A Case Study of Incremental Concept Induction

By Jeffrey C. Schlimmer, D. Fisher. 1986

Subfields: Representation & Self-Supervised Learning

How researchers see this paper

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

In Representation & Self-Supervised 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.”
FeTrIL: Feature Translation for Exemplar-Free Class-Incremental Learning (2022, neutral): “Incremental learning [36] was introduced to reduce the memory and computational costs of machine learning algorithms.”
Continual Learning with Adaptive Weights (CLAW) (2019, background): “Data belonging to different tasks could potentially be non i.i.d. (Schlimmer & Fisher, 1986; Sutton & Whitehead, 1993; Ring, 1997; Schmidhuber, 2013; Nguyen et al., 2018; Schmidhuber, 2018).”
Incremental Learning of Object Detectors without Catastrophic Forgetting (2017, background): “The problem of incremental learning has a long history in machine learning and artificial intelligence [6,29,36,37].”
Incremental Induction of Decision Trees (1989, methodology): “Previously, Schlimmer and Fisher (1986a) designed an algorithm named ID4 to address the incremental construction of decision trees.”
Instance-Based Learning Algorithms (1991, background): “Several researchers have shown how to decrease these costs significantly ( Schlimmer & Fisher, 1986; Utgoff, 1989; Van de Velde, 1989).”
Online Learning with Random Representations (1993, background): “There are n inputs denoted x1, x2, . . . , xn (xi ∈ {0, 1}) and m hidden units.”
Partitioned Variational Inference: A unified framework encompassing federated and continual learning (2018, background): “…learning or life-long learning or incremental learning) is the ability to learn continually and adapt quickly to new experiences without catastrophically forgetting previously seen experiences [Schlimmer and Fisher, 1986, McCloskey and Cohen, 1989, Sutton and Whitehead, 1993, Ratcliff, 1990].”
Hybrid decision tree (2002, methodology): “Schlimmer and Fisher (1986) proposed an algorithm named ID4 which”
Decision Tree Induction Based on Efficient Tree Restructuring (1997, methodology): “The idea of maintaining such information for symbolic variables was demonstrated by Schlimmer and Fisher (1986).”
The Impact of Diversity on Online Ensemble Learning in the Presence of Concept Drift (2010, unknown): “The well-known SEA Concepts [11] and STAGGER Concepts [26] data sets are considered to have a gradual drift and abrupt drifts, respectively.”
Monotonicity Maintenance in Information-Theoretic Machine Learning Algorithms (1995, methodology): “TDIDT algorithms that were not studied here, such as C4.5 (Quinlan, 1987), ID4 ( Schlimmer & Fisher, 1986 ), and ID5R (Utgoff, 1989), may also be adapted to deal with MOC problems using a similar approach.”
When Is the Right Time to Refresh Knowledge Discovered from Data? (2013, methodology): “For classification, algorithms such as ID4 (Schlimmer and Fisher 1986) and ID5 (Utgoff 1988, 1989) were developed for revising a decision tree induced from old data as new data were added in.”
Dynamic Weighted Majority: An Ensemble Method for Drifting Concepts (2007, methodology): “In previous work, we evaluated DWM using two synthetic data sets, the STAGGER concepts (Schlimmer and Granger, 1986) and the SEA concepts (Street and Kim, 2001), achieving the best published results (Kolter and Maloof, 2003).”
How catastrophic can catastrophic forgetting be in linear regression? (2022, background): “The goal is to adapt the model to new tasks while preserving its performance on previously-learned tasks [53, 62, 70].”