Lifelong Learning: A Case Study.

By Sebastian Thrun. 1995

How researchers see this paper

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

Citations in the corpus

Net2Net: Accelerating Learning via Knowledge Transfer (2015, background): “These machine learning systems need to continue to function for long periods of time and continually experience new training examples as these examples become available.”
Lifelong Generative Modeling (2017, background): “Abstract Lifelong learning [1, 2] is the problem of learning multiple consecutive tasks in a sequential manner where knowledge gained from previous tasks is retained and used for future learning.”
Clustering Learning Tasks and the Selective Cross-Task Transfer of Knowledge (1998, unknown): “Applying the TC algorithm to more sophisticated image encodings, and applying algorithms that can discover more sophisticated commonalities between multiple tasks such as those surveyed in [34], is subject to ongoing research.”
Learning to Learn (1998, unknown): “It might seem that sequential transfer [Pratt & Mostow 1991; Pratt 1992; Sharkey & Sharkey 1992; Thrun & Mitchell 1994; Thrun 1995] would be easier.”
Knowledge Transfer in Deep convolutional Neural Nets (2007, methodology): “They include discriminability-based transfer (Pratt 1993), multitask learning (Caruana 1997), explanation based neural nets (Thrun 1995), knowledge based cascade correlation (Schultz & Rivest 2000) and internal representation learning (Baxter 2000).”
Indian Buffet Neural Networks for Continual Learning (2019, background): “Firstly, models need to leverage transfer learning from previously learned tasks during the learning of a new task at time t [5, 1, 6, 7].”
Multitask Learning (1997, methodology): “It might seem that sequential transfer (Pratt & Mostow 1991; Pratt 1992; Sharkey & Sharkey 1992; Thrun & Mitchell 1994; Thrun 1995) would be easier.”
Big data classification: problems and challenges in network intrusion prediction with machine learning (2014, methodology): “Hence the machine lifelong learning (ML3) techniques should be used [17-19].”
Constructive neural-network learning algorithms for pattern classification (2000, background): “• Lifelong Learning: Recent research in lifelong learning [48] has proposed training networks that learn to solve multiple related problems by applying the knowledge acquired from the simpler problems to learn the more difficult ones.”