

arXiv:2112.10017공식 코드 · Papers with Code

Continual Learning of a Mixed Sequence of Similar and Dissimilar Tasks

Existing research on continual learning of a sequence of tasks focused on dealing with catastrophic forgetting, where the tasks are assumed to be dissimilar and have little shared knowledge. Some work has also been done to transfer previously learned knowledge to the new task when the tasks are similar and have shared knowledge. To the best of our knowledge, no technique has been proposed to learn a sequence of mixed similar and dissimilar tasks that can deal with forgetting and also transfer knowledge forward and backward. This paper proposes such a technique to learn both types of tasks in the same network. For dissimilar tasks, the algorithm focuses on dealing with forgetting, and for similar tasks, the algorithm focuses on selectively transferring the knowledge learned from some similar previous tasks to improve the new task learning. Additionally, the algorithm automatically detects whether a new task is similar to any previous tasks. Empirical evaluation using sequences of mixed tasks demonstrates the effectiveness of the proposed model.

arXiv:2112.08906

On the Uncertain Single-View Depths in Colonoscopies

Estimating depth information from endoscopic images is a prerequisite for a wide set of AI-assisted technologies, such as accurate localization and measurement of tumors, or identification of non-inspected areas. As the domain specificity of colonoscopies -- deformable low-texture environments with fluids, poor lighting conditions and abrupt sensor motions -- pose challenges to multi-view 3D reconstructions, single-view depth learning stands out as a promising line of research. Depth learning can be extended in a Bayesian setting, which enables continual learning, improves decision making and can be used to compute confidence intervals or quantify uncertainty for in-body measurements. In this paper, we explore for the first time Bayesian deep networks for single-view depth estimation in colonoscopies. Our specific contribution is two-fold: 1) an exhaustive analysis of scalable Bayesian networks for depth learning in different datasets, highlighting challenges and conclusions regarding synthetic-to-real domain changes and supervised vs. self-supervised methods; and 2) a novel teacher-student approach to deep depth learning that takes into account the teacher uncertainty.

arXiv:2112.08654공식 코드 · Papers with Code

Learning to Prompt for Continual Learning

The mainstream paradigm behind continual learning has been to adapt the model parameters to non-stationary data distributions, where catastrophic forgetting is the central challenge. Typical methods rely on a rehearsal buffer or known task identity at test time to retrieve learned knowledge and address forgetting, while this work presents a new paradigm for continual learning that aims to train a more succinct memory system without accessing task identity at test time. Our method learns to dynamically prompt (L2P) a pre-trained model to learn tasks sequentially under different task transitions. In our proposed framework, prompts are small learnable parameters, which are maintained in a memory space. The objective is to optimize prompts to instruct the model prediction and explicitly manage task-invariant and task-specific knowledge while maintaining model plasticity. We conduct comprehensive experiments under popular image classification benchmarks with different challenging continual learning settings, where L2P consistently outperforms prior state-of-the-art methods. Surprisingly, L2P achieves competitive results against rehearsal-based methods even without a rehearsal buffer and is directly applicable to challenging task-agnostic continual learning. Source code is available at https://github.com/google-research/l2p.

arXiv:2112.07774

Assessing Human Interaction in Virtual Reality With Continually Learning Prediction Agents Based on Reinforcement Learning Algorithms: A Pilot Study

Artificial intelligence systems increasingly involve continual learning to enable flexibility in general situations that are not encountered during system training. Human interaction with autonomous systems is broadly studied, but research has hitherto under-explored interactions that occur while the system is actively learning, and can noticeably change its behaviour in minutes. In this pilot study, we investigate how the interaction between a human and a continually learning prediction agent develops as the agent develops competency. Additionally, we compare two different agent architectures to assess how representational choices in agent design affect the human-agent interaction. We develop a virtual reality environment and a time-based prediction task wherein learned predictions from a reinforcement learning (RL) algorithm augment human predictions. We assess how a participant's performance and behaviour in this task differs across agent types, using both quantitative and qualitative analyses. Our findings suggest that human trust of the system may be influenced by early interactions with the agent, and that trust in turn affects strategic behaviour, but limitations of the pilot study rule out any conclusive statement. We identify trust as a key feature of interaction to focus on when considering RL-based technologies, and make several recommendations for modification to this study in preparation for a larger-scale investigation. A video summary of this paper can be found at https://youtu.be/oVYJdnBqTwQ .

