CHILD: A First Step Towards Continual Learning

By Mark B. Ring. 1997

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Reinforcement Learning

Citations in the corpus

Toward Understanding Catastrophic Forgetting in Continual Learning (2019, background): “Continual learning (or life-long learning) [32, 37, 44] is the ability of a machine learning model to continuously learn from a stream of data, which could possibly be non-iid or come from different but related tasks.”
Overcoming catastrophic forgetting in neural networks (2016, background): “Notably, previous reinforcement learning approaches to continual learning have either relied on either adding capacity to the network [Ring, 1998, Rusu et al., 2016] or on learning each task in separate networks, which are then used to train a single network that can play all games[Rusu et al.,…”
Revisiting the Arcade Learning Environment: Evaluation Protocols and Open Problems for General Agents (2017, background): “One traditional goal in reinforcement learning is for agents to continually improve their performance as they obtain more data (Hutter, 2005; Ring, 1997; Singh, Barto, & Chentanez, 2004; Sutton et al., 2011; Thrun & Mitchell, 1993; Wilson, 1985).”
Unifying Count-Based Exploration and Intrinsic Motivation (2016, background): “The idea also finds roots in continual learning (Ring, 1997).”
Continual Lifelong Learning with Neural Networks: A Review (2018, background): “An early attempt to realize lifelong learning through transfer learning was proposed by Ring (1997) through the use of a hierarchical neural network that solves increasingly complex reinforcement learning tasks by incrementally adding neural units and encode a wider temporal context in which…”
Experience Replay for Continual Learning (2018, background): “An intelligent system that interacts with its environment faces the problem of continual learning, in which new experience is constantly being acquired, while old experience may still be relevant [25].”
Gradient Episodic Memory for Continual Learning (2017, background): “This learning setting led to implementations [Carlson et al., 2010, Ruvolo and Eaton, 2013, Ring, 1997], and theoretical investigations [Baxter, 2000, Balcan et al., 2015, Pentina and Urner, 2016], although the latter ones have been restricted to linear models.”
Policy Consolidation for Continual Reinforcement Learning (2019, methodology): “…that grow the amount of neural resources, some utilise the deﬁnition of task boundaries (Rusu et al., 2016; Xu & Zhu, 2018), while others do not (Ring, 1997; Zhou et al., 2012), but since the computational complexity grows with the size of the network, it is not clear that they are scalable for…”
Continual Normalization: Rethinking Batch Normalization for Online Continual Learning (2022, background): “Continual learning (Ring, 1997) is a promising approach towards building learning systems with human-like capabilities.”
Continual Sequence Generation with Adaptive Compositional Modules (2022, methodology): “Continual learning (CL) (Ring, 1998; Thrun, 1998) has been introduced to improve model’s ability to learn tasks in a stream by mitigating forgetting Figure 1: Comparison between previous methods (a and b) and our proposed method (c), from a multi-layer transformer model perspective.”
Kernel Continual Learning (2021, background): “Continual learning (Ring, 1998; Lopez-Paz & Ranzato, 2017; Nguyen et al., 2018), also known as life-long learning, was introduced to deal with catastrophic forgetting.”
Continual Learning for Text Classification with Information Disentanglement Based Regularization (2021, background): “As a result, Continual Learning (CL) (Ring, 1998; Thrun, 1998) has received more attention recently as it can enable models to perform positive transfer (Perkins et al., 1992) as well as remember previously seen tasks.”
Gradient Projection Memory for Continual Learning (2021, background): “In stark contrast, Artificial Neural Networks (ANNs) under such Continual Learning (CL) paradigm (Ring, 1998; Thrun & Mitchell, 1995; Lange et al., 2021) forget the information learned in the past tasks upon learning new ones.”
Continuous Coordination As a Realistic Scenario for Lifelong Learning (2021, background): “While lifelong learning has been identiﬁed as an important and challenging problem decades ago (Thrun, 1998; Ring, 1998), it has recently seen a surge of interest (Lopez-Paz & Ran-zato, 2017; Chaudhry et al., 2018b; 2019; Kirkpatrick et al., 2017; Aljundi et al., 2018) with the success of deep…”
Towards Continual Reinforcement Learning: A Review and Perspectives (2020, background): “This includes the process of building complicated behaviors on top of those already developed (Ring, 1997) while being able to reapply, adapt, and generalize previously learned abilities to new situations.”
IIRC: Incremental Implicitly-Refined Classification (2020, background): “Lifelong Learning is a broad, multi-disciplinary, and expansive research domain with several synonyms: Incremental Learning [42], Continual Learning [48], and Never Ending Learning [35].”