Iterative Machine Teaching

Abstract

In this paper, we consider the problem of machine teaching, the inverse problem of machine learning. Different from traditional machine teaching which views the learners as batch algorithms, we study a new paradigm where the learner uses an iterative algorithm and a teacher can feed examples sequentially and intelligently based on the current performance of the learner. We show that the teaching complexity in the iterative case is very different from that in the batch case. Instead of constructing a minimal training set for learners, our iterative machine teaching focuses on achieving fast convergence in the learner model. Depending on the level of information the teacher has from the learner model, we design teaching algorithms which can provably reduce the number of teaching examples and achieve faster convergence than learning without teachers. We also validate our theoretical findings with extensive experiments on different data distribution and real image datasets.

Published in ICML 2017

References in corpus (7)

• Distilling the Knowledge in a Neural Network
• FitNets: Hints for Thin Deep Nets
• Do Deep Nets Really Need to be Deep?
• The Teaching Dimension of Linear Learners
• Online Optimization in Dynamic Environments
• Becoming the Expert - Interactive Multi-Class Machine Teaching
• Analysis of a Design Pattern for Teaching with Features and Labels

Cited by in corpus (28)

• An Overview of Machine Teaching
• Causative Cyberattacks on Online Learning-based Automated Demand Response Systems
• Teaching Categories to Human Learners with Visual Explanations
• Towards Black-box Iterative Machine Teaching
• Fair Machine Guidance to Enhance Fair Decision Making in Biased People
• Angular Visual Hardness
• Teaching Multiple Concepts to a Forgetful Learner
• An Optimal Control Approach to Sequential Machine Teaching
• Teacher Improves Learning by Selecting a Training Subset
• An Optimal Control View of Adversarial Machine Learning
• Using Machine Teaching to Investigate Human Assumptions when Teaching Reinforcement Learners
• Machine Teaching by Domain Experts: Towards More Humane,Inclusive, and Intelligent Machine Learning Systems
• Unlearn What You Have Learned: Adaptive Crowd Teaching with Exponentially Decayed Memory Learners
• Interaction-limited Inverse Reinforcement Learning
• Curriculum Design for Teaching via Demonstrations: Theory and Applications
• Privacy-preserving Stochastic Gradual Learning
• Learning Data Teaching Strategies Via Knowledge Tracing
• An Investigation of Data Poisoning Defenses for Online Learning
• The Teaching Dimension of Kernel Perceptron
• The Sample Complexity of Teaching-by-Reinforcement on Q-Learning
• Iterative Teaching by Label Synthesis
• Teaching an Active Learner with Contrastive Examples
• Towards Scheduling Federated Deep Learning using Meta-Gradients for Inter-Hospital Learning
• Average-case Complexity of Teaching Convex Polytopes via Halfspace Queries
• Preference-Based Batch and Sequential Teaching
• Iterative Classroom Teaching
• Learning to Reweight with Deep Interactions
• Improving Generalization via Attribute Selection on Out-of-the-box Data