arXiv:2112.07066공식 코드 · Papers with Code

Continual Learning In Environments With Polynomial Mixing Times

The mixing time of the Markov chain induced by a policy limits performance in real-world continual learning scenarios. Yet, the effect of mixing times on learning in continual reinforcement learning (RL) remains underexplored. In this paper, we characterize problems that are of long-term interest to the development of continual RL, which we call scalable MDPs, through the lens of mixing times. In particular, we theoretically establish that scalable MDPs have mixing times that scale polynomially with the size of the problem. We go on to demonstrate that polynomial mixing times present significant difficulties for existing approaches, which suffer from myopic bias and stale bootstrapped estimates. To validate our theory, we study the empirical scaling behavior of mixing times with respect to the number of tasks and task duration for high performing policies deployed across multiple Atari games. Our analysis demonstrates both that polynomial mixing times do emerge in practice and how their existence may lead to unstable learning behavior like catastrophic forgetting in continual learning settings.

arXiv:2112.05779

Quantum Architecture Search via Continual Reinforcement Learning

Quantum computing has promised significant improvement in solving difficult computational tasks over classical computers. Designing quantum circuits for practical use, however, is not a trivial objective and requires expert-level knowledge. To aid this endeavor, this paper proposes a machine learning-based method to construct quantum circuit architectures. Previous works have demonstrated that classical deep reinforcement learning (DRL) algorithms can successfully construct quantum circuit architectures without encoded physics knowledge. However, these DRL-based works are not generalizable to settings with changing device noises, thus requiring considerable amounts of training resources to keep the RL models up-to-date. With this in mind, we incorporated continual learning to enhance the performance of our algorithm. In this paper, we present the Probabilistic Policy Reuse with deep Q-learning (PPR-DQL) framework to tackle this circuit design challenge. By conducting numerical simulations over various noise patterns, we demonstrate that the RL agent with PPR was able to find the quantum gate sequence to generate the two-qubit Bell state faster than the agent that was trained from scratch. The proposed framework is general and can be applied to other quantum gate synthesis or control problems -- including the automatic calibration of quantum devices.

arXiv:2112.06511공식 코드 · Papers with Code

Ex-Model: Continual Learning from a Stream of Trained Models

Learning continually from non-stationary data streams is a challenging research topic of growing popularity in the last few years. Being able to learn, adapt, and generalize continually in an efficient, effective, and scalable way is fundamental for a sustainable development of Artificial Intelligent systems. However, an agent-centric view of continual learning requires learning directly from raw data, which limits the interaction between independent agents, the efficiency, and the privacy of current approaches. Instead, we argue that continual learning systems should exploit the availability of compressed information in the form of trained models. In this paper, we introduce and formalize a new paradigm named "Ex-Model Continual Learning" (ExML), where an agent learns from a sequence of previously trained models instead of raw data. We further contribute with three ex-model continual learning algorithms and an empirical setting comprising three datasets (MNIST, CIFAR-10 and CORe50), and eight scenarios, where the proposed algorithms are extensively tested. Finally, we highlight the peculiarities of the ex-model paradigm and we point out interesting future research directions.

arXiv:2112.05677

Concept Representation Learning with Contrastive Self-Supervised Learning

Concept-oriented deep learning (CODL) is a general approach to meet the future challenges for deep learning: (1) learning with little or no external supervision, (2) coping with test examples that come from a different distribution than the training examples, and (3) integrating deep learning with symbolic AI. In CODL, as in human learning, concept representations are learned based on concept exemplars. Contrastive self-supervised learning (CSSL) provides a promising approach to do so, since it: (1) uses data-driven associations, to get away from semantic labels, (2) supports incremental and continual learning, to get away from (large) fixed datasets, and (3) accommodates emergent objectives, to get away from fixed objectives (tasks). We discuss major aspects of concept representation learning using CSSL. These include dual-level concept representations, CSSL for feature representations, exemplar similarity measures and self-supervised relational reasoning, incremental and continual CSSL, and contrastive self-supervised concept (class) incremental learning. The discussion leverages recent findings from cognitive neural science and CSSL.

arXiv:2112.05025

Gradient-matching coresets for continual learning

We devise a coreset selection method based on the idea of gradient matching: The gradients induced by the coreset should match, as closely as possible, those induced by the original training dataset. We evaluate the method in the context of continual learning, where it can be used to curate a rehearsal memory. Our method performs strong competitors such as reservoir sampling across a range of memory sizes.

arXiv:2112.05095공식 코드 · Papers with Code

Provable Continual Learning via Sketched Jacobian Approximations

An important problem in machine learning is the ability to learn tasks in a sequential manner. If trained with standard first-order methods most models forget previously learned tasks when trained on a new task, which is often referred to as catastrophic forgetting. A popular approach to overcome forgetting is to regularize the loss function by penalizing models that perform poorly on previous tasks. For example, elastic weight consolidation (EWC) regularizes with a quadratic form involving a diagonal matrix build based on past data. While EWC works very well for some setups, we show that, even under otherwise ideal conditions, it can provably suffer catastrophic forgetting if the diagonal matrix is a poor approximation of the Hessian matrix of previous tasks. We propose a simple approach to overcome this: Regularizing training of a new task with sketches of the Jacobian matrix of past data. This provably enables overcoming catastrophic forgetting for linear models and for wide neural networks, at the cost of memory. The overarching goal of this paper is to provided insights on when regularization-based continual learning algorithms work and under what memory costs.

arXiv:2112.04215공식 코드 · Papers with Code

Self-Supervised Models are Continual Learners

Self-supervised models have been shown to produce comparable or better visual representations than their supervised counterparts when trained offline on unlabeled data at scale. However, their efficacy is catastrophically reduced in a Continual Learning (CL) scenario where data is presented to the model sequentially. In this paper, we show that self-supervised loss functions can be seamlessly converted into distillation mechanisms for CL by adding a predictor network that maps the current state of the representations to their past state. This enables us to devise a framework for Continual self-supervised visual representation Learning that (i) significantly improves the quality of the learned representations, (ii) is compatible with several state-of-the-art self-supervised objectives, and (iii) needs little to no hyperparameter tuning. We demonstrate the effectiveness of our approach empirically by training six popular self-supervised models in various CL settings.

arXiv:2112.04467

CoMPS: Continual Meta Policy Search

We develop a new continual meta-learning method to address challenges in sequential multi-task learning. In this setting, the agent's goal is to achieve high reward over any sequence of tasks quickly. Prior meta-reinforcement learning algorithms have demonstrated promising results in accelerating the acquisition of new tasks. However, they require access to all tasks during training. Beyond simply transferring past experience to new tasks, our goal is to devise continual reinforcement learning algorithms that learn to learn, using their experience on previous tasks to learn new tasks more quickly. We introduce a new method, continual meta-policy search (CoMPS), that removes this limitation by meta-training in an incremental fashion, over each task in a sequence, without revisiting prior tasks. CoMPS continuously repeats two subroutines: learning a new task using RL and using the experience from RL to perform completely offline meta-learning to prepare for subsequent task learning. We find that CoMPS outperforms prior continual learning and off-policy meta-reinforcement methods on several sequences of challenging continuous control tasks.

arXiv:2112.03271공식 코드 · Papers with Code

Adapting BERT for Continual Learning of a Sequence of Aspect Sentiment Classification Tasks

This paper studies continual learning (CL) of a sequence of aspect sentiment classification (ASC) tasks. Although some CL techniques have been proposed for document sentiment classification, we are not aware of any CL work on ASC. A CL system that incrementally learns a sequence of ASC tasks should address the following two issues: (1) transfer knowledge learned from previous tasks to the new task to help it learn a better model, and (2) maintain the performance of the models for previous tasks so that they are not forgotten. This paper proposes a novel capsule network based model called B-CL to address these issues. B-CL markedly improves the ASC performance on both the new task and the old tasks via forward and backward knowledge transfer. The effectiveness of B-CL is demonstrated through extensive experiments.

arXiv:2112.02995공식 코드 · Papers with Code

TaskDrop: A Competitive Baseline for Continual Learning of Sentiment Classification

In this paper, we study the multi-task sentiment classification problem in the continual learning setting, i.e., a model is sequentially trained to classifier the sentiment of reviews of products in a particular category. The use of common sentiment words in reviews of different product categories leads to large cross-task similarity, which differentiates it from continual learning in other domains. This knowledge sharing nature renders forgetting reduction focused approaches less effective for the problem under consideration. Unlike existing approaches, where task-specific masks are learned with specifically presumed training objectives, we propose an approach called Task-aware Dropout (TaskDrop) to generate masks in a random way. While the standard dropout generates and applies random masks for each training instance per epoch for effective regularization, TaskDrop applies random masking for task-wise capacity allocation and reuse. We conducted experimental studies on three multi-task review datasets and made comparison to various baselines and state-of-the-art approaches. Our empirical results show that regardless of simplicity, TaskDrop overall achieved competitive performances for all the three datasets, especially after relative long term learning. This demonstrates that the proposed random capacity allocation mechanism works well for continual sentiment classification.

arXiv:2112.03252

CSG0: Continual Urban Scene Generation with Zero Forgetting

With the rapid advances in generative adversarial networks (GANs), the visual quality of synthesised scenes keeps improving, including for complex urban scenes with applications to automated driving. We address in this work a continual scene generation setup in which GANs are trained on a stream of distinct domains; ideally, the learned models should eventually be able to generate new scenes in all seen domains. This setup reflects the real-life scenario where data are continuously acquired in different places at different times. In such a continual setup, we aim for learning with zero forgetting, \IE, with no degradation in synthesis quality over earlier domains due to catastrophic forgetting. To this end, we introduce a novel framework that not only (i) enables seamless knowledge transfer in continual training but also (ii) guarantees zero forgetting with a small overhead cost. While being more memory efficient, thanks to continual learning, our model obtains better synthesis quality as compared against the brute-force solution that trains one full model for each domain. Especially, under extreme low-data regimes, our approach outperforms the brute-force one by a large margin.