PyTorchKR Papers

최신 논문

총 2,786편
태그·기간 등 상세 필터

카테고리 탐색

로그인 후 이 태그 구독

arXiv:2307.05741

Towards Robust and Efficient Continual Language Learning

As the application space of language models continues to evolve, a natural question to ask is how we can quickly adapt models to new tasks. We approach this classic question from a continual learning perspective, in which we aim to continue fine-tuning models trained on past tasks on new tasks, with the goal of "transferring" relevant knowledge. However, this strategy also runs the risk of doing more harm than good, i.e., negative transfer. In this paper, we construct a new benchmark of task sequences that target different possible transfer scenarios one might face, such as a sequence of tasks with high potential of positive transfer, high potential for negative transfer, no expected effect, or a mixture of each. An ideal learner should be able to maximally exploit information from all tasks that have any potential for positive transfer, while also avoiding the negative effects of any distracting tasks that may confuse it. We then propose a simple, yet effective, learner that satisfies many of our desiderata simply by leveraging a selective strategy for initializing new models from past task checkpoints. Still, limitations remain, and we hope this benchmark can help the community to further build and analyze such learners.

arXiv:2307.05747공식 코드 · Papers with Code

Integrating Curricula with Replays: Its Effects on Continual Learning

Humans engage in learning and reviewing processes with curricula when acquiring new skills or knowledge. This human learning behavior has inspired the integration of curricula with replay methods in continual learning agents. The goal is to emulate the human learning process, thereby improving knowledge retention and facilitating learning transfer. Existing replay methods in continual learning agents involve the random selection and ordering of data from previous tasks, which has shown to be effective. However, limited research has explored the integration of different curricula with replay methods to enhance continual learning. Our study takes initial steps in examining the impact of integrating curricula with replay methods on continual learning in three specific aspects: the interleaved frequency of replayed exemplars with training data, the sequence in which exemplars are replayed, and the strategy for selecting exemplars into the replay buffer. These aspects of curricula design align with cognitive psychology principles and leverage the benefits of interleaved practice during replays, easy-to-hard rehearsal, and exemplar selection strategy involving exemplars from a uniform distribution of difficulties. Based on our results, these three curricula effectively mitigated catastrophic forgetting and enhanced positive knowledge transfer, demonstrating the potential of curricula in advancing continual learning methodologies. Our code and data are available: https://github.com/ZhangLab-DeepNeuroCogLab/Integrating-Curricula-with-Replays

arXiv:2307.05784공식 코드 · Papers with Code

EgoAdapt: A multi-stream evaluation study of adaptation to real-world egocentric user video

In egocentric action recognition a single population model is typically trained and subsequently embodied on a head-mounted device, such as an augmented reality headset. While this model remains static for new users and environments, we introduce an adaptive paradigm of two phases, where after pretraining a population model, the model adapts on-device and online to the user's experience. This setting is highly challenging due to the change from population to user domain and the distribution shifts in the user's data stream. Coping with the latter in-stream distribution shifts is the focus of continual learning, where progress has been rooted in controlled benchmarks but challenges faced in real-world applications often remain unaddressed. We introduce EgoAdapt, a benchmark for real-world egocentric action recognition that facilitates our two-phased adaptive paradigm, and real-world challenges naturally occur in the egocentric video streams from Ego4d, such as long-tailed action distributions and large-scale classification over 2740 actions. We introduce an evaluation framework that directly exploits the user's data stream with new metrics to measure the adaptation gain over the population model, online generalization, and hindsight performance. In contrast to single-stream evaluation in existing works, our framework proposes a meta-evaluation that aggregates the results from 50 independent user streams. We provide an extensive empirical study for finetuning and experience replay.

arXiv:2307.05399공식 코드 · Papers with Code

Domain-Agnostic Neural Architecture for Class Incremental Continual Learning in Document Processing Platform

Production deployments in complex systems require ML architectures to be highly efficient and usable against multiple tasks. Particularly demanding are classification problems in which data arrives in a streaming fashion and each class is presented separately. Recent methods with stochastic gradient learning have been shown to struggle in such setups or have limitations like memory buffers, and being restricted to specific domains that disable its usage in real-world scenarios. For this reason, we present a fully differentiable architecture based on the Mixture of Experts model, that enables the training of high-performance classifiers when examples from each class are presented separately. We conducted exhaustive experiments that proved its applicability in various domains and ability to learn online in production environments. The proposed technique achieves SOTA results without a memory buffer and clearly outperforms the reference methods.

arXiv:2307.04869

Fed-CPrompt: Contrastive Prompt for Rehearsal-Free Federated Continual Learning

Federated continual learning (FCL) learns incremental tasks over time from confidential datasets distributed across clients. This paper focuses on rehearsal-free FCL, which has severe forgetting issues when learning new tasks due to the lack of access to historical task data. To address this issue, we propose Fed-CPrompt based on prompt learning techniques to obtain task-specific prompts in a communication-efficient way. Fed-CPrompt introduces two key components, asynchronous prompt learning, and contrastive continual loss, to handle asynchronous task arrival and heterogeneous data distributions in FCL, respectively. Extensive experiments demonstrate the effectiveness of Fed-CPrompt in achieving SOTA rehearsal-free FCL performance.

arXiv:2307.04094

Class-Incremental Mixture of Gaussians for Deep Continual Learning

Continual learning models for stationary data focus on learning and retaining concepts coming to them in a sequential manner. In the most generic class-incremental environment, we have to be ready to deal with classes coming one by one, without any higher-level grouping. This requirement invalidates many previously proposed methods and forces researchers to look for more flexible alternative approaches. In this work, we follow the idea of centroid-driven methods and propose end-to-end incorporation of the mixture of Gaussians model into the continual learning framework. By employing the gradient-based approach and designing losses capable of learning discriminative features while avoiding degenerate solutions, we successfully combine the mixture model with a deep feature extractor allowing for joint optimization and adjustments in the latent space. Additionally, we show that our model can effectively learn in memory-free scenarios with fixed extractors. In the conducted experiments, we empirically demonstrate the effectiveness of the proposed solutions and exhibit the competitiveness of our model when compared with state-of-the-art continual learning baselines evaluated in the context of image classification problems.

arXiv:2307.04722

Advances and Challenges in Meta-Learning: A Technical Review

Meta-learning empowers learning systems with the ability to acquire knowledge from multiple tasks, enabling faster adaptation and generalization to new tasks. This review provides a comprehensive technical overview of meta-learning, emphasizing its importance in real-world applications where data may be scarce or expensive to obtain. The paper covers the state-of-the-art meta-learning approaches and explores the relationship between meta-learning and multi-task learning, transfer learning, domain adaptation and generalization, self-supervised learning, personalized federated learning, and continual learning. By highlighting the synergies between these topics and the field of meta-learning, the paper demonstrates how advancements in one area can benefit the field as a whole, while avoiding unnecessary duplication of efforts. Additionally, the paper delves into advanced meta-learning topics such as learning from complex multi-modal task distributions, unsupervised meta-learning, learning to efficiently adapt to data distribution shifts, and continual meta-learning. Lastly, the paper highlights open problems and challenges for future research in the field. By synthesizing the latest research developments, this paper provides a thorough understanding of meta-learning and its potential impact on various machine learning applications. We believe that this technical overview will contribute to the advancement of meta-learning and its practical implications in addressing real-world problems.

arXiv:2307.04345

Continual Learning as Computationally Constrained Reinforcement Learning

An agent that efficiently accumulates knowledge to develop increasingly sophisticated skills over a long lifetime could advance the frontier of artificial intelligence capabilities. The design of such agents, which remains a long-standing challenge of artificial intelligence, is addressed by the subject of continual learning. This monograph clarifies and formalizes concepts of continual learning, introducing a framework and set of tools to stimulate further research.

arXiv:2307.02615공식 코드 · Papers with Code

Human Inspired Progressive Alignment and Comparative Learning for Grounded Word Acquisition

Human language acquisition is an efficient, supervised, and continual process. In this work, we took inspiration from how human babies acquire their first language, and developed a computational process for word acquisition through comparative learning. Motivated by cognitive findings, we generated a small dataset that enables the computation models to compare the similarities and differences of various attributes, learn to filter out and extract the common information for each shared linguistic label. We frame the acquisition of words as not only the information filtration process, but also as representation-symbol mapping. This procedure does not involve a fixed vocabulary size, nor a discriminative objective, and allows the models to continually learn more concepts efficiently. Our results in controlled experiments have shown the potential of this approach for efficient continual learning of grounded words.

arXiv:2307.02738공식 코드 · Papers with Code

RecallM: An Adaptable Memory Mechanism with Temporal Understanding for Large Language Models

Large Language Models (LLMs) have made extraordinary progress in the field of Artificial Intelligence and have demonstrated remarkable capabilities across a large variety of tasks and domains. However, as we venture closer to creating Artificial General Intelligence (AGI) systems, we recognize the need to supplement LLMs with long-term memory to overcome the context window limitation and more importantly, to create a foundation for sustained reasoning, cumulative learning and long-term user interaction. In this paper we propose RecallM, a novel architecture for providing LLMs with an adaptable and updatable long-term memory mechanism. Unlike previous methods, the RecallM architecture is particularly effective at belief updating and maintaining a temporal understanding of the knowledge provided to it. We demonstrate through various experiments the effectiveness of this architecture. Furthermore, through our own temporal understanding and belief updating experiments, we show that RecallM is four times more effective than using a vector database for updating knowledge previously stored in long-term memory. We also demonstrate that RecallM shows competitive performance on general question-answering and in-context learning tasks.

arXiv:2307.02435

Exploring Continual Learning for Code Generation Models

Large-scale code generation models such as Codex and CodeT5 have achieved impressive performance. However, libraries are upgraded or deprecated very frequently and re-training large-scale language models is computationally expensive. Therefore, Continual Learning (CL) is an important aspect that remains underexplored in the code domain. In this paper, we introduce a benchmark called CodeTask-CL that covers a wide range of tasks, including code generation, translation, summarization, and refinement, with different input and output programming languages. Next, on our CodeTask-CL benchmark, we compare popular CL techniques from NLP and Vision domains. We find that effective methods like Prompt Pooling (PP) suffer from catastrophic forgetting due to the unstable training of the prompt selection mechanism caused by stark distribution shifts in coding tasks. We address this issue with our proposed method, Prompt Pooling with Teacher Forcing (PP-TF), that stabilizes training by enforcing constraints on the prompt selection mechanism and leads to a 21.54% improvement over Prompt Pooling. Along with the benchmark, we establish a training pipeline that can be used for CL on code models, which we believe can motivate further development of CL methods for code models. Our code is available at https://github.com/amazon-science/codetaskcl-pptf

arXiv:2307.01430공식 코드 · Papers with Code

Continual Learning in Open-vocabulary Classification with Complementary Memory Systems

We introduce a method for flexible and efficient continual learning in open-vocabulary image classification, drawing inspiration from the complementary learning systems observed in human cognition. Specifically, we propose to combine predictions from a CLIP zero-shot model and the exemplar-based model, using the zero-shot estimated probability that a sample's class is within the exemplar classes. We also propose a "tree probe" method, an adaption of lazy learning principles, which enables fast learning from new examples with competitive accuracy to batch-trained linear models. We test in data incremental, class incremental, and task incremental settings, as well as ability to perform flexible inference on varying subsets of zero-shot and learned categories. Our proposed method achieves a good balance of learning speed, target task effectiveness, and zero-shot effectiveness. Code will be available at https://github.com/jessemelpolio/TreeProbe.

arXiv:2307.02251공식 코드 · Papers with Code

RanPAC: Random Projections and Pre-trained Models for Continual Learning

Continual learning (CL) aims to incrementally learn different tasks (such as classification) in a non-stationary data stream without forgetting old ones. Most CL works focus on tackling catastrophic forgetting under a learning-from-scratch paradigm. However, with the increasing prominence of foundation models, pre-trained models equipped with informative representations have become available for various downstream requirements. Several CL methods based on pre-trained models have been explored, either utilizing pre-extracted features directly (which makes bridging distribution gaps challenging) or incorporating adaptors (which may be subject to forgetting). In this paper, we propose a concise and effective approach for CL with pre-trained models. Given that forgetting occurs during parameter updating, we contemplate an alternative approach that exploits training-free random projectors and class-prototype accumulation, which thus bypasses the issue. Specifically, we inject a frozen Random Projection layer with nonlinear activation between the pre-trained model's feature representations and output head, which captures interactions between features with expanded dimensionality, providing enhanced linear separability for class-prototype-based CL. We also demonstrate the importance of decorrelating the class-prototypes to reduce the distribution disparity when using pre-trained representations. These techniques prove to be effective and circumvent the problem of forgetting for both class- and domain-incremental continual learning. Compared to previous methods applied to pre-trained ViT-B/16 models, we reduce final error rates by between 20% and 62% on seven class-incremental benchmarks, despite not using any rehearsal memory. We conclude that the full potential of pre-trained models for simple, effective, and fast CL has not hitherto been fully tapped. Code is at github.com/RanPAC/RanPAC.

arXiv:2307.00123공식 코드 · Papers with Code

How Do Human Users Teach a Continual Learning Robot in Repeated Interactions?

Continual learning (CL) has emerged as an important avenue of research in recent years, at the intersection of Machine Learning (ML) and Human-Robot Interaction (HRI), to allow robots to continually learn in their environments over long-term interactions with humans. Most research in continual learning, however, has been robot-centered to develop continual learning algorithms that can quickly learn new information on static datasets. In this paper, we take a human-centered approach to continual learning, to understand how humans teach continual learning robots over the long term and if there are variations in their teaching styles. We conducted an in-person study with 40 participants that interacted with a continual learning robot in 200 sessions. In this between-participant study, we used two different CL models deployed on a Fetch mobile manipulator robot. An extensive qualitative and quantitative analysis of the data collected in the study shows that there is significant variation among the teaching styles of individual users indicating the need for personalized adaptation to their distinct teaching styles. The results also show that although there is a difference in the teaching styles between expert and non-expert users, the style does not have an effect on the performance of the continual learning robot. Finally, our analysis shows that the constrained experimental setups that have been widely used to test most continual learning techniques are not adequate, as real users interact with and teach continual learning robots in a variety of ways. Our code is available at https://github.com/aliayub7/cl_hri.

arXiv:2307.00183

Long-Tailed Continual Learning For Visual Food Recognition

Deep learning based food recognition has achieved remarkable progress in predicting food types given an eating occasion image. However, there are two major obstacles that hinder deployment in real world scenario. First, as new foods appear sequentially overtime, a trained model needs to learn the new classes continuously without causing catastrophic forgetting for already learned knowledge of existing food types. Second, the distribution of food images in real life is usually long-tailed as a small number of popular food types are consumed more frequently than others, which can vary in different populations. This requires the food recognition method to learn from class-imbalanced data by improving the generalization ability on instance-rare food classes. In this work, we focus on long-tailed continual learning and aim to address both aforementioned challenges. As existing long-tailed food image datasets only consider healthy people population, we introduce two new benchmark food image datasets, VFN-INSULIN and VFN-T2D, which exhibits on the real world food consumption for insulin takers and individuals with type 2 diabetes without taking insulin, respectively. We propose a novel end-to-end framework for long-tailed continual learning, which effectively addresses the catastrophic forgetting by applying an additional predictor for knowledge distillation to avoid misalignment of representation during continual learning. We also introduce a novel data augmentation technique by integrating class-activation-map (CAM) and CutMix, which significantly improves the generalization ability for instance-rare food classes to address the class-imbalance issue. The proposed method show promising performance with large margin improvements compared with existing methods.

arXiv:2307.00497

Don't Memorize; Mimic The Past: Federated Class Incremental Learning Without Episodic Memory

Deep learning models are prone to forgetting information learned in the past when trained on new data. This problem becomes even more pronounced in the context of federated learning (FL), where data is decentralized and subject to independent changes for each user. Continual Learning (CL) studies this so-called \textit{catastrophic forgetting} phenomenon primarily in centralized settings, where the learner has direct access to the complete training dataset. However, applying CL techniques to FL is not straightforward due to privacy concerns and resource limitations. This paper presents a framework for federated class incremental learning that utilizes a generative model to synthesize samples from past distributions instead of storing part of past data. Then, clients can leverage the generative model to mitigate catastrophic forgetting locally. The generative model is trained on the server using data-free methods at the end of each task without requesting data from clients. Therefore, it reduces the risk of data leakage as opposed to training it on the client's private data. We demonstrate significant improvements for the CIFAR-100 dataset compared to existing baselines.

arXiv:2306.17165

An Efficient General-Purpose Modular Vision Model via Multi-Task Heterogeneous Training

We present a model that can perform multiple vision tasks and can be adapted to other downstream tasks efficiently. Despite considerable progress in multi-task learning, most efforts focus on learning from multi-label data: a single image set with multiple task labels. Such multi-label data sets are rare, small, and expensive. We say heterogeneous to refer to image sets with different task labels, or to combinations of single-task datasets. Few have explored training on such heterogeneous datasets. General-purpose vision models are still dominated by single-task pretraining, and it remains unclear how to scale up multi-task models by leveraging mainstream vision datasets designed for different purposes. The challenges lie in managing large intrinsic differences among vision tasks, including data distribution, architectures, task-specific modules, dataset scales, and sampling strategies. To address these challenges, we propose to modify and scale up mixture-of-experts (MoE) vision transformers, so that they can simultaneously learn classification, detection, and segmentation on diverse mainstream vision datasets including ImageNet, COCO, and ADE20K. Our approach achieves comparable results to single-task state-of-the-art models and demonstrates strong generalization on downstream tasks. Due to its emergent modularity, this general-purpose model decomposes into high-performing components, efficiently adapting to downstream tasks. We can fine-tune it with fewer training parameters, fewer model parameters, and less computation. Additionally, its modularity allows for easy expansion in continual-learning-without-forgetting scenarios. Finally, these functions can be controlled and combined to meet various demands of downstream tasks.

arXiv:2306.17091

The Importance of Robust Features in Mitigating Catastrophic Forgetting

Continual learning (CL) is an approach to address catastrophic forgetting, which refers to forgetting previously learned knowledge by neural networks when trained on new tasks or data distributions. The adversarial robustness has decomposed features into robust and non-robust types and demonstrated that models trained on robust features significantly enhance adversarial robustness. However, no study has been conducted on the efficacy of robust features from the lens of the CL model in mitigating catastrophic forgetting in CL. In this paper, we introduce the CL robust dataset and train four baseline models on both the standard and CL robust datasets. Our results demonstrate that the CL models trained on the CL robust dataset experienced less catastrophic forgetting of the previously learned tasks than when trained on the standard dataset. Our observations highlight the significance of the features provided to the underlying CL models, showing that CL robust features can alleviate catastrophic forgetting.

arXiv:2306.16817공식 코드 · Papers with Code

Improving Online Continual Learning Performance and Stability with Temporal Ensembles

Neural networks are very effective when trained on large datasets for a large number of iterations. However, when they are trained on non-stationary streams of data and in an online fashion, their performance is reduced (1) by the online setup, which limits the availability of data, (2) due to catastrophic forgetting because of the non-stationary nature of the data. Furthermore, several recent works (Caccia et al., 2022; Lange et al., 2023) arXiv:2205.13452 showed that replay methods used in continual learning suffer from the stability gap, encountered when evaluating the model continually (rather than only on task boundaries). In this article, we study the effect of model ensembling as a way to improve performance and stability in online continual learning. We notice that naively ensembling models coming from a variety of training tasks increases the performance in online continual learning considerably. Starting from this observation, and drawing inspirations from semi-supervised learning ensembling methods, we use a lightweight temporal ensemble that computes the exponential moving average of the weights (EMA) at test time, and show that it can drastically increase the performance and stability when used in combination with several methods from the literature.

arXiv:2306.15117공식 코드 · Papers with Code

Continual Learning for Out-of-Distribution Pedestrian Detection

A continual learning solution is proposed to address the out-of-distribution generalization problem for pedestrian detection. While recent pedestrian detection models have achieved impressive performance on various datasets, they remain sensitive to shifts in the distribution of the inference data. Our method adopts and modifies Elastic Weight Consolidation to a backbone object detection network, in order to penalize the changes in the model weights based on their importance towards the initially learned task. We show that when trained with one dataset and fine-tuned on another, our solution learns the new distribution and maintains its performance on the previous one, avoiding catastrophic forgetting. We use two popular datasets, CrowdHuman and CityPersons for our cross-dataset experiments, and show considerable improvements over standard fine-tuning, with a 9% and 18% miss rate percent reduction improvement in the CrowdHuman and CityPersons datasets, respectively